Second conference of the

International Network for

Didactic Research in University Mathematics

INDRUM 2018
April 5-7

KRISTIANSAND
University of Agder

PROCEEDINGS

[Link]

INDRUM 2018 is an ERME topic conference

IM AG
INSTITUT MONTPELLIERAIN
ALEXANDERORDTHENDIECK
PROCEEDINGS of INDRUM 2018
Second conference of the
International Network for Didactic
Research in University Mathematics
April 5-7, 2018
Kristiansand, Norway

Editors: Viviane Durand-Guerrier, Reinhard Hochmuth,

Simon Goodchild and Ninni Marie Hogstad
Publisher: University of Agder and INDRUM
Place: Kristiansand, Norway
Year: 2018
ISSN: 2496-1027 (online)
Table of contents

INDRUM 2018 Editorial 7

Plenary talk 9

Lessons for mathematics higher education from 25 years of mathematics sup-

port, Duncan Lawson [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Panel 11

Education and professional development of University Mathematics Teachers,

Winsløw Carl [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

TWG 1: Calculus and Analysis 13

Are all denumerable sets of numbers order-isomorphic?, Branchetti Laura [et al.] 14

An empirical study of the understanding of formal propositions about sequences,

with a focus on infinite limits, Chorlay Renaud . . . . . . . . . . . . . . . . . . . 24

Students’ understanding of -statements involving equality and limit, Ghedamsi

Imène [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

”A function is continuous if and only if you can draw its graph without lifting
the pen from the paper” – Concept usage in proofs by students in a topology
course, Hanke Erik . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Learning complex analysis in different branches – Project Spotlight-Y for future

teachers, Hanke Erik [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

On students’ understanding of Riemann sums of integrals of functions of two

variables, Trigueros Marı́a [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Teaching and learning continuity with technologies, Vandebrouck Fabrice [et al.] 74

1
TWG 2: Mathematics for engineers - Mathematical Modelling - Mathematics
and other disciplines 84

The ecological relativity of modelling practices: adaptations of an study and

research path to different university settings, Barquero Berta . . . . . . . . . . . 85

C’ (x) = C(x+1)-C(x)? - Students’ connections between the derivative and its

economic interpretation in the context of marginal cost, Feudel Frank . . . . . . 95

Weekly homework quizzes as formative assessment for Engineering students are

a fair and effective strategy to increase learning?, Gaspar Martins Sandra . . . . 105

The use of integrals in Mechanics of Materials textbooks for engineering students:

the case of the first moment of an area, González-Martı́n Alejandro S. [et al.] . . 115

Predictors of performance in engineering mathematics, Gradwohl Julia [et al.] . . 125

”Making connections” in the mathematics courses for engineers: the example of

online resources for trigonometry, Gueudet Ghislaine [et al.] . . . . . . . . . . . . 135

Engineering students’ engagement with resources in an online learning environ-

ment, Kanwal Shaista . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

Central dialectics for mathematical modelling in the experience of a study and

research path at university level, Monreal Galán Nacho [et al.] . . . . . . . . . . . 155

Task design for Engineering Mathematics: process, principles and products, Schmidt
Karsten [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165

Mathematical modelling and activation - a study on a large class, a project-based

task and students’ flow, Vos Pauline [et al.] . . . . . . . . . . . . . . . . . . . . . 175

Using schematic representation of resource systems to examine how first year

engineering students use resources in their studies of mathematics,
Hillesund Eivind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185

Mathematics Teaching for Economics Students, But How?, Landgärds Ida . . . . 187

Non-standard Problems in an Ordinary Differential Equations Course, Rogovchenko

Svitlana [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

Students’ understanding of mathematical models in three norwegian biology courses,

Tetaj Floridona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

2
TWG 3: Number, Algebra, Logic 193

Contemporary Research Practices in Discrete Mathematics - A way to enrich the

understanding of Discrete Mathematics at University Level, Abdallah Elise . . . 194

Analysing regressive reasoning at university level, Barbero Marta [et al.] . . . . . 204

A Study of Students’ Reasoning About ”There exists no ...”, Brown Stacy . . . . 214

Students’ problems in the identification of subspaces in Linear Algebra, Fleis-

chmann Yael [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

A TDS analytical framework to study students’ mathematical activity An exam-

ple: linear transformations at University, Lalaude-Labayle Marc [et al.] . . . . . . 234

Leveraging Specific Contexts and Outcomes to Generalize in Combinatorial Set-

tings, Lockwood Elise [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244

Discrete mathematics at university level. Interfacing mathematics, computer sci-

ence and arithmetic, Ouvrier-Buffet Cécile [et al.] . . . . . . . . . . . . . . . . . . 255

Tasks for enriching the understanding of the concept of linear span, Papadaki
Evi [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

Delineating Aspects of Understanding Eigentheory through Assessment Develop-

ment, Wawro Megan [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

TWG 4: Students’ practices 285

The complexity of knowledge construction in a classroom setting,

Dreyfus Tommy [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

Analysing tasks for the Flipped Classroom from the perspective of Realistic Math-
ematical Education, Fredriksen Helge . . . . . . . . . . . . . . . . . . . . . . . . . 296

Engaging with feedback: How do students remediate errors on their weekly

quiz, Howard Emma [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

Student Partners in Task Design in a computer medium to promote Foundation

students’ learning of mathematics, Jaworski Barbara [et al.] . . . . . . . . . . . . 316

Theoretical and Empirical Description of Phases in the Proving Processes of Un-

dergraduates, Kirsten Katharina . . . . . . . . . . . . . . . . . . . . . . . . . . . 326

Student use of resources in Calculus and Linear Algebra, Kock Zeger-Jan [et al.] 336

3
 Supporting self-efficacy and self-regulation skills in undergraduate mathematics:
comparison of two different instructional designs, Lahdenperä Juulia . . . . . . . 346

Situational interest in university mathematics courses: similar for real-world prob-

lems, calculations, and proofs?, Rach Stefanie [et al.] . . . . . . . . . . . . . . . . 356

Determining your own grade – Student perceptions on self-assessment in a large

mathematics course, Häsä Jokke [et al.] . . . . . . . . . . . . . . . . . . . . . . . 366

Seminar groups as part of first-semester mathematics teaching: What did the

students learn?, Pettersson Kerstin [et al.] . . . . . . . . . . . . . . . . . . . . . . 368

Students’ usage patterns of video recorded lectures in an undergraduate mathe-

matics course, Wold Margrethe . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370

TWG 5: Teachers’ practices 372

Helping lecturers address and formulate teaching challenges: an exploratory study,

Bosch Marianna [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

An Analysis of University Mathematics Teaching using the Knowledge Quar-

tet, Breen Sinéad [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383

Knowledge of the Practice in Mathematics in University Teachers, Delgado Re-

bolledo Rosa [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393

Exploration of new post-secondary mathematics teachers’ experiences: prelimi-

nary results of a narrative inquiry, Mathieu-Soucy Sarah [et al.] . . . . . . . . . . 403

Inquiry-based learning and pre-service teachers education, Nicolás Pedro [et al.] . 412

A model for instructional design in mathematics implemented in teacher educa-

tion, Strømskag Heidi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422

University Teachers-Researchers’ Practices: the Case of Teaching Discrete Math-

ematics, Tabchi Theresia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432

Characteristics of participation - A mathematician and a mathematics educator

collaborating on a developmental research project, Viirman Olov . . . . . . . . . 442

A case study of a university teacher of calculus 1, Alzubaidi Ibrahim [et al.] . . . 452

Who Becomes U.S. Mathematics Teachers: A Longitudinal Analysis of First-time

Praxis
R II Mathematics Content Knowledge Examinees, Glassmeyer David [et al.]454

4
 Cross-linking maths – using keynotes to structure a curriculum for future teach-
ers, Hamann Tanja [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456

BiMathTutor – The program for further training of new tutors in University

Mathematics, Huget Judith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458

Getting language awareness: A curriculum for language and language teaching

for pre-service studies for teachers of mathematics, Schmidt-Thieme Barbara . . 460

A university lecturer’s role and interactions in a flipped classroom,

Syrdalen Erlend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462

Programming in mathematics teacher education – a collaborative teaching ap-

proach, Viirman Olov [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464

TWG 6: Transition to and across university 466

Different goals for pre-university mathematical bridging courses – Comparative

evaluations, instruments and selected results, Biehler Rolf [et al.] . . . . . . . . . 467

From single to multi-variable Calculus: a transition?, Brandes Hadas [et al.] . . . 477

A study of transitions in an undergraduate mathematics program, Broley Laura [et

al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487

Bridging probability and calculus: the case of continuous distributions and inte-
grals at the secondary-tertiary transition, Derouet Charlotte [et al.] . . . . . . . . 497

Affective variables in the transition from school to university mathematics, Geisler

Sebastian [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507

Discussing Mathematical Learning and Mathematical Praxeologies from a Subject

Scientific Perspective, Hochmuth Reinhard . . . . . . . . . . . . . . . . . . . . . . 517

Evaluating Innovative Measures in University Mathematics – The Case of Affec-

tive Outcomes in a Lecture focused on Problem-Solving, Kuklinski Christiane [et
al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527

Defining as discursive practice in transition – Upper secondary students reinvent

the formal definition of convergent sequences, Schüler-Meyer Alexander . . . . . 537

Fostering Heuristic Strategies in Mathematics Teacher Education, Stender Pe-

ter [et al.] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547

Sponsors and participants 557

5
Author Index 559

6
 Editorial
Viviane Durand-Guerrier1, Reinhard Hochmuth2, Simon Goodchild3, Ninni Marie Hogstad3
University of Montpellier, [Link]-guerrier@[Link]
University of Hannover, hochmuth@[Link]
University of Agder, [Link]@[Link], [Link]@[Link]

The INDRUM 2018 conference was held in Kristiansand, 5-8 April 2018. The INDRUM
conferences fall within the activities of INDRUM (International Network for Didactic
Research in University Mathematics), which has been initiated by an international team of
researchers in didactics of mathematics at university level. This network aims to contribute to
the development of research in didactics of mathematics at all levels of tertiary education,
with particular focus on support for young researchers in the field and for dialogue with
mathematicians.
The idea for the network and biennial conferences was first discussed in Paris (France),
November 2014 and then Oberwolfach (Germany), December 2014. Following these
discussions, a scientific committee with 19 scholars from 12 countries was established. The
decision for organising the first conference in Montpellier (France), 31 March - 2 April 2016,
with Elena Nardi and Carl Winslow as chair and co-chair was taken during CERME 9 in
Prague, in February 2015.
Following the success of INDRUM 2016, the decision for organising the second INDRUM
conference in Kristiansand in April 2018 with Simon Goodchild as Chair of the local
committee, was taken during INDRUM 2016 in Montpellier. The International Scientific
Committee held a meeting in Dublin during CERME 10, and nominated the INDRUM 2018
International Programme Committee: Viviane Durand-Guerrier (Montpellier, France) Chair;
Reinhard Hochmuth (Hannover, Germany) Co-chair ; Marianna Bosch (Barcelona, Spain);
Simon Goodchild (Kristiansand, Norway) ; Thomas Hausberger (Montpellier, France) ; Ninni
Marie Hogstad (Kristiansand, Norway) ; Elena Nardi (Norwich, United Kingdom) ; Chris
Rasmussen (San Diego, United States) ; Carl Winsløw (Copenhagen, Denmark). The Local
Organising Committee was composed of Simon Goodchild (Kristiansand, Norway) Chair;
Lillian Egelandsaa (Kristiansand, Norway); Ninni Marie Hogstad (Kristiansand, Norway);
Thomas Hausberger (Montpellier, France); Elisabeth Rasmussen (Kristiansand, Norway). As
for INDRUM 2016, INDRUM 2018 was an ERME Topic Conference.
A total of 53 papers and 14 posters was accepted for presentation. The final number of papers
and posters presented at the conference and included in these proceedings (51 full papers and
14 posters, with the latter represented in the Proceedings as two-page short papers) varied
slightly as a small number of delegates withdrew submissions or cancelled attendance for
personal reasons. Discussion of the accepted papers and posters was organised in six thematic
working groups (TWG1-TWG6), based on a classification of contents. Two members of the
INDRUM Scientific Committee were invited to lead each of the five TWGs:
TWG1: Calculus and Analysis (María Trigueiros, Fabrice Vandebrouck)
TWG2: Mathematics for engineers; Mathematical Modelling; Mathematics and other
disciplines (Alejandro S. Gonzáles Martín, Ghislaine Gueudet)
TWG3: Number, Algebra, Logic (Faïza Chellougui, Viviane Durand-Guerrier)
TWG4: Students' practices (Chris Rasmussen, Elena Nardi)
TWG5: Teachers’ practices (Marianna Bosch, Simon Goodchild)
TWG6: Transition to and across university (Thomas Hausberger, Reinhard Hochmuth)

7
The scientific programme comprised: A plenary talk by Duncan Lawson (United Kingdom):
Lessons for mathematics higher education from 25 years of mathematics support; a
presentation of posters and of thematic working groups; a plenary panel chaired by Carl
Winsløw: Preparation and training of university mathematics teachers: Panelists: Rolf Biehler
(Germany), Barbara Jaworski (United Kingdom), Frode Rønning (Norway), Megan Wawro
(United States). The accepted papers were presented in two parallel sessions and discussed in
four thematic working group (TWG) sessions. A report was prepared and presented in plenary
on April 8th.
The conference was attended by a total of about 120 registered participants. In the light of the
volume and quality of submissions, and substance of exchanges during the sessions, we are
happy to conclude that the second INDRUM conference turned out as a further eminent
success.
Papers appear in these Proceedings in a version chosen by the authors following the (optional)
possibility to upload a final version of their paper soon after the conference.
Very special thanks are due to the organising committee, chaired by Simon Goodchild and co-
chaired by Ninni Marie Hogstad, for their tireless work over many months towards this event.
Ninni Marie was responsible for the website, and received continuous support from Thomas
Hausberger. Administrative support was offered by Lillian Egelandsaa and Elisabeth
Rasmussen. These colleagues worked unstintingly before, during and after the conference to
ensure that every participant had a smooth, productive and enjoyable INDRUM experience.
They have set the bar high for the conferences to follow and we are indebted to them all.
The organizers are grateful to MatRIC, Centre for Research, Innovation and Coordination of
Mathematics Teaching for financial support covering the work of the local organizing
committee and some other conference arrangements, also the University of Agder for
technical and domestic services and conference accommodation.
INDRUM follow-up
Strengthening the Network through publications is an important goal of the INDRUM
network. Apart the INDRUM conferences proceedings, two publications were planed after
INDRUM 2016.
First, an International Journal for Research in Undergraduate Mathematics (IJRUME) Special
Issue has been guest-edited by Elena Nardi and Carl Winsløw, with support from IJRUME
Editor Chris Rasmussen and reviewers including members of the INDRUM2016 Scientific
Committee, was published in time for the 2018 conference and participants were able to
download a copy of the issue without charge.
Second, a book reporting from INDRUM 2016 and INDRUM 2018 will be published in the
Routledge ERME Series. It will be based on the scientific work developed in the TWGs
during both conferences. A TWGs session of INDRUM 2018 was devoted to provide input to
the book, in addition to input provided during INDRUM2016. Carl Winslow, Viviane
Durand-Guerrier, Elena Nardi and Reinhard Hochmuth will be the editors.
Third, during INDRUM 2018, 13 colleagues from 10 countries have accepted an invitation to
be members of the INDRUM International Scientific Committee, that now comprises 31
colleagues form 15 countries.
Finally, we are very happy to be able to announce that the Faculty of Sciences of Bizerte
(Tunisia) will host INDRUM 2020; 27-29 March 2020, with Faïza Chellougui as chair of the
Local Organising Committee and Rahim Kouki as co-chair. The conference will be chaired by
Thomas Hausberger and Marianna Bosch.
We now invite you to carry on reading this volume and we hope that the promise of its
contents will encourage you to join, or continue to be part of, the ambitious, bold enterprise
that is INDRUM!

8
Plenary talk
 Lessons for mathematics higher education from 25 years of
mathematics support
Duncan Lawson1 and Tony Croft²
1
Newman University, United Kingdom; ²Loughborough University, United Kingdom

INDRUM KEYNOTE PRESENTATION

The scale and scope of mathematics support within UK universities have grown
significantly since the 1990s. Mathematics support has evolved from a ‘cottage
industry’ initiated by enthusiasts into a main-line student support provision overseen
by institutional senior managers. Over this 25+ year period, the importance of the
mathematical sciences in other disciplines has similarly boomed. No longer is it just
engineering and physics undergraduates who need to acquire highly developed
mathematical skills. Today geographers, bioscientists, sociologists and political
scientists (to name but a few) have to be more skilled than ever before with
understanding mathematical and statistical models and methods, particularly if they
are to be able to access the international research literature and compete in the
international employment market. Just as in the 1980s and 1990s, the Engineering
Council produced reports warning of ‘the mathematics problem’, so in the 2000s and
2010s, the British Council and Royal Society of Arts have done the same. This
presentation will outline how mathematics support has developed throughout the UK
to meet this increasing demand.
Whilst the initial impetus for mathematics support came from a desire to improve the
mathematical learning of students from other disciplines, it is an indisputable fact that
a significant proportion of the users of mathematics support has been, and remains,
mathematics undergraduates. This gives us cause to reflect: why is mathematics
support so attractive to mathematics undergraduates? To answer this question, we
explore the views of mathematics undergraduate students as expressed through the
National Student Survey and in focus groups and individual interviews. The views
the students express shed light on the reasons why many of them find mathematics
support to be an attractive resource to support their learning.

10 [Link]:indrum2018:199857
Panel
 Education and professional development of University Mathematics
Teachers

Panellists: Rolf Biehler (Paderborn University, Germany)

Barbara Jaworski (Loughborough University, United Kingdom)
Frode Rønning (Norwegian University of Science and Technology,
Norway)
Megan Wawro (Virginia Tech, United States)
Chair: Carl Winsløw (University of Copenhagen, Denmark)

ABSTRACT
The theme of this panel may surprise some, as university teachers of mathematics
typically hold a PhD in mathematics or some adjacent field, and in many places some
“pedagogical training” is also foreseen. However, university teaching presents still
more challenges (in many places: more inhomogeneous or different student groups to
teach), and opportunities (including new technology, and – we hope – useful
resources from research on UME). For all of these reasons, the panel will address the
following questions:
1. What is the current, typical preparation of University Mathematics Teachers
for their function as teachers? What “in-service” opportunities for teacher
development exist? - naturally, answers will depend both on countries and
institutions, but sharing experiences could help to provide an updated picture
of how the “professional knowledge of UME teachers” is currently built and
sustained.
2. Do the current preparation and opportunities for development meet the
demands that exist or can be foreseen? Could the preparation and development
opportunities be improved, for instance by giving university teachers (more)
access to selected parts of current research on UME, and possibly also
participate in research and development projects? What initiatives exist, and
which could be imagined as beneficial – both to increase the impact and quality
of research on UME, and of UME itself?

12 [Link]:indrum2018:200035
TWG 1: Calculus and Analysis
 Are all denumerable sets of numbers order-isomorphic?
Laura Branchetti1 and Viviane Durand-Guerrier2
1
Department of Mathematical, Physical and Computer sciences, Univ. of Parma,
Italy, [Link]@[Link]
2
Institut Montpelliérain Alexander Grothendieck, CNRS, Univ. Montpellier, France
In this paper we study cognitive conflicts on the issue of number sets being dense,
ordered and denumerable. We first provide historical-epistemological background
related to these notions. Then we consider the cognitive conflicts under the lenses of
concept image and concept definition, which we use to analyse empirical data
collected in order to understand better the didactical and cognitive issues at stake.
Keywords: density, ordered set, denumerable set, concept image versus definition.
INTRODUCTION
Our interest in the title question comes from our teaching experiences in the first year
of the Master degree in Mathematics course in Italy for the first author, and in first-
year university courses in France for the second author. In both cases, the focus was
on the distinction between density and continuity for an ordered set of numbers. Both
authors were surprised by the following students’ questions concerning the
denumerable sets N and Q.
Q1. How is it possible to find an order in Q if Q is dense, i.e. when the consecutive
number of a rational does not exist?
Q2. How is it possible that there is a bijection between N and Q, but N is discrete and Q
is dense?
The students’ questions highlighted potential conflicts which emerge when making
explicit the properties of density1 and denumerability of the set of rational numbers at
the same time. This motivates a research investigation into the didactical
transposition of the objects that underlie the two properties, namely order on a set,
properties of both discrete2 and denumerable sets, bijection, ordered isomorphism,
difference between cardinal and ordinal numbers, and enumeration3. Our general
research question is: how should we deal with these questions in classroom activities
in order to help students overcome the apparent contradictions? In this paper, we will
focus on a less ambitious sub-question concerning the first question posed by
students (Q1).

1
From the point of view of order, an order dense set is a linearly ordered set (X,<) with the property that if x<y then
there exists z∈X with x<z<y (Jech, 2003). Here with the term “density” we refer to order density.
2
A set S is discrete in a topological space X if every point x∈X has a neighbourhood U such that 𝑆 ∩ 𝑈 = {𝑥} (points
are said to be isolated) (Krantz, 1999, p.63).
3
An enumeration is a complete, ordered listing of all the items of a set. An enumeration for an infinite set is a one-to-
one correspondence between this set and the set of positive integers.

14 [Link]:indrum2018:174129
 RQ: What are university students’ and teachers’ concept images and concept
definitions of dense, ordered and denumerable set? How do they connect them?
We analysed students’ and teachers’ answers from two perspectives: i. a historical-
epistemological analysis of the topics which emerged in the students’ questions: ii.
concept image and concept definition (Tall & Vinner, 1981). The first point is
addressed in the first section: we refer to historical works in which infinite sets were
studied from the point of view of cardinality, ordering and enumeration (Galilei,
1638; Lolli, 2013; Peano, 1889). The second point is addressed in the second section;
we use our historical-epistemological analysis, together with results in mathematics
education (Tirosh & Tsamir, 1996; Bergé, 2010; Durand-Guerrier, 2016; Branchetti,
2016), as resources to identify a priori possible students’ concept images that could
conflict with each other. In the third section, we describe the contexts and
methodology of data collection and analysis, carried out in parallel in France and
Italy which involved university teachers and Master degree students. Finally, we
provide a brief overview of the way the concepts are introduced in scholastic stages
prior to university studies in both countries, as a relevant background for our
conclusions and starting point for further developments.
HISTORICAL EPISTEMOLOGICAL ISSUES
Galileo’s view on numbers and their squares: an issue about cardinality
In one of the most famous books by Galileo Galilei (1564 - 1642), Dialogues
concerning Two New Sciences (1638), the Italian physician, mathematician and
philosopher introduced the one-to-one correspondence between natural numbers and
their squares. We report here a brief summary of the main ideas. The dialogue
concerns the difficulty that appears when trying to compare the number of points
contained in two segments, one being longer than the other. Salviati, the voice of
Galilei, states; “This is one of the difficulties which arise when we attempt, with our
finite minds, to discuss the infinite, assigning to it those properties which we give to
the finite and limited” (English translation, 1914, p.31). He then moves to numbers
and develops an argument on the impossibility of comparing the totality of all
numbers with the numbers of squares since they are both infinite:
“neither is the number of squares less than the totality of all numbers, nor the latter
greater than the former; and, finally, the attributes "equal," "greater," and "less" are not
applicable to infinite, but only to finite, quantities.” (p. 32-33)
Density of Q, cardinal and ordinal numbers: Cantor's contribution
Cantor (1845 - 1918), working on trigonometric series and their convergence, moved
on to the creation of a new theory of transfinite numbers, and the perspectives of
Number Theory and Set Theory. Cantor came firstly to the definitions of a derived set
– the set of limit points – and of a dense set and then of a dense-in-itself set, like the
rational numbers set (Lolli, 2013). Studying infinite sets, he introduced the diagonal

15 [Link]:indrum2018:174129
argument to prove that not only do the squares have the same cardinality as natural
numbers, but so also does the Cartesian product of the set of natural numbers by it-
self, and hence the set of rational numbers, thanks to the existence of a surjection
from the Cartesian product to this latter set. In other words, he had to find a way to
enumerate the ordered pairs of natural numbers, being sure to consider every pair
once and once only, following an ordering principle. To do this, he moved from the
usual linear image of order to a 2-dimensional image. A crucial distinction,
introduced by Cantor when he was facing such problems, is between cardinal
numbers and ordinal numbers (Lolli, 2013). While in the set of natural numbers with
its standard structure (formalized by Peano in 1889) the relation of order is strictly
connected to the problem of ordering and with the induction principle, this is no
longer the case in Q. Indeed, the standard order on Q (i.e. that consistent with
measurement of magnitudes onto the line) is not consistent with Cantor’s diagonal
3 5 3
ordering principle, i.e. in the resulting order, is listed before but after .
4 4, 2

Peano's formalization of Arithmetic and the issue of order in natural numbers

In Peano’s Arithmetic, the ideas developed by Cantor were formalized and used as
principles to grasp the “essence” of natural numbers: the injective function that
establishes Cantor’s “first generation principle” (Lolli, 2013) for the consecutive
element of a natural number is strictly linked to the operation of addition (the
consecutive of n, being S(n), is equal to n+1) and to the comparison between natural
numbers, if we consider the standard order. In this structure, the consecutive element
is always greater than its precedent in respect to the standard order.
RESEARCH FRAMEWORK
According to Tall & Vinner (1981) every concept, from a cognitive point of view, is
associated to different concept images. A relevant cognitive feature of
conceptualization concerns the introduction of formal definitions: it often happens
that the concept definition is not introduced appropriately by teachers in relation to
the concept images. According to the authors, “a teacher may give the formal
definition and work with the general notion for a short while before spending long
periods in which all examples are given by formulae. In such a case the concept
image may develop into a more restricted notion, only involving formulae, whilst the
concept definition is largely inactive in the cognitive structure” (p. 3). Some students’
concept images may be recognized as conflicting and inconsistent from an expert
point of view, but they can coexist in their mind until a conflict is shown evoking
them together simultaneously (p. 2). Such cognitive conflicts are occasions for
learning and advancing in the process of conceptualization, but if not recognized and
suitably overcome, they can become obstacles in the learning processes. We
hypothesised that the students who asked our two questions (Q1 and Q2) were facing
cognitive conflicts and trying to manage the following apparent conflicting images of

16 [Link]:indrum2018:174129
Q: i. a dense set; ii. an ordered set; iii. a set with the same cardinality as a discrete set,
like N.
A priori identification of concept images
The concept definitions relevant to our topic are the following: linear ordering;
enumeration; dense set; discrete set; denumerable set; cardinality; bijection;
isomorphism. Relying on literature review and historical-epistemological analysis, we
identified possible features of concept images that may cause cognitive conflicts:
CI1) Consecutive element is greater: generalizing an association that is typical of
N, reinforced by the spatial image of the oriented line (a greater element is on the
right as the consecutive element). Students may think that a consecutive element
in a list must be greater than the previous one (considering the standard order).
CI2) A dense set cannot be enumerated: Students might have a concept image of
ordered sets as sets in which the elements are “one after the other” on the line: an
enumeration must move from left to right consistently with the standard order.
CI3) Dense not discrete: Q might be said to be dense as opposed to N, which is
not. The difference is “shown” either on the line or with numerical examples as an
absolute difference (to have elements in “between” or not), independent of the
particular order. Density may thus be considered an absolute property of a set and
the “visual contrary” of discrete.
CI4) Linear or bidimensional representation: the cardinalities of Q and N are
shown to be equal, “re-ordering” Q and constructing a bijection between the two
sets. The ordering procedure is usually represented in two dimensions (the
“dovetail” counting method) while the standard order is represented using the line.
CI5) Bijection is identification: students may associate the term “bijection” with a
total identification between the structures (A=B), not just in terms of cardinality.
CI6) Finite versus infinite: we represent indeed finite quantities of corresponding
integer and rational numbers. This may lead the students to implicitly compare
images of finite subsets of N and Q (maybe in the same graphic representation of
intervals), concluding that N and Q cannot be composed by the same quantity of
elements, since “rationals are more than integers” (see Galilei, 1638; finite
reasoning applied to infinite sets, Tirosh & Tsamir, 1996).
METHODOLOGY AND PRELIMINARY DATA ANALYSIS
To answer our research question, we collected the following data: i. the Italian
student’s explanation of her doubt in written form; ii. answers to a similar
questionnaire by an Italian university teacher and Master degree students in France
preparing the selection procedure exam to become mathematics secondary teachers.
The questionnaire was written in English and then translated into Italian and French;
it is based on the historical-epistemological analysis and on the hypothesized student

17 [Link]:indrum2018:174129
concept images. We created different versions of the same questionnaire, modifying
them according to the two different countries’ curricula for high school and
university syllabi, and to the target of the questionnaire (university teachers or Master
students).
At the beginning of both questionnaires, we present a realistic didactical situation,
and we pose appropriate questions (see below).
We analyse one excerpt from a written interview to the Italian student, then we
comment on excerpts of the answers from an Italian university teacher. Finally, we
provide a brief summary of the answers to a version of the questionnaire submitted to
Master students in France. To carry out the analysis of the students’ interviews,
consistently with our research question, we searched for their concept images and
concept definitions before comparing the two source groups for connections and
potential conflicts: in the case of the students, we consider personal concept images,
while in the case of the teacher we look for examples that can reveal the hypothesized
student concept images and concept definitions.
Analysis of the first student’s comment (Italy)
The student who asked the first question explained her doubt as follows:
“The doubt arose when [1] looking at the schemata that is used to [2] find a bijection
between N and Q. [image] N is [3] ordered by definition, because, starting from 0, every
element has a consecutive element, while [4] Q is not, since it is dense. But what prevents
me from saying that 3/2 is the consecutive of 4/1? According to Peano’s axiomatization,
N is an [5] abstract structure that we can apply to natural numbers but also to Q, thanks to
the bijection. I could thus say: [6] there are infinite rational numbers between 1/3 and ½,
so there are infinite elements between the natural numbers associated with 1/3 and ½,
using the [7] bijection in the reverse way. Reflecting more deeply, I realize that the
problem is caused by the fact that [8] the bijection between N and Q is not “ordered” like
that between N and P (pair numbers): 2 < 3, 2/1 > ½. I still do not understand why Q is
dense and N is not, since we said that Peano’s axioms can be used for several models and
not just the natural numbers we already knew, but [9] if I enumerate Q with the natural
numbers, it no longer has any sense to say < or > in Q.”
The student was reasoning according to images (1,2) rather than definition and in
mentioning the definition of order (3) she said: “every element has a consecutive
element”, revealing how she is not really using a formal definition but an image of
ordering where each one is set after the other [CI2]; she was surprised that the
consecutive could not be greater [CI1], so much so as to claim that in Q the meaning
of < and > disappears. Also, she uses a representation of the bijection that she herself
mentions as an identification (5, 7), reasoning on the schemata; indeed, she said that
we can use it in both directions, identifying completely a couple of elements, one in
the first and one in the second set [CI5]. She also used the image of “infinite elements
in between” to say that N must be dense since we can image infinite elements

18 [Link]:indrum2018:174129
between the natural numbers associated to a couple of rational numbers [CI3]. The
student tried to connect different concept images and conflicts which emerged. The
conflict that is expressed in the question, as she said, is caused by the use of images
according to which [CI2] natural numbers and rational numbers are completely
identified by means of a bijection, resulting in contradictions. It is impossible for the
following to be simultaneously true: 1. Q is dense and N is not dense and N and Q are
“the same structure”; 2. in the identification, Q loses the property that the consecutive
element is greater, while N conserves this same property.
Analysis of one university teacher’s questionnaire (Italy)
For this instance of the questionnaire, we provided the following realistic situation:
during the lesson, a teacher is interrupted by a student asking the first question. The
university teachers were explained what students are expected to have been taught
before on density, infinite cardinality and the problem of “consecutive numbers” in
Q. We asked the teachers to interpret the students’ doubts and to propose how to deal
with said situation in the classroom. We report and comment some excerpts from the
answers to the questions (1 refers to Q1, and 2 refers to Q2):
1a. How would you answer the questions? How would you explain it to the whole class?
I. “The [1] order is not linked to the state of consecutive-ness: when she speaks of
consecutive numbers we are in the domain of the [2] induction principle, which is
only valid in N. In N we have much more: it is [3] well-ordered so there are no lower-
unlimited subsets. Once the properties of N had been observed, i.e. the [4]
“smallness” of the set that must satisfy all these features, I would move to the [5]
differences between N and Q. Finally, I would observe that everything results from
the fact that we are able to say that [6] one number is greater than another and this
definition is also valid in N, as N is a subset of Q. I would say that [7] we cannot
compare two properties linked to different definitions. Also, I would say that [8] in N
we can give a definition of order that is linked to the induction principle but NOT
generalized to bigger sets, but I would avoid going indepth into this issue, so as not
to create confusion for the weaker students. Here, [9] the problem is that the sets are
infinite”.
II. “To be [10] dense is very different from being denumerable, I would [11] remind
them of the definitions.”
2. Do you find a possible connection between the two questions, whereby these chance
episodes could be used to help deal with some important topics in mathematics? What
further examples or explanations would you propose to the student or what activities
would you design in order to deal with such a question?
“I would [12] show that {1/n} becomes more and more dense as it approaches 0. Then I
would point out that the student should not be surprised when one proves that there is a
[13] bijection between pair numbers and N, as well as with odds, the multiples of 3, the
prime numbers, the negative integers. This is to [14] see (and prove, showing the
application) examples of sets that are in a bijection with N.”

19 [Link]:indrum2018:174129
 3. Would you take the opportunity to explain something to students that could help them
in trying to answer these kinds of questions on their own in the future?
“What we learn from these questions is that science proceeds with [15] analogies and
differences, that we [16] must always consider the definitions, and that it’s important that
these are very precise.”
On several occasions the teacher mentions definitions (6, 7, 8, 10, 11, 16), with
different goals: we can’t compare properties linked to different definitions; to be
dense and to be denumerable are not linked because they concern different definitions
(6, 7, 8, 10); we must use the definitions (11, 16).
When the teacher proposes examples, though, he uses words like show and see, and
he uses images that may cause conflicts: he refers to “greater” and “smaller” related
to infinite sets [CI5] and uses cardinality to compare N and Q. He mentions the
bijection between N and its subsets and Q by showing the application, identifying it
thus without stressing the issue of ordering [CI5]. In one case [CI2], the ordering is
consistent with the linear order and in the other not, but it is not stressed. He shows
that {1/n} becomes increasingly denser near 0, encouraging the use of images of
density [CI3, CI5]. If we think about the student’s comments, these answers would
not have clarified the point she was “struggling” with: what she thought of as
identification did not identify N and Q exactly. He mentions the definitions but, in the
examples, he uses images and never seems to connect the definition to the images.
Preliminary analysis of the data collected in France
In France, an adapted version of the questionnaire was submitted to 30 first-year
Master’s students on October 12th, 2017, in the first half-hour of a teaching session on
didactic and epistemology of mathematics. No epistemological or didactic work on
this topic had been done before with these students. The answers were then discussed
later in the fall as a starting point in a session devoted to epistemological and
didactical aspects of numbers construction. For both Q1 and Q2, we asked students
(in French): 1. Have you ever asked yourself this question? If so, in what context and
how did you answer this question yourself? 2. Imagine that a student of a lyceum or
of a preparatory class for the “grandes écoles” is asking you this question. How
would you answer? Finally, the last question was: Do you find a possible connection
between the two questions, whereby these chance episodes could be used to help deal
with some important topics in mathematics? The questionnaire was anonymous. The
students were asked to indicate their previous university studies.
To the first question, 6 students answered “yes” and commented on their answer; 8
answered “no” and commented on their answers; 16 answered “no” without
comments. Some Master’s students claimed that the notion of density was still
unclear for them. In answering the second question, some of these students proposed

20 [Link]:indrum2018:174129
an incorrect explanation for hypothetical younger students, which relied on concept
images of successor without any reference to definition:
M1-18 “the successor of an integer exists, so in Q there are elements having successors.
Thus, we can say that the successor of a rational does not ‘always’ exist rather than ‘does
not exist’”.
Several students explained that thanks to the definition of density, it is possible to
define an order, while in order to define density-in-itself, it is necessary to already
have an order, as in the example below:
M1-15 - in first year university when the set theory was introduced - I told myself that in
a dense set like Q, the notion of successor as may be imagined on N [1] does not exist,
but thanks to the definition of density [2] of a set, it was possible to define an ordering [3]
of this set and consequently order this set [4]
The student begins with reference to an image [1], then refers to the definition [2],
and concludes with the possibility to define an ordering [3], which is an inversion of
the definitions between density and order. It is noticeable that the student makes a
distinction between “ordering” and “order”, while there is no reference to the already-
known order of Q. Among the 30 answers, only one student relies on the existence of
the standard order in Q to justify that it is possible to define an order on Q. This brief
summary of the students’ answers accounts for the weakness of their knowledge of
the concept of density, and of its link with the concept of order.
Insights into the Italian Curricula and traditional didactical practices
In the Italian high school curricula from grade 9 to 13, order and density are never
mentioned explicitly; teachers are, however, advised to introduce the concept of
infinite, showing the connection between mathematics and philosophy, in grades 11
or 12 while introducing transcendental numbers. Natural, integer and rational
numbers are mentioned, but attention is focused on computation techniques,
representations of numbers (fractions, decimal numbers, points of a line) and
approximation. In the curriculum for primary school, both “sequential” and
“cardinal” sense of numbers appears. In middle and high school, students are taught
that between two rational numbers you can always find a rational number, and that it
follows that a rational number has no consecutive element. What is generally not
made explicit by high school teachers is that this is not an absolute property of Q, but
depends on the order chosen in Q. Also, numbers are usually represented in Italy on a
number line, so the discrete and the dense are distinguished using more visual than
theoretical considerations (the existence or not of “something in between”). Discrete
is often counter-posed to continuous. At the end of high school and/or in the first year
of university, the concept of accumulation point is introduced for dealing with limits
and discontinuity of real functions with real variables. The existence of an infinite
quantity of real numbers “between” two real numbers is said to be due to the density

21 [Link]:indrum2018:174129
property of R. This kind of practice is likely to reinforce the concept images we
identified and to reinforce the habit of reasoning in absolute terms while referring
implicitly to the standard order, without stressing the dependence of the properties on
the choice of order relation.
The French context
In France, the situation is not very different. Durand-Guerrier (2016, p.341) presented
it briefly, and provides evidence of the weakness of fresh university students’
knowledge about numbers, which could be related to the curriculum. Briefly, high
school students deal with approximations, mainly with the use of calculators. In grade
12, they learn the mean value theorem for derivatives without a proof, and without
discussion about the fact that this theorem holds in the set of real numbers but no
longer applies in the sets of decimal or rational numbers. Consequently, students
beginning university generally have no idea of the differences and interplay between
finite decimal numbers, rational numbers and non-terminating decimal expansions,
and thus are not prepared for what they will be taught at university. Indeed, in many
French universities, in first-year mathematical courses, an axiomatic definition of the
set of real numbers is given, most often via “the supremum property”, without any
explicit construction. In some cases, the representation of real numbers as non-
terminating decimal expansions and the corresponding characterization of the type of
numbers are introduced, and improper expansions such as 0.9 are discussed with
students (Durand-Guerrier, 2016, p.341). A topological course is generally offered,
but it is mainly theoretical, and students have very few opportunities to connect the
theoretical concepts with their interpretation in the ordered field of real numbers.
CONCLUSIONS AND DEVELOPMENTS
A first relevant result is that the framework is suitable to interpret our data. Our
epistemological investigation and empirical data analysis do indeed help to formulate
interpretations of the conflicts which appeared in the first question, as confirmed by
the Italian student’s interview analysis. Also, we observed a total identification
between structures due to the constructions of bijections between elements of the sets,
which is implicitly present in the high school practices. University teachers
mentioned merely definitions, but, as Tall & Vinner (1981) showed in the case of
discontinuity of functions, concept definition may be largely inactive in the cognitive
structure and concept images may be used instead of the definition in order to grasp
better its meaning. In this case, in the first question (that we analysed in depth here),
the conflict emerged at the level of concept images, so definitions would not have
been sufficient to solve the students’ doubts. For the French Master students who
answered our questionnaire, this lack of awareness of the links between the concepts
of density-in-itself and order of Q might prevent them from designing appropriate
learning situations, once they pass the selection procedure exam and become
teachers. We hypothesise that, even if such questions emerge in university courses,

22 [Link]:indrum2018:174129
the reasoning and consequent conflicts can be due to the lack of explicit reference to
the dependence of Q properties on the order relation which still exists in the high
school, a lack which calls for epistemological and didactic clarification in teacher
training. As developments, we consider it crucial to identify the curricula issues
where non-recognition could generate such conflicts, and to look for suitable teaching
strategies in high school and university to deal appropriately with these concepts.
“When the teacher is aware of the possible concept images, it may be possible to
bring incorrect images to the surface and, by discussion, rationalise the problem”
(Tall & Vinner, 1981, p. 17).
REFERENCES
Bergé, A. (2010). Students’ perceptions of the completeness property of the set of
real numbers. International Journal of Mathematical Education in Science and
Technology, 41(2), 217–227.
Branchetti, L. (2016). Teaching real numbers in the high school: an onto-semiotic
approach to the investigation and evaluation of the teachers' declared choices.
Ph.D. dissertation, Università degli studi di Palermo.
Durand-Guerrier V. (2016) Conceptualization of the continuum an educational
challenge for undergraduate students. International Journal of Research in
Undergraduate Mathematics Education, 2, 338–361.
Galilei, G. (1638). Dialogues concerning Two New Sciences, English translation by
Crew, H., de Salvio, A. (1914). New York: Mac Millan. Available online at URL:
[Link]
Jech, T. (2003) Set theory. The third millennium edition, revised and expanded.
Springer Monographs in Mathematics.
Krantz, S. G. (1999). Discrete sets and isolated points. Handbook of Complex
Variables. Boston, MA: Birkhauser.
Lolli, G. (2013). Nascita di un'idea matematica. La teoria degli insiemi da Cantor a
Zermelo. Pisa: Edizioni della Normale.
Peano, G. (1889). Arithmetics principia, nova methodo exposita [The principles of
arithmetic, presented by a new method]. Bocca: Torino.
Tall, D., Vinner, S. (1981). Concept images and concept definition in mathematics
with particular reference to limits and continuity. Educational Studies in
Mathematics, 12, 151–169.
Tirosh, D., Tsamir, P. (1996). The role of representations in students’ intuitive
thinking about infinity. Journal of Mathematical Education in Science and
Technology, 27 (1), 33–40.

23 [Link]:indrum2018:174129
 An empirical study of the understanding of formal propositions
about sequences, with a focus on infinite limits

Renaud Chorlay
ESPE de Paris, Uni. Paris-Sorbonne, France & LDAR (EA4434)
[Link]@[Link]

In this paper, we analyze the answers of one group of high-school students and
two groups of first-year University students to a questionnaire designed to test
their level of recognition and understanding of the formal definition of the
concept of infinite limit. Although this empirical study is ancillary to a larger
project centered on didactic engineering, its analysis sheds light on the key issue
of the logical prerequisites for the learning of the fundamental concepts of
analysis. It also provides a new tool to investigate students’ concept-image of
limits, and assess the impact of teaching contexts and teaching paths.
Keywords: Teaching and learning of analysis and calculus, teaching and
learning of logic, reasoning and proof, definitions, limits.
CONTEXT AND RATIONALE
At the INDRUM 2016 conference, Cécile Ouvrier-Buffet and Renaud Chorlay
presented a poster outlining a medium-scale project on definitions in analysis
(Chorlay & Ouvrier-Buffet, 2016), with a focus on the formal definition of the
limit of a numerical sequence. This topic lied at the intersection of the research
interests of the two researchers: Cécile Ouvrier-Buffet is a maths-education
researcher with a strong epistemological background, whose work bears mainly
on definitions, their use, and the conditions for their genesis in teaching-contexts
(Ouvrier-Buffet 2011). Since most of her former work bore on discrete
mathematics, she wanted to investigate the extent to which the theoretical tools
she had developed in this context had to be adapted to deal with a teaching
context with very different mathematical (continuous vs discrete) and didactical
(transition from calculus to analysis) features. Renaud Chorlay is a historian of
mathematics and teacher educator with a long-standing interest in the history
(Chorlay, 2011) and didactics of analysis.
We selected the topic of limits because we felt many years of didactical
investigations had made it a mature topic; a topic about which knowledge has
accumulated to form a sound and coherent body of knowledge. Indeed, we know
a lot about limits in terms of conceptions and misconceptions (Robert, 1982);
also in terms of obstacles (Sierpinska, 1985). As far as the genesis or
rediscovery of the (or a) definition is concerned, many attempts have been made
and reported upon in details, whether in the framework of didactic engineering 1
(Robert, 1983) (Bloch & Gibel, 2011) or with other research tool-boxes
(Mamona-Downs, 2001) (Przenioslo, 2005) (Swinyard, 2011) (Lecorre, 2016)

24 [Link]:indrum2018:170161
(Roh & Lee, 2017). The tricky logical aspects were studied, in particular, in
(Arsac & Durand-Guerrier, 2005).
On this solid basis, our work on the genesis and use of definitions has so far
been engaged along three different lines of investigation; we will distinguish
between ex-ante studies – before students’ first encounter with formal
definitions of limits – and ex-post studies.
 Ex-ante 1: For year 12 (final year of secondary education), the French
curriculum requires that students majoring in mathematics and the
sciences study a definition of limits (finite or infinite) of numerical
sequences. Students are not really expected to use this definition on their
own; rather, the teacher is expected to use these definitions on a few
occasions, to show that some properties of limits can actually be proved
mathematically (in particular: any unbounded and increasing sequence
tends to +). The underlying idea is that early encounter with a few
rigorous definitions and proofs should ease the transition between high-
school calculus – with its combination of algorithmic procedures and
graphical intuition – and university analysis. This classroom work on the
formal definition of limits is connected to another requirement of the
current curriculum, namely that throughout high-school, the basic notions
and the standard notations of mathematical logic be gradually made
explicit. In this context, the discovery of a definition for limit, with its
specific sequence of nested quantifiers, is supposed to be the culmination
of this gradual process. In 2016, one of us (Chorlay) designed a teaching-
session in the spirit of didactic engineering, for students to gradually
formulate a formal definition of the infinite limit. We will report on this in
detail in another context.
 Ex-post 1: in 2015-2016 we studied how – if at all – prospective maths-
teachers made use of the definition of limits in order to identify and
analyze vague, informal or erroneous statements regarding limits. We
reported on this in a poster presented at the INDRUM 2016 conference.
 Ex-post 2: in 2016-2017 we designed a questionnaire in order to assess the
level of recognition and understanding of the formal definition of the
infinite limit. This questionnaire, and the answers collected with three
groups of students are be the topic of this paper.

QUESTIONNAIRE - DATA COLLECTION

The questionnaire was of the True/False type, divided in two parts. We give
below an English translation, along with indications on the correct answers.
Part I. For each one of the implications below, circle either “True” or “False”. If
you circle “False”, justify your answer.

25 [Link]:indrum2018:170161
#1 If lim un = + then  A  R  nA  N such that A
T-F Justification (if “False”):
#2 If lim un = + then  A  R  n  N un  A
T-F
#3 If lim un = + then  A  R  nA  N such that A
T-F
#4 If lim un = + then  A  R  n  N , un  A
T-F
#5 If lim un = + then AR  nA  N such that for any integer
n greater than nA A
T-F
Correct answers:
#1 True: Here the consequent means “not bounded above”.
#2 False: Here the consequent is a property which never holds; hence the
implication is always invalid.
#3 True: Here the consequent is always valid, hence the implication is always
valid.
#4 True: Here the consequent means “bounded below”.
#5 True: Here the consequent is the definition, worded semi-formally.
Part II. The four implications below are taken from part I. For each one of them,
first state its converse, then circle “True” or “False” regarding the converse.
Justify if “False”.
#1 If lim un = + then  A  R  nA  N such that A
Converse :
T-F Justification (if “False”):
#3 If lim un = + then  A  R  nA  N such that A
Converse :
T-F
#4 If lim un = + then  A  R  n  N , un  A
Converse :
T-F
#5 If lim un = + then AR  nA  N such that for any integer
n greater than nA A
Converse:
T-F

Conv. of #1 False: standard counter-examples are , …

C of #3 False: the antecedent being always true while the consequent can be
false, the implication is invalid.
C of #4 False: being bounded below does not imply lim = +.

26 [Link]:indrum2018:170161
C of #5 True: definition.
The specific form of the questionnaire derives from its original intended use. It
was first designed to assess the didactic engineering, which focused on the
formal definition of the infinite limit. Other forms of assessment of the ability to
recognize, and of the level of understanding of the formal definition were ruled
out, in particular interviews (as in (Robert, 1982)) or proof-writing (as in (Roh
& Lee, 2017)). We felt this questionnaire would give us feedback regarding two
key features of the engineering, namely (1) the role of logic, hence the flood of
formulae with nested quantifiers in this questionnaire; (2) the fact that “not
bounded above” is a necessary condition for lim un = + but not a sufficient
condition, hence the importance of item #1 and its converse.
We did not ask for justifications when the item was deemed “True” by the
students, mainly to save time and keep the questionnaire feasible in about 20
minutes. In addition, the justificatory task for True statements could vary a lot
across teaching-contexts and would not easily lend itself to comparison. For
instance, considering item #4 (if lim un = + then the sequence is bounded
below): in some contexts citing a theorem studied in class would suffice whereas
in other contexts students would have to devise and write a non trivial proof. We
also chose to drop the converse of item #2, since the fact that an implication
whose antecedent is False is considered valid is a purely logical matter.
In the spring of 2017, the questionnaire was administered to three groups of
students: Group 1 is one of the two French Year-12 classes which had
experienced the engineering; Group 2 and 3 are first-year university students in
Mons University (Belgium), with high-achieving maths majors in Group 2 and
medium-achieving2 computer science majors in Group 3. In all three cases, the
questionnaire was given several months after the course on limits had been
taught, and students had not been asked to revise anything in particular. They
were told the questionnaire was given for research purposes, and would not be
graded. They were given between 20 and 30 minutes. The number of students
was: 31 (group 1), 50 (group 2), and 17 (group 3).
We originally hoped a comparison between the three groups would enable us to
study the effects of three teaching units: our engineering (group 1), a “standard”
maths-lecturer course (group 2), and Robert’s engineering (group 3, as reported
upon in (Bridoux, 2016)). Unfortunately, we were not able to do that, since other
factors seemed to have had a more significant impact.
FINDINGS
Result #1
A first result is that this questionnaire is not unfeasible. In group 2, 14 of the 50
questionnaires were answered perfectly correctly, with relevant counter-
examples for the False statements. Some of these counter-examples had been

27 [Link]:indrum2018:170161
studied in class (such as for the converse of #1); in these cases,
students managed to interpret “ A  R  nA  N  A” as “not bounded
above” and selected a relevant counter-example in a memorized repertoire. In
other cases, counter-examples had not been studied in the course on limits –
because they had nothing to do with limits – and students crafted ad-hoc
counter-examples, displaying some command of logic (for instance, to prove
that the negation of “ A  R  nA  N  A” always holds).
Result #2
A second set of results sheds light on the role of an explicit teaching of logic.
When we collected the data we first engaged in quantitative analysis, and were
pretty unhappy about the following result: in group 1 (our engineering), only
26% of the students considered #4 to be “True”, compared to 86% in group 2
and 71% in group 3. A closer look at the answers showed that in group 1, a
significant number of students had actually engaged in another task than the
prescribed task. In Fig. 1 and 2 we translated extracts of answer-sheets from
group 1:

Figure 1. Student 29 of group 1

Figure 2. Student 3 of group 1

In these answer-sheets, the students did not engage in an assessment of the
logical implications but in a comparison between the formal statements given as
consequents (in part I) and the definition of lim un = +. In these examples the
comparison can be clumsy (as for #2 for student 29, or the “and  A” for student
3). Nevertheless, it rests on the fact that the definition is known (correct answers
for #5 and its converse), and is seen as the relevant template against which other
quantified formulae ought to be contrasted. Moreover, the comparison is not
purely syntactical: in her assessment of #4, student 29 did not only spot that “

28 [Link]:indrum2018:170161
A  R  n  N , un  A” is not the definition, but also elicited in her own
words why it could not be, namely “ n  N un  A” does not capture “beyond
a certain rank”, which is a key element of the definition. The reinterpretation of
the prescribed task is typical of at least one third of the questionnaires from
group 1.
By contrast, only one of the 67 students from Mons University reinterpreted the
implication-assessment task as a comparison-with-the-definition task. A key
difference between group 1, on the one hand, and groups 2 and 3, on the other
hand, is that at Mons University students had studied logic in the first term,
whereas the French high-school students had only occasionally been exposed to
logic. The French students were familiar with the notion of converse, and had
some knowledge of the meaning of quantifiers  and , but were not familiar
with sequences of quantifiers; much less with the negation of such sequences.
These formal aspects were not problematic for a large majority of the Mons
students. This does not mean that all the logical aspects were mastered by the
Mons students. In particular, when it came to proving that some formal
statement was valid, many answer-sheets showed misconceptions regarding the
use of  and .
This sheds some light on the standard but thorny issue of prerequisites: since the
formal definition of limits involves a sequence of nested quantifiers, how much
logic should be taught (either beforehand or along the way) for students to be
able to do anything with it? Our results suggest that the answer depends on how
“do” something with a definition is construed. Using the formal definition to
design and write proofs probably requires some know-how regarding the
interpretation of hitherto unknown sequences of quantifiers, and the negation of
such sequences; for a significant proportion of the French student, their
occasional and in-context encounters with logical notations did not allow them
to acquire such know-how. However, if “do” is taken to mean “remember the
definition” and even “understand the definition”, then for a large majority of the
French students, their command of logic was adequate. For instance, we
consider the work of student 29 of group 1 (fig.1) to display some degree of
conceptual understanding of definition, namely some understanding of the
specific role of each of the three quantifiers. Student 3 is clearly able to interpret
“ A  R  n  N un  A”. This understanding does not rest on a general
ability to make sense of and formally manipulate logical formulae, but is limited
to the context of the definition of limits. Since it relies on the specific
connections between the concept-image and concept-definition of “limit”
targeted (and, apparently, stabilized) in the didactically engineered teaching-
session, this understanding is probably not only context-dependent but also path-
dependent.

29 [Link]:indrum2018:170161
Result #3.
In the a priori analysis for the engineering, we studied the relations between
three mathematical properties of numerical sequences:
(1) lim un = +
un) is not bounded above;
(3) un) is strictly increasing, at least from a certain rank onward.
Our hypothesis was that properties (2) and (3) were part of the concept-image of
(1) for most students; of a concept-image2 in which all three properties are
considered to “go together”, without any specific and explicit logical
connections being part of the cognitive structure. This hypothesis was based on
the didactical literature (Robert 1982) (Mamona-Downs 2001) (Swinyard 2011),
and was perfectly confirmed during the two implementations of the engineering.
For this reason, our design aimed for conceptual differentiation, to be achieved
first through the study a few well-chosen sequences, and then through the formal
explicitation of the logical connections between (1), (2), and (3). Consequently,
we wanted our post-experiment questionnaire to help us assess to what extent
students knew that (1)  (2) is valid, while its converse is not.
Due to the significant level of reinterpretation of the prescribed task in group 1,
the data gathered do not easily lend themselves to quantitative comparison.
However, the fact that “not bounded above” (2) is a key component of the
concept image of lim un = + (1) is again confirmed beyond doubt. Let us first
compare groups 2 and 3. In group 3 – the medium-achieving computer science
majors – all 17 students deemed the converse of #1 to be True. Leaving out 3
students whose answer-sheets show an inadequate command of the logical
aspects, it seems that Aline Robert’s engineering (which targeted the definition
of finite limits) had no impact on the belief that if a sequence (un) takes on
arbitrarily large values, then lim un = +. In group 2, that of high-achieving
maths majors, the results were not as striking; they were telling just as well.
Among the 50 answer-sheets, let us focus on the subpopulation of those for
which all the answers to part I were correct (including relevant counter-
examples for #2), and all the converses were stated correctly. Among these 33
students, 17 deemed the converse of #1 to be False – which is the correct answer
– and all but one provided a relevant counter-example (usually
which – as the lecturer confirmed – had been studied in detail). Student 25 even
wrote: “ A  R  nA  N such that  A means that the sequence is not
bounded above, but it doesn’t mean it tends to +, it may oscillate. Let’s
consider (…)”. However, the other 16 students ticked “True” for the
converse of #1. The resistance of this belief, even among students with a
reasonable command of logic, who know the definition of lim un = +(item #5
and its converse), and who had been exposed to a teaching which explicitly
tackled this issue suggest that the conflation of (1) and (2) is an epistemological

30 [Link]:indrum2018:170161
obstacle (Chorlay & de Hosson, 2016). It is probably not independent from the
belief that all sequences are monotonous, at least after a certain rank (Robert
1982), but our questionnaire offers no new insight as to this.
This confirms – in hindsight – that we were justified to take (2) into account
when designing a teaching-session on the formal definition of (1). However, it
does not tell us whether targeting the formulation of the definition of (1) through
a process fostering the conceptual differentiation between (1) and (2) was
didactically relevant – as standard constructivist tenets suggest – or just
foolhardy.
The results of group 1 allow us to be cautiously optimistic. From a purely
quantitative viewpoint, 58% of the students of group 1 deemed the converse of
#1 to be “False” – which is the correct answer – but no conclusions can be
drawn from this fact beyond that this 58% stands in sharp contrast with the 0%
of “False” on the subpopulation of OK-answer-sheets of group 3. In group 1, for
instance, the third of the students who clearly reinterpreted the task as “compare
with the definition” ticked “False”, but this does not indicate that they are aware
of the connections between properties (1) and (2), or that they were able to
reformulate “ A  R  nA  N such that  A” as “not bounded
above”. Answer-sheet 30 of group 1 shows, again, that some conceptual
understanding can be achieved in a formal context in spite of a poor level of
command of symbolic logic. This student systematically stated BA as
converse of AB; hence one has to study her assessment of the converse of #4
– instead of #1 – to see if she mistakes (2) for (1); which she does not, actually.
Of the 31 students of group 1, only two interpreted the task correctly and
provided relevant correct answers for the converse of #1, either with a formulaic
counter-example or with a graphical counter-example (of the
type). However, about one fourth of the students deemed the converse of
#1 to be false, interpreted the task as “assess the implications” and provided
arguments which we could be indicative of some conceptual understanding. In
these cases, they justified their assessment not by displaying a counter-example,
but by explaining why the antecedent was not strong enough to warrant the
consequent: under the hypothesis “ A  R  nA  N such that  A”,
the sequence can oscillate; or: the antecedent does not imply that the sequence is
increasing. Our empirical data does not enable us to tell which of the following
is the case: either, students argue on the basis of the fact that if a sequence is
increasing and not-bounded above then it tends toward + (a theorem they are
familiar with); or, students conflate (1) and (3).
CONCLUSIONS AND RESEARCH PERSPECTIVES
While the questionnaire studied in this paper was originally designed to compare
the effectiveness of three teaching-modules on the definition of limits of
sequences, it turned out that they could not serve that purpose due to the

31 [Link]:indrum2018:170161
decisive impact of another factor, namely the level of familiarity with predicate
calculus – both in terms of syntactic command, and in terms of ability to make
sense of logical formulae involving nested quantifiers. Nevertheless, we claim
that meaningful conclusions or insights can be gained from the analysis of our
empirical results.
For students with some command of logic – a command which cannot be gained
through an occasional and in-context use of logical formalism – this
questionnaire does provide insight into the connections between concept-image
and concept-definition for limits, thus providing a new investigative tool to
study this issue; a tool which does not involve conducting interviews or studying
students ability to use the definition in proofs. As far as students are concerned,
the comparison between group 2 and group 3 suggest that not all teachings on
limits are equivalent in this respect; the case of group 2 shows that – under
circumstances which call for further investigation – first-year university students
can display a reasonable command of the concept of limit.
As far as group 1 is concerned, the result show that the prerequisites in logic
may not be as high as one might expect, if what is targeted is the ability to
memorize the formal definition, and the ability to display understanding of some
key features of the concept. As far as our didactic engineering is concerned,
these results show that (1) it was not a complete failure, (2) some of its guiding
principles – such as the importance of the conceptual differentiation between
infinite-limit and not-bounded-above, or the use of logical formalism – seem
relevant. However, in this context, this questionnaire is probably not the best
tool for a fine-grained assessment of what the actual impact of this engineering
is.
1. For introduction to didactic engineering as task-design oriented research method, see (Bosch & Barquero
2015).

2. This assessment of the overall level of the groups is that of the team of maths lecturers at Mons University, as
communicated to us by Stéphanie Bridoux, who is both a member of that team and a mathematics education
researcher (LDAR). Many thanks to her for her collaboration on this project.

3. D. Tall and S. Vinner introduced the distinction between the image and the definition of a concept to stress the
difference between mathematics as a mental activity and as a formal system. “We shall use the term concept
image to describe the total cognitive structure that is associated with the concept, which includes all the mental
pictures and associated properties and processes. (…) it needs not be coherent (…).” (Quoted in (Tall 1991, 7)).

REFERENCES
Arsac, G., & Durand-Guerrier, V. (2005). An epistemological and didactic study
of a specific calculus reasoning rule. ESM, 60, 149-172.
Bloch, I., & Gibel, P. (2011). Un modèle d’analyse des raisonnements dans les
situations didactiques. Etude de niveaux de preuves dans une situation
d’enseignement de la notion de limite. RDM, 31(2), 191-228.

32 [Link]:indrum2018:170161
Bosch, M., & Barquero, B. (2015). Didactic Engineering as a Research
Methodology: From Fundamental Situations to Study and Research Paths. In
A. Watson, M. Ohtani (eds.) Task Design in Mathematics Education – an
ICMI Study 22. Springer International Publishing.
Bridoux, S. (2016). Introduire la notion de convergence avec une ingénierie des
années 1980 : rêve ou réalité didactique pour l’enseignant ? In E. Nardi, C.
Winslow, T. Hausberger (Eds.), Proceedings of INDRUM 2016 (pp.53-62).
Montpellier: Université de Montpellier.
Chorlay, R. (2011). “Local – Global”: The First Twenty Years. Archive for
History of Exact Sciences 65(1), 1-66.
Chorlay, R. & Ouvrier-Buffet, C. (2016). Comparaison de schémas de genèses
didactiques de définitions, le cas de la limite d’une suite. In E. Nardi, C.
Winslow, T. Hausberger (Eds.), Proceedings of INDRUM 2016 (pp.175-176).
Montpellier: Université de Montpellier.
Lecorre, T. (2016). Rationality and concept of limit. In E. Nardi, C. Winslow, T.
Hausberger (Eds.), Proceedings of INDRUM 2016 (pp.83-92). Montpellier:
Université de Montpellier.
Mamona-Downs, J. (2001). Letting the intuitive bear of the formal; A didactical
approach for the understanding of the limit of a sequence. ESM 48, 259-288.
Ouvrier-Buffet, C. (2011). A mathematical experience involving defining
processes: in-action definitions and zero-definitions. Educational Studies in
Mathematics, 76(2), 165-182.
Przenioslo, M. (2005). Introducing the concept of convergence of a sequence in
secondary school. ESM 60, 71-93.
Robert, A. (1982). L’acquisition de la notion de convergence des suites dans
l’enseignement supérieur. RDM, 3(3), 307-341.
Robert, A. (1983). L’enseignement de la notion de convergence des suites
numériques en DEUG. Bulletin de l’A.P.M.E.P., 340, 431-449.
Roh, K.H., & Lee, Y.H. (2017). Designing tasks of introductory real analysis to
bridge the gap between students’ intuition and mathematical rigor: the case of
the convergence of a sequence. IJRUME, 3(1), 34-68.
Sierpinska, A. (1985). Obstacles épistémologiques relatifs à la notion de limite.
RDM, 6(1), 5-67.
Swinyard, C. (2011). Reinventing the formal definition of limit: The case of
Amy and Mike. Journal of Mathematical Behavior, 30, 93-114.
Tall, D. (1991). The Psychology of Advanced Mathematical Thinking. In D.
Tall (ed.) Advanced Mathematical Thinking. Dordrecht: Kluwer.

33 [Link]:indrum2018:170161
 Students’ understanding of ɛ-statements involving equality and limit
Imène Ghedamsi1 and Thomas Lecorre2
1
University of Tunis, Tunisia, ighedamsi@[Link]; 2Cergy-Pontoise University,
Paris, France, [Link]@[Link]
The purpose of this paper is to explore how students may understand the link between
the formalisations through ɛ-statements of infinite processes in the Archimedean
continuum. These processes illustrate either equality or limit. In particular, we focus
on the extent to which students perceive the formalisation of the infinite closeness
notion in the two processes. The data is collected from an extensive design research
carried out at the transition between Calculus course and Analysis course. TDS
construct of milieu is deployed to build and to analyse exploratory teaching-
experiments. The results put forward how ɛ-statements may assist students to
reconsider their informal understanding of limit.
Keywords: ɛ-statement, process, limit, equality, milieu.
INTRODUCTION
In the transition between Calculus and Analysis courses, formal definition of the limit
is needed not only to establish precise definitions of fundamental notions such as
differential, integral, and series, but more importantly, to build up and use formal
statements for making formal proofs. Yet, the key question of how to create a rigorous
understanding of infinite processes and initiate the use of formal statements remains a
challenging issue for researchers in the field of Calculus education.
Considerable research has been conducted on students’ difficulties to encapsulate the
infinite processes of limit into the formal limit (Tall & Vinner, 1981; Przenioslo, 2004;
Roh, 2008; Oehrtman, 2009). Most of this research highlights the impact of students’
previous use of informal statements to represent infinite processes both graphically and
numerically. Those statements usually involve expressions related to successive
computation of terms and closeness such as: the more is x close to infinity, the more is
f(x) close to l, and inversely. Several other studies have explored the complex structure
of the formal limit and have shown multiple aspects that may not help students develop
efficient interpretations of formal statements (Cottrill et al., 1996; Durand-Guerrier &
Arsac, 2005; Mamona-Downs, 2001; Roh, 2010; Oehrtman et al., 2014). Those aspects
fundamentally refer to the role of quantifiers and their order, the arbitrariness of ɛ and
its relation to the other parameter, and the connection between the statements
expressing changes in the variables. Some other research have designed tasks to assist
students connecting informal and formal statements related to limit (for an overview
of the concerned literature, see Bressoud et al, 2016). Specifically, Swinyard (2011)
has demonstrated that students are able to reinvent limit using formal statements.
Drawing on this study and on the genetic decomposition of limit of Cottrill et al. (1996),
Swinyard & Larsen (2012) develop a six steps model of how students come to
understand the formal definition of limit. This model provides consistent arguments of
how students reason about two infinite processes: 1) the process of finding limit which

34 [Link]:indrum2018:174484
is associated to the first three steps (as x gets closer to a, f (x) gets closer to L), and 2)
the process of validating limit which is described by the formal limit and encapsulated
in the last step through the formalisation of the infinite closeness notion (gets closer
to) via the concept of arbitrary closeness. Swinyard & Larsen hypothesized that limit
at infinity may assist students to focus first on variation of the dependent variable and
to shift to the validating process. In addition, the focus on the variation of a single
variable may improve students’ reasoning on the infinite closeness notion. Although
we agree with those hypotheses, the empirical data does not outline how students
connect informal statements of the first process to the formal statement of the second
process. These processes encompass the cornerstone notion of infinite closeness, so
why students do not feel the need to formalise the finding process and to emphasize its
difference with the validating process and by the way, to understand why quantifiers
should be described in such a way? However, Swinyard & Larsen call for research that
could investigate students’ formalisation of infinite processes in the context of whole-
class teaching experiments and beyond the context of reinvention (p.492).
In this paper, we focus on the formalisation of two infinite processes in the
Archimedean continuum by using formal statements that we call ɛ-statements. These
processes involve the formalisation of infinite closeness notion and illustrate either
equality or limit of function at infinity1. The research of Swinyard & Larsen has served
to structure our thought and to rigorously address our central question: to what extent
the formalizations of the infinite closeness involved in these two processes and their
link may assist students’ understanding of formal limit? The aim of this paper is to give
some insights on the potency of this link in way that somewhat guarantee students’
making sense of formal limit beyond restrictive contexts; this is why we deploy the
Theory of Didactic Situations (TDS) constructs to conduct exploratory whole-class
teaching experiments in the transition between Calculus and Analysis courses.
THEORETICAL FRAME
The TDS is a model of learning mathematical notions founded on an optimization of
the interactions taking place within the system of relationships between students, a
teacher, and a mathematical milieu which includes mathematical knowledge
(preconstructed tasks, tools, graphs, symbols, etc.), students’ prior knowledge, and
students’ informal understanding. The situation refers to the actual implementation in
a classroom of this ideal model (noted Situation with capital S) in accordance with a
targeted mathematical notion. The students’ work and the teacher management are
modelized at several levels according to the nature of the milieu. The expected
interactions are materialized through the role of both students and teacher specifically
in three particular levels: milieu for action, milieu for formulation and milieu for
validation. The efficiency of the interactions among peers is ensured by teacher’s
enrichment of the mathematical milieu. Depending on the complexity of the targeted

1
In this paper, the formal statement related to limit of function at infinity is: The limit of a function f is L at infinity if for
all ε > 0, there exists A > 0 such that for all x > A, L - ε < f(x) < L + ε.

35 [Link]:indrum2018:174484
notion, the teacher may ask questions and provide some others auxiliary mathematical
knowledge without minimizing students’ responsibility in producing knowledge (this
is the case for example of the formal limit). As mentioned by González-Martin et al.
(2014): "It is important to stress that the central object of TDS is not the cognizing
individual, but the Situation, which shapes and constrain the adaptive processes
students can develop, and thus the mathematical knowing which can be constructed."
(p.118). However, the robustness of a Situation depends fundamentally on the
mathematical milieu. The elaboration of this milieu is based upon a consistent
epistemological analysis of the targeted notion; this analysis should allow students to
experiment motivating questions and to reconsider their informal understanding – test
and make conjectures, provide examples and non-examples, and refute formulations.
The aim of this research is to design situations to explore how students understand the
formalization of two infinite processes that are strongly connected to the natural root
of limit idea, and the extent to which the link between those processes may provide
some insights on the formal limit. The starting point for the building of the
mathematical milieu is the fundamental historical idea of validating equalities by
means of infinite processes. The use of those processes provides results (for example,
the area of parabolic segments, the sum of infinite geometric progressions, etc.) that
would now be dealt with by means of limits and initiates the shift towards the formal
limit. If we look to the nowadays structure of equality: a = b if for all ε > 0, -ε < a – b
< ε, we may notice that the link between this structure and formal limit is modest. But,
this equality can be applied to a function f using the property P: There exists A >0, for
all ε>0 such that for all x > A, L - ε < f(x) < L + ε. If f verifies P then for all x > A, f(x)
= L. However, if we exchange the quantifiers in P, we obtain Q: For all ε>0, there
exists A >0 such that for all x > A, L - ε < f(x) < L + ε. Yet, if f verifies Q then the limit
of f in plus infinity is L. In this Situation, we focus on the role of ε in ε-statements: it
leads "at most" to equality and "at least" to limit depending whether the involved
statement contains there exists A > 0 for all ε > 0 or for all ε > 0 there exists A > 0. In
the following, we explain how the constructed milieu concentrates on the formalization
of infinite closeness in order to help students to recognize the utilities of formalizing
infinite processes through ε–statements. In this milieu, the use of finite limit at infinity
helps students to focus on the dependent variable and on the specific role of quantifiers
in each ε–statement.
METHODOLOGY
Whole-class teaching experiments
This study is based on extensive design research carried out from 2013 to 2015 at the
last year of secondary school in France involving a succession of eight situations
related to the limit notion (Lecorre, 2016). The teaching experiments were conducted
by one of the two authors serving both as the classes teacher (this author was the official
mathematics instructor of those classes) and as a researcher. The teaching experiments
took place inside classes’ allowed time; each class contains about thirty 17-18 years
old students. The teacher-researcher provided the whole-class with preconstructed

36 [Link]:indrum2018:174484
tasks and gradually enriched the mathematical milieu asking questions to assist the
progression of students’ work and giving tools that would help students to address
problems. Data consisted of audiotape recordings and copies of students’ written work.
In this paper, we focus on the transcripts of four successive class sessions that are
related to the fifth and sixth situations of the whole design; each class session lasted
two hours. The fifth situation is based upon the graph of the monster (fig.1) and it is
supposed to destabilize students’ informal understanding of the infinite closeness
notion in the limit process and to trigger the need to formalize this process using ε–
statements. The sixth situation is designed in way that students face: 1) the problems
of validating equalities and limits candidates using two infinite processes; and 2) the
individual subtle formalization via ε–statement of each process depending on whether
the statement contains there exists…for all or contains for all…there exists. Prior to
taking part in the selected class sessions, the students had participated to the preceding
teaching experiments concerning the first four situations of the whole design, and they
were already familiar with whole-class discussions. Specifically, the students had
constructed informal understanding of limit of function at infinity. Building on the
graphs of paradigmatic functions (for example 1/x), they had investigated limits at
infinity by using expressions such as close to infinity and gets closer to. In addition,
they had explored double quantified statements and that double quantifications should
be differenced according to the order of the quantifiers and to the convention of
interpretation. However, the formal definition of limit is still not introduced to them.
A priori analysis of the monster situation
Students’ previous work on the statement f(x) is upper bounded by g(x) "in infinity" led
the teacher-researcher to formalize "in infinity" by means of There exists A > 0 such
that for all x > A. This formalization which is one of the fundamental elements of the
mathematical milieu of this situation is part of prior students’ knowledge. The central
element of the mathematical milieu is the monster (fig.1).

Figure 1: The monster

The monster is the graphical representations of two functions f and g such that f
remains below g (g which "soon" becomes a constant) except in rare but regular peaks
(every 106) where f is over g on small intervals (less than 10-6). In addition, this milieu
contains the conjecture C3: Given two functions f and g having no infinite limits in
infinity. If the limit of f is strictly less than the limit of g in plus infinity then there exists
A >0 such that for all x > A, f(x) <g(x). The students are asked to answer the core
question of this situation: The monster is an example, a counter-example, a non-
example of C3? The use of graphs allows students to create ideas about the limit
process; it also provides them with helpful feedbacks -even if not formal- that may

37 [Link]:indrum2018:174484
contribute to reconsider their informal understanding (p.124). The graphs of
paradigmatic functions are given by the teacher-researcher during the debate to
reinforce the doubt about the meanings that students give to the infinite closeness in
limit process. Students’ formulations are based on their informal understanding of this
closeness. The class discussion may lead to broad agreement about the validity of C3
but some students may remain uncertain considering that it has not been proven yet.
This inquiry is not exactly a request for the formalization of the infinite closeness in
limit process, but it is the beginning of the awareness that prior understanding about
limit process are fragile and have to be formally structured.
A priori analysis of the ε-statements situation
The arbitrariness of ε is the keystone idea of the ε-statements; it founds the equality
process and the limit process through the decreasing of ε towards 0. This situation
contains three phases, they are planned in way that: the first phase focuses on the
formalization of infinite closeness using statements with the only ε; the second phase
deals with the formalization of infinite closeness using statements with ε and other
variables; and the third phase highlights the formalization of infinite closeness involved
in the limit process by emphasizing the role of quantifiers.
- 1st phase: The mathematical milieu contains the property P for A = 50 and L = 2 (P1:
For all ε>0, for all x > 50, 2 - ε < f(x) < 2 + ε), the conjecture C4-1: If f verifies P1
then for all x > 50, f(x) = 2 and the conjecture C4-2: If f verifies P1 then
𝑙𝑖𝑚 𝑓(𝑥) = 2. The students are firstly asked to say what can be concluded if f verifies
→
P1. Then, the teacher-researcher has the responsibility to enrich the milieu by asking
the students whether or not: The function f(x) = 2 + 1/x is an example, a counter-
example, a non-example of C4-1? Depending on the evolution of the debate among the
class students, the use of the same function should permit to study C4-2. More
precisely, students are familiar with the use of graphs to give examples in order to make
or to verify conjectures. Graphical representations of functions may lead to the
visualization of the closeness of f(x) to 2 by using several values of ε. It is expected
that students’ validation of C4-1 via a reductio ad absurdum reasoning permits to focus
on the arbitrariness of ε as formalizing infinite closeness involved in P1. The study of
the function f(x) = 2 + 1/x which does not fit the hypothesis of C4-1 -instead of
verifying there exists A (A=50), for all ε>0 […], this function verifies for all ε>0, there
exists A > 0 […]- should support students’ formulations about infinite closeness
involved in limit process. The discussion of C4-2 should reinforce those formulations
and assists students on thinking about the link between the two processes through the
notion of infinite closeness. It is rather probable that the choice of A (50 in P1) will be
questioned: does any other A>0 and A ≠50 exist in way that the function f(x) = 2+1/x
fits the hypothesis? The issue related to the values of A will be discussed in the
following phase.
- 2nd phase: The mathematical milieu contains the property P which is given for
unknown A (P2: There exists A >0, for all ε>0 such that for all x > A, 2 - ε < f(x) < 2
+ ε), the conjecture C4-3: If f verifies P2 then for all x >A, f(x) = 2 and the conjecture

38 [Link]:indrum2018:174484
C4-4: If f verifies P2 then 𝑙𝑖𝑚 𝑓(𝑥) = 2. The students are firstly asked to give some
→
properties of functions verifying P2. Students’ formulations may lead to the
establishment and the discussion of C4-3 and C4-4. A validation of C4-3 based on
reductio ad absurdum proof is not expected; however, the validation will inevitably
highlight the formalization of infinite closeness within statement containing ε and
another variable (A in P2). The study of C4-4 is supposed to improve students’
formulations about infinite closeness in the limit process using the closeness involved
in P2. It isn’t expected that at this stage students will feel the need to talk about the role
of quantifiers; but, when we inverse P2 into P3: For all ε>0, there exists A >0 such
that for all x > A, 2 - ε < f(x) < 2 + ε, fruitful discussions about the double quantification
statements may arise among students; the third phase deals with this inversion.
- 3rd phase: The milieus of the above phases are planned to bring into focus the use of
ε-statement of equality to validate a limit candidate and so to stimulate students
thinking about the formalization of limit process. In this phase, the milieu focuses on
the ε-statement of limit to validate a limit candidate. This milieu contains P3, the
conjecture C5-1: If f verifies P3 then there exists B>0 for all x>B such that f(x) = 2
and the conjecture C5-2: If f verifies P3 then 𝑙𝑖𝑚 𝑓(𝑥) = 2. The students are firstly
→
asked to say whether f(x) = 2+1/x is an example, a counter-example, neither an
example nor a counter-example of C5-1. Then the teacher-researcher has in charge to
add C5-2 and to ask the following question: what do you think about this conjecture?
It is expected that the starting point of class discussion concerns the question related to
whether f(x) = 2+1/x verifies or not the hypothesis of C5-1. Students’ formulations
may concentrate on the finding of the target A given a specific value of ε; the validation
emerges from the necessity to generalize this argument for each ε. The discussion of
C5-1 emphasizes the need to elucidate the link between the formalizations of infinite
closeness in both P2 and P3. It is expected that the use of f(x) = 2+1/x helps students
to catch the subtleties of this link through the inversion of quantifiers. Students’
formulations about C5-2 are supposed to concentrate on the formalization of infinite
closeness involved in limit process.
Brief description of data analysis method
In the TDS frame, the a priori analysis is important not only to control the data analysis
of the experimental situations but mainly to highlight what does not happen as expected
specifically by focusing on how students’ understanding assist them to progress or not
as planned by the situation. In the case of this research, the data analysis is conducted
in this spirit and it is organized through two major levels. The global level of the data
analysis involved reviewing transcripts paying attention to the potency of the situations
to tackle the research question. The global data analysis shows that the situations give
students the opportunity to enter on the problem and to test their understanding through
actions, formulations and even validations. The planned milieus incite students to
express and share their understanding of limit process and to progress towards formal
understanding. The social dimension of these situations succeeds to stimulate valuable
discussions among students who acted to convince their peers or to be convinced by

39 [Link]:indrum2018:174484
them2. We take advantage of those discussions to engage on the local level of analysis
which focus on students’ understanding of infinite closeness and its formalizations in
the equality and limit processes. This study is based on the evolution of students’ work
through the three levels of the milieu (action, formulation and validation) and on the
arguments they used to explain their work. In the following section, some results of the
local level are exposed and exemplified by generic3 students’ utterances that are
translated verbatim from French. We mainly focus on students’ shared understanding;
however, the individual student’s understanding is underlined when it is awkward and
deep.
RESULTS
Students’ understanding of infinite closeness in the limit process
As expected, class’s discussion about the monster put forward the diversity of students’
informal understanding of limit process. Students’ actions involve the use of
expressions related to closeness such that approaching more and more, from below,
from above, gets closer to. Yet, Students’ argumentations strengthen the need to give
more precisions about those expressions. Building on the graphics of prototypical
functions (fig.2), the following formulation gains broad agreement about how the
infinite closeness should be stated in the limit process: For this kind of sinus curve no
limit, the second, it is sometimes above and sometimes under […] always going closer
to the limit […] the third function […] the peaks are shrinking and the values of the
function are getting closer to the limit each time.

Figure 2: Infinite closeness through 3 graphics

The visualization of infinite closeness through graphical representations helps students
not only to share the same meanings but also to get aware of the fragility of their
informal understanding. Of course, this is not enough to ensure their engagement in the
formalization of closeness in limit process mostly because they have not yet felt the
need to validate their limit candidates.
Students’ understanding of the formalization of infinite closeness in equality
process
To examine students’ understanding of the formalization of infinite closeness in
equality, we mainly focus on the data analysis related to students’ work on C4-1 and
C4-3. In the following, the results are organized into two steps depending on whether
the statements used refer to the only ε (P1) or not (P2). In both cases, students’ work
concentrates simultaneously on the process involved in the statement as a way to verify
equality as well as a way to validate equality.

2
Due to space constraints, the results of the global level of analysis are limited to this description.
3
By generic we mean that it is representative of whole class utterances.

40 [Link]:indrum2018:174484
Students’ work involving ε-statement with the only ε
The use of several graphical representations assists students’ formulations about the
role of ε in the statement: little epsilon, change the value of epsilon, etc. At this stage,
the arbitrariness of ε as formalizing infinite closeness is strongly highlighted and it
constitutes the starting point for the shift towards validating the equality. The validation
is based upon a graphical reductio ad absurdum starting naturally from x > 50 (fig.3):
To show that it's true… show that f(x) can't be different from two [...].

Figure 3: Graphical reductio ad absurdum

Students’ understanding of the formalization of infinite closeness involved in P1 (fixed
A = 50) is aided by graphical arguments and emerges from the necessity of both
verifying and validating equality.
Students’ work involving ε-statement with ε and A
Students’ work on P2 (any A) leads to the discussion of C4-3. The students argue on
the validity of this conjecture on the basis of the graphical reduction ad absurdum
specified for a fixed A (50): It is exactly the same statement with A instead of 50. This
generalization is not yet a proof that students’ understanding of the formalization of
infinite closeness in equality takes account of quantifiers in P2 statement.
Students’ understanding of the infinite closeness in limit process through its
formalization in equality process
To study students’ understanding of infinite closeness in limit through its formalization
in equality, we mainly focus on the data analysis related to students’ work on C4-2,
C4-4 and the case of f(x) = 2 + 1/x. Students’ work about whether this function fits or
not P1 and P2 is supposed to pave the way for linking equality and limit processes as
well as to underline the quantifiers and their order in P2 statement.
Students’ work involving ε-statement with the only ε
Students’ formulations about whether f(x) = 2 + 1/x fits or not P1 are based on
numerical computations and lead soon to the necessity to invalid this example by using
the case of x = 51 and ε = 0.001. Students’ actions on C4-2 are mostly based on the
already stated validation of C4-1: f equals 2 and this result does not give information
about the limit of f. The use of f(x) = 2 + 1/x reinforces the doubt on the validity of C4-
2 and some students’ formulations about this case permit progressively to highlight the
specificities of the relationship between ε and A = 50 in P1: here for all ε there is the
same A equal to 50 from which f(x) equal 2 thus f(x) is between 2 – ε and 2+ ε […] and
so the limit is two. Yet, the involved argument does not provide successful feedbacks
among peers. However, the necessity to validate a limit candidate through the use of
the equality ε-statement (P1) compels students to reorganize their understanding of the
infinite closeness in limit process by taking into account the arbitrariness of ε.
Students’ work involving ε-statement with ε and A
Students’ validation of the statement f(x) = 2 + 1/x is a non-example of C4-4 is based
on a numerical argumentation which is expressed as follows: I would like to ask those

41 [Link]:indrum2018:174484
who think it is true, to choose an A, any A, and I will be able each time to find a counter
example (an epsilon in fact). This argument emphasizes the order of quantifiers in the
equality process but students’ work on C4-4 is inconclusive mostly because they do
not succeed to draw upon the arbitrariness of ε to formalize the limit process. However,
their understanding of the infinite closeness is enhanced by the use of P2 as an ε-
statement firmly consent with the limit process.
Students’ understanding of the role of quantifiers in the formalization of infinite
closeness in limit process
These results are mainly based on students’ work on C5-1, C5-2 and the case of f(x) =
2 + 1/x. They are splitted into two sections: 1) students’ interpretations of quantifiers’
orders; and 2) the potential sum up of limit process into the formalized ε–statement P3.
Interpretations of quantifiers’ orders
Students’ work on whether f(x) = 2+ 1/x verifies P3 or not highlights their difficulties
to perceive the distinction between P2 and P3 and progressively emphasizes the
necessity to take care of quantifiers’ orders. Students’ firstly act as for P2 to interpret
the quantifiers in P3 before focusing on a peer intervention: […] the question is written
as for all there exists he must give us an epsilon and we have to find an A. Students’
discussions highlight the inversion of quantifiers issue and the need for convention of
interpretations. The teacher-researcher intervenes in order to help students finding the
targeted A for every given ε and to confirm the invalidity of C5-1.
Sum up of limit process via ε-statement
During the debate concerning quantifiers, students’ work with ε-statements is strongly
connected to the necessity to answer those questions: given epsilon, how to give A?
Given A, how to give epsilon? In addition, students’ work on the validation of the limit
candidate involved in C5-2 leads to the formalization of the infinite closeness in limit
process. Students’ further formulations put forward the need to explore additional
question: to what extent this formalization is sufficient to sum up the formal limit?
CONCLUSION
This study examines students’ understanding of the formalisations of two infinite
processes in the Archimedean continuum by using ɛ-statements. These formalizations
concern with the infinite closeness notion and refer to equality and limit of function at
infinity. The aim of this paper is to give data on how these formalizations and their link
may support students’ understanding of formal limit. We deploy TDS constructs to
design situations in which the milieu is built on students’ informal understanding of
limit and concentrates on the role of quantifiers to differentiate the formalizations given
to the infinite closeness in the process of equality and limit, respectively. This study
highlights three main results concerning students’ understanding of ε-statements: 1) It
is possible to organize a milieu that leads students to question their informal
understanding of limit process: in this study, the doubt emerges through several ways
used by students to perceive the infinite closeness in the limit process; 2) The only
formalization of the infinite closeness in the equality process cannot provide insights
on its formalization in the limit process. The focus on the quantifiers’ orders is crucial

42 [Link]:indrum2018:174484
to achieve this formalization; 3) The formalization of the infinite closeness in limit
process does not ensure students’ sum up of limit process into the formalized ε–
statement, this issue needs further investigation. The social dimension of TDS helps
students to progressively construct meanings that will constitute the bricks of the
meaningful argument which tends to be collectively adopted. In the end, students can
admit the irrefutability of the reasoning when all their reluctances are taken into
account by their peers.
REFERENCES
Bressoud, D., Ghedamsi, I., Martinez-Luaces, V., & Törner, G. (2016). Teaching and
Learning of Calculus. Cham: Springer International Publishing.
Cottrill, J., Dubinsky, E., Nichols, D., Schwinngendorf, K., Thomas, K., & Vidakovic,
D. (1996). Understanding the limit concept: Beginning with a coordinated process
schema. Journal of mathematical behavior, 15, 167–192.
Durand-Guerrier, V., &Arsac, G. (2005) An epistemological and didactic study of a
specific calculus reasoning rule, ESM, 60(2),149–172.
González-Martín, A,, Bloch, I., Durand-Guerrier, V., & Maschietto, M.
(2014). Didactic Situations and Didactical Engineering in University Mathematics:
Cases from the Study of Calculus and Proof. RME, 16 (2), 117–34.
Lecorre, T. (2016). Des conditions de conception d’une ingénierie relative à la
définition de limite – Elaboration d’un cadre basé sur un modèle de rationalité.
[Conditions for the conception of a design of limit concept – A framework on
rationality]. Thesis, University of Grenoble-Alpes (France).
Mamona-Downs, J. (2001) Letting the intuitive bear on the formal; a didactical
approach for the understanding of the limit of a sequence. ESM, 48, 259–288.
Oehrtman, M., Swinyard, C., & Martin, J. (2014). Problems and solutions in students’
reinvention of a definition for sequence convergence. Journal of Mathematical
Behavior, 33, 131–148.
Oehrtman, M. (2009). Collapsing dimensions, physical limitations, and other student
metaphors for limit concepts. Journal for RME, 40(4), 396–426.
Przenioslo, M. (2004). Images of the limit of function formed in the course of
mathematical studies at the university. ESM, 55(1/3), 103–132.
Roh, K. H. (2008). Students’ images and their understanding of definitions of the limit
of sequence. ESM, 69, 217–233.
Roh, K. (2010). An empirical study of students’ understanding of a logical structure in
the definition of the limit of a sequence via the ε-strip activity. ESM, 73, 263–279.
Swinyard, C. (2011). Reinventing the formal definition of limit: The case of Amy and
Mike. The Journal of Mathematical Behavior, 30(2), 93–114.
Swinyard, C., & Larsen, S. (2012). What does it mean to understand the formal
definition of limit? Insights gained from engaging students in reinvention. Journal
for Research in Mathematics Education, 43(4), 465–493.
Tall, D., & Vinner, S. (1981). Concept image and concept definition in mathematics
with particular reference to limits and continuity. ESM, 12(2), 151–169.

43 [Link]:indrum2018:174484
“A function is continuous if and only if you can draw its graph without
lifting the pen from the paper” – Concept usage in proofs by
students in a topology course
Erik Hanke
University of Bremen, Faculty 3 of Mathematics and Computer Science, Germany,
[Link]@[Link]
Many students enter university having learned that the graph of a continuous function
is “in one piece” and “can be drawn without lifting the pen from the paper.”
Rigorously, a function ℝ → ℝ is continuous if and only if its graph is path-connected.
In this article, I examine proofs of this fact by students in a topology course. Based on
Moore (1994), concept usage of continuity and path-connectedness is analysed
through recognition and building-with of the RBC-model of epistemic actions (Dreyfus
& Kidron, 2014) in combination with a refinement of Oerter’s (1982) contextual layers
of objects. A “propositional” layer to describe relationships between objects used in
proofs is introduced and used to perform case studies of students’ solutions.
Keywords: Teaching and learning of specific topics in university mathematics,
teaching and learning of analysis and calculus, topology, continuity, epistemic actions.

INTRODUCTION
The concept image of a continuous real function of one real variable as one whose
graph is “in one piece” or “can be drawn without lifting the pen from the paper”
(provided that the function is defined on an interval) is held by many students in school
or university (Tall & Vinner, 1981; Hanke & Schäfer, 2017).
This piece of research is intended to expand the viewpoint from students’ concept
usage of continuity from first-year analysis to higher courses and sensitise for some
students’ thinking processes. To a large extent, this paper is of philosophical nature
and suggests a refinement of Schäfer’s (2010) approach to describe epistemic actions
with the RBC-model, namely by introducing a new specific layer called propositional
layer. This is relevant for the specifically mathematical procedure of deduction from
theorems about relationships between objects. This refinement is then applied to
students’ (partial) proofs of the fact that a real function defined on ℝ is continuous if
and only if it has a path-connected graph. Thus, this study begins to fill a gap in the
literature on this highly recurring concept image and the difficulties in finding a
rigorous proof which requires some topological knowledge.
The general research question before starting this investigation was which mental
images students use in a proof to a prevalently, vividly acceptable theorem. Here, the
aim of this article is to display students’ proofs and moot a way to dissect these
according to the levels of concreteness of the objects the students used.

44 [Link]:indrum2018:171761
The task
The exercise in question was (translation E.H.; see Ross (2013, p. 182)):
“In school, one often says ‘A function is continuous if you can draw its graph without
lifting the pen.’ Prove the following exact version of this proposition: A function 𝑓: ℝ →
ℝ is continuous if and only if its graph Γ𝑓 = {(𝑥, 𝑓 (𝑥)): 𝑥 ∈ ℝ} ⊂ ℝ2 is path-connected.”
The intuition of “without lifting the pen” has to be translated into a valid statement
carefully. Path-connectedness really is required instead of connectedness, and the
theorem is no longer valid for functions from an arbitrary path-connected space to the
real numbers. It is tacitly assumed in this task that the topologies for ℝ and ℝ2 are
Euclidean and Γ𝑓 inherits the induced topology. Note that the graph Γ𝑓 is the image of
the function id × 𝑓: ℝ → ℝ2 , 𝑥 ↦ (𝑥, 𝑓(𝑥)). Thus, the forward implication in the
exercise follows from the facts that functions into products of topological spaces are
continuous (with respect to the product topology) if their components are continuous,
and continuous images of path-connected sets are path-connected. For the reverse
direction, continuity of 𝑓 at 𝑝 ∈ ℝ can be proven by contraposition via the 𝜀-𝛿-
definition using the existence of a path of the form 𝛾 = (𝜑, 𝑓 ∘ 𝜑) between (𝑢, 𝑓(𝑢))
and (𝑣, 𝑓(𝑣)) in Γ𝑓 for some 𝑢 < 𝑝 < 𝑣.
Since this paper has theoretical aims next to the empirical investigation and due to page
restrictions, I omit an à-priori-analysis of different possibilities of proving this theorem
and prerequisites.

FRAMEWORK: THEORY AND METHODOLOGY

Concept usage, object layers and the model of nested epistemic actions
The concept image of a learner for a mathematical object, class of objects or procedures
is “the total cognitive structure that is associated with the concept, which includes all
the mental pictures and associated properties and processes” (Tall & Vinner, 1981,
p. 152). In contrast, (personal) concept definitions are students’ attempts to specify a
concept. Moore (1994) claims that besides “mathematical language and notation” and
“getting started” one of the major difficulties in proving is “concept understanding”
(mix of concept images, concept definition and concept usage) (p. 249): This means,
many students who fail in a proof “lack intuitive understanding of concepts”, “cannot
use concept images to write a proof”, “cannot state the definitions [properly, E.H.]” or
“do not know how to structure a proof from a definition” (Moore, 1994, p. 253). The
term concept usage “refers to the ways one operates with the concept in generating or
using examples or in doing proofs” (Moore, 1994, p. 252).
The Abstraction in Context Methodology (AiC) (Dreyfus & Kidron, 2014) offers a
theory about learning, originating in the need for a new mathematical construct, its
construction of knowledge and its consolidation taking into account a specific form of
context. In this setting, construction of knowledge means performing nested epistemic
actions (RBC-model): Recognising, building-with and constructing. Recognising is

45 [Link]:indrum2018:171761
“seeing the relevance of a specific previous knowledge construct to the problem at
hand”, building-with “comprises the combination of recognized constructs, in order to
achieve a localized goal such as the actualization of a strategy, a justification, or the
solution of a problem” and construction “consists of assembling and integrating
previous constructs by vertical mathematization to produce a new construct” (Dreyfus
& Kidron, 2014, p. 89). The actions are nested in the sense that building-with
something requires its recognition and construction requires building-with. Hence, the
word construction is meant globally within the AiC methodology and locally in the
RBC-model. It is noteworthy that flexibility and availability of a construct does not
stem from construction itself but consolidation.
Based on Oerter’s (1982) theory of activity, Schäfer (2010) differentiated between
three layers which help to describe recognition processes of objects more precisely.
Next to objectification [orig. Vergegenständlichung] which is the creation of objects,
objectual concern [orig. Gegenstandsbezug] towards previously constructed objects
can be classified on three layers (Oerter, 1982): On the singular layer objects are not
distinguishable from the action of an individual itself (e. g. recognising numbers in a
table); for the actor, the objects are not yet seen as objects and do not need to have
names. On the contextual layer objects are characterised by their usage – not anymore
restricted to an individual but shared within a community – and the usage is performed
within a specific contentual context and similarity of situations; the objects gain
persistence beyond singular action. Finally, on the formal layer objects are disengaged
from specific actions or context (Oerter, 1982; Schäfer, 2010). The objectual concern
of previously constructed objects (of a learning process) is reflected in the way
someone can use this object. In this article, this activity theory oriented standpoint is
specified to mathematics practice in students’ attempts to prove a topological fact.
The approach to consider object layers and actions on them seems related to Sfard’s
(1991) idea of the duality of operational and structural conceptions (views of an
individual of a concept) within the process of concept formation; concept here means
a “mathematical idea […] in its ‘official’ form” (p. 5). Structural conception considers
concepts as “abstract objects” and its dual form, the operational conception, is about
“processes, algorithms and actions”, not the notion as an object itself (Sfard, 1991,
p. 4; emph. orig.). Both of these views of conception can be seen in mathematical
practice: In the action of formal recognition as well as propositional recognition and
building-with (see below) objects are conceptualised as structural and through
operational conceptions a conclusion is achieved. Propositions themselves are
structural, and their scope is reflected in their use, allowing a deduction or justification
(not necessarily mathematically correct though).
Founded in the Anthropological Theory of the Didactic, Hausberger (2018) described
“structuralist praxeologies” which characterise mathematical justification practices
that are oriented towards replacing a statement about the particular with one about the
general: “Structuralist thinking is characterized by reasoning in terms of classes of
objects, relationships between these classes and (structural) stability of properties

46 [Link]:indrum2018:171761
under operations on structures” (pp. 82f.; emph. orig.). What I describe here as
propositional building-with action would fall most likely under his levels 2 and 3 of
“structuralist dimension” of a proof (Hausberger, 2018, p. 81; emph. orig.) which
describe the application of theorems to a task at hand, therefore reasoning on
structuralist rather concrete object level. For example, the task of showing that an
object O (e. g. ℤ) possesses property A (e. g. unique factorisation domain) can be
changed to the task of showing that O belongs to a class of objects C (e. g. Euclidean
rings) in order to apply a theorem which states that each member of C possesses A
(e. g. every Euclidean ring is a unique factorisation domain). The identification of O’s
membership to C resembles in the proofs in this paper to recognition actions (of
different layers) that O possesses property C. This procedure is “illuminating as to the
‘root causes’ behind the result” (Hausberger, 2018, p. 83).
Mathematical notions such as “continuous real function” are on the formal layer for
experienced students and mathematicians, and instances thereof can be recognised as
having the general properties of elements of the class they belong to. However,
mathematical theorems can also be seen as objects on the formal layer. If they come
into use, e. g. by specialisation to a concrete situation in an exercise, they become an
object one builds-with. In fact, since the usage of objects is of particular interest here,
I claim that a new object layer should be included, the propositional layer: On the
propositional layer we find theorems as objects that describe properties of objects on
the formal layer. Thus, it does not merely contain abstract mathematical objects but
relationships between objects as own objects. These relationships can then be applied
in a propositional building-with action towards objects on the formal or the contextual
layer. [1]
In practice, the decision for which object layer occurs at which place can be guided by
the instantiation of objects (“let 𝑓 be given by 𝑓(𝑥) = 𝑥 2 + 1”), which mostly suggests
the singular or contextual layer, or their declaration (“let 𝑓 be a function such that …”),
which highlights an object rather on formal layer. In proofs, the identification of
propositional recognition and building-with actions may be assisted by the writer, e. g.
referencing the lecture, the number of a theorem etc. However, somebody’s “personal
mathematical toolkit” may determine which layer really is in use. The analyses below
reflect an interpretation of the written product, not the way of finding the proof.
Examples for the identification of the object layers
If one wants to show that the unit circle 𝑆 1 is compact, one can directly use the
definition of compactness by taking any open cover of the circle, assuming there was
no finite subcover and, using concrete, contextual properties of 𝑆 1 , evoking a
contradiction. This way, propositional objects are not necessarily involved (except for
the definition of course, and depending on the particular argumentation). On the other
hand, identifying 𝑆 1 as continuous image of [0,1] → ℝ2 , 𝑡 ↦ (cos(2𝜋𝑡) , sin(2𝜋𝑡)),
is a contextual recognition of 𝑆 1 to the context of the map, and together with the
propositional recognition of the fact that continuous images of compact spaces are

47 [Link]:indrum2018:171761
compact the propositional building-with action of this fact to the situation at hand
yields the compactness of 𝑆 1 . Structurally, these two proofs are completely different.
Next, consider the above proof that the graph of a continuous function is path-
connected. First, the graph is recognised as the image of a certain map id × 𝑓
(contextual layer), and this map is recognised as a product of continuous maps
(contextual layer). The theorem that products of continuous maps are continuous is
recognised on the propositional layer and used to deduce the continuity of the map
written down in the proof (building-with on the propositional layer to further recognise
the continuity of id × 𝑓 on formal layer: The concrete map is no longer important,
simply its continuity). Finally, the theorem that continuous images of path-connected
spaces are path-connected (recognition on the propositional layer) is used in a building-
with action on the propositional layer yielding that the graph of 𝑓 is path-connected by
recognising that the theorem is applicable to the situation at hand. Even though this
proof is very short, many recognition and building-with actions had to be completed,
brought into order and were compressed in only two sentences.
The “if”-direction of the given task can be proved in the formal/propositional manner
as described and is a special case of the theorem “If 𝐹: 𝑋 → 𝑌 is any function between
a path-connected topological space 𝑋 and any topological space 𝑌, then the graph of 𝐹
is path-connected if 𝐹 is continuous.” The main ingredient is the fact that for any path-
connected space 𝑈 and any function 𝑔: 𝑈 → 𝑉 between topological spaces, continuity
of 𝑔 implies the path-connectedness of 𝑔(𝑈). Since the reverse directions of these two
statements are not true [2], the reverse direction of the students’ task cannot be proved
(completely) in the formal/propositional manner considering solely (path-connected)
topological spaces, continuous maps and their properties as above. Nevertheless, it is
surely possible to argue with propositional objects, e. g. using the intermediate value
theorem or the intermediate value property of functions (Ross, 2013, p. 182f.).

DATA COLLECTION
The data collection for this study took place during the spring semester 2017 at the
University of Bremen within the topology class for Bachelor students in pure
mathematics. I was not involved in this class but was informed by the lecturer about
the contents. During the third week of the semester the students had to solve (besides
others) the task presented above. Construction is not directly observable in the analyses
below because all notions involved are not new to the students. The context of the
contextual layer is understood very locally, depending on the objects already available
in the particular solution, for instance those that have been declared before.

RESULTS
The following cases are supposed to illustrate the work with the above framework (for
groups 2 and 3 only one implication is shown). Overall, the solutions were very
different regarding the approaches used. More details cannot be included here. All of
the following transcripts were translated from German respecting (unusual) syntax.

48 [Link]:indrum2018:171761
Notational and language errors are mostly ignored. Abbreviations of German words
were often not abbreviated. Small notational errors like forgetting a closing bracket
were corrected. An open ball of radius 𝜔 centred at 𝑎 is denoted by 𝑈𝜔 (𝑎).
Case study: Group 1
The following is a transcript of the solution of group 1 with a sketch redrawn by myself.
1 “⇐” Let 𝜀 > 0
2 Since Γ𝑓 is path-connected, there exists 𝛾: [0,1] → Γ𝑓
3 with 𝛾 (0) = (𝑥 − 𝛿1 , 𝑥 − 𝜀)𝑇 and 𝛾 (1) = (𝑥 + 𝛿2 , 𝑥 + 𝜀)𝑇
4 Choose 𝛿 = max{𝛿1 , 𝛿2 }, then it holds that
5 |𝑥 − 𝑦| < 𝛿: |𝑓 (𝑥) − 𝑓 (𝑦)| < 𝜀 𝑦 ∈ ℝ
6 and thus 𝑓 is continuous because 𝜀 arbitrary.
As a first step, the students declare a positive 𝜀 (line 1) which indicates that they would
like to test the 𝜀-𝛿-definition of continuity. It looks like a recognition on the formal
layer, but it is not stated explicitly at which point they want to check continuity; most
probably it is “𝑥”. Afterwards, the students assume that there exists a function (most
likely a path, even though not stated) [0,1] → Γ𝑓 connecting the points (𝑥 − 𝛿1 , 𝑥 − 𝜀)𝑇
and (𝑥 + 𝛿2 , 𝑥 + 𝜀)𝑇 . It is not clear why these points should lie on the graph of 𝑓 but
this should be the case since the path lies in Γ𝑓 by their assumption (recognition on
contextual layer) (lines 2-3). Interpreting the students’ sketch of the graph, which they
do not refer to, I hypothesise that they actually mean the points (𝑥 − 𝛿1 , 𝑓(𝑥) − 𝜀)𝑇
and (𝑥 + 𝛿2 , 𝑓(𝑥) + 𝜀)𝑇 , and 𝛿1 and 𝛿2 are chosen such that 𝑥 − 𝛿1 and 𝑥 + 𝛿2 are
preimages of the corresponding second components of the points on the graph under
𝑓. Thus, the path is recognised on contextual layer using false assumptions. Then, the
students choose the minimal of these 𝛿s to implicate that 𝑓 fulfils the 𝜀-𝛿-definition of
continuity in line 5 (lines 4-6). This is a building-with action on contextual layer since
taking the minimum of the 𝛿s is quite a standard technique in analysis. However, if the
function is not “nice enough” between 𝑥 − 𝛿1 and 𝑥 + 𝛿2 , for example monotonically
increasing on [𝑥 − 𝛿1 , 𝑥 + 𝛿2 ] as indicated by the students’ figure, then the interval
(𝑥 − 𝛿, 𝑥 + 𝛿) is not necessarily mapped into the interval (𝑓(𝑥) − 𝜀, 𝑓(𝑥) + 𝜀).
Hence, the building-with action does not lead to the needed conclusion in line 5. Even
if there were preimages of 𝑓(𝑥) ± 𝜀 (only to the left or right of 𝑥 according to the sign
of 𝜀), one had to choose 𝛿1 = inf{δ > 0: 𝑓(𝑥 − 𝛿) = 𝑓(𝑥) − 𝜀} and 𝛿2 = inf{δ >
0: 𝑓(𝑥 + 𝛿) = 𝑓(𝑥) + 𝜀}, and would need to show that these are different from 0. Since
functions often encountered are “nice enough” or monotonically increasing, this wrong
argumentation might originate in students’ “met-befores” (McGowen & Tall, 2010).
7 “⇒” Let Γ𝑓 = {(𝑥, 𝑓(𝑥)): 𝑥 ∈ ℝ} ⊂ ℝ2 be the graph of the continuous function 𝑓.
8 Let (𝑥, 𝑓 (𝑥))𝑇 , (𝑦, 𝑓(𝑦))𝑇 ∈ Γ𝑓 (wlog 𝑥 < 𝑦)
9 To show ∃𝛾: [0,1] → Γ𝑓 path.
10 Let 𝑧 ∈ [𝑥, 𝑦]. Since 𝑓 is continuous, it holds that
11 ∀𝜀 > 0 ∃𝛿 > 0 ∶ |𝑧 − 𝑎| < 𝛿: |𝑓 (𝑧) − 𝑓(𝑎)| < 𝜀 (𝑎 ∈ ℝ)

49 [Link]:indrum2018:171761
 12 No matter how small the neighbourhood 𝑈𝜀 (𝑓 (𝑧)) is chosen, the values of 𝑓(𝑎) ∈
𝑈𝜀 (𝑓 (𝑧)) for 𝑎 ∈ 𝑈𝛿 (𝑧).
13 ⇒ Γ𝑓′ = {(𝑎, 𝑓 (𝑎)): 𝑎 ∈ (𝑧 − 𝛿, 𝑧 + 𝛿 )} ⊂ 𝑈𝛿 (𝑧) × 𝑈𝜀 (𝑓(𝑧))
14 Since ℝ is connected, the neighbourhoods are also connected.
15 ⇒ It exists a path 𝛾: [0,1] → Γ𝑓
16 with 𝛾 (0) = (𝑥, 𝑓 (𝑥))𝑇 𝛾 (1) = (𝑦, 𝑓(𝑦))𝑇
The solution starts with the definition of the graph of 𝑓 and the students choose two
points on the graph (lines 7-8). They want to show the existence of a path in Γ𝑓 (line
9), presumably that links the two points, although they do not state it. This is contextual
recognition since the path is adjusted to the concrete setting and formal recognition
would be hypothetical because the definition of path-connectedness is not completely
adapted correctly to the given problem. Afterwards, they recognise the definition of
continuity of 𝑓 at some point 𝑧 between the first components of the given points on the
graph (lines 10-11) on the formal layer (independent of a concrete 𝑓). Next, they
recognise on the formal layer a topological version of continuity via neighbourhoods
(line 12) and built-with on the contextual layer the implication that a part of the graph
lies inside the product of the neighbourhoods 𝑈𝛿 (𝑧) and 𝑈𝜀 (𝑓(𝑥)) (nevertheless, one
has to mention that the neighbourhoods have never been instantiated because 𝜀 and 𝛿
only appear within quantifiers) (line 13). Afterwards, the recognition of ℝ as a
connected space is formal (proven property in the lecture) but the deduction of the
connectedness of the neighbourhoods (likely those in lines 12-13), or their product, is
not justified (line 14) (possibly, the students also mixed up connectedness with path-
connectedness here). As a last step, the group now directly concludes that there exists
a path in Γ𝑓 joining the points on the graph chosen in the beginning (lines 15-16). It can
be interpreted that the students believe that subspaces of path-connected sets are path-
connected and apply this to Γ𝑓′ ⊂ 𝑈𝛿 (𝑧) × 𝑈𝜀 (𝑓(𝑧)) without making clear that the
neighbourhoods are path-connected, not only connected. Under this interpretation, the
building-with action would be propositional, but since the group’s claimed implication
does not seem to be logically connected to their previous proof steps, their thinking
cannot be ascertained.
Case study: Group 2
17 “⇒” Let 𝑓: ℝ → ℝ be continuous. Note that ℝ is path-connected.
18 Now let (𝑎, 𝑓(𝑎)) and (𝑏, 𝑓 (𝑏)) ∈ Γ𝑓 .
19 Since ℝ is connected, there is a path 𝛾 such that 𝛾(0) = 𝑎, 𝛾(1) = 𝑏.
20 Then, 𝑔 = (𝛾, 𝑓 ∘ 𝛾) is continuous because the composition of continuous maps is
continuous.
21 𝑔 is also a path from (𝑎, 𝑓 (𝑎)) to (𝑏, 𝑓(𝑏)) because:
22 𝑔(0) = (𝛾(0), (𝑓 ∘ 𝛾 )(0)) = (𝑎, 𝑓(𝑎))
23 𝑔(1) = (𝛾(1), (𝑓 ∘ 𝛾 )(1)) = (𝑏, 𝑓 (𝑏))
24 thus, Γ𝑓 is path-connected since 𝑔 is continuous.
On the formal layer, the students recognise that ℝ is path-connected (note that no

50 [Link]:indrum2018:171761
additional information of ℝ is used) (line 17) and that they have to find a path between
two arbitrary points in the graph which is indicated by the declaration of two points in
Γ𝑓 (line 18). On the propositional layer, they build (which means they state its
existence) a path 𝛾 between the first components of the chosen points using that ℝ is
connected (in fact, they should have used that it is path-connected; it is not clear
whether this is just a notational error since they recognised before that ℝ is path-
connected) (line 19). Next, the students recognise on the propositional layer that
compositions of continuous functions are continuous and build-with this fact on the
propositional layer that the product function 𝑔 = (𝛾, 𝑓 ∘ 𝛾), contextually recognised as
a product of continuous maps, is continuous (their argument lacks the fact that the
continuity of the components implies the continuity of the product map) (line 20).
Recognising that they have to plug in 0 and 1 for verification (formal layer of a part of
the definition of path-connectedness), they conclude that 𝑔 is in fact a path between
the initial points (lines 21-23) (singular/contextual layer in lines 22-23 because the
concrete form of 𝑔 is used). The last line contains again a propositional building-with
action since the graph of 𝑓 is (presumably) shown to have the property that any two of
its points can be linked with a path (even though not explicitly stated).
Case study: Group 3
25 “⇐” Approach: If one can show, let it be supposed that there exists a path between
two points of the graph of a function, then there also exists an injective path, it follows
(*)
26 Suppose, 𝑓 not continuous at 𝑥 ∈ ℝ ⇒ ∃𝜀 > 0: ∀𝛿 > 0 ∃𝑥̃ ∈ ℝ: 𝑥̃ ∈ 𝑈𝛿 (𝑥), 𝑓(𝑥̃ ) ∉
𝑈𝜀 (𝑓 (𝑥))
27 By assumption there is a path 𝜇: [0,1] → Γ𝑓 with 𝜇 (0) = (𝑥̃, 𝑓(𝑥̃ )), 𝜇(1) =
(𝑥 ′ , 𝑓 (𝑥 ′ )).
28 Wlog 𝑥̃ < 𝑥 < 𝑥 ′ .
29 By (*) there exists an injective path 𝜇̃. Now let 𝜏: [𝑎, 𝑏] → [0,1], 𝜏(𝑥) ∶= (𝑥 −
𝑎)/(𝑏 − 𝑎), is continuous!
30 ⇒ 𝜋2 ∘ 𝜇̃ ∘ 𝜏 = 𝑓|[𝑎,𝑏] , where 𝜋2 is the projection of the second component.
31 Namely, for 𝑥 ∈ [𝑎, 𝑏] it holds:
32 𝜋2 (𝜇̃(𝜏(𝑥))) = 𝜋2 (𝜇̃((𝑥 − 𝑎)/(𝑏 − 𝑎)) = 𝜋2 (𝑥, 𝑓 (𝑥)), since there is only one
possibility for an injective continuous map in the first component of 𝜇̃ in
33 question. 𝜋2 (𝑥, 𝑓 (𝑥)) = 𝑓 (𝑥). // ↯ to 𝑓 not continuous
The group begins a proof by contradiction in line 26 and therefore the students
recognise the negation of the 𝜀-𝛿-definition of continuity of 𝑓 at some 𝑥 on the formal
layer, relying on the notation with neighbourhoods from the lecture. The students built-
with on the contextual layer a path from two points on the graph whose first coordinates
𝑥̃ and 𝑥 ′ surround 𝑥, the point where the function is assumed to be discontinuous (lines
26-28) (however, 𝑥 ′ is not explicitly declared). Taken for granted that one can construct
an injective path linking two points given any path between these two (line 25) –
admittedly, the group does neither argue how this may work nor state explicitly that
this injective path has to have the same start and end point or domain – the students use

51 [Link]:indrum2018:171761
such a path 𝜇̃ and compose it with the above 𝜏 which is recognised on contextual level
as continuous (line 29), to perform a building-with action deducing that 𝜋2 ∘ 𝜇̃ ∘ 𝜏 is
equal to 𝑓 restricted to the interval [𝑎, 𝑏] (lines 29-30). The map 𝜏 is used here to
transform the path defined on the unit interval to a path 𝜇̃ ∘ 𝜏 in Γ𝑓 defined on [𝑎, 𝑏]
which shall function as the domain where 𝑓 can be applied (likely, the students actually
meant 𝑎 = 𝑥̃ and 𝑏 = 𝑥′). In lines 30ff., the students try to justify that the second
component of 𝜇̃ ∘ 𝜏 is 𝑓|[𝑎,𝑏] . Here, the students believe to recognise the first component
of 𝜇̃ ∘ 𝜏 as the identity on [𝑎, 𝑏] because there shall be only one possibility for an
injective path between two real numbers (lines 32-33). This is wrong. However, this
mistake could formally be resolved by performing a “velocity change of paths” which
makes the first component of 𝜇̃ ∘ 𝜏 equal to the identity on [𝑎, 𝑏]. The second equal
sign in line 32 is then only justified by this erroneous recognition of the only injection
[𝑎, 𝑏] → [𝑎, 𝑏] being the identity. Obtaining 𝜋2 (𝑥, 𝑓(𝑥)) as solution of the calculation
in line 32 is a building-with action on the singular/contextual layer; singular here refers
to the special situation – the assumption in line 25 – the students find themselves in.
As a last step of the proof, I hypothesise that the students recognise 𝑓|[𝑎,𝑏] (they write
𝑓 in line 33 though, which clearly agrees with 𝑓 on [𝑎, 𝑏]) to be continuous (at 𝑥) (a
contradiction to their assumption in line 26). Their justification is however not directly
observable; they may have used the composition of the continuous maps 𝜏, 𝜇̃ and 𝜋2 .
This is recognised as a contradiction to the assumption of discontinuity of 𝑓 at 𝑥 on
formal layer (line 33). Finally, the end of the proof is obtained as the result of the
propositional building-with action that finding a contradiction to the hypothesis of the
contraposition of the statement to prove is equivalent to the original statement.

DISCUSSION & CONCLUSION

The notion of propositional layer of objects refines the three layers of objects Schäfer
(2010) used to analyse the epistemic action of recognising. In particular, this new layer
describes the building-with action of applying a proposition about relationships
between objects on the formal layer. In the case studies, it turned out that the
recognition and building-with actions usually succeeded when the prerequisites of a
definition or theorem in use had been successfully recognised. However, the
recognition of the non-satisfaction of necessary conditions for the application of a
theorem failed several times because of insufficient mental imagery of continuity (e. g.
“local niceness” in the “only if”-proof of Group 1) or paths (e. g. injectivity of paths
with group 3) and wrong properties attributed to the objects to be acted on.
Nevertheless, subsequent building-with actions on the propositional layer were often
carried out coherently based on these wrong assumptions.

ACKNOWLEDGEMENT
I would like to thank the German Federal Ministry of Education and Research (BMBF)
for its support in the subproject “Spotlights Lehre” within the project “Schnittstellen
gestalten” at the University of Bremen.

52 [Link]:indrum2018:171761
NOTES
1. Of course, in definitions previously defined concepts are usually specialised, thus definitions may also be seen as
objects on propositional layer. Seeing what makes up a definition or whether something satisfies a definition is regarded
as a recognition action here, and objects which are recognised to satisfy a definition will be on formal layer.

2. The graph and the image of arg: 𝑆 1 → ℝ, 𝑒 𝜃𝑖 ↦ 𝜃 (with 0 ≤ 𝜃 < 2𝜋) are path-connected but arg is not continuous.

REFERENCES
Dreyfus, T. & Kidron, I. (2014). Introduction to abstraction in context (AiC). In A.
Bikner-Ahsbahs & S. Prediger (Eds.), Networking of theories as a research practice
in mathematics education (pp. 85–96). New York: Springer. doi:10.1007/978-3-
319-05389-9_6
Hanke, E. & Schäfer, I. (2017). Students' view of continuity: An empirical analysis of
mental images and their usage. In: T. Dooley & G. Gueudet (Eds.), Proceedings of
the Tenth Congress of the European Society for Research in Mathematics Education
(CERME10, February 1-5, 2017) (pp. 2081–2088). Dublin, Ireland: DCU Institute
of Education and ERME.
Hausberger, T. (2018). Structuralist praxeologies as a research program on the teaching
and learning of abstract algebra. International Journal of Research in
Undergraduate Mathematics Education, 4, 74–93. doi:10.1007/s40753-017-0063-4
McGowen, M. & Tall, D. (2010). Metaphor or met-before? The effects of previouos[!]
experience on practice and theory of learning mathematics. Journal of Mathematical
Behaviour, 29, 169–179. doi:10.1016/[Link].2010.08.002
Moore, R. C. (1994). Making the transition to formal proof. Educational Studies in
Mathematics, 27(3), 249–266. doi:10.1007/BF01273731
Oerter, R. (1982). Interaktion als Individuum-Umwelt-Bezug. In E. D. Lantermann
(Ed.), Wechselwirkungen (pp. 101–127). Göttingen, Germany: Verlag für
Psychologie.
Ross, K. (2013). Elementary analysis. The theory of calculus (2nd ed.). New York:
Springer. doi:10.1007/978-1-4614-6271-2
Schäfer, I. (2010). Towards a refinement of epistemic actions: A case study of
processes of recognition on spirolaterals. In M. F. Pinto & T. F. Kawasaki (Eds.),
Proceedings of the 34th Conference of the International Group for the Psychology
of Mathematics Education, Vol. 4 (pp. 161–168). Belo Horizonte, Brazil: PME.
Sfard, A. (1991). On the dual nature of mathematical conceptions: Reflections on
processes and objects as different sides of the same coin. Educational Studies in
Mathematics, 22(1), 1–36. doi:10.1007/BF00302715
Tall, D. & Vinner, S. (1981). Concept image and concept definition in mathematics
with particular reference to limits and continuity. Educational Studies in
Mathematics, 12(2), 151–169. doi:10.1007/BF00305619

53 [Link]:indrum2018:171761
 Learning complex analysis in different branches
– Project Spotlight-Y for future teachers
Erik Hanke1 and Ingolf Schäfer2
1
University of Bremen, Faculty 3 of Mathematics and Computer Science, Germany,
[Link]@[Link];
²University of Bremen, Faculty 3 of Mathematics and Computer Science, Germany,
[Link]@[Link]
At the University of Bremen in teaching complex analysis, we split the last part of the
lecture into two branches according to profession: While future mathematicians
deepen their understanding in a branch for them, future teachers take a branch of the
lecture where they prepare a task with dynamical geometry software for pupils which
is based on phenomena of complex analysis. Here, we describe the design of the
course, some general aims and first results obtained from the branch for future
teachers.
Keywords: Novel approaches to teaching, teaching and learning of specific topics in
university mathematics, specialised content knowledge, teacher training, complex
analysis.

INTRODUCTION
More than one hundred years ago, Felix Klein (2016) acknowledged that there is a
“double discontinuity” in teacher education in mathematics. When a pupil enters
university he or she does not see the connection of elementary mathematics taught in
school with the formal mathematics in university, and when teaching in school he or
she does not see how the university maths informs his or her practice (Hefendehl-
Hebeker, 2013; Vollstedt, Heinze, Gojdka, & Rach, 2014). Within mathematics
courses at university, it is thus necessary to highlight connections between the
mathematics future teachers are taught in lectures at university and the mathematics
they will teach in schools (Prediger, 2013).
The project “Spotlight-Y” aims for an institutional link of mathematics and
mathematics education and at providing students with the experience to relate content
from the lecture in complex analysis to their future teaching practice. By helping
them to design tasks for pupils with mathematics from a “higher standpoint” (Klein,
2016), the students get the opportunity to try out these tasks with pupils from local
schools.

THEORETICAL CONSIDERATIONS
Ball, Thames, and Phelps (2008) proposed to distinguish the mathematical content
knowledge (CK) of teachers further into common content knowledge (CCK) and
specialised content knowledge (SCK). The first is mathematical knowledge that does
not depend on a specific profession and is known to pure mathematicians as well as

54 [Link]:indrum2018:171762
teachers and others, and the latter is mostly useful only for teaching mathematics. So
for a teacher, not only pedagogical content knowledge (PCK) is important and unique
for the profession but also SCK.
The international study TEDS-FU established that PCK is the most important factor
to judge the didactical problems and opportunities of a classroom situation and that
CK is the most important factor to assess the utterances of pupils (Blömeke et al.,
2014).
Thus, teaching mathematics requires an interplay of different kinds of knowledge for
the teacher and it seems worthwhile to try to specifically address SCK when doing
teacher education programs at university.
Winsløw and Grønbӕk (2014) proposed to study the phenomenon of the second
discontinuity by the study of different mathematical praxeologies between university
and high school. In their paper, they are able to identify certain challenges for this
transposition. For university students, one of these challenges is the establishment of
a school praxeology for a problem typically solved with a university maths
praxeology (their example is the method of least squares for simple linear regression).
The authors claim that the main related difficulties lie in spawning capacities for
students’ autonomous research and handling (ir-) relevant literature. This is where
Spotlight-Y is situated. Our approach can be seen as complementary: Instead of
focusing on how the mathematical praxeologies may change and what challenges
may hinder, we try to focus on how the didactical praxeologies are combined with the
mathematical praxeologies from university.
Research question
In this paper, we want to foster the understanding of the combination of PCK and
SCK by the Y-model that structured the lecture (see next section). We try to address
the following question: “In what ways do students combine specialised and
pedagogical content knowledge within the preparation and implementation of their
learning environments?” For this purpose, we will describe example projects.

OVERVIEW OF THE PROJECT

At the University of Bremen the lecture in complex analysis is the final course in
mathematics for future teachers. Historically, many teacher students in Bremen are
known to view this as a course you have to pass to get through your studies, but
which is thought of having no direct connection to what is taught in school. From the
standpoint of the designer of the curriculum of teacher training in mathematics,
complex analysis was seen as a course which brings different aspects of mathematics
together and may help to get a more holistic view on various areas of the subject, e. g.
polynomials, trigonometric functions, power series or conformal mappings. As such,
it should be a good starting point to get a “higher standpoint” as Klein (2016) put it.
“Spotlight-Y” is a design research project within the project “Schnittstellen gestalten”
at the University of Bremen, funded by the German Federal Ministry of Education

55 [Link]:indrum2018:171762
and Research (BMBF) that started in fall 2016. Its general aim is the interlock of the
scientific disciplines of mathematics and mathematics education. Our students shall
see that elements of school mathematics, in particular all relevant classes of
functions, can be understood from the complex setting which explains more than the
real picture. In three design cycles we develop
1) the structure of the lecture on complex analysis,
2) the specific branch for future teachers with a focus on identifying phenomena
of complex analysis relevant for teaching mathematics in upper classes of
secondary schools and creating exploratory learning environments, and
3) a day for eXperimental Mathematics (XMaSII) which local pupils from upper
classes of secondary school (Sekundarstufe II) attend to work on the learning
environments by the teacher students.
It was mandatory for our students to use the free dynamical geometry software (DGS)
GeoGebra ([Link] for implementation because we assumed that
this would make difficult concepts easier to present and provide a suitable technology
to really be able to explore a phenomenon. Besides, our students get an opportunity to
create authentic materials, plan and structure group work with pupils from secondary
schools in 11th or 12th grade and gain experience in working with them. In particular,
they get experience with pupils outside internships and their practical semester
teaching mathematics for rather gifted learners.
Some of the general research questions in the project are: How do students combine
mathematics from the complex analysis course, their general knowledge on
mathematics education taught earlier in their studies and simultaneously to the course
on complex analysis (a seminar on task design)? In which ways do they handle their
learning environments and the material for the pupils? How do the future teachers see
mathematics education, do they consider it as a scientific discipline or rather a
collection of methods to teach school mathematics?
After about two years, we will report on our design principles, experiences and
empirical data we collect during the design project to establish a transfer package for
other mathematics lectures. In the second year, we perform a first transfer to the
stochastics lecture in the Bachelor for future mathematicians and mathematics
teachers.
Structuring a mathematics lecture: The Y-model
In “Spotlight-Y” we adopt a Y-shaped model for the course in complex analysis. The
course takes place in the fall semester and is attended by pure/applied mathematics
students in the Bachelor as well as teacher students in their first year of the Master
programme. The course takes one semester and the participants are split up around
Christmas after two thirds of the semester (see Figure 1a). The part for everyone
covers the introduction of holomorphic and conformal mappings, line integration and
Cauchy’s theorem up to the residue theorem. The specific branch for future

56 [Link]:indrum2018:171762
mathematicians continues with more advanced topics like invariance of integrals with
respect to homotopy or homology of paths, analytic continuation, the Riemann
mapping theorem or the prime number theorem etc., depending on the lecturer. In the
branch for teachers we cover geometric properties of complex functions and an
introduction to Riemann surfaces. Furthermore, we give the students time to work on
their learning environments. The course is finished with a written or an oral exam.

Figure 1: a. Lecture with Y-model and seminar on task design (left), b. The different
components that support the task design process (right)
In fall 2016/2017, the first design cycle in the course on complex analysis started and
the next semester was used to start analysing data we gathered during the first
implementation. In fall 2017/2018, the second design cycle lead to improvements
with respect to the profession-specific branch for the future teachers. In fall
2018/2019, the last cycle will start: A third run during the course on complex analysis
will be devoted to the final answers to our research questions and in summer 2019 a
second transfer to the course on stochastics will be performed.
Elements of the task design process
In order to prepare the tasks for the pupils, the students have to bring input from three
different components together (cf. Figure 1b). They make use of the mathematical
CK from the general branch of the lecture, they activate their SCK from the branch
for future teachers and bring those branches together with their PCK, i. e. heuristics
on task design from their seminar on mathematics education.
Methodology
For the whole project we collected several data. First, the students had to write a
“preflection”, a short reflection which is not after an action on an action but rather the
anticipation of actions, aims and sequences: They should fix which phenomenon they
want to discuss in their learning environment, what the pupils are supposed to
discover, the planned sequence of events, expected difficulties and the mathematics
from the lecture that is directly used. Immediately after the implementation of
XMaSII, the students filled in a questionnaire named “Ad-hoc notes” to write down
own executed actions, recapitulate the execution and match it with expected or

57 [Link]:indrum2018:171762
anticipated occurrences. In order to pass the course, the students also needed to write
a reflection on their project and their learning. They had to discuss their topic and
how they found it, describe detailed the phenomenon to be explored by the pupils, the
organisation of their tasks, the schedule of the implementation etc. In a final section,
we explicitly asked whether there were certain elements of the course on complex
analysis that changed the relationship of the students to mathematics as a science and
mathematics as a school subject. In total, 19 students participated in the study. We
also conducted guided interviews with two students of four groups each to get more
insight to the aspects above (and some more). We omit details here since this paper
deals mainly with the learning environments the student groups created.

EXAMPLE PROJECTS
We describe two case examples. Group A created a learning environment called
“Differentiation as linearisation” and Group B worked on “Polynomials of infinite
degree”. Two members of Group A have also been interviewed. Two other groups
created tasks for spherical geometry, another one introduced complex numbers and a
sixth group also worked on power series. As mentioned above, the tasks themselves
focus on phenomena that are rooted in complex analysis and can be explored by the
pupils from high school by the means of a DGS.
“Differentiation as linearisation”
Group A looks at real differentiation, i. e. differentiable functions defined on
(subsets) of the real numbers to the real numbers. The students concentrate on the
image that a differentiable function locally looks like a straight line by magnifying
the graph around a given point, say (p, f(p)). The students emphasise that the
derivative at a point can be interpreted as the slope of the function, respectively of the
tangent at the graph of the function, at this point (which is CCK). They create a
sequence of tasks and subtasks within their learning environment that aim at
magnifying the graph of a function (use of PCK) and lead to answer the question
which exponential function has itself as a derivative. However, the students do not
explicitly clarify what this has to do with linearisation besides the optical appearance.
These two “Grundvorstellungen” (“basic mental models”, Greefrath, Oldenburg,
Siller, Ulm, & Weigand (2016, p. 101)) of the derivative at a point, “tangent slope”
and “local linearity”, are usually distinguished in mathematics education literature
(see e. g. Greefrath et al. (2016) who also discuss “local rate of change” and
“amplification factor”). However, we do not see this as a lack of understanding of
different Grundvorstellungen with the students. Rather, group A focuses on a
geometric point of view and many of their tasks focus on the idea of zooming into the
graph of a function to get a more and more straight line (PCK, SCK) on the computer
screen (e. g. “Plot with GeoGebra the function f(x) = x³ - 0.75x² - 9x + 1. Then draw
the point A(1, f(1)) […] Zoom in the neighbourhood of point A into the drawing.
What do you notice? Try to describe your observation.” In a later task: “Describe
with this magnified image how to calculate the slope in point A approximately (Hint:

58 [Link]:indrum2018:171762
Remember the slope of a secant).” (own translation)). In other subtasks, group A
provides the pupils with GeoGebra worksheets they have prepared before, where they
had plotted several functions and their “slope function”, and wants the pupils to make
observations about these. In another task at the end of the task sequence, they display
an exponential function f(x) = ax and its slope function with a parameter a which can
be changed with a slider and ask for which a these functions coincide.
The definition of derivative at a point with limits of difference quotients is not
stressed in detail but the GeoGebra worksheets nicely implements that when zooming
into the graph two points on it approach the point (p, f(p)) (one from left, another
from right) to show that the slopes of the secants through these points eventually
yield the slope of the tangent, no matter if one approaches from left or right.
However, since the students do not provide counterexamples, it does not seem very
likely that the concept of derivative can be fully understood, since the class of
functions the students used are pretty standard and “friendly” in the way that they all
really look locally straight (e. g. s given by s(x) = x²∙sin(1/x), s(0) = 0, is
differentiable at 0 but oscillates very much around the origin and therefore hardly is
locally straight).
An additional task in the learning environment is about rotation-dilation, which is an
interpretation of the complex derivative discussed in detail in the first third of the
lecture and exercise classes (SCK). A hexagon and two points are displayed in the
Cartesian plane and pushing around the points changes the hexagon according to the
multiplication by the complex number associated to the points (see Figure 2a, SCK).
However, such a multiplication is not made explicit in the task. Rather, the pupils
shall measure distances and angles, and the task seems rather unrelated to
differentiation. Not even a function is in play. The complex interpretation of
derivative is transformed into a task about rotation-dilation, but not related to the
previous tasks on differentiation. This is not surprising since the idea of magnifying
pursued in the real approach does not shed any light on dilation-rotation. In the
interview with two of the three group members we asked for an intuitive, vivid
meaning of a holomorphic function. The students were unable to argue and did not
even respond with the keyword of (local) rotation-dilation.
Thus, in terms of our framework, the connection of SCK and PCK did not happen as
was hoped for. One problem seems to be that the necessary SCK for the complex
derivate could not be utilised by the students, while they argue very well about
prerequisites the pupils should have for their learning environment such as secant,
tangent, linear functions etc. (SCK for the real derivative). The students of group A
did not take part in the parallel seminar on task construction. Therefore, we
hypothesise that they followed an alternative strategy to construct tasks: They
scrolled through mathematics education literature to find a suggestion on how to fill
the aspect of local linearity with life. In the interview, one of the students said,
roughly speaking, “From the earlier lectures [in maths education] one knew at least

59 [Link]:indrum2018:171762
where to research, because it is clear that one cannot have everything stored [in one’s
mind] and fetch it up immediately, so one needs to do research” (own translation).
From a real point of view in terms of concept formation, the learning environment
seems rather unproblematic. We believe this is the case since didactic aspects of real
functions and derivatives are very present in the courses on mathematics education
the students in Bremen are required to take. However, there does not seem a thorough
idea on how to relate the complex interpretation of the derivative of a complex
function to the real setting. Thus, intuition of differentiation is not coherently
transported from one to the other setting. Also, group A does not argue whether their
real images of derivatives find their counterpart in the complex setting.

Figure 2: a. Dilation-rotation of a hexagon (left, from group A’s GeoGebra worksheet),

b. -ln(1-x) and its sixth Taylor polynomial at 0 (right, from group B’s GeoGebra
worksheet).
“Polynomials of infinite degree”
Group B has two general aims with its learning environment about power series
which they steadily call “polynomials of infinite degree”. Firstly, in school
polynomial functions form a very frequently used class of functions (PCK) and
power series seem a quite direct generalisation of it (CCK). Secondly, nearly all
functions that appear in school can be represented by power series (even though this
is usually not made explicit in schools, PCK). For a motivation why power series can
be useful for pupils, group B argues that using power series one may be able to
explain easily why the derivative of the exponential function is itself and the
derivative of the sine function is the cosine function etc. (PCK, SCK).
As necessary mathematical background from the complex analysis course group B
recognises holomorphic functions, power series and their derivatives, entire and non-
entire functions, radii of convergence of power series and the power series expansion
theorem. They relate all these ideas in a correct manner to each other. Nevertheless,
the language the students use in their written reflection and on the task sheets for the
pupils is quite tenuous: The words polynomial and power series are not clearly

60 [Link]:indrum2018:171762
differentiated. It may be possible that the students wanted to use the same word
“polynomial” with the addition of “finite degree” or “infinite degree” because it is
known to the pupils. Group B recognises that the distinction of entire and non-entire
functions is relevant for school and illustrates this with examples like -ln(1-x) and
(x+1)1/2 (see Figure 2b, PCK and SCK).
In their preparation and reflection the students raise the question to what extend
convergence issues of power series or why Taylor approximation works (CCK)
should be covered (PCK), but they consider this too difficult and not anymore
relevant for the discovery of the phenomenon for the pupils that previously known
functions – and basically all encountered in school – can be expressed differently. In
fact, the group recognises that arguing in whatever way that Taylor expansion
requires differentiation and defining elementary functions by their series expansions
may lead to circular reasoning (CCK).
The mathematics background is well explained with a few mistakes mostly in
language and notation. Group B’s learning environment really allows to discover one
of the fundamental phenomena of complex analysis, namely that every holomorphic
is at least locally a power series, in the real and school-related setting. In the
GeoGebra environment, several functions and their Taylor polynomials of varying
degree, adjustable with a slider, are presented. This shows a creative part in the work
of our student group, since it visualises very nicely the approximation of a function
by its Taylor polynomials. However, the distinction of entire and non-entire functions
is mathematically deep and relevant for school (SCK). The question of why the
representation of a function by a power series is relevant for pupils is answered with a
deepening of their insights into higher mathematics and the application to calculate
derivatives similar to the procedure with polynomials. Unsurprisingly, as described
above, the problem of convergence is left aside.
Nevertheless, the task sheets for the pupils lack rigor and do not provide consistent
language at certain places. For example, in a “remember box” group B writes “A
function f(x) is an entire function if it is a polynomial of degree n or is a function
which can be written as a polynomial of infinite degree P(x) for every x∈ℝ
[grammatical errors deleted, E.H.]”, and in another box they write “A function f(x) is
a non-entire function if it is NO polynomial of degree n but can be approximated by a
polynomial of infinite degree. However, this polynomial of infinite degree
approaches the function f(x) only in some interval. It [the function] is not defined for
all x∈ℝ or does not have a tangent everywhere [slight grammatical adaptions, E.H.]”.
The absence of the domain of f is usual in school and probably left out for this reason.
It could have been stressed more detailed that a power series expression for an entire
function is globally valid, i. e. that P equals the given function on the whole of ℝ. In
the definition of non-entire functions, on the other hand, the local approximation by a
power series is mentioned correctly (assuming that the group meant that the sequence
of Taylor polynomials approximates the original function) but it is not clarified that

61 [Link]:indrum2018:171762
this power series actually equals the function in some interval containing more than
one point. Also, it should have been stressed that this local phenomenon is valid for
every point in the domain of the non-entire function but the corresponding power
series may change. Another problem is the mentioning of tangent. Tangents do never
show up in the learning environment nor in the students’ reflection again and it is
very unclear what the students wanted to say.
In this case the connection between SCK and PCK seems to be working much better,
although there are some flaws in the formulation of the definition. These problems
may be rooted in misconceptions in the CCK.

FINDINGS
The learning environments differ heavily in mathematical rigour, creativity and
recognisability of complex phenomena in either a real setting or else. In the section
on the example projects we described that Group A showed little understanding of the
geometric meaning of the derivative of holomorphic functions (SCK) (or, at least,
they did not implement this well) but otherwise provided a nice example driven and
exploratory environment. Group B, on the other hand, translated the notion of power
series representation of entire and non-entire functions to real examples and made a
complicated phenomenon accessible to pupils (PCK and SCK). Due to the
complexity, arguments on why and where power series converge and how to find
Taylor series expansions have not been dealt with (CCK). The written report is
mostly correct, however the material for the pupils needs improvement with respect
to consistency in language, mathematical exactness on the level of pupils and
correctness from a formal point of view.
Still in evaluation phase, i. e. coding the interviews, we are nevertheless already able
to state first results regarding our design process:
1. Most students seemed more engaged in creating the tasks than in the lecture.
2. One has to make connections between SCK and PCK explicit, many students
fail to see them on their own.
3. It has been evident from the interviews that basic mathematical content
knowledge in complex analysis was not available two months after the exam,
but the content of the tasks they created was.
4. One needs to clarify and guide how to implement a topic from complex
analysis when complex numbers are not available for pupils or shall not be
introduced.
5. A possible improvement for XMaSII could be to restructure the day into
different sessions or workshops: 1) Introduction to complex numbers and
geometry of the complex plane either by the lecturers or some student groups
and then 2) working groups dealing with different phenomena from complex
analysis.

62 [Link]:indrum2018:171762
ACKNOWLEDGEMENT
We would like to thank the German Federal Ministry of Education and Research
(BMBF) for its support in the subproject “Spotlights Lehre” within the project
“Schnittstellen gestalten” at the University of Bremen.

REFERENCES
Ball, D., Thames, M. H., & Phelps, G. (2008). Content knowledge for teaching: What
makes it special? Journal of Teacher Education, 59(5), 389–407.
doi:10.1177/0022487108324554
Blömeke, S., König, J., Busse, A., Suhl, U., Benthien, J., Döhrmann, M., & Kaiser,
G. (2014). Von der Lehrerausbildung in den Beruf – Fachbezogenes Wissen als
Voraussetzung für Wahrnehmung, Interpretation und Handeln im Unterricht.
Zeitschrift für Erziehungswissenschaft, 17(3), 509–542. doi:10.1007/s11618-014-
0564-8
Greefrath, G., Oldenburg, R., Siller, H. S., Ulm, V., & Weigand, H. G. (2016).
Aspects and “Grundvorstellungen” of the concepts of derivative and integral.
Journal für Mathematik-Didaktik, 37(1), 99–129. doi:10.1007/s13138-016-0100-x
Hefendehl-Hebeker, L. (2013): Doppelte Diskontinuität oder die Chance der
Brückenschläge. In Chr. Ableitinger, J. Kramer, & S. Prediger (Eds.), Zur
doppelten Diskontinuität in der Gymnasiallehrerbildung. Ansätze zu
Verknüpfungen der fachinhaltlichen Ausbildung mit schulischen Vorerfahrungen
und Erfordernissen (pp. 1–15). Wiesbaden: Springer. doi:10.1007/978-3-658-
01360-8_1
Klein, F. (2016). Elementary mathematics from a higher standpoint (Vol. 1). Berlin:
Springer. doi:10.1007/978-3-662-49442-4
Prediger, S. (2013): Unterrichtsmomente als explizite Lernanlässe in fachinhaltlichen
Veranstaltungen. In Chr. Ableitinger, J. Kramer, & S. Prediger (Eds.), Zur
doppelten Diskontinuität in der Gymnasiallehrerbildung. Ansätze zu
Verknüpfungen der fachinhaltlichen Ausbildung mit schulischen Vorerfahrungen
und Erfordernissen (pp. 151–168). Wiesbaden: Springer. doi:10.1007/978-3-658-
01360-8_9
Vollstedt, M., Heinze, A., Gojdka, K., & Rach, S. (2014). Framework for examining
the transformation of mathematics and mathematics learning in the transition from
school to university. In S. Rezat, M. Hattermann, & A. Peter-Koop (Eds.),
Transformation - A Fundamental Idea of Mathematics Education (pp. 29–50).
New York, NY: Springer. doi:10.1007/978-1-4614-3489-4_2
Winsløw, C., & Grønbӕk, N. (2014). Klein’s double discontinuity revisited.
Recherches En Didactique Des Mathématiques, 34(1), 59–86.

63 [Link]:indrum2018:171762
On students’ understanding of Riemann sums of integrals of functions
of two variables
María Trigueros1 and Rafael Martínez-Planell2
1
Instituto Tecnológico Autónomo de México, mtriguerosg@[Link]; 2University
of Puerto Rico at Mayagüez
APOS (Action-Process-Object-Schema) Theory is used to pose and test a conjecture
of mental constructions that may be used to understand the relation between integrals
of two variable functions over rectangles and corresponding Riemann sums.
Interviews with ten students who had just finished a multivariable calculus course
showed that the conjectured mental constructions are necessary.
Keywords: Functions of two variables, APOS, integral calculus
Multivariable functions and multivariable Calculus are important in engineering and
the natural sciences as a tool for modelling. Their learning has received more
attention lately from the Mathematics Education community. Starting with the
analysis of students’ understanding about two-variable functions (for example:
Trigueros and Martínez-Planell, 2010; Martínez-Planell and Trigueros, 2012)
researchers have documented students’ difficulties and have shown that the transition
from one-variable Calculus to multivariable Calculus is far from being smooth.
There are few studies in the literature that deal with students’ difficulties and
understanding of the integral multivariable Calculus (Jones and Dorko, 2015;
Martínez-Planell and Trigueros, 2017). In one of these few studies, McGee and
Martínez-Planell (2014) showed that a course based on lectures did not promote
students’ understanding, while activities introducing the use of semiotic chains and
the development of synergy among representations helped students understand this
concept. The research questions are:
What constructions relating double integrals and Riemann sums are evidenced by
students who finished a Multivariable Calculus course based on lectures?
What constructions may be needed to relate double integrals and Riemann sums?

THEORETICAL FRAMEWORK
APOS theory (Arnon et al. 2014) is used in this study to analyse possible mental
constructions by students who have already taken a course on multivariable calculus.
We only summarize the main structures of this theory. An Action in APOS Theory is
a transformation of a previously constructed mathematical object that the individual
perceives as external in the sense that students need some guidance and are not able
to justify what they do. When an Action is repeated, and the individual reflects on
what he or she does, it may be interiorized into a Process. A Process is perceived as
internal in the sense that it has meaningful connections to other mathematical

64 [Link]:indrum2018:174222
knowledge of the individual. A Process allows the individual to imagine doing the
Actions without actually doing them, to omit steps, anticipate results, and to justify
the Process. Different Processes may be coordinated to form new Processes. When
the individual needs to apply Actions on a Process, it can be encapsulated into an
Object. When an individual shows a Process or Object conception of a mathematical
notion we say that the individual “understands” the notion. A Schema is a coherent
collection of Actions, Processes, Objects, and other Schemas, that the individual uses
to work with problems related to some mathematical notions. Schemas are not used
in this paper.
To analyse students’ work using APOS Theory, a conjecture of those constructions
that may be used to understand a specific mathematical notion is designed. This
model, called a genetic decomposition (GD), does not pretend to be unique and needs
to be tested with research data. The GD may be revised and expanded in successive
cycles of research, teaching material development, and implementation. This research
cycle makes it possible to use APOS theory to be better suited to future research
needs to study the multivariable integral calculus.

GENETIC DECOMPOSITION
We only present a portion of the GD of integrals of functions of two variables over
rectangles. Its development is based on mathematics, on the researchers’ teaching
experience, and data from the research literature: mainly, ideas about representation
registers (Duval, 2006) described in the study by McGee and Martínez-Planell (2014)
and the ideas of “orienting pre-layer” and “product layer” described by Sealy (2014),
which stress the need of attending to the individual meaning of the product f ( xi )x
and its components in the construction of integrals of one-variable functions.
The GD starts with pre-requisite constructions which include: a Process conception of
two-variable functions and volume of prisms as Object.
Actions are performed on a given two-variable function in any representation with
domain restricted to a rectangle, to produce the geometric representation of the
restricted domain as a subset of 3D space. Actions are performed on the same
function to obtain values of the function on the given domain and to represent them in
the space as points and/or curves in the graph of the function. These Actions are
interiorized into a treatment or conversion Process to represent the graph of the
function over the given rectangle together with the rectangle so that the student can
imagine the relation between function and rectangular domain as a graph in space.
Actions of evaluating the given function of two variables at a specific point of a given
sub-rectangle of its domain, multiplying it by the length and width of the rectangle to
form a product of the form f (a, b)xy are done. These Actions are interiorized into a
Process which can be coordinated with conversion Processes between different
representations of function, rectangle, and given point, to imagine the product as the
volume of a rectangular prism in space.

65 [Link]:indrum2018:174222
Given a continuous function in different representations defined on a rectangle, with
the function simple enough so that its maximum and minimum values on the
rectangle may be recognized without doing any explicit computation, the Action of
obtaining an overestimate and an underestimate of the product f (a, b)xy is taken.
These Actions may be interiorized into a Process that enables to imagine the
existence of points (a, b) where underestimate and/or overestimate of the product
f (a, b)xy are attained. This Process is coordinated with a treatment or conversion
Process to draw a rectangular prism corresponding to over and/or underestimate in
space. Actions are performed to change the chosen point to construct a prism that
better approximates a given exact value of the integral. These Actions are interiorized
into a Process that enables the recognition that for such continuous function, there is a
point somewhere on the rectangle that will produce the exact value of the volume
between the graph of the function and its rectangular domain.
Given two small specific positive integer numbers, n and m, the Action of
subdividing given intervals [a,b] and [c,d] into subintervals of equal length both
numerically and geometrically is done to obtain a subdivision of the rectangle
[a, b][c, d ] . These Actions are interiorized into the corresponding Process. Given a
continuous function f defined on the rectangle, the Action of choosing a prescribed
point ( xi , y j ) on each sub-rectangle of the given partition and producing the products
f ( xi , y j )xy , and the corresponding sum, interpreting this sum geometrically,
numerically, symbolically as an extended sum, symbolically using sigma notation,
and verbally, may be interiorized into a Process that enables imagining forming such
sums in different representations for the collection of sub-rectangles in any partition
of any given rectangle.

METHOD
Ten students were chosen by their professor to be interviewed at the end of a
multivariable calculus course selecting four over-average, three average, and three
under-average. The course was completely based on lectures. The interviews lasted
46 minutes on average. Students answered a set of questions designed in terms of the
GD and also related to what was covered during the course, and produced a written
response while sharing their thoughts out-loud. The interviews were recorded,
transcribed, individually analysed, and results were negotiated by the two researchers.
Students’ responses were analysed according to the GD, while keeping notes on
unexpected responses and other difficult to classify observations. These were the
questions used:
1a. The following is the complete graph of function z  f ( x, y) . Represent the domain
of f in the figure (Figure 1).
1b. Let g(x, y) = x2 + y be a function with domain restricted to 0 ≤ x ≤ 2 and 1 ≤ y ≤ 2. Use
the coordinate system given in the following figure to represent the domain in three-
dimensional space [An empty drawing of the first quadrant was given].

66 [Link]:indrum2018:174222
1c. The above functions f and g are the same [Figure 1 was given again here and in
the rest of the problems]. If x  2 and y  1 , what is the numerical value of
f (0,1)xy ? What does it represent geometrically?

Figure 1: (repeated in each part of problem 1 except 1b)

1d. Let x  2 and y  1 . How does f (0,1)xy compare with  f ( x, y)dA ?
D
[No

numerical computations are needed in parts d, e, f, and g.]

1e. How does f (2, 2)xy compare with  f ( x, y)dA ?
D

1f. Is there any point (a,b) in the domain D of f such that f (a, b)xy is equal to
 f ( x, y)dA ?
D

1g. Let x  1 and y  1/ 2 . Consider the Riemann sum f (0,1)∆x∆y + f (0,1.5)∆x∆y +

f (1,1)∆x∆y + f (1,1.5)∆x∆y of the integral  f ( x, y)dA . What does the Riemann sum
D

represent geometrically and how does its value compare to that of  f ( x, y)dA ?
D

Note that problems 1a and 1b are essentially the same in different representations.
They both test the portion of the GD dealing with recognition of rectangle and
function. Problem 1c gives information on the portion of the GD dealing with
forming one term of a Riemann sum. Problems 1d, 1e, and 1f relate to the portions of
the GD dealing with underestimate, overestimate, and exact value. Problem 1g gives
information on the portion of the GD dealing with a partition and Riemann sum.

RESULTS
On function and domain of a function
Many students showed they had not constructed the concept of two-variable function.
They gave evidence of considering these functions in terms of a correspondence rule,
and showed difficulty interpreting functions given graphically. Moreover, these
students also showed not to have constructed the concept of domain of the function.
Most of them considered that the domain of a two-variable function should include

67 [Link]:indrum2018:174222
information about the function, since it had to be represented in 3D space. Eight
students showed difficulties similar to those of Luis, as exemplified in the following
discussion with the interviewer (in Problem 1a):
Luis: I can tell you what the domain is but if I don’t have a function I don’t think
I can tell you the exact point where each of the points in the graph is…
Interviewer: So, is the graph part of the domain?
Luis: No, the domain is obtained from the graph. I can obtain the domain having
the function but to do so I have to define the function.
After some discussion:
Interviewer: So the domain, is it only x and y or may it also include z?
Luis: The domain may include the z.
This example shows how Luis needs a correspondence rule to determine the domain
of the function. It also evidences that he considers the function itself should be part of
the domain of the function. Other students showed this difficulty.
Students’ responses pointed to a need to pay attention to the different representations
of functions in 3D space and to have students do treatments and conversions between
representations. In Problem 1a, some of them quickly represented the rectangular
domain as part of the given figure in 3D space. However, when the function was
given symbolically their notion of domain seemed to change. This shows that
recognizing the domain of a two-variable function is a construction that needs the
interiorization of Actions on functions given in different representations. These
difficulties as well as the counterfactual belief of teachers that students may easily
generalize concepts for one-variable functions to multivariable functions have been
reported before (Martinez-Planell and Trigueros, 2012).
Area and volume
Students also showed an unexpected confusion between area and volume when they
described graphs of functions in 3D-space. This difficulty surfaced in Problem 1c.
All students were able to calculate the value of that product; however, they were in
trouble when explaining its geometrical meaning. Brian, for example, explained:
Brian: … this part, f(0,1) would be a point in this graph here. Change in x, change
in y, I am not a hundred percent sure… that would be an area then, of the
surface, or the entire function…
And later:
Interviewer: Can you tell me what does the double integral of f(x,y)dA represents?
Brian: dA is the area of the function, the area of this figure,
Interviewer: The area of the surface?
Brian: Yes… of the surface on the given domain….
Other students, as Luis, showed confusion:
Luis: The area of the figure, that is, the area of the function which in this case is that

68 [Link]:indrum2018:174222
 figure [referring to the graph in Figure 1].
Interviewer: Like the area of a surface?
Luis: Exactly
Interviewer: So if it had units would it be like square inches or square centimetres? What
units would the double integral have if x, y, and z had units?
Luis: Cubic
Interviewer: Cubic; then, would it be area?
Luis: It would be volume...
After some discussion:
Interviewer: …Let’s suppose that this other paper that I am raising here is the graph of
the function [He raised a sheet of paper] What volume are we talking
about?
Luis: …The volume is the one of this paper…since I have a function and I’m
integrating in the values of the function then what I’m going to get are z,
small z’s of what the function is, I’d be getting the volume of the figure.
Interviewer: ... So you pointed to the paper that is floating. But, does it have a volume?
Luis: ... Yes, it has a change in x, it has a change in y, and the z is the one from
the function, so I say that it has a volume.
Other five students showed the same confusions. The above excerpt exemplifies that
a student can describe the individual components of f (0,1)xy but might not be able
to do the Action of putting them together to interpret it as the volume of a rectangular
prism, even if they can calculate the result of the product by doing the Action of
substituting the given values in the expression, as conjectured in the GD. This
difficulty is possibly related to the fact that these students have not constructed space
as an Object, which does not allow them to imagine what their teachers mean when
they talk about a surface in space and the double integral as related to the volume
under a surface (Trigueros and Martínez-Planell, 2010).
As considered in the GD, these difficulties make it impossible for students to do the
necessary Actions on the function restricted to a rectangle to represent the domain
geometrically as a subset of space, and to interiorize the Process to imagine the
relation between a restricted region on the domain and the function. The lack of all
these constructions becomes an obstacle to understand other related concepts,
including double integrals, as will be shown below. Students who do not show these
constructions may not follow teachers’ explanations; they would be confused and
resort to memorization to respond to exam questions.
Only one student, Farid, gave evidence of the pre-requisite constructions described in
the GD. He was successful in explaining Riemann sums and double integrals.

69 [Link]:indrum2018:174222
Riemann sums, underestimates, overestimates and double integrals
Most students had many difficulties working with problems dealing with Riemann
sums and their relation to double integrals. Even after the interviewer explained to
some of them that f (0,1)xy was a volume and drew it, their difficulties did not
enable them to make the whole construction as Brian showed in Problems 1d, 1e, and
1g:
Brian: So the Riemann sum would be the approximation of the area [sic] under this
figure [referring to the graph in Figure 1], obviously it wouldn’t be as
precise as the value of the integral. Let’s see... so geometrically 0,1, x, y,
let’s draw a square here like this [he is now evaluating and drawing
rectangular prisms]... 0, 1.5, maybe another square closer this way, higher...
1,1 we are still at x 1 and even higher here… like this, change in x change in
y... change in y being 1/2, I don’t think we get from 1 to 2 with 1/2 [He
seems to believe that since y  1/ 2 the prisms will be restricted to the
region 1  y  1.5 . He might think of y as “change in y” were the “change”
is taken from the initial y value in point (0,1).], so the integral would give
this area [sic] here, a figure more or less like this...
Interviewer: You said area...
Brian: [Interrupting] Volume, I mean volume, sorry... yes, volume of the integral.
This would give us something more stepwise... let’s see if I can draw it here
like this... 1,2,3,4, like this, a series of cubes like this, stepwise,
approximating, not all of this, but only this half here… [See Figure 2].
Interviewer: Do you mean the left hand part of the solid?
Brian: Yes, the left hand part of the solid would be what is approximated with this
Riemann sum.
Although Brian was able to construct the meaning of volume, his construction was
not right, the boxes he drew filled only the left-hand side of the rectangle. It seems
that Brian could do the Actions to construct the prisms but he did not interiorize those
Actions into the Process that would enable him to imagine all the constructions
needed to relate Riemann sums and double integrals.
An interesting result of this experience was that even though students showed many
difficulties during the interview, some of them, like Brian, showed evidence of doing
some of the expected constructions during the interview. Others reflected during the
interview and constructed meaning. This was the case of Victor who had considered
f (0,1)xy as an area. When discussing Riemann sums, and after being told that this
product represents a volume he explained:
Interviewer: So you drew a little box.
Victor: Exactly a little box as we know that delta x would be 2 and delta y 1; a
rectangle with width 2, eh, length 2 with 2 and height 1.

70 [Link]:indrum2018:174222
 Interviewer: Then how do you compare those volumes.
Victor: Ok, now I understand, this f(0,1) delta x delta y is only the volume up to
this point, I mean up to a certain height, and then, the double integral on
that same area that we put on xy is, let’s say, the same box but with a height
that varies with the function. Now this is it!

Figure 2: Brian’s drawing for Problem 1d, 1e, 1g (respectively)

Interviewer: But which is larger, what is smaller, are they equal?

Victor: No, no, they are completely different, the larger is that obtained from the
double integral since the height is higher.
Victor: f(2,2) delta x delta y is the box with dimensions over D…and then this is the
same equation as before, but f(2,2) is higher so the volume there is larger.
Victor could do the Actions needed to compare volumes of prisms obtained from
different values of the function. Another student, Farid, evidenced he could imagine
forming one term of a Riemann sum, as discussed before. He also showed to have
done the constructions necessary to imagine volumes of prisms and their role in
Riemann sums. When comparing the volume of the prism in Problem 1d with the
double integral:
Interviewer: And what does that represent? [Referring to the double integral.]
Farid: That represents the volume between the surface and the plane, the domain...
Interviewer: Then, how do those two numbers there compare?
After some doubts:
Farid: Represented this part [referring to the product], now, this product would be
smaller, than the double integral, because this here is a, I represented it as a
cube, given the value of f at that point, while this is the double integral of
everything, of all the function x,y over D, so this value seems bigger
Interviewer: Which one?
Farid: The value of the double integral over D of f(x,y)dA
He was able to compare the volume of the prisms with the double integral. When he
had to decide if there would be a prism with volume equal to the value of the double
integral, a problem that was impossible for all the other students, he explained:

71 [Link]:indrum2018:174222
 Farid: … The thing that comes to my mind when thinking on an inequality is the
sandwich theorem…that there... must exist then a value for x and y that
could be named a and b that is equal to the double integral on dA.
When discussing Riemann sums with a specific partition, most students could not
work with the problem, even with help from the interviewer. As was shown by Brian
in the previous example, some students, including Victor, imagined drawing several
prisms or boxes that shared the base, and only had different heights. Those students
showed they could do the Action of changing the height of a given prism but not that
of partitioning the domain into small areas of the same size. Victor could describe the
sum of the prisms’ volume, at first he said that the Riemann sum was always an
approximation to the volume under the surface, although later he reconsidered:
Victor: No, the Riemann sum is an approximation and if you take more
subintervals, ah! If you take more subintervals, those were the squares, that
one uses, the Riemann sum is a closer approximation and that
approximation would be closer with more subintervals, and the double
integral is the exact value.
Only Victor and Farid seemed to have interiorized the Action of forming a partition
into a Process they coordinated with the Process of selecting heights for each
subrectangle into the Process of calculating the volume corresponding to the prisms
to approximate the volume under the surface.
DISCUSSION AND CONCLUSION
Results from this experience show that most of these students demonstrate a very
limited understanding of two-variable functions and of those concepts associated to
the construction of the double integral of a two-variable function and its geometrical
interpretation. Only two students showed some understanding, although one of them
relied mostly in memorized facts that he could use appropriately in most cases. This
student seems to have constructed meaning for some of those facts during the
interview. Students’ responses show the importance of the predicted constructions
included in the Genetic Decomposition. In this investigation we related observed
difficulties with specific mental constructions in the GD that students seemed to lack.
The importance of the pre-requisite constructions in learning this difficult topic was
underscored. Its lack became an insurmountable obstacle to understand even the most
basic ideas leading to the learning of the double integral.
A more encompassing understanding of function in different representation registers
proved to be indispensable. Results indicate that students who could only perform
Actions constructed a confusing network of concepts where the properties learnt
about one-variable function are not well differentiated from those of two-variable
functions. This inhibits their possibility to make those constructions involved in the
understanding of 3D space, functions, and their domains. Fluency in operating within
and across different representations plays an important role in the construction of
two-variable functions as an Object, instead of considering them as simple
correspondence rules containing one or more variables.

72 [Link]:indrum2018:174222
These results emphasize once more the importance of spending more time on helping
students to construct the notion of two-variable function. But, even when two-
variable function has been constructed as a Process, the notions of volume under the
surface and the role of the Riemann sum in the construction of the double integral
constitute fundamental constructions in the learning of double integrals.
The genetic decomposition proved useful in determining and underscoring those
mental constructions that are needed to learn double integrals with meaning. It also
reveals the subtleties involved in learning the double integral. After classroom use of
specially designed activities, future studies may reformulate the same interview
problems and also extend them to explore other ideas of the integral calculus.
REFERENCES
Arnon, I., Cotrill, J., Dubinsky, E., Octaç, A., Roa Fuentes, S., Trigueros, M., &
Weller, K. (2014). APOS Theory: A framework for research and curriculum
development in mathematics education. Springer Verlag: New York.
Duval, R. (2006). A cognitive analysis of problems of comprehension in a learning of
mathematics. Educational Studies in Mathematics 61, 103–131.
Jones, S. R. & Dorko, A. (2015). Students’ generalizations of single-variable
conceptions of the definite integral to multivariate conceptions. In (Eds.) T.
Fukawa-Connelly, N. Engelke Infante, K. Keene, and M. Zandieh, Proceedings of
the 18th Annual Conference on Research in Undergraduate Mathematics
Education, Pittsburgh, Pennsylvania.
Martínez-Planell, R. & Trigueros Gaisman, M. (2012). Students' understanding of
the general notion of a function of two variables. Educational Studies in
Mathematics, 81 (3), 365-384.
Martínez-Planell, R. & Trigueros, M. (2017). Students’ relating integrals of functions
of two variables and Riemann sums. In Kaur, B., Ho, W.K., Toh, T.L., & Choy,
B.H. (Eds.). Proceedings of the 41st Conference of the International Group for the
Psychology of Mathematics Education, Vol. 3, pp. 225-232. Singapore: PME.
McGee, D. & Martínez-Planell, R. (2014). A study of semiotic registers in the
development of the definite integral of functions of two and three variables.
International Journal of Science and Mathematics Education, 12,883-916.
Sealy, V. (2014). A framework for characterizing student understanding of Riemann
sums and definite integrals. The Journal of Mathematical Behavior, 33, 230-45.
Trigueros, M. & Martinez- Planell, R. (2010). Geometrical representations in the
learning of two variable functions. Educational Studies in Mathematics, 73, 3-19.

73 [Link]:indrum2018:174222
 Teaching and learning continuity with technologies
Sghaier Salem Béchir1 and Fabrice Vandebrouck²
1
ISEFC, Tunisie, sghaierb@[Link] ; ²Université Paris Diderot, LDAR, France
vandebro@[Link]
We developed a digital tool aiming at introducing the concept of – local - continuity
together with its formal definition for Tunisian students at the end of secondary
school. Our approach is a socioconstructivist one, mixing conceptualisation in the
sense of Vergnaud together with Vygotski’s concepts of mediation and ZPD. In the
paper, we focus on the design of the tool and we give some flashes about students’
productions with the tool and teachers’ discourses in order to foster students’
understanding of the continuity.
Keywords: teaching and learning of analysis and calculus, novel approaches to
teaching, continuity, digital technologies
The definition of continuity of functions at a given point, together with the concept
of continuity, remains a major difficulty in the teaching and learning of analysis.
There is a dialectic between the definition and the concept itself which make
necessary the introduction of the two aspects together.
The definition of continuity brings FUG aspects in the sense of Robert (1982). This
means first that it permits to formalize (F) the concept of continuity. But it also
allows to unify (U) several different images (or situations) of continuity encountered
by students: in Tall and Vinner (1981), several emblematic situations of continuity
are established (see below) and the definition aims at unifying all these different
kinds of continuity. Moreover, the definition of continuity allows generalisations (G)
to all other numerical functions, not already encountered and not necessarily with
graphical representations, or more general functions inside other spaces of functions.
As Robert (1982) stresses for the definition of limit of sequences, notions which
bring FUG aspects must be introduced with a specific attention to mediations and
especially the role of the teacher.
Our ambition is then to design a technological tool which allows on one hand
students activities concerning the two aspects of continuity and, on the other hand,
allows the teacher to introduce the concept of continuity with its formal definition,
referring to the activities developed on the technological tool. As it was noticed in
the first INDRUM conference, papers about introduction of technologies in the
teaching of analysis remain very few.
We first come back to well-known concept images and concept definitions of
continuity. Then, we explain our theoretical frame about conceptualisation and
mathematical activities. This theoretical frame leads us to the design of the
technological tool which brings most of the aspects we consider important for the
conceptualisation of continuity. Due to the text constraints, the results of the paper

74 [Link]:indrum2018:174515
are mostly in term of the design itself and the way the tool encompasses our
theoretical frame and our hypotheses about conceptualisation (with tasks, activities
and opportunities for mediations). Then, we can give some flashes about students’
activities with the software and also teachers’ discourses to introduce the definition
of continuity, based on students’ mathematical activities on the software.

CONCEPT IMAGES AND CONCEPT DEFINITIONS OF CONTINUITY

No one can speak about continuity without referring to Tall and Vinner’s paper about
concept images and concept definitions in mathematics, whose particular reference is
about limits and continuity (Tall and Vinner, 1981). Tall considers that the concept
definition is one part of the total concept image that exists in our mind. Additionally,
it is understood that learners enter their acquisition process of a newly introduced
concept with preexisting concept images.
Sierpinska (1992) used the notion of epistemological obstacles regarding some
properties of functions and especially the concept of limit. Epistemological obstacles
for continuity are very close to those observed for the concept of limit and they can
be directly relied to students’ concept images, as a specific origin of theses
conceptions (El Bouazzaoui, 1988). One of these obstacles can be associated to what
we call a primitive concept image: it is a geometrical and very intuitive conception of
continuity, related to the aspects of the curve. With this concept image, continuity
and derivability are often mixed and continuity means mainly that the curve is
smooth and have no angles. Historically, this primitive conception leads Euler to
introduce a definition of continuity based on algebraic representations of functions.
This leads to a second epistemological obstacle: a continuous function is given by
only one algebraic expression, which can be called the algebraic concept image of
continuity. This conception has led to a new obstacle with the beginning of Fourier’s
analysis. Then, a clear definition is necessary. This definition comes with Cauchy
and Weierstrass and it is close to our actual formal definition.
We also refer to Bkouche (1996) who identifies three points of view about continuity
of functions which are more or less connected to the epistemological obstacles we
have highlighted. The first one is a cinematic point of view. Bkouche says that the
variable pulls the function with this dynamic concept image. The other one is an
approximation point of view: the desired degree of approximation of the function
pulls the variable. This last point of view is more static and leads easily to the formal
definition of continuity. These two points of view are also introduced by Robert
(1982) when she studies the introduction of the formal definition of limit (for
sequences). A third point of view is also identified by Bkouche that is the algebraic
point of view, which is about algebraic rules, without any idea of the meaningful of
these rules.
At last, we refer to more recent papers and specifically the one of Hanke and Schafer
(2017) about continuity in the last CERME congress. Their review of central papers

75 [Link]:indrum2018:174515
on concept images about students’ conceptions of continuity leads to a classification
of the eight possible mental images that are reported in the literature: I : Look of the
graph of the function : “A graph of a continuous function must be connected” - II :
Limits and approximation : “The left hand side and right hand side limit at each
point must be equal” - III : Controlled wiggling : “If you wiggle a bit in x, the values
will only wiggle a bit, too” - IV : Connection to differentiability : “Each continuous
function is differentiable” - V : General properties of functions : “A continuous
function is given by one term and not defined piecewise”- VI : Everyday language :
“The function continues at each point and does not stop” - VII : Reference to a
formal definition : “I have to check whether the definition of continuity applies at
each point” -VIII : Miscellaneous
We can recognize some of the previous categories, even if some refinements are
brought. Mainly, concept images I, II, IV and VI can be close to the primitive
concept image whereas VII refers to the formal definition and V seems to refer to the
algebraic approach of continuity.

CONCEPTUALISATION OF CONTINUITY
We base our research work on these possible concepts image and concepts definition
of continuity. However, we are more interested in conceptualisation, as the process
which describes the development of students’ mathematical knowledge.
Conceptualisation in our sense has been mainly introduced by Vergnaud (1990) and
it has been extended within an activity theoretical frame developed in the French
didactic of mathematics. These developments articulate two epistemological
approaches: that of mathematics didactics and that of developmental cognitive
psychology as it is discussed and developed in Vandebrouck (2018).
Broadly, conceptualisation means that the developmental process occurs within
students’ actions over a class of mathematical situations, characteristic of the concept
involved. This class of situations brings technical tasks – direct application of the
concept involved - as well as tasks with adaptations of this concept. A list of such
adaptations can be found in Horoks and Robert (2007): for instance mix between the
concept and other knowledge, conversions between several registers of
representations (Duval 1995), use of different points of view, etc. Tasks that require
these adaptations of knowledge or concepts are called complex tasks. These ones
encourage conceptualisation, because students become able to develop high level
activities allowing availability and flexibly around the relevant concept.
A level of conceptualisation refers to such a class of situations, in a more modest
sense and with explicit references to scholar curricula. In this paper, the level of
conceptualisation refers to the end of scientific secondary school in Tunisia or the
beginning of scientific university in France. It supposes enough activities which can
permit the teacher to introduce the formal definition of continuity together with the
sense of the continuity concept. The aim is not to obtain from students a high

76 [Link]:indrum2018:174515
technicity about the definition itself – students are not supposed to establish or to
manipulate the negation of the definition for instance. However, this level of
conceptualisation supposes students to access the FUG aspects of the definition of
continuity.
Of course, we also build on instrumental approach and instrumentation as a sub
process of conceptualisation (Rabardel, 1995). Students’ cognitive construction of
knowledge (specific schemes) arise during the complex process of instrumental
genesis in which they transform the artifact into an instrument that they integrate
within their activities. Artigue (2002) says that it is necessary to identify the new
potentials offered by instrumented work, but she also highlights the importance of
identifying the constraints induced by the instrument and the instrumental distance
between instrumented activities and traditional activities (in paper and pencil
environment). Instrumentation theory also deals with the complexity of instrumental
genesis.
We also refer to Duval’s idea of visualisation as a contribution of the
conceptualisation process (even if Duval and Vergnaud have not clearly discussed
this point inside their frames). However, the technological tool brings new dynamic
representations, which are different from static classical figures in paper and pencil
environment. These new representations lead to enrich students’ activities – mostly
in term of recognition - bringing specific visualization processes. Duval argues that
visualization is linked to visual perception, and can be produced in any register of
representation. He introduces two types of visualization, namely the iconic and the
non-iconic, saying that in mathematical activities, visualization does not work with
iconic representations (Duval, 1999).
At last, we refer on Vygotsky (1986) who stresses the importance of mediations
within a student’s zone of proximal developmental (ZPD) for learning (scientific
concepts). Here, we also draw on the double approach of teaching practices as a part
of French activity theory coming from Robert and Rogalski (2005). The role of the
teacher’ mediations is specifically important in the conceptualisation process,
especially because of the FUG aspects of the definition of continuity (as we have
recalled above).
First of all, we refine the notion of mediation by adding a distinction between
procedural and constructive mediations in the context of the dual regulation of
activity. Procedural mediations are object oriented (oriented towards the resolution
of the tasks), while constructive mediations are more subject oriented. We also
distinguish individual (to pairs of students) and collective mediations (to the whole
class).
Secondly, we use the notion of proximities (Bridoux, Grenier-Boley, Hache and
Robert, 2016) which are discourses’ elements that can foster students’ understanding
– and then conceptualisation - according to their ZPD and their own activities in

77 [Link]:indrum2018:174515
progress. In this sense, our approach is close to the one of Bartolini Bussi and
Mariotti (2008) with their Theory of Semiotic Mediations. However, we do not refer
explicitly at this moment to this theory which supposes a focus on signs and a more
complex methodology than ours. According to us, the proximities characterize the
attempts of alignment that the teacher operates between students’ activities (what has
been done in class) and the concept at stake. We therefore study the way the teacher
organizes the movements between the general knowledge and its contextualized
uses: we call ascending proximities those comments which explicit the transition
from a particular case to a general theorem/property; descending proximities are the
other way round; horizontal proximities consist in repeating in another way the same
idea or in illustrating it.

DESIGN OF THE TECHNOLOGICAL TOOL

The technological tool called “TIC-Analyse” is designed to grasp most of the aspects
which have been highlighted above. First of all, it is designed to foster students’
activities about continuity aspects in the two first points of view identified by
Bkouche: several functions are manipulated – continuous or not – and for each of
them, two windows are in correspondence. In one of the window, the cinematic-
dynamical point of view is highlighted (figure 1) whereas in the second window the
approximation-static point of view is highlighted (figure 2).

Figure 1: two windows for a function, the dynamic point of view about continuity
The correspondence between the two points of view is in coherence with Tall’s idea
of incorporation of the formal definition into the pre-existing students’ concept
images. It is also in coherence with the importance for students to deal with several
points of view for the conceptualisation of continuity (adaptations).

78 [Link]:indrum2018:174515
Figure 2: two windows for a function, the static points of view about continuity
In second, the functions at stake in the software are extracted from the categories of
Tall and Vinner (1981). For instance, we have chosen a continuous function which is
defined by two different algebraic expressions, to avoid the algebraic concept image
of continuity and to avoid the amalgam between continuity and derivability. We also
have two kinds of discontinuity, smooth and with angle.
There is an emphasis not only on algebraic representations of functions in order to
avoid algebraic conceptions of functions. Three registers of representations of
functions (numerical, graphical and algebraic) are coordinated to promote students’
activities about conversions between registers (adaptations).

Figure 3: example of commentary given by a pair of students in the dynamic window

79 [Link]:indrum2018:174515
The design of the software is coherent with the instrumental approach mostly in the
sense that the instrumental distance between the technological environment, the
given tasks, and the traditional paper and pencil environment is reduced. However
the software produces dynamic new representations – a moving point on the curve
associated to a numerical table of values within the dynamic window; two static
intervals, one being included or not in the other, for the static window – occurring
non iconic visualisations which intervene in the conceptualisation process.

Figure 4: example of commentary given by a pair of students in the static window

The software promotes students’ actions and activities about given tasks: in the
dynamic window, they are supposed to command the dynamic point on the given
curve – corresponding to the given algebraic expression. They can observe the
numerical values of coordinates corresponding to several discrete positions of the
point and they must fill a commentary with free words about continuity aspects of
the function at the given point (figures 1, 3). In the static window, they must fill the
given array with values of α, the β being given by the software (figures 2, 4). Then,
they have to fill a commentary which begins differently according to the situation
(continuity or not) and the α they have found (figures 4, 5).
As we have mentioned in our theoretical frame, students are not supposed with these
tasks and activities to get the formal definition by themselves. However, students are
supposed to have developed enough knowledge in their ZPD so that the teacher can
introduce the definition together with the sense and FUG aspects of continuity.

STUDENTS ACTIVITIES AND TEACHER’S PROXIMITIES

The students work by pair on the tool. The session is a one hour session but four
secondary schools with four teachers are involved. Students have some concept
images of continuity but nothing has been thought about the formal definition. The

80 [Link]:indrum2018:174515
teacher is supposed to mediate students’ activities on the given tasks. Students are
not supposed to be in a total autonomy during the session according to our socio
constructivist approach.

Figure 5: example of commentary given by a pair of students in the static window

We have collected video screen shots, videos of the session (for each schools) and
recording of students’ exchanges in some pairs. Students’ activities on each tasks are
identified, according to the tasks’ complexity (mostly kinds of adaptations), their
actions and interactions with computers and papers (written notes), the mediations
they receive (procedural or constructive mediations, individual or collective, from
the tool, the pairs or the teacher) and the discourses’ elements seen as “potential”
proximities proposed by the teacher.
It appears that the teacher mostly gives collective procedural mediations to introduce
the given tasks, to assure an average progression of the students and to take care of
the instrumental process. Some individuals mediations are only technical ones (“you
can click on this button”). Some collective mediations are most constructive such as
“now, we are going to see a formal approach. We are going to see again the four
activities (ie tasks) but with a new approach which we are going to call formal
approach...”. The constructive mediations are not tasks oriented but they aim at
helping students to organize their new knowledge and they contribute to the aimed
conceptualisation according to our theoretical approach.
As examples of students’ written notes (as traces of activities), we can draw on
figure 3 and 4. A pair of students explains the dynamic non-continuity with their
words “when x takes values more and more close to 2 then f(x) takes values close to -
2,5 and -2. It depends whether it’s lower or higher” (figure 3) which is in coherence
with the primitive concept image of continuity. The same pair of students explains
the non-continuity in relation to what they can observe on the screen: “there exists β

81 [Link]:indrum2018:174515
positive, for all α positive – already proposed by the tool in case of non-continuity -
such that f(i) not completely in j… f is not continuous”. We can note that the students
are using “completely” to verbalize that the intersection of the two intervals is not
empty. However, the inclusiveness of an interval into another one is not expected as
a formalized knowledge at this level of conceptualisation. Their commentary is
acceptable. Students are expressing what they have experimented several times : for
several values of β (β = 0,3 in figure 4), even with α very small (α = 0,01 in figure
4), the image of the interval ]2- α, 2+ α[ is not included in ]-2,5- β, -2,5+ β[.
Concerning a case of continuity, the students are also able to write an acceptable
commentary (figure 5) “for all β positive, their exists α positive – already proposed
by the tool in case of continuity – such that f(i) is included in j.”
Students’ activities on the given tasks are supposed to help the teacher to develop
proximities with the formal definition. It is really observed that some students are
able to interact spontaneously with the teacher when he wants to write the formal
definition on the blackboard. This is interpreted as a sign that the teacher’s discourse
encounters these students’ ZPD. Then the observed proximities seem to be horizontal
ones: the teacher reformulates several times the students’ propositions in a way
which lead gradually to the awaited formal definition, for instance “so, we are going
to reformulate, for all β positive, their exists α positive, such that if x belong to a
neighbour of α … we can note it x0 – α, x0 + α….”
Of course, it is insufficient to ensure proof and effectiveness of our experimentation.
The conceptualisation of continuity is an ongoing long process with is only initiated
by our teaching process. However, we want to highlight here the important role of
the teacher and more generally the importance of mediations in the conceptualisation
process of such a complex concepts. We only have presented the beginning of our
experimentation. It is completed by new tasks on the tool which are designed to
come back on similar activities and to continue the conceptualisation process.

REFERENCES
Artigue, M. (2002). Learning mathematics in a cas environment: The genesis of a
reflection about instrumentation and the dialectics between technical and
conceptual work. International Journal of Computers for Mathematical Learning.
7(3). pp 245 -274.
Bartolini Bussi, M. G. and Mariotti, M. A. (2008). Semiotic Mediation in the
Mathematics Classroom: Artifacts and Signs after a Vygotskian Perspective. In
L.D. English (ed), Handbook of International Research in Mathematics Education
(pp. 750-787). Mahwah, NJ: LEA.
Bkouche, R. (1996). Points de vue sur l’enseignement de l’analyse : des limites et de
la continuité dans l’enseignement. Repère IREM, 24, 67–76

82 [Link]:indrum2018:174515
Bridoux, S., Grenier-Boley, N., Hache, C., Robert, A. (2016). Les moments
d’exposition des connaissances : analyses et exemples. Annales de didactique et
de sciences cognitives, 21, 187–233
Duval, R. (1995) Sémiosis et pensée humaine: registres sémiotiques et
apprentissages intellectuels. Berne: Peter Lang.
Duval, R. (1999). Representation, vision and visualisation: cognitive functions in
mathematical thinking. Basic issues for learning, In F. Hitt et M. Santos (Eds.)
Actes de 21st North American PME Conference. 1. 3-26
El Bouazzoui, H. (1988). Conceptions des élèves et des professeurs à propos de la
notion de continuité d’une fonction. Thèse de l’Université Laval, janvier 1988.
Hanke, X., Shafer, X. (2017). Students view of continuity: an empirical analysis of
mental imagesand their usage, CERME10, [Link]
[Link]/[Link]?conferenceID=5118&action=prog_list&session=378
84
Horoks, J. and Robert, A. (2007). Tasks Designed to Highlight Task-Activity
Relationships. Journal of Mathematics Teacher Education, 10(4–6), 279–287
Rabardel, P. (1995). Les hommes et les technologies, approche cognitive des
instruments contemporains. Armand Colin
Robert, A. (1982). L'acquisition de la notion de convergence des suites dans
l'enseignement supérieur. Recherche en Didactique des Mathématiques. 3(3),
307–341.
Robert, A., Rogalski, J. (2005). A cross-analysis of the mathematics teacher’s
activity. An example in a French 10th-grade class. Educational studies in
mathematics, 59, 269–298
Sierpinska, A. (1992). On understanding the notion of function. Dans G. Harel et E.
Dubinsky (Eds.) The Concept of Function: aspects of Epistemology and
Pedagogy. Mathematical Association of America Notes, volume 25.
Tall, D., Vinner, S. (1981). Concept image and concept definition in mathematics,
with special reference to limits and continuity. Educational Studies in
Mathematics. Vol 12. pp 151-169.
Vandebrouck F. (2018) Activity Theory in French Didactic Research. In: Kaiser G.,
Forgasz H., Graven M., Kuzniak A., Simmt E., Xu B. (eds) Invited Lectures from
the 13th International Congress on Mathematical Education. (pp 679-698) ICME-
13 Monographs. Springer, Cham
Vergnaud, G. (1990). La théorie des champs conceptuels. Recherches en Didactique
des Mathématiques. Vol 10 (2-3). pp 133-169.
Vygotski, L. (1986). Thought and Language. Cambridge MA: MIT Press.

83 [Link]:indrum2018:174515
TWG 2: Mathematics for engineers
- Mathematical Modelling
- Mathematics and other disciplines
 The ecological relativity of modelling practices: adaptations of an
study and research path to different university settings
Berta Barquero1
1
University of Barcelona, Faculty of Education, Spain, bbarquero@[Link]

This paper focuses on the problem of the ecology of mathematical modelling

practices at university level through the systematic variation of teaching institutions.
Our aim is to deal with the variety of constraints appearing when modelling is
implemented in university classrooms, and to study the way new teaching proposals
can overcome them. Within the framework of the anthropological theory of the
didactic, a teaching and learning proposal in terms of study and research paths in
tertiary education shows new possibilities to surmount some of these constraints. The
paper presents the design and successive adaptations of an SRP about an urban bike-
sharing system according to the specificities of different university institutions and
the reactions obtained by the students and lecturers.
Keywords: Modelling; anthropological theory of the didactic; research and study
path; ecology; institutional relativity.
INTRODUCTION
The starting point of this research is delving into the problem of studying the variety
of constraints appearing when mathematical modelling proposals are implemented in
university classrooms, impeding their regular development, and to study the way new
teaching proposals can overcome them. Several research projects have highlighted
the existence of strong constraints impinging on the large-scale dissemination of
mathematics as a modelling activity in current educational systems at all school
levels (Doerr & Lesh, 2011; Kaiser & Maaβ, 2007). We use the term ecology to refer
to the institutional conditions allowing and the constraints hindering the way a given
activity is produced, transposed, taught and learned in a given educational setting.
In previous research developed in the framework of the anthropological theory of the
didactic (ATD), we propose the use of a general frame to detect and place the
institutional constraints hindering the possible large-scale dissemination of modelling
activities based on a hierarchy of levels of didactic co-determinacy (Chevallard,
2002). In Barquero, Bosch and Gascón (2013), we use this general frame to detect
constraints appearing at different levels, from the specific ones related to how
mathematical contents are proposed to be taught at school, to the more general ones
regarding the general organisation of school activities and the role assigned to
schools in our societies. This ecological analysis shows how institutional constraints
are anchored in deep-rooted practices and are difficult – for teachers and also for
researchers – to notice since they appeared as “the natural way of doing”. For
instance, Barquero et al. (2013) characterise and empirically contrast the
predominance of “applicationism” as the dominant way of interpreting, describing

85 [Link]:indrum2018:174869
and conceptualizing mathematical modelling in natural sciences university degrees.
Under its influence, modelling is understood as a mere application of previously
constructed knowledge, as if the construction of knowledge were independent of its
use. At a more general level, in many schools the prevailing pedagogy is still strongly
influenced by the paradigm of “visiting works” (Chevallard, 2015), according to
which school knowledge organisations are presented as interesting monuments to
visit, instead of as useful tools to provide answers to problematic questions.
In this paper, we focus on going one-step to study the ecological relativity of
modelling practices in university institutions. As it is described in Castella (2004) and
Sierra (2006), each institution endures an institutional relation with knowledge, in
particular, with mathematical knowledge. Consequently, each institution establishes a
set of specific conditions and constraints that can favour or, on the contrary, prevent
certain teaching and learning processes and knowledge constructions to be
appropriately developed. It is in this aspect where we want to look more carefully.
Therefore, we focus on analysing the emergence, persistence and scope of the
conditions and constraints for development of modelling through a variation of
university institution. In our research, we work on the use of the study and research
paths (SRP) as epistemological and didactic model (Chevallard, 2015; Winslow et
al., 2013; Barquero et al., 2018) where mathematics are conceived as a modelling tool
for the study of problematic questions. We here present an SRP based on an urban
bike-sharing system inaugurated in Barcelona in 2007 that has been experimented in
three different university settings. The starting point of this SRP is the difficulty to
get a homogeneous distribution of bicycles in a city with many sloping streets. We
present the successive transformations of the SRP to three different university
settings, according to the specificities of each institution, and to the reactions from
students and lecturers. Some of the commonalities found show the stable constraints
hindering the development of the SRP, whereas the differences detected bring new
insights about the conditions to surmount them.

DESIGN OF AN SRP ABOUT A SHARING-BIKE SYSTEM

In the following we describe the initial design of the study and research path (SRP)
about the sharing-bike system whose starting point is the generating questions (Q0)
about how to improve the distribution of bikes in the ‘Bicing’ system to provide a
better service to users. When working with the a priori design of the SRP, there are
foreseen several derived questions from Q0 that needs from a progressive modelling
process. In general terms, the modelling project was organised around the following
questions that structured the two phases the Bicing project:
Q(A): How can we describe the daily flow of bikes between stations? What is the
natural behaviour of the system when it is left alone (without redeployment)?
Q(B): How can we predict the bikes’ redeployment needs? Which changes can be
proposed to improve the current policy of bikes redeployment in the city?

86 [Link]:indrum2018:174869
Linked to these questions, we consider real data from Bicing about the distribution of
bikes among the different bikes’ stations. We, the researchers and the experts who
collaborated with us, agreed to organise these data in certain city areas according to
the similarities different stations shared on the pattern of daily bikes trips and routes
followed. Finally, we decided to present the data organised in six areas (as shown in
Table 1), which corresponds to the origin-destination matrix (OD matrix) containing
the potential number of daily bikes’ uses. Each number {odij} means the average of
the amount of bike traveling in a day from area j and arriving to area i.

Table 1: Origin-Destination matrix with daily bikes’ trips

To face the first question Q0(A), and going beyond the descriptive analysis of the data
contained in the OD matrix, models based on recurrent sequences of order d > 1 can
be considered, which are equivalent to matrix recurrent sequences X (n)= f (X(n-1)))
where X (n) = (x1 (n), x2 (n), …, x6 (n)) is the vector with the bike distribution in each
of the six areas at time n. Next we summarize the a priori design in terms of
hypothesis (H), questions (Q) and answers (A) delimited by the researchers about the
models that might be used in an implementation of the SRP.
One of the easier assumptions we can work with is considering that:
H(A)1: There is no redeployment of bikes in the system and the bike flows between
stations is the same every day.
Q(A)1.1: Then, if we deploy different amounts of bikes in each station, what will be the
distribution of bikes after 1, 2, 3,…, n days?
The model that can be considered under these assumptions is:
X (n) = M.X (n-1) à X (n) = Mn. X (0) for n > 0 (1)
where M is the transition matrix (or transition probability matrix) obtained from the OD
matrix, where {mij} is the percentage of transition between two areas. That is, the potential
number of daily travels with origin in j and arriving to i {(od ij)} divided by total amount of
departures from j (d(j)). When working with this first model, several questions can appear:
Q(A)1.2: Working with the transition matrix and with different X(0) at the beginning of the
day, which traits from the trajectory of X(n) can be underlined?

87 [Link]:indrum2018:174869
 Q(A)1.3: Does it exist any fixed point X f to which the sequence X(n) converges to? Do all
X(n) converge towards a fixed point X f ? Is it possible to calculate X f in advance?
Q(A)1.4 : Which relation there exist between X f and the n-power of the transition matrix?
And, it can easily appear questions about the limitations of the hypothesis assumed and
models built, such as:
Q(A)1.5 : How can include other factors that are important for Bicing, such as: the total
amount of trips made by a bike, the potential demand of bikes, the available bikes?
Introducing questions about how to improve our hypothesis and the models to be more
realistic with the system we want to analyse can open many possibilities. One possible new
reformulation of the hypothesis we can work with is:
H(A)2: We assume that (1) each bike trip takes about t minutes, (2) the entire fleet of bikes
does not move every t min, (3) the total number of bikes that moves in period t depends
on: (a) the potential demand for bike trips, and (b) the amount of bikes available.
At this point, there appear more complex models where it is important to frame the time t,
for instance, t = 30 minutes (which it is the average of a bike trip in Bicing). Then, we can
define Bi (t) as the number of bikes in an area at time t and
B(t) = (B1(t), B2(t), B3(t), B4(t), B5(t), B6(t)) as the vector with the bikes distribution in each
area. Then, if we define the departures as D(t) = (D1(t), D2(t), D3(t), D4(t), D5(t), D6(t)) and
the arrival as A(t) = (A1(t), A2(t), A3(t), A4(t), A5(t), A6(t)), B(t) can be modelled by:
B(t +1) = B(t) - D(t) + A(t +1) (2)
where D(t) = min [demand_trips(30 min), B(t)] and A(t+1) = M·D(t), with M the transition
matrix in time periods t. When this second model is considered, several questions can guide
the study process:
Q(A)2.1 : Using this model (2), and considering different initial distribution of bikes at the
beginning of the day B(0), which will be the bike distribution B(t) at the end of the day?
And, if the system is left alone, after 2, 3, 4, …, 30 days?
Q(A)2.2 : Which traits can we underline about the trajectory of B(t) through the simulation
of model (2)? Are there also some fixed points to which the sequence B(t) converge?
Q(A)2.3 : Is there any relation between the fixed points X f we reach with the ones detected
with model (1)?
Q(A)2.4 : Which relationship is there between the first and second models, defined in (1)
and by (2)? Which of the two models do integrate more realistic conditions about Bicing?
In the next section we retake this a priori design of the SRP in terms of Q0 and the likely
hypothesis and derived questions Q(A).n to analyse the particular implementation of the SRP
about Bicing project in the different university institutions. Besides underlying the
adaptations that were necessary to the SRP in each university institution, we focus on the
most important conditions (common or not) that favour the development of the SRP, and
consequently of the modelling practice. In most of the occasions, these conditions and

88 [Link]:indrum2018:174869
constraints were phrases by the students and lecturers involved in the implementations or by
the survey and interview done at the end of each implementation.

ECOLOGICAL ANALYSIS OF THE SRP IN DIFFERENT UNIVERSITY

INSTITUTIONS
First SRP adaptation: The ‘Bicing project’ at the University of Copenhagen
The first implementation of the SRP about the bikes’ distribution in the Bicing system
took place in the University of Copenhagen (UC). Twenty-three students participated
in this implementation. They were taking the course called MathMod (Mathematical
Modelling), which was an optional course in the third year of the Mathematics
degree. The course run over seven weeks, plus two extra weeks to prepare their final
team project. The course had three weekly sessions of two hours each. In general
terms, the first session was a lecture, the second was a practical or exercise-based
session to practice the content introduce in the previous lecture and, the third one, to
work in teams in the computer room to simulate by Mapple some models introduced
along the course or to work on the team final project. The teaching course proposal
was based on the realization of four short projects (mini-projects), linked to some
practical activities. These mini-projects mostly consisted of being introduced to some
pre-existing models in the lectures sessions to then asked students to put them into
practice in the practical sessions. Some example of the project composing the course
are: “Mini-project 1: Using the Malthusian and logistic models to predict population
evolution” or “Mini-project 4: The Lotka-Volterra models”.
In the academic year 2009/10, the author of the paper participated in this course as
researcher and the lecturer offered the opportunity of implementing the SRP about
Bicing. It was integrated as the fifth (and last) project of the course. The SRP
implementation ran over two weeks, with six sessions of two hours. At the end of
each week, students working in teams had to deliver a report with their temporary
results of the Bicing project. It was necessary to break with the above-mentioned
organisation of the course sessions and to set up time for the presentations by the
lecturer-researcher and for students’ presentation. There, students could compare
their proposals and to collectively agree how to follow. During the first week, once
the generating question Q0 was presented by the lecturer-researcher, students agreed
to firstly focused on Q0(A) from where students developed most of the path described
in the previous section about model (1). In the second week, we (students and
instructors) worked on how to reformulate the H(A)2 and Q(A)2, as most of the groups
noticed that in model 1 there were considered some unrealistic assumptions. Due to
time restrictions, we could not go further the second model. Finally, each team had to
deliver a report one week later the ending of the project with some suggestions for
Bicing about how to improve their bike replacing system, Q0(B). Figure 1 summarizes
the path followed in this first implementation.

89 [Link]:indrum2018:174869
Figure 1: Summary of the path follow in the first implementation of the SRP at UC
We counted on different conditions that favour that the SRP progress fruitfully. First,
as it was the fifth project of the course, and the course was explicitly focus on
modelling, students and lecturers shared a common discourse to refer to modelling.
This was an important condition for modelling to be noticed (Barquero et al. 2013).
Secondly, the second mini-project was about Leslie matrices and transition matrices.
It thus facilitated that students autonomously posed many new questions, such as
Q(A)1.3 and Q(A)1.4 and, thanks to the previous work developed with Maple, students
easily worked on calculating and simulating sequences and studying their
convergence. On the contrary, there were also some constraints that were manifested
by students mostly at the beginning of the SRP implementation. When we started
with the Bicing project, students were astonished by the new responsibilities that they
were asked, such as: formulating hypothesis, looking for and building models, testing
models’ appropriateness, formulating new questions, writing a report without any
predetermined structure, etc. Although their initial confusion, consequence of a big
rupture with the didactic contract established in the course, they started assuming
these new responsibilities. In the previous activities of the course, students were only
asked to “apply” the models they had been introduced to. So that, breaking some
rules of the didactic contract and make students responsible of several new tasks in
the modelling process were the main constraints we had to surmount. In fact, the
course organisation shown many traits (and constraints) derived from
“applicationism” (Barquero et al. 2013). For instance, it was assumed (throughout the
course organisation) that the mathematical models had to be introduced in advanced
and then applied to different situation, models that are rarely questioned and hardly
reformulated. When the Bicing project started, many students’ resistances appeared
that reflected the implicit assumptions about what modelling was suppose to be and
what we (as students and as lecturer) were asked to do. At the end of the course,
when students were asked through a survey and with the interviews with some of
them, they stated how interesting it was this last project for several reasons. Some of
main reasons mentioned by the students were: the openness of the questions, the
possibility to delimit the questions to face, the necessity of clearly understanding the
modelling process (the hypothesis assumed, the models’ construction and their
validation), the possibility to compare teams’ proposals and results with the rest of
the groups who could have been working differently, possibilities to discuss the
limitation of the models proposed and make them evolve.

90 [Link]:indrum2018:174869
Ecological relativity of the second SRP adaptation implemented at UAB
The second implementation of the SRP was the following academic year at the
Universitat Autònoma de Barcelona (UAB). There was a course called
“Mathematical modelling workshop” which started in 2009/10 with second-year
students of Mathematics degree. It was the first edition of the course, which was
compulsory, with a total of 45 students participating. The didactic organisation of the
course was different from the previously described at the UC. The main aim of the
course was to develop a project in working teams (composed of 4-5 students) that
students selected from a list provided by the lecturers of the course. Running in
parallel, there were planned some short activities about modelling. The first year this
course was implemented, one of the modelling activities planned was the ‘Bicing
project’. It ran over 5 weeks, with two 2-hour sessions per week. We invested more
than the double of time than in its first implementation. Similarly, students were
asked that at the end of each week they had to deliver a report with a synthesis of
their advances in term of: (a) questions they had focused on, (b) hypothesis assumed
and mathematical models considered, (c) temporary answers and (d) new questions to
follow with). At the end of the Bicing project, each working team had to deliver a
final report as summary of the whole modelling work developed. In general terms,
the modelling process students and instructors followed in this occasion was not so
different concerning Q0(A), although now none of the students’ working team tackled
the second phase of the project with Q0(B), or posed any questions about the properties
of the n-power of transition matrices, such as: Q(A)1.4 or Q(A)2.3.

Figure 2: Summary of the path follow in the second adaptation of the SRP in UAB
One important novelty (and extension of the SRP) was that some students asked
about the possibility of working with partial matrices, for instance, by considering
different OD matrix to describe differently the bikes’ flow in the morning and in the
afternoon. Students had checked in the web how many bikes were available at
different time frames and they had concluded that there were different patterns of
bikes disposition depending in the daily time frame. The instructors asked to the
experts we worked with about the possibility of having these new data. The external

91 [Link]:indrum2018:174869
experts provided us two new matrices: one for the morning pattern, from 05:00h to
14:30h, and the other for the afternoon, from 14:30h to 00:00h. With this new data,
the modelling process concerning Q0(A) was extended towards the construction of a
third model, built upon the two previous ones (1) and (2), and taking into account
these two different OD matrices. Figure 2 summarizes the path followed in this
occasion and the extension it supposed for the first phase of the Bicing project.
If we focus on analysing the conditions and constraints we detected in the second
implementation of the SRP, we have to mention that in this occasion it was the
lecturer of the course who expressed more clearly some important constraints. He
expressed, in an interview at the end of the implementation, that we had invested too
much time with the project. He manifested that students needed to work more
independently and there was no need of planning common discussions among all the
working groups. His main request was to let students work independently and ask
them to present their finding at the end of the course. Reactions that were on an
opposite sense than the ones expressed by the Danish lecturer, who expressed that the
activity was too open and too less guided for students. We can say that these
reactions corresponded to their spontaneous teaching models that both lecturers
implicitly defended. In this second implementation, it shared traits of a modernist
teaching model (Gascón, 2001), by considering knowledge construction as an
individual process, also private. That is why the lecturer preferred not planning any
teaching device where to share and collectively talk about the modelling work
developed, and where to question, debate and agree about the questions, tools and
strategies to follow along the modelling process. As the course organisation at the
UAB showed, each team was supposed to work most of the time independently in
their project, and it was not until the end of the course when they explained their
results. We could observe several inconveniences, linked to important constraints,
which were more evident in the following courses when the lecturers planned short
modelling activities as complement to the working group project of the course. First,
students showed a lack of terminology and of a common discourse (shared with
lecturers) to talk and write about the modelling activity developed. Second, the main
outcome from the students modelling work was their final presentation of the project
at the end of the course. It was delivered as a report that mostly contained the final
models and models simulation, as if all the intermediate modelling work may remain
in the private space of each group. Consequently, most of final reports showed a poor
progression of the models considered and of the tools to contrast and validate them.

CONCLUSIONS AND DISCUSSION

It has to be highlighted that the two adaptations of the SRP presented in this paper
were done under advantageous conditions. First, it was experienced with students of
the Mathematics degree who were taking a course on mathematical modelling and
with lecturers who are experts on modelling. Second, in both cases, the schedule and
programme of the course were flexible and we had longer sessions (2-hour sessions
two or three times per week) than the prevailing university settings use to offer.

92 [Link]:indrum2018:174869
Nevertheless, one could think that we may detect similar conditions and constraints in
these two university setting, but it is important to see how different institutions
established different relations with the knowledge at the stake, in this case, with the
teaching of mathematics modelling. Then, for example, some conditions that appear
in the first implementation can become strong constraint for the second one. For
instance, it was the case of the necessity of sharing a common discourse to talk about
and analyse modelling practices, which was an important condition underlined in the
first implementation, becoming a constraints in the second one.
But, if we move away from these “optimal” university conditions, do we find similar
constraints? Which of them are sensitive to be surmounted? How to overcome some
of the most important constraints? To face these questions, and follow enquiring into
the institutional relativity of the conditions favouring and the constraints hindering
modelling practices, we proceeded with the third adaptation of the SRP. It was
redesigned and later implemented with first-year university students of business and
administration degree (4-year programme) in IQS School of Management of
Universitat Ramon Llull in Barcelona (Spain) during the entire academic year
2013/14. In this occasion, the Bicing project was extended (called now “Cycling
project”) to become the central project developed along the three terms of the
mathematics first-year course. The SRP was broken into three branches. The one
described in this paper (in section 2) was implemented during the third term, only
focusing on the first model (1). During the entire course, not only the initial structure
of the SRP was extended, but also we pay special attention to which teaching devices
and strategies could help to overcome some of the most common constraints for
modelling and to create appropriate conditions for modelling and for the SRP. We are
in the process of analysing them in depth with the aim of extending our knowledge
about the ecology of the SRP and its institutional relativity.

ACKNOWLEDGMENTS
The research leading to these results has received funding from the Spanish R & D project: EDU2015-69865-C3-1-R
and EDU2015-64646-P.

REFERENCES
Barquero, B., Bosch, M., & Gascón, J. (2013). The ecological dimension in the
teaching of modelling at university level. Recherches en Didactique des
Mathématiques, 33(3), 307-338.
Barquero, B., Monreal, N., Ruiz-Munzón, N., & Serrano, L. (2018). Linking
transmission with inquiry at university level through study and research paths: The
case of forecasting Facebook user growth. International Journal of Research in
Undergraduate Mathematics Education, 4(1), 8-22.
Castela, C. (2004). Institutions influencing mathematics students’ private work: a
factor of academic achievement. Educational Studies in Mathematics, 57(1), 33-
63.

93 [Link]:indrum2018:174869
Doerr, H. & Lesh, R. (2011). Models and Modelling Perspectives on Teaching and
Learning Mathematics in the Twenty-First Century. In G. Kaiser, W. Blum, R.
Borromeo Ferri, & G. Stillman (Eds.), Trends in Teaching and Learning of
Mathematical Modelling (pp. 247-268). Dordrecht, The Netherlands: Springer.
Chevallard, Y. (2002). Organiser l’étude 3. Ecologie & régulation [Organizing the
study 3. Ecology & regulation]. In J.-L. Dorier, M. Artaud, R. Berthelot, & R.
Floris (Eds.), Actes de la XIe École d’Été de Didactique des Mathématiques (pp.
41-56). Grenoble, France: La Pensée Sauvage.
Chevallard, Y. (2006). Steps towards a new epistemology in mathematics education.
In M. Bosch (Ed.), Proceedings of the 4th Conference of the European Society for
Research in Mathematics Education (pp. 21-30). Barcelona, Spain: Fundemi-IQS.
Chevallard, Y. (2015). Teaching Mathematics in Tomorrow’s Society: A Case for an
Oncoming Counter Paradigm. In S.J. Cho (Ed.), Proceedings of the 12th
International Congress on Mathematical Education (pp. 173-187). Dordrecht:
Springer.
Kaiser, G. & Maaβ, K. (2007). Modeling in lower secondary mathematics classroom
– problems and opportunities. In W. Blum, P.L. Galbraith P. L., H. Henn, & M.
Niss (Eds.), Modelling and applications in mathematics education: The 14th
ICMI study (pp. 99-107). New York, NY: Springer.
Sierra, T. (2006). Lo matemático en el diseño y análisis de organizaciones didácticas.
Los sistemas de numeración y la medida de magnitudes. PhD dissertation.
Departamento de Didáctica de las Matemáticas, Facultad de Educación,
Universidad Complutense de Madrid, Spain.
Winsløw, C., Matheron, Y., & Mercier, A. (2013). Study and research courses as an
epistemological model for didactics. Educational Studies in Mathematics, 83(2),
267-284.

94 [Link]:indrum2018:174869
 C’ (x) = C(x+1)-C(x)? - Students’ connections between the derivative
and its economic interpretation in the context of marginal cost
Frank Feudel
Humboldt-University of Berlin, feudel@[Link]
The derivative concept plays a major role in economics. Therefore, students of
economics should have a proper understanding of the concept and its application in
economics. One important competence for these students is to interpret the
derivative in economic contexts. In books of economics the derivative is commonly
interpreted as amount of change while increasing the production by one unit.
However, from a mathematical point of view, this interpretation does not directly
correspond to the derivative. In the study presented here, it was investigated to what
extent students can make an adequate connection between the derivative as a
mathematical concept and its economic interpretation mentioned above.
Keywords: derivative, students of economics, economic interpretation, concept
image, marginal cost.

INTRODUCTION
The derivative plays an important role in economics. It is used to solve optimization
problems, to describe and characterize economic functions, and in marginal analysis,
in which the impact of small changes from the current state is examined (example:
the effect of small changes in the price of a product on the demand) in order to make
optimal decisions. Hence, students of economics should have an adequate
understanding of the derivative concept in order to be able to use it in economics in a
reflective manner. The study presented here focuses on students’ of economics
understanding of the derivative after their Calculus course with special emphasis on
its economic interpretation, which is essential for the ability to apply the concept in
economics. It is part of a larger research project about the understanding of the
derivative in mathematics for students of economics (my PhD-thesis, supervisor:
Rolf Biehler) at the Centre for Higher Mathematics Education in Germany (khdm).

LITERATURE REVIEW AND EMBEDDING OF THE RESEARCH

There is a lot of research about students’ understanding of the derivative. Concerning
the interpretation of the slope and the derivative in contexts, different difficulties are
documented. Typical mistakes are the slope/height confusion or the graph-as-picture
error (Beichner, 1994; Çetin, N., 2009; Carlson, M., Oehrtman, M. & Engelke, N.,
2010). An interesting study, which included a task to interpret the derivative in the
context of motion explicitly, was conducted by Bezuidenhout (1998). He asked
students to interpret S '(80)  1.15 if S (v) is the stopping distance of a vehicle in metres
in dependence of the velocity in km/h. Many students overgeneralized that the
derivative is the acceleration or the velocity itself. Many also had problems with the

95 [Link]:indrum2018:174442
units. These students did not understand the derivative as rate of change of the given
function S properly.
While there is some research of students’ understanding of the derivative in physical
contexts (examples mentioned above), there is little research related to economic
contexts. Wilhelm & Confrey (2003) showed that students cannot automatically
transfer their knowledge about rate of change from a physical context to the context
of money. Hence, an economic interpretation of the derivative should be explicitly
taught in a Calculus course for students of economics. But even if an economic
interpretation of the derivative was covered in the students’ Calculus course,
problems occur. Mkhatshwa & Doerr (2015) showed that many students talked about
marginal cost (the derivative of a cost function) as amount of change when solving
economic problems, although it was underlined in the Calculus course that the
derivative is a rate. This indicates a rather superficial understanding of the
connection between the derivative as a mathematical concept and its economic
interpretation. A similar result is also found in Feudel (2017). Students’ answers in a
GE
task to interpret P '(73)  0.2 (GE = units of money, ME = units of quantity) of a
ME
profit function P economically indicated that many students were not aware of the
numerical differences and the differences in the unit between the derivative and its
economic interpretation as additional profit. However, in these two studies the
students were not obliged to reveal their ideas about the connection between the
derivative as a mathematical concept and its economic interpretation explicitly. The
study presented here directly focuses on this connection.

THEORETICAL BACKGOUND OF THE STUDY

The economic interpretation of the derivative
To be able to use the derivative in economics, students need to be able to interpret its
values in economic contexts. However, understanding the interpretation of the
derivative commonly used in economics is a special challenge for students because it
does not directly correspond to any of the usual representations of the derivative as
limit of the difference quotient, slope of the tangent line, local rate of change, or as
instantaneous velocity. If C :[0; )  [0; ) is a cost function (the variable x represents
the output of a product), the derivative C '( x) , called marginal cost, is often
interpreted as additional cost of the next unit. However, if one takes this
interpretation literally it corresponds to the difference C ( x  1)  C ( x) , which differs
from the derivative in its unit and in its numerical value. Since the students already
have previous knowledge about the derivative from school, e.g. as slope of a
function at one point, this might confuse them. Hence, the economic interpretation of
the derivative should be carefully connected to the students’ previous knowledge,
and justified for economic contexts in the students’ of economics Calculus course. A
typical justification is via the approximation formula C ( x  h)  C ( x)  C '( x)  h for

96 [Link]:indrum2018:174442
h  0. Since h  1 can be considered as small in economics, the numerical values of
C '( x)and C ( x  1)  C ( x) are close to each other, and they can be identified (a more
detailed explanation can be found in Feudel (2016)).
The above mentioned perspective on marginal cost as being defined as derivative
and interpreted as additional cost of the next unit coincides with what is taught in
mathematics courses for students of economics (see e.g. (Sydsæter and Hammond,
2013)) and with what is presented in some books of economics like Breyer (2015).
However, marginal cost can also be defined as additional cost of the next unit like in
Blum (2003). In this case the derivative is viewed as method of calculation of the
additional cost. Nevertheless, the problem to justify the identification of the two
different mathematical objects in economic contexts also remains in this approach.
The notion of concept image to describe students’ conceptual knowledge
The economic interpretation of the derivative and its connection to the pure
mathematical concept, as it was explained above, should be part of students’ of
economics conceptual knowledge of the derivative concept. To describe students’
conceptual knowledge I will refer to the notion of concept image by Tall & Vinner
(1981), which describes the total cognitive structure associated to a concept. This
includes all mental pictures, properties and associated processes. In the case of the
derivative students’ of economics concept image should contain its representations,
the differentiation rules, its connection to the concepts of monotonicity and
convexity, its use as a tool for optimization problems, and in particular an adequate
economic interpretation of the derivative. Since the common economic interpretation
of the derivative as amount of change while increasing the production by one unit is
a different mathematical object, it should be in particular carefully connected to the
rest of the students’ concept image, called synthesizing in literature (Dreyfus, 2002).
Knowledge concerning the derivative covered in the students’ Calculus course
In the Calculus course for students of economics in which the study took place
(University of Paderborn 2015, Germany), the sessions involving the derivative
began with the definition of the derivative as limit of the difference quotient.
Alongside with the symbolic definition, its representations as slope of the tangent
line (tangent line introduced as limit of secant lines) and as rate of change were
introduced. Afterwards, the unit of the derivative in the case of a cost function C
was discussed and justified via the symbolic definition of the derivative. In the
second lecture the economic interpretation of the derivative in the context of
marginal cost was introduced, which is essential for the study presented here. Two
possible economic interpretations of the derivative were presented in the lecture:
1. Interpretation as approximation of the additional cost of the next unit
This interpretation was justified via the approximation formula C  C '( x)  x , which
was derived from the definition of the derivative by using the approximation aspect

97 [Link]:indrum2018:174442
of the limit. The terms  C and C '( x)  x were also visualized on the board with the
help of the tangent line.
2. Interpretation as additional cost of the next marginal unit
It was visualized that the mistake between  C and C '( x)  x becomes smaller if
x  0 . This results in the asymptotic equation dC  C '( x)dx in which the lecturer
called the “fictive infinitely small quantities” dx and dC marginal units.
Some lectures later the concepts of monotonicity and convexity and their connection
to the derivative were discussed. The sessions finished with optimization problems.
All the topics covered in the lecture were also practised in small groups, in which the
students had to solve problems. Relevant for the study presented here is, that these
problems also included a task to interpret the value C '(5) of the cost function
C( x)  8x2  10x  700 in an economic context (in the way presented above).

METHODOLOGY OF THE STUDY

The study aimed to find out to what extent students of economics can make an
adequate connection between the derivative and its economic interpretation after
their Calculus course. Hence, eight economics students who successfully completed
their Calculus Course at the University of Paderborn were interviewed. Each
interview lasted about 30 minutes. The interviews were structured by four tasks:
1. Consider the cost function C that is given by the following equation:
1 3 1 2
C ( x)  x  x  21x  500, x  0 . The output x is given in units per quantity, the
1000 4
cost C ( x) is given in Euro. Determine the marginal cost for an output of x  100
units of quantity. Determine the unit (of the marginal cost), too.
2. Is the derivative C '( x) the same like the additional cost while increasing the
production from x units of quantity by one unit? Justify yours answer.

Figure 1: Graphs of the cost function from task 1 and the cost function from the task,
in which the additional cost for the 11th unit had to be determined

98 [Link]:indrum2018:174442
3. Justify with the picture (left graph of figure 1) why the numerical values of C '( x)
and C ( x  1)  C ( x) of the function C from task 1 are almost identical (for x  100 ).
4. Justify with the definition of the derivative why it can often be assumed in
economics that the values of C '( x) and C ( x  1)  C ( x) are almost identical.
The tasks were not always presented to the students in written explicitly, but were
sometimes given by the interviewer verbally during the interview process.
Task 1 was an introductory task with the aim to find out which of the two
mathematical objects ( C '( x) or C ( x  1)  C ( x) ) the students associated first with the
term “marginal cost”. Task 2 was central in the interview. It aimed to find out to that
extent students knew the differences between the derivative and the additional cost
of the next unit. If the students claimed both objects to be equal they got the
additional task to determine the additional cost of the 11 th unit of a function, which
was only given graphically (see right graph of figure 1). Its aim was that the students
could no longer use their algorithm to determine the derivative and would use the
difference C ( x  1)  C ( x) instead. This should make the students’ concept image of
marginal cost incoherent and provoke a cognitive conflict to make them rethink their
ideas about the identity of the derivative and the additional cost of the next unit, and
to reorganize their concept image. With the interviewer’s help the students were then
led to the differences between the derivative and the additional cost of the next unit.
Task 3 had the aim to find out if the students could justify the identification of C '( x)
and C ( x  1)  C ( x) within the graphical representation with the help of the tangent line
(similar to the visualization presented in the lecture). Task 4 finally aimed to find out
to what extent the students have internalized the justification of the identification of
C '( x) with C ( x  1)  C ( x) in economics symbolically via the approximation formula
C ( x  h)  C ( x)  C '( x)h for h  0 (as it was taught in their Calculus course). The results
of tasks 3 and 4 are not discussed in the paper in detail due to limited space.
The interviews were recorded, transcribed, and interpreted. First, individual cases
were interpreted line by line. Later the results between different individuals were
compared. To ensure reliability, the author’s interpretations were discussed with
colleagues of the Centre of Higher Mathematics Education in Germany (khdm).

SOME DETAILED RESULTS

Altogether eight students were interviewed. Due to limited space, two students
(Holger and Lisa) with different understandings of marginal cost were chosen, whose
interview parts referring to the tasks 1 and 2 are presented here in detail.
Holger’s understanding of marginal cost
Holger immediately solved the first task to determine the marginal cost at x  100 for
1 3 1 2
the function C ( x)  x  x  21x  500, x  0 by calculating C '(100) . This means his
1000 4

99 [Link]:indrum2018:174442
first association of marginal cost was the derivative and not its economic
interpretation. His result was C '(100)  1 . He did not mention a unit himself. After
having been asked for the unit by the interviewer explicitly he mentioned “Euro”, but
could not justify it. This shows that he had an incoherent concept image of marginal
cost: he associated the derivative for calculations but did not think of marginal cost
as a rate (otherwise the unit would have to be Euro per unit of quantity).
In the next part of the interview Holger was explicitly confronted with the question if
a definition of marginal cost as additional cost of the next unit would represent the
same mathematical object like the derivative. He started thinking about it and then
agreed. Therefore, the interviewer tried to provoke a cognitive conflict by giving
Holger the task to determine the additional cost while increasing the production from
10 units by one unit (see right graph of figure 1). Holger first wanted to use the
derivative again, but then started to think about the task again:
32 Holger: You would have to imagine the derivative. Then you would see the
additional cost. The derivative is nothing else than the slope at a point.
If we take any point.
33 Interviewer: Here is one explicitly given. We search for it at a particular point.
34 Holger: Here is one given, 10 units of quantity. So we have x  10 where the
total cost is 21. Now we need the cost if one more is produced. Well,
but we do not need this because if we are at 11 the cost is 22 point
something. So the additional cost has to be one point, yes 1.1.
Holger now used the difference C ( x  1)  C ( x) to solve the task. To the following
question of the interviewer if the value would have been the same by using C '( x) ,
Holger agreed. He then determined the slope at x  10 graphically and got the
solution “round about one” (correct value: 1, see right graph in figure 1). After the
interviewer emphasized that Holger just said “round about”, Holger claimed that he
cannot determine the value exactly by graphical means. Hence, the interviewer asked
Holger afterwards to determine the cost difference for the cost function C from task
1 given by an equation. He now got the result 1.051 and justified the “error”
compared to C '(100)  1 as follows:
77 Holger: One nearly gets [1], but only nearly. This is probably due to rounding.
One can see that Holger was really convinced that the derivative and the additional
cost of the next unit are exactly the same, even if the calculated values differed. The
interviewer now emphasized that there was no rounding involved. He then pointed to
the graph of the cost function (left graph in figure1) and underlined that one can see
the error in the graph, too. After the interviewer had asked Holger again to determine
the value of the derivative by graphical means, now for the cost function of task 1
(left graph in figure 1), Holger found for himself a resolution of the conflict:
94 Holger: Oh, the reason is, because it is not exact. The origin of the derivative was to
determine the slope at a point. To achieve this you take one point left and
one point right of it, which have the same distance, and the slope between.

100 [Link]:indrum2018:174442
 Afterwards you try to make this distance as small as possible, as you could
think of, but we cannot reach the one point, but in our mind we want to
reach it. And I assume this the very small rounding mistake, no, not
rounding mistake, but this small difference is due to the fact that you do not
reach the point exactly.
Two misconceptions of the derivative occurred here. Holger did not imagine the
derivative as limit of slopes of secants through ( x0 , f ( x0 )) and ( x0  h, f ( x0  h)) for
h  0 but as limit of secants through ( x0  h, f ( x0  h)) and ( x0  h, f ( x0  h)) for h  0 .
But the important misconception that now prevented him from questioning the
identity of C '( x) and C ( x  1)  C ( x) was his opinion that the “true slope” at a point x0
was not reached by taking the limit. He imagined the derivative to be the slope of a
secant through ( x0  h, f ( x0  h)) and ( x0  h, f ( x0  h)) with a very small h  0 (he
repeated this several times later, even more explicitly than in the lines above).
To sum up, at the start of the interview, Holger identified the derivative C '( x) with
its economic interpretation as additional cost of the next unit. During the interview a
conflict occurred due to different numerical values of these two. But instead of
questioning the identity between the derivative and the additional cost of the next
unit he made his concept image coherent again by attributing this error to an error
between the “true slope at a point” and the derivative as result of a limiting process.
Lisa’s understanding of marginal cost
Similar to Holger, Lisa also immediately solved the task to determine the marginal
cost at the output x  100 by using the derivative. Unlike Holger she stated as unit
“Euro per unit of quantity”, a unit of a rate. So Lisa also associated the derivative
with the notion of marginal cost first. When confronted with the definition of
marginal cost as additional cost of the next unit she replied:
20 Lisa: Yes, I really thought about this last semester. In the economic subjects we
really learn it this way. […] And I always had to say: If you increase x by one
unit, y increases by these many units, eh? This is really the case. But I have,
since I had mathematics last semester, always thought that you learn it
differently in mathematics. In mathematics you say, if you increase x by one
marginal unit, y increases by these many marginal units.
Unlike Holger, Lisa did not identify the derivative with the cost of the next unit. She
even felt a conflict between the knowledge about marginal cost she learned in her
maths course and the actual use of marginal cost in econometrics. Her remarks point
out that she was of the opinion that marginal cost is not the additional cost of the
next unit but of the next marginal unit. In which way she understood the term
“marginal unit” was not clear yet. Therefore, the interviewer asked about this term:
21 Interviewer: Now the question, what is a marginal unit?
22 Lisa: Well, a marg/ Shall I draw it?
The interviewer gave Lisa the graph of the function of task 1 (left graph in figure 1).

101 [Link]:indrum2018:174442
 32 Lisa: Well, one unit could be from 100 to 101 we said. So you increase x by
one unit from 100, eh? From the actual output. Then I would be here
[pointing on C (101) ].
After the interviewer’s comment to determine the accurate value by calculation she
got C(101)  C (100)  1.051 (compared to C '(100)  1 ), and continued as follows:
48 Lisa: And a marginal unit I imagine very, very small. Here I would go a
right very, very little bit to the right and then a very little, little bit
upwards.
We see here that Lisa understood a marginal unit as a very small, but finite unit. This
understanding is also found in books of economics (and differs from the way a
marginal unit was taught in the maths course as a “fictive infinitely small quantity”).
After Lisa’s explanation of the “marginal unit”, the interviewer tried to induce a
cognitive conflict by asking Lisa if the additional cost of such a small unit should not
be close to zero. She then explained the following:
52 Lisa: This has to do with the slope you have. For the derivative you
calculate the slope of the tangent line. The slope of the tangent line is
what you calculate, isn’t it?
53 Interviewer: Right, the slope of the tangent line, yes.
54 Lisa: Yes, and this is also what I get if you increase x by a marginal unit,
starting at 100. Ah, what do I get? No, if you increase x by one
marginal unit, the marginal cost still increase by one. I think the slope
still remains one, right?
This shows that Lisa understood the additional cost of a marginal unit as C '(100  dx)
with dx being a very small, but finite unit. This is in her opinion numerically the
same like C '(100) because the slope stays the same at 100  dx . The interviewer again
asked if additional cost and slope are the same whereat Lisa agreed.
To sum up, Lisa knew that the derivative is not the additional cost of the next unit.
She remembered the interpretation of the derivative to be the additional cost of a
marginal unit from her Calculus course. She imagined a marginal unit dx to be a
very small, but finite unit (and not in the way it was taught in the course as a “fictive
infinitely small quantity” in the asymptotic equation dC  C '( x)dx ) and identified the
additional cost of a marginal unit in her mind with C '(100  dx) , which is in her
opinion the same like C '(100) . She did not recognize the different nature of the
derivative being a rate of change and the additional cost being an amount of change.

SUMMARY AND DISCUSSION

Just like in the cases of Holger and Lisa presented in detail here, the study showed
that the students had difficulties to make an adequate connection between the
derivative as a mathematical concept and its economic interpretation as additional
cost. The author considered a connection as adequate if the students were aware of
the differences between C '( x) and the additional cost of the next unit, and were able

102 [Link]:indrum2018:174442
to justify their identification in economics (it did not matter if they associated
marginal cost with the derivative or the additional cost of the next unit first).
At the beginning of the interview no student could make an adequate connection
between the derivative and the additional cost of the next unit. The majority just
declared these objects to be exactly the same (like Holger).
During the interview, all students recognized the question concerning the identity of
the derivative and the additional cost of the next unit as very relevant and became
aware of differences between these two with the interviewer’s help. However, in the
process of leading them to these differences on the graphical level via the tangent
line, several problems occurred (not discussed here in detail due to limited space):
1. Misconceptions concerning the derivative concept (like Holger’s misconception
about the derivative being not the “true slope” at the point)
2. Incomplete concept images (example: knowledge of the geometric representation
of the derivative as slope at a point but no association of the tangent line)
3. Problems in determining the slope of a linear function
Only one student could make a connection on the symbolic level via the formula
C ( x  h)  C ( x)  C '( x)  h for h  0 like presented in the Calculus course.

Furthermore, most participants of the study had not thought about the differences and
the connection between the derivative and its economic interpretation as additional
cost before the interview, although these were presented in their Calculus course.
A solution to these problems in a traditional Calculus course for students of
economics, in which the concept of derivative is taught first, an economic
interpretation afterwards, could be to confront the students with the two different
notions of marginal cost in the tutorials of the course directly, and to provoke a
cognitive conflict, just like in the interview. Afterwards, one could let them try to
connect the derivative and the additional cost of the next unit by themselves or in
small groups, and help them individually if misconceptions or incomplete concept
images of the derivative occur. Another solution could be to start with the concept of
marginal cost as additional cost first, which can be approximated by the derivative as
linear approximation. However, an understanding of the mathematical concept of the
derivative as slope of the tangent line is also necessary in this approach.
Concerning future research, one has to emphasize that the study presented here relied
on one perspective on the connection between the derivative as a mathematical
concept and its use in economics: additional cost as interpretation of the derivative.
This perspective is important because students of economics are often confronted
with it their maths course and in courses about economic theory (it can be found in
respective books of economics). But as explained in the theoretic part of the paper: it
is not the only one. Other perspectives may further enrich the knowledge about
students’ of economics understanding of the derivative and its use in economics.

103 [Link]:indrum2018:174442
REFERENCES
Beichner, R. J. (1994). Testing student interpretation of kinematics graphs. American
journal of Physics, 62(8), 750-762.
Bezuidenhout, J. (1998). First-year university students’ understanding of rate of
change. International journal of mathematical education in science and
technology, 29(3), 389-399.
Blum, U., Karmann, A., Lehmann-Waffenschmidt, M., Thum, M., Wälde, K., &
Wiesmeth, H. (2003). Grundlagen der Volkswirtschaftslehre. Berlin, Germany:
Springer.
Breyer, F. (2015). Mikroökonomik. Berlin, Heidelberg, Germany: Springer.
Carlson, M., Oehrtman, M., & Engelke, N. (2010). The precalculus concept
assessment: A tool for assessing students’ reasoning abilities and understandings.
Cognition and Instruction, 28(2), 113-145.
Çetin, N. (2009). The Ability of Students to Comprehend the Function-Derivative
Relationship with Regard to Problems from Their Real Life. PRIMUS, 19(3), 232-
244.
Dreyfus, T. (1991). Advanced mathematical thinking processes. In D. Tall (Ed.),
Advanced mathematical thinking (pp. 25-41). Dordrecht, The Netherlands:
Springer.
Feudel, F. (2016). Relevant knowledge concerning the derivative concept for
students of economics – A normative point of view and students' perspectives. In
E. Nardi, C. Winsløw, & T. Hausberger (Eds.), Proceedings of the First
conference of International Network for Didactic Research in University
Mathematics. Montpellier, France: University of Montpellier and INDRUM.
Feudel, F. (2017). Students’ interpretation of the derivative in an economic context.
In T. Dooley & G. Gueudet (Eds.), Proceedings of the 10th Congress of the
European Society for Research in Mathematics Education. Dublin, Ireland: DCU
Institute of Education & ERME.
Mkhatshwa, T., & Doerr, H. (2015). Students' understanding of marginal change in
the context of cost, revenue, and profit. In K. Krainer & N. Vondrová (Eds.),
Proceedings of the 9th Congress of the European Society for Research in
Mathematics Education. Prague, Czech Republic: Charles University in Prague,
Faculty of Education and ERME.
Sydsæter, K., & Hammond, P. J. (2015). Mathematik für Wirtschaftswissenschaftler:
Basiswissen mit Praxisbezug. Hallbergmoos, Germany: Pearson.
Tall, D., & Vinner, S. (1981). Concept image and concept definition in mathematics
with particular reference to limits and continuity. Educational Studies in
Mathematics, 12(2), 151-169.
Wilhelm, J. A., & Confrey, J. (2003). Projecting rate of change in the context of
motion onto the context of money. International journal of mathematical
education in science and technology, 34(6), 887-904.

104 [Link]:indrum2018:174442
 Weekly homework quizzes as formative assessment for Engineering
students are a fair and effective strategy to increase learning?
Sandra Gaspar Martins1
1
Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia, Portugal,
sicgm@[Link]
A strategy to apply online weekly homework quizzes as formative assessment for
Engineering students was designed and tested in order to study if it increases
student’s learning. The strategy was to make optional weekly online quizzes with
questions not randomly generated that students may retry over and over again until
to reach the correct answer, they contribute to 10% of grade but only if students get
45% or more in usual pencil and paper assessment.
The quizzes were applied to two different mathematics courses (Single and
Multivariable Calculus) of two different Engineering degrees, each one to around
100 students and during a semester. Student’s adherence was very high, nearly all
students refer quizzes as fair and useful to learning. Students’ grades were compared
with several other years.
Keywords: The role of digital and other resources in university mathematics
education, Assessment practices in university mathematics education, Teaching and
learning of analysis and calculus.

INTRODUCTION
Frequent online quizzes have been suggested as a strategy to enhance learning by
several institutions and researchers. The National Centre for Public Policy and Higher
Education in the U.S.A (Twigg, 2005) consider computer based continuous
assessment and feedback to be a key strategy for quality improvement in learning.
According to Gibbs (2000), student assessment is an effective way to increase
understanding and online quizzes force students to spend more time working
productively outside of class. Tuckman (1998) refers this as being especially valuable
to procrastinators. One method that can be used to address the crisis in college
mathematics, according to Thiel, Peterman, and Brown (2008), is to ‘provide regular
assessment of progress’ and they state that ‘online homework and quizzes with online
grading provide students with immediate feedback, the opportunity to correct their
homework mistakes, and ongoing assessment of their success in the course’. Booth
(2012) considers that homework should be given out at regular times, over regular
intervals, on a weekly basis; proposing that learning is work and students should
develop regular work habits in order to succeed. Feedback is crucial for student
success but giving adequate feedback with large class sizes is difficult and therefore
automated systems are a useful solution to the large class size problem.
Quizzes are part of several successful approaches with different kinds of students,
both in top universities and in other higher education institutions. Examples include:

105 [Link]:indrum2018:174631
TEAL (Dori & Belcher, 2004) at Massachusetts Institute of Technology (MIT);
SCALE-UP (Beichner, et al, 2007) at North Carolina State University; Peer Teaching
(Lasry, Mazur, & Watkins, 2008) at Harvard University.
Particularly, in higher education mathematics teaching, several approaches have been
raised but literature is not yet in agreement about the effectiveness of quizzes to
enhance learning (Siew, 2003; Varsavsky, 2004; Myers & Myers, 2007; Blanco,
Estela, Ginovart & Saà, 2009; Lim, Thiel & Searles, 2012; Broughton, Robinson &
Hernandez-Martinez, 2013; Shorter & Young, 2011).

CONTEXT
This research took place in two mathematics’ courses to Engineering students of
Instituto Superior de Engenharia de Lisboa, Portugal, each during a semester. In those
semesters, weekly online quizzes on Moodle (the learning management system of the
institute) were made available for a week each. The AM2 course in 2013/14 was
about Multivariable Calculus, the MAE course in 2015/16 was about Single Variable
Calculus. Around 100 students and 3 teachers were involved in each course.
The quizzes were called ‘Mini-tests’ to reinforce their relevance. The ‘regular’
assessment involved two face-to-face tests or the First Exam or the Second Exam.
For AM2, the quizzes scored up to two values proportional to the best 12 (out of 14)
grades in the quizzes and it was added if the student scored more than 9.0 values (out
of 20) in ‘regular’ assessment. For MAE, it was slightly different: the quizzes valued
10% of the grade if the student scored more than 9.0 values (out of 20) in the
‘regular’ assessment and if this grade was better than the ‘regular’ grade. In both
cases the quizzes were optional.
The aim of the quizzes was not to assess students, it was to make them study more,
not to postpone, not to study first the other subjects that were naturally more pleasant
for them (since they belong to their study area); to make students more aware of their
level of understanding (often students only realise that they cannot solve the exercises
when they go to the first test, in the middle of the semester). Students are usually
optimistic about their capabilities (Wandel, 2015). It was written in Moodle and
teachers repeatedly reminded students that the aim of the quizzes was to make
students study more and be aware of their level of understanding; that students could
copy all quizzes but, probably would not get the 9.0 values required in ‘regular’
assessment and therefore, it not be worthwhile.

THE QUIZZES
The quizzes were produced through the ‘Moodle activity: test’. It allows the
introduction of images and mathematical symbols using LaTeX (see Fig. 1). The
possibility of creating questions with different instances for each student was
considered, but it would take much more time to create questions and students also
know how to solve a problem with a constant instead of a number, so it did not seem
worthwhile.

106 [Link]:indrum2018:174631
Figure 1. Multiple-choice questions including a figure and mathematical text, MAE
and AM2 example. (Translated)
Whenever it was possible, we used numeric or short answers instead of multiple-
choice answers since in multiple-choice answers, with a few tries, students could get
the correct answer. The type of questions that we most used was ‘embedded
answers’, because this enables a teacher to embed more than one sub-question and
those sub-questions may be chosen from all the different question types: numeric,
short answers, multiple-choice, true or false, etc. The ‘embedded answer’ question
type allows the teacher to evaluate the student through their pathway and not only
their final result (see Fig. 3). The feedback does not show the correct answer.

Figure 2. A question with multiple embedded questions along the path (including
numerical answers), an AM2 and MAE example. (Translated)

RESEARCH DESIGN, DATA AND RESULTS

This research design is a quasi-experience (not an experience since not all variables
could had been controlled) where two sets of quizzes were applied to two
mathematics courses. The research question of this study is: are the quizzes (applied
with this strategy) a fair and effective tool to increase students’ learning? The strategy
for application of the quizzes is that they are weekly, online, non-mandatory, count
towards grades if students achieve a certain level on traditional assessment, are not
randomly generated and students may resubmit without penalty. This research
question was split into four sub-questions:

107 [Link]:indrum2018:174631
 • RQ1: Did the students adhere to the quizzes?
• RQ2: What was students’ perception of the quizzes?
• RQ3: Did students felt quizzes as unfair?
• RQ4: Did the quizzes increase students’ grades?
The instruments utilized were: a students’ survey about the quizzes; data from the
answers to the quizzes; and course grades over several semesters. The quizzes were
applied to two mathematics courses: AM2 with 104 subscribed students and MAE
with 108.
The anonymous survey on Moodle was addressed to all students for each edition. The
sample of students who answered the survey was reasonable. From the 104 students
subscribed to AM2, all subscribed to Moodle, 65 answered the survey. From the 108
students subscribed to MAE, 94 in Moodle, 61 answered the survey. Moreover, by
splitting the students by their grade at the first test (the survey was applied before the
second test), the number of students answering the survey with a given grade
reasonably correlates to the number of students in general who achieved that grade.
Pearson correlation coefficients are ρ = 0.6 and ρ = 0.5 respectively.
Students of the institute do not have precedencies among courses and may be
subscribed to a large number of courses, so it is usual that students subscribe to many
courses where, in fact, they do not attempt to achieve success. We may verify this, for
example, by noticing that from the 108 students subscribed to MAE only 94 were
subscribed to Moodle, so the 14 remaining students did not access anything from the
course: syllabus, slides, quizzes etc. Since there is no simple and fair way of
identifying these students, in this research we always use the subscribed students to
make measures. However, it is relevant to have in mind that it includes those ‘ghost
students’.

RQ1: DID THE STUDENTS ADHERE TO THE QUIZZES?

AM2 had 104 subscribed students, 79 attempted regular assessments and 76 students
attempted at least one quiz. All but one of the approved students answered at least
one quiz. The final quiz grade was the average of the best 10 out of 14 grades in
quizzes, so it was natural that the last four quizzes had lower attendance (and for this
reason we modified this rule for MAE, where the best 12 grades were chosen).

108 [Link]:indrum2018:174631
Chart 1. The number of students that answered AM2 quizzes split by grade.
MAE had 108 subscribed students, 103 completed regular assessment and 93 students
attempted at least one quiz. All approved students answered at least one quiz. The
final quiz grades were the average of the best 12 out of 14 grades in quizzes, so it is
natural that the last two quizzes had a lower attendance (this rule changed from
AM2). It is important to note that, for example, in Q5 the number of students with a
total grade was lower than in the other quizzes and the number of attempts to solve
the quiz was higher than in the others (326). This shows that students were, in fact,
trying to reach the correct answers (this test was particularly large and complex).

Chart 2. The number of students that answered MAE quizzes split by grade. The
number of attempts to answer the quiz, registered by Moodle, is in parenthesis.
A large portion of students got a very high grade, but this was natural since students
may retry without penalty and the questions were equal to all students, so it was
expected that students talk to each other and reach the correct answer.
The quizzes were not mandatory and improved the grade if the student got more than
9 out of 20 values in regular assessment, so it could be expected that many students
decided not to take it. However, on a regular basis, nearly half of the subscribed
students answered the quizzes.
An objective result was, despite of the optional policy, that students strongly adhered
to quizzes. The percentage of subscribed students that answered one quiz was
93/108=86% and 76/104=73%. All the quizzes had a high rate of attendance. Among
the students that undertook ‘regular’ assessment, almost all took a quiz and a large
percentage got high average grades on the quizzes.

RQ2: WHAT WAS STUDENTS’ PERCEPTION OF THE QUIZZES?

Table 1 shows that, according to the survey, none of the students thought that the
quizzes were of no interest and did not care about the quizzes, while a large
percentage believed that the quizzes reminded them to study, showed them the level
that they were reaching and encouraged them to learn new things; some of those
things they thought they understood but in fact they did not.

109 [Link]:indrum2018:174631
 AM2 MAE
Total 65 100% 61 100%
Quizzes remind me to study the subject every week. 55 85% 50 82%
Quizzes show me there are things I thought I knew but
48 74% 53 87%
I didn’t.
Quizzes help me to have a better perception of the
47 72% 38 62%
level I'm reaching.
I learn new things answering to quizzes. 33 51% 35 57%
Quizzes have no interest. 0 0% 0 0%
I do not care for quizzes, I just copy the results. 0 0% 1 2%
I do not care for quizzes, I not even copy the results. 1 2% 0 0%

Table 1. Students’ answers to ‘Select ALL the statements that you agree with’ in both
surveys.

Chart 3. Percentage of students answers to ‘The quizzes were…’ in both surveys.

Summarising, more than 90% of students found quizzes useful (Chart 3); that they
study more due to the quizzes. Students agree that quizzes remind them to study,
show them that there were things that they thought that had understood but did not,
encouraged them to learn new things and gave them a better perception of level that
they were reaching.

RQ3: DID STUDENTS FELT QUIZZES AS UNFAIR?

In daily life as a teacher, teachers tell several times that one reason why they do not
use online quizzes is because students may be cheating and it may generate
unfairness. To avoid that problem, it was strongly emphasised to students that quizzes
were much more relevant as formative assessments than summative assessments;
students could resubmit the quiz without penalty to stimulate them to try to answer by
themselves without fear of being penalised; and a clause was included that the
quizzes only count towards grades if students get 9.0 values (out of 20) in regular
assessments, as in Varsavsky (2004). As result, the answers in the survey to the
question ‘Quizzes generate unfairness?’ show that very few students perceive quizzes
as unfair (see Chart 5).

110 [Link]:indrum2018:174631
Chart 4. Percentage of students answers to ‘How do you answer to quizzes?’ in both
surveys.

Chart 5. Percentage of students answers to ‘Quizzes generate unfairness?’ in both

surveys.
When questioned in the survey, no student stated that they had copied the results (see
Chart 4), despite it being reinforced in that question that the survey was automatically
anonymous.
Therefore, with this approach, the level of unfairness of quizzes is not considered as
relevant.

RQ4: DID THE QUIZZES INCREASE STUDENTS’ GRADES?

Since the goal was that all students achieve a total score in all quizzes, is was
expected that quiz grades would not correlate to final grades. This did occur and it
was verified using the non-parametric Spearman Rho for AM2 ( ρ = 0.34, N = 54, p =
0.01) and for MAE ( ρ = 0.28, N = 61, p = 0.03), since data were not normal (Kolmogorov-
Smirnov, p < 0.01).
According to Chart 6, around 70% of students that answered the survey, believe that
quizzes helped them achieve a higher grade.

111 [Link]:indrum2018:174631
Chart 6. Percentage of students’ answers to ‘Without quizzes, I’ve scored…’ in both
surveys.
The data of Tables 2 and 3, relate to six responsible teachers/approaches and ten
different teachers. The syllabus was essentially the same across the semesters but the
approaches were naturally different. In the intervention semesters, the responsible
teachers were also different. So, the quizzes were not the only different variable in
that semester, thus we cannot attribute grade differences directly to the quizzes. For
AM2, the pass rate nearly doubled in that semester, the average grade also increased
significantly.
AM2 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16

S1 S2 S1 S2 S1 S2 S1 S2 S1 S2 S1 S2

Subscribed students 101 200 128 153 90 123 80 104 56 66 56 108

Approv. students 27 38 31 41 20 23 12 54 10 19 16 33

Average appr. grade 11.7 11.8 12.3 11.7 13.9 12.4 11.5 11.7 11.5

Pass/Subscribed 27% 19% 24% 27% 22% 19% 15% 52% 18% 29% 29% 31%

F+G J+K+ J+K+

Professors A+… A+… A+… A+… A+B A+C D+E D+F D +I
+H IL I

Table 2. Grades of AM2 students across ten semesters, the letter representing the
coordinator teacher is underlined and the experimental semester is shaded.
The MAE course had, in some editions, five or six quizzes in class. It is curious to
note that in the year that there were no quizzes, the pass rate was much lower. And
the MAE pass grade and the average grade had the highest value in the experimental
semester. However, it may have been a coincidence, we do not have enough data to
reach any conclusions, it is just a positive indication.
MAE 2011/12-SI 2012/13-SI 2013/14-SI 2014/15-SI 2015/16-SI

Subscribed students 73 109 121 125 108

Pass students 17 30 58 56 61

Average pass grade 12.7 12.2 13.5 12.7 13.5

Pass/Subscribed 23% 28% 48% 45% 56%

Number of quizzes 0 6 in class 5 in class 5 in class 14 online

Professors A A A A+B B+A

Table 3. Grades of MAE students across five semesters, the letter representing the
coordinator teacher is underlined and the experimental semester is shaded
Summarising, as expected, quiz grades do not correlate to final grades; around 70%
of respondents to the survey state that due to the quizzes they achieved a better grade.

112 [Link]:indrum2018:174631
The pass rate and the average grade increased significantly in the semesters that the
quizzes were applied, which is a positive indicator but cannot be directly attributed to
quizzes.

CONCLUSIONS
Two sets of 14 weekly quizzes on Moodle were available to all the engineering
students on two mathematics courses (Single and Multivariable Calculus). The online
quizzes were not mandatory, counted to grading if the student had more than 9 out of
20 values on traditional assessments, were not randomly generated and students could
resubmit without penalty. The research question is ‘Are the quizzes (applied with this
strategy) a fair and effective tool to increase students’ learning?’
In the answers to the survey, more than 90% of students found quizzes useful; more
than 60% stated that studied more due to the quizzes; students agreed that quizzes
reminded them to study; showed them that there were things that they thought they
understood but did not; made them learn new things and gave them a better
perception of the level that they were reaching.
The quizzes were not mandatory so students may have just ignored them. Although a
large proportion of students attempted quizzes and kept answering them until the last
ones.
Quiz questions were not randomly generated, so all students got the same questions
and naturally, students shared the solutions with each other. To avoid unfairness, it
was strongly emphasised that quizzes were important to students’ formative
assessment, to allow them to test themselves and get feedback on their level of
understanding. Moreover, quizzes only contributed to grades if the students got more
than 9 out of 20 values in ‘traditional’ assessments. Moreover, if a student copied
many quiz results, probably would not achieve the minimum grade and it would not
be worthwhile. The result was that, in the answers to the surveys, very few students
stated it as being unfair. Over 70% of respondents to the surveys stated that due to the
quizzes they achieved a better grade. The pass rate and the average grade increased
significantly in the semesters that the quizzes were applied, which is a positive
indicator, but it cannot be directly attributed to the quizzes.
This research suggests that these quizzes, with this strategy, are a fair and useful tool
to increase students learning.

REFERENCES
Beichner, R., et al. (2007). The SCALE-UP project. Research-based reform of
university physics, 1(1), 2-39.
Blanco, M., Estela, M., Ginovart, M., & Saà, J. (2009). Computer assisted assessment
through moodle quizzes for calculus in an engineering undergraduate course.
Quaderni di Ricerca in Didattica, 19(2), 78-83.

113 [Link]:indrum2018:174631
Booth, D. J. (2012). Managing Mathematics with CALMAT. International Journal of
Innovation in Science and Mathematics Education, 2(1).
Broughton, S., Robinson, C., & Hernandez-Martinez, P. (2013). Lecturers’
perspectives on the use of a mathematics-based computer-aided assessment
system. Teaching mathematics and its applications, 23, 1-14.
Dori, Y. & Belcher, J. (2004). Improving Students’ Understanding of
Electromagnetism through Visualizations-A Large Scale Study. In Paper submitted
to the 2004 NARST Annual Meeting, Vancouver.
Gibbs, G. (2000) Changing student learning behaviour outside of class. The
Professional & Organizational Development Network in Higher Education. 11.
Lasry, N., Mazur, E., & Watkins, J. (2008). Peer instruction: From Harvard to the
two-year college. American Journal of Physics, 76(11), 1066-1069.
Lim, L., Thiel, D. & Searles, D. J. (2012). Fine tuning the teaching methods used for
second year university mathematics. International Journal of Mathematical
Education in Science and Technology, 43(1), 1-9.
Myers, C., & Myers, S. (2007) Assessing assessment: The effects of two exam
formats on course achievement and evaluation. Innovative Higher Educ., 31, 227-
36.
Shorter, N., & Young, C. (2011). Comparing assessment methods as predictors of
student learning in an undergrad. mathematics course. IJMEST, 42(8), 1061-67.
Siew, P. F. (2003). Flexible on-line assessment and feedback for teaching linear
algebra. I. J. of Mathematical Education in Science and Technology, 34(1), 43-51.
Thiel, T., Peterman, S. & Brown, M. (2008) Addressing the Crisis in College
Mathematics: Designing Courses for Student Success, Change: The Magazine of
Higher Learning, 40:4, 44-49.
Tuckman, B. W. (1998). Using tests as an incentive to motivate procrastinators to
study. The Journal of Experimental Education, 66(2), 141-147.
Twigg, C. A. (2005). Course Redesign Improves Learning and Reduces Cost. Policy
Alert. National Center for Public Policy and Higher Education.
Varsavsky, C. (2004). Can online weekly quizzes contribute to learning in
mathematics. In Proceed. of 9th Asian Tech. Conf. in Mathematics, Singapore.
Wandel, A. P., Robinson, C., Abdulla, S., Dalby, T., Frederiks, A., & Galligan, L.
(2015). Students’ mathematical preparation: Differences in staff and student
perceptions. I. J. of Innovation in Science and Math. Education, 23(1).

114 [Link]:indrum2018:174631
 The use of integrals in Mechanics of Materials textbooks for
engineering students: the case of the first moment of an area
Alejandro S. González-Martín and Gisela Hernandes-Gomes
Université de Montréal, Département de Didactique, Canada
[Link]-martin@[Link]; [Link]@[Link]
Research has reported on the difficulties engineering students face in relating the
content of their mathematics courses to what is taught in their professional courses.
One way to address these difficulties is by better understanding how mathematical
notions are used in professional engineering courses. This paper analyses how the
notion of first moment of an area – which is defined as an integral – is used in civil
engineering courses. Basing our analysis on elements from the anthropological
theory of the didactic, we are currently analysing a classic Mechanics of Materials
book. Our findings indicate that although first moments are introduced as an
integral, the textbook’s tasks do not require students to use techniques typically
introduced in a traditional calculus course.
Keywords: Mathematics for engineers, teaching and learning of analysis and
calculus, textbooks, anthropological theory of the didactic, first moment of an area.

INTRODUCTION
Engineering courses are usually organized into two groups: basic science courses
(which are taught in the first two years, including foundational skills in mathematics
and physics), and technical courses (which appear later in the programme and are
more specific to each field of engineering). However, research in engineering
education and mathematics education indicates that engineering students encounter
many difficulties in their mathematics courses in the first years of study, which can
lead to high failure rates, and in many cases, result in students dropping out of
engineering programmes (Ellis, Kelton, & Rasmussen, 2014). In this sense, “poor
mathematics skills are a major obstacle to completing […] engineering programs”
(Fadali, Johnson, Mortensen, & McGough, 2000, p. S2D-19).
Researchers have identified some negative situations for students who pass these
mathematics courses. One situation is that these students often find it difficult to
relate the learned mathematical content to the content of the professional courses.
For Flegg, Mallet, and Lupton (2011, p. 718) “without the explicit connection
between theory and practice, the mathematical content of engineering programs may
not be seen by students as relevant”. Another situation is that in spite of having
passed the mathematics courses (with a rather rigid structure and rare concrete
applications relevant to engineering), students must apply mathematics in their
engineering courses, where many new mathematical notions appear without having
been encountered in the previous mathematics courses (Hochmuth, Biehler, &

115 [Link]:indrum2018:172539
Schreiber, 2014, p. 694). Faced with these problems, the mathematics and
engineering education communities have been engaged in research and discussion on
“how to improve engineering students’ mathematics learning, and hence their service
teaching” (Bingolbali, Monaghan, & Roper, 2007, p. 764).
Our current research program investigates how calculus notions are used in
engineering courses, aiming at identifying possible ruptures between how notions are
first introduced and used in calculus, and how they are later used in professional
courses. First, we analyse how engineering textbooks present these notions, working
under the principle that most tertiary instructors organise their teaching using
textbooks as an important resource (e.g., Mesa & Griffiths, 2012). The manner in
which mathematics notions are used in professional courses has not been the subject
of much research. However, we believe this type of research could help bridge the
gap between two communities. On the one hand, mathematics lecturers in
engineering programs could benefit from knowing how their course content is used
in professional courses; on the other hand, professional course instructors could
benefit from a critical analysis of their use of mathematics, to help their students
make connections between the content of mathematics and professional courses. For
example, our analysis of the way integrals are used to define bending moments for
beams in strength of materials textbooks for civil engineering reveals different uses
of “the same” object (González-Martín & Hernandes Gomes, 2017a). Although
bending moments are defined as an integral, the tasks, techniques, and justifications
used in calculus courses are very different from the ones presented in professional
engineering courses; this may result in students not recognising “the same” object in
two different courses, and they may question the relevance of integration techniques
that are not used in tasks concerning bending moments. In this paper we develop the
content of González-Martín & Hernandes-Gomes (2017b) as we explore the use of
integrals to introduce another engineering notion: first moment of an area. We aim to
address two questions: how is the content related to integrals used in engineering to
work with first moments of an area, and how does this use relate to the content in
calculus courses?
Defining first moment of an area
Moments of areas are topics commonly taught in engineering courses that cover
strength of materials. Due to space limitations, in this paper we focus on the first
moment of an area. In civil engineering, for example, to solve bending problems one
must take into account some specific geometrical characteristics of cross-sections of
a bar, which is the general term for structures that include beams (Feodosyev, 1973).
In this situation, the notion of first moment of an area is used to calculate the
centroid of an area and the shearing stresses in transverse bending. The centroid of
an area A is its geometrical barycentre and is the point C of coordinates and such
that the following relationships hold true: A x dA  Ax and A y dA  Ay .

116 [Link]:indrum2018:172539
 Let A be an area situated in the xy plane (Figure 1), using x
and y as the coordinates of an element of area dA.
According to Beer, Johnston, DeWolf, and Mazurek
(2012, p. A2), the first moment of an area A with respect
to the x axis (resp. y axis) is mathematically defined as the
integral Qx  A y dA (resp. Q y  A x dA ). In both integrals,
the index A in the integral sign indicates that the integral is
Figure 1: General area
calculated over the whole cross-sectional area. Both
A with infinitesimal area
integrals characterize the sum of the products of each
dA in the xy plane (Beer
et al., 2012, p. A2).
element of area dA and its distance to the respective axis
(x or y) and are measured in cubic units (Beer et al., 2012).
When an area possesses an axis of symmetry, the first moment with respect to that
axis is zero, since every element of area dA of abscissa x (resp. ordinate y)
corresponds to an element of area dA’ of abscissa –x (resp. ordinate –y). This implies
that when an area possesses an axis of symmetry, its centroid is located on that axis.
For instance, in a rectangular cross-section (two axes of symmetry), its centroid C
coincides with its geometric centre. Determining the position of the centroid is
important, since several forces in a bar pass through its centroid.
To illustrate these definitions and their calculation with an
example, let us consider the case of a bar with a
rectangular cross-section (Figure 2). If we consider the
expression above, Qx  A y dA , we can take dA as the area
of the grey rectangle, whose dimensions are b and dy.
Substituting dA in the integral, we have that
Q x   y dA   y b dy . Calculating this integral throughout
A A

all the vertical extension of the rectangular cross section,

h Figure 2: Determination
we obtain: Qx  0 y b dy . Calculating the integral, we of the first moment with
y2
h
h2 02 h2 respect to the x-axis of
obtain: Qx  b b  b , therefore Q x  b . an area with rectangular
2 0
2 2 2
cross-section.
THEORETICAL FRAMEWORK
As stated above, we are interested in analysing how calculus notions are used in
professional engineering courses, aiming at identifying possible breaks from the
content in calculus courses. For our research, we use tools from the anthropological
theory of the didactic (ATD – Chevallard, 1999) because it considers human
activities as institutionally situated. In this sense, knowledge about these activities
and their raison d’être is also institutionally situated (Castela, 2016, p. 420). In
particular, ATD offers a general epistemological model of mathematical knowledge,

117 [Link]:indrum2018:172539
where mathematics is seen as a human activity through which various types of
problems are studied (Barbé, Bosch, Espinoza, & Gascón, 2005, p. 236).
The key element we use in our analysis is the notion of praxeology (or, in our case,
mathematical organisation or mathematical praxeology – MO hereinafter), which is
formed by a quadruplet [T/τ/θ/Θ] consisting of a type of task T to perform, a
technique τ which allows the task to be completed, a discourse (technology) θ that
explains and justifies the technique, and a theory Θ that includes the discourse. The
first two elements [T/τ] are the practical block (or know-how), whereas the
knowledge block [θ/Θ] describes, explains, and justifies what is done. These two
blocks are important elements of the ATD model of mathematical activity that can be
used to describe mathematical knowledge. Furthermore, ATD distinguishes different
types of MO: punctual, which are associated with a specific type of task; local, which
integrate multiple punctual MOs that can be explained using the same technological
discourse; and regional, which integrate local MOs that accept the same theoretical
discourse (Barbé et al., 2005, pp. 237-238).
Praxeologies, like knowledge in general, may move from the institution where they
emerge to other institutions that find them useful (Castela & Romo Vázquez, 2011).
This is the case, for instance, of mathematical notions that are used to solve
engineering problems. In this process, there are transposition effects on the
concerned praxeologies (Castela & Romo Vázquez, 2011; Chevallard, 1999). We
consider the work of Castela (2016), who identified that “when a fragment of social
knowledge, produced within a given institution I, moves to another one IU in order to
be used, the ATD’s epistemological hypothesis states that such boundary crossing
most likely results in some transformations of knowledge, called transpositive
effects” (p. 420). In this boundary-crossing process, some (or all) elements of the
original praxeology may evolve. Therefore, it is important to analyse the types of
tasks and techniques as well as the discourses and theories employed. To that end,
our research identifies specific local MOs present in professional courses; we analyse
how calculus notions are used (practical block) and whether this use relates to the
way the notions are usually presented in calculus courses (knowledge block).

METHODOLOGY
It is worth noting that, in order to understand how calculus notions are used in
engineering courses, we have had several exchanges with an engineering teacher
holding bachelor’s and master’s degrees in civil engineering, with more than 28
years of experience teaching a variety of professional engineering courses at
Brazilian universities. This teacher has explained notions related to his field and has
helped us identify course content in which first-year calculus notions are used.
At this teacher’s university, first moments of area are introduced during the third
semester of the programme (second year), in the Strength of Materials for Civil
Engineering course (students take calculus in their first two semesters). The course’s

118 [Link]:indrum2018:172539
reference book is Beer et al. (2012). First moments are initially cited in chapter 4
(4.2. Stresses and deformations in the elastic range). We proceeded in two stages:
 First, we analysed the general structure of the content related to integrals in the
calculus courses. We identified the main tasks proposed to students, grouping
them according to the technological elements needed, identifying therefore the
main local MOs that structure this content.
 Second, we started our analyses of the reference book for the Strength of
Materials course. We identified all instances where first moments appear in the
book (using key words to search in an electronic version of this book). For
each occurrence of this notion, we are currently analysing the tasks presented
in the book where first moments are used. For each task, we are analysing the
techniques and discourses (technologies) the textbook uses. As the notion of
first moment is used in different chapters of the book, where different
professional notions are introduced and explained, the technological
discourses are quite varied, giving place to various MOs. The next section
provides specific details of our analysis.

DATA ANALYSIS AND DISCUSSION

Calculus is taught in the first year of the program over two semesters in two courses:
Calculus I and Calculus II. Up until a few years ago, integrals appeared only in
Calculus II, but some content was moved to Calculus I because Physics II (a course
in the second semester) requires a knowledge of integrals. Integrals appear towards
the end of the first course and are the main topic in the second course (the second
author of this paper has taught Calculus I for 15 years and Calculus II for two years).
The content covers indefinite integral (antiderivative of f), Riemann sum and definite
integral, applications for the calculation of areas, integration by substitution,
volumes (Calculus I), and integration techniques, improper integrals, and arc length
(Calculus II). The main source for the calculus courses is the classic book by Stewart
(2012). The content concerning integrals is basically structured using two local MOs.
The first, MOM1, introduces techniques for calculating indefinite integrals
(immediate integration to begin with, followed by various integration techniques);
however, theoretical elements justifying the different integration methods are mostly
absent and those present are explained without a proof. The second MO, MO M2,
introduces Riemann sums to formally define integrals and interpret them as areas,
and leads to the Fundamental Theorem of Calculus and the calculation of definite
integrals using Barrow’s rule; this leads to some applications of the integral (area,
volume …). Many of the techniques used in MOM2 are derived from MOM1.
We are currently analysing the use of first moments and centroids in the engineering
textbook (Beer et al., 2012), which numbers more than 800 pages. So far, our first
analyses indicate that although this content is introduced as an integral, the

119 [Link]:indrum2018:172539
techniques employed do not call for integration. Our ongoing analyses of the use of
first moments in the textbook are summarised in Figure 3.
Description of use Terms used Chapter – Sections
The term appears in a theoretical First moment;
4.2 (p. 245); 4.2 (p. 245); 4.4 (p. 262); 4.6 (p. 274)
explanation. It appears with an First moment
expression using the integral sign, of an area; Q;
6.1 (p. 421); 6.1 (p. 421); 6.3 (p. 437)
but no calculation is required. centroid
4.3 (p. 262)
First moment;
The term appears in a theoretical 6.1C (p. 424); 6.1C (p. 424); 6.4 (p. 440); 6.6 (p. 454; pp. 459-460);
First moment
explanation. It appears without an Review (p. 467)
of an area; Q;
expression using the integral sign. 8.1 (p. 559); Review (p. 591)
centroid
9.5A (p. 651); 9.5A (p. 651); 9.5A (p. 654); 9.5A (p. 654); 9.6B (p. 666)
4.2 (p. 247; p. 248)
Concept application: It is First moment;
6.1 (p. 422); 6.1 (p. 422); 6.3 (p. 438); 6.6 (pp. 456-457)
involved in some calculations, but First moment
8.3 (pp. 577-578)
no calculation of integrals is of an area; Q;
9.5A (p. 652); 9.5A (p. 653); 9.5A (p. 655); 9.5A (p. 656);
required. centroid
9.6B (p. 667); 9.6C (p. 669)
Sample problem: It is involved in 4.3 (p. 251); 4.5 (p. 265); 4.10 (p. 326)
First moment; 6.2 (p. 429); 6.5 (pp. 443-446); 6.6 (p. 462)
some calculations, but no
Q; centroid
calculation of integrals is required 8.3 (p. 583)

Figure 3: Synthesis of uses of first moments in Beer et al. (2012).

Here, due to space limitations, we describe our analysis of two MOs present in the
textbook at points where first moments come into play. It is worth noting the book
advises students that they should already have completed a course in statics, that the
properties of moments and centroids are explained in Appendix A, and that this
material can be used to reinforce the discussion of the determination of normal and
shearing stresses in beams in chapters 4, 5, and 6 (Preface, p. x).
First case: MOE1
The initial use of first moments, MOE1, concerns stresses and deformations in the
elastic range (section 4.2 of the book). Its aim is to calculate the maximum stress that
beams can resist, resulting in some recommendations about the size and shape of
beams. Using some formulae, the book arrives at  ydA  0 and concludes: “This
equation shows that the first moment of the cross section about its neutral axis must
be zero. Thus, for a member subjected to pure bending and as long as the stresses
remain in the elastic range, the neutral axis passes through the centroid of the
section” (p. 245, italics in the original). This is the first apparition of first moments in
the book; however, they are not explained and the authors refer readers to Appendix
A. In Appendix A, first moments and their link with the centroid are introduced in a
similar manner as in this paper, using implicitly theoretical elements from MO M2
(namely, the interpretation of an integral). However, the book makes the connection
with the centroid and deduces many integrals using geometric considerations (and
the properties of the centroid), and adds “Centroids of common geometric shapes are
given in a table inside the back cover” (p. A3). Therefore, this content is justified
vaguely through some basic integral content (present in MO M2), but mostly by using
geometric considerations. The tasks in MOE1 calculate stresses and bending moments

120 [Link]:indrum2018:172539
 in known geometrical shapes. In
the case of a rectangle (Figure 4-
left), the coordinates of the
centroid are deduced using
geometry (and not techniques
derived from MOM1 or MOM2); the
same approach is used in the case
of a semicircular cross-section
Figure 4: Left: The centroid is placed calculating (Figure 4-right).
the half measure of each side of the rectangle Thus, although the notions of first
(Beer et al., 2012, p. 247). Right: The centroid is moment and centroid are necessary
placed using geometric formulae (p. 248). to solve tasks in MOE1, the
techniques employed are not based
on elements derived from MOM1 or MOM2. Students can solve the tasks present in
this MOE1 without using any of the techniques learned in MOM1 or MOM2, or hardly
any of the technological elements present in them.
Second case: MOE2
First moments and centroids are also used in chapter 6. In section 6.1A, Shear on the
horizontal face of a beam element, MOE2 seeks to determine the horizontal shear per
unit length (or shear flow) on a beam. Defining Δx as the length of a section of the
beam, V as the shear force, ΔH as the horizontal shearing force exerted on the lower
face of the element, Q as the first moment, and I as the centroidal moment of inertia,
and using techniques and technological elements covered in this and previous
H VQ
chapters, the horizontal shear per unit length (q) is deduced as: q =  . It is
x I
worth noting that the techniques used to arrive at this expression involve integrals,
but they are referred to in terms of notions belonging to MO E2. The above expression
is used to solve tasks such as the
A beam is made of three planks, 20 by 100 mm in cross-
one in Figure 5.
section, and nailed together. Knowing that the spacing
between nails is 25 mm and the vertical shear in the beam The resolution of the task is
is V = 500N, determine the shearing force in each nail. based on the determination, via
different expressions, of Q and I
(since V = 500N is provided) to
find the horizontal force exerted
on the lower face of the upper
plank. For the first moment, Q,
the following technique is
presented: “Recalling that the
Figure 5: Task and diagrams used concerning first element of an area with
horizontal shear (Beer et al., 2012, p. 422). respect to a given axis is equal to
the product of the area and of the

121 [Link]:indrum2018:172539
distance from its centroid to the axis, Q = A y ” (p. 422). The area of the cross-section
of the upper plank is calculated as 0.020m × 0.1m, and the coordinate of the centroid
of this horizontal plank with respect to the axis of symmetry of the cross-section is
0.05m + 0.01m (that is, half the measure of the central plank, plus half the measure
of the horizontal plank). Q is thus obtained as: Q = A y = (0.020m ×
0.100m)(0.060m) = 120 × 10-6m3. We see that, once more, the tasks to solve in this
MOE2 involve cross-sections with geometrical shapes that make use of geometrical
considerations, thus avoiding techniques belonging to MO M1 or MOM2.
Although the technological elements of MO E2 refer to elements that imply the use of
integrals, tasks are presented in such a way that previously deduced formulae can be
used and magnitudes can be deduced using these formulae and geometrical
considerations. The book later provides a table with values (Figure 6). Therefore, it
is possible for students to simply memorise the formulae or use the tables to solve
the given tasks without actually using any technical or technological element derived
from MOM1 or MOM2.

Figure 6: Areas and centroids of common shapes (Beer et al., 2012, p. 654).

FINAL CONSIDERATIONS
The data presented here, together with the data from González-Martín & Hernandes
Gomes (2017a), indicate that two notions used in civil engineering (bending moment
and first moment) are defined as integrals. This may often be used to justify the fact
that “engineers need to learn integrals”. However, our analyses show that the types
of tasks and the techniques developed are not actually derived from praxeologies
explored in a calculus course. In the two cases presented in this paper, both MOE1
and MOE2 have their own set of tasks and techniques, and both develop their own
technological discourse, which uses the notion of integral to define their own notions
and deduce properties. As Figure 3 shows, this seems to be the general situation
throughout the textbook.
As Castela (2016) states, when a fragment of knowledge (in this case, the notion of
integral) produced within a given institution moves to and is used by another

122 [Link]:indrum2018:172539
institution, this process results in a transformation of knowledge. In the case
analysed here, it is clear that all the technological discourse proper to a calculus (or
even an analysis) course pertaining to the notion of integral is transformed when it is
used to define first moments (and centroid) in a professional course, where
explanations mostly rely on basic geometric considerations. In this case, it seems that
the transpositive effects cause the notion of integral to be used very differently in
both courses. The techniques presented in the Strength of Materials course make use
of given formulae and geometric considerations, rendering the techniques introduced
in the calculus course irrelevant for the use of first moments in MOE1 and MOE2. This
may result in students not recognising the object “integral” when they move from
MOM1 and MOM2 to MOE1 and MOE2. Students may encounter many difficulties in
learning MOM1 and MOM2, but this knowledge is not necessary to solve tasks in
engineering courses, so students may question the need to learn these MOs.
It is therefore important that mathematics lecturers in engineering programs become
aware of how the notions they teach are used in professional courses. Once they
develop a better understanding of the techniques and tasks used in professional
courses, mathematics instructors may be prompted to reflect on the mathematical
praxeologies developed in their own courses and make stronger connections with the
techniques used in professional courses. This could help students transition from
mathematics courses to professional courses, enabling them to relate mathematical
content to the content of their professional courses and better understand its
relevance (Flegg et al., 2011).
We plan to continue analysing the use of integrals in professional courses in
engineering. This will be the source of future papers.

ACKNOWLEDGEMENTS
The authors wish to thank Januário Pellegrino Neto, the engineer who participated in
this project, for sharing his knowledge and experience and for making himself
available for the interview and our many follow-up questions. The research presented
here was funded by grant 435-2016-0526 of the Social Sciences and Humanities
Research Council (SSHRC) of Canada’s Insight program.

REFERENCES
Barbé, J., Bosch, M., Espinoza, L., & Gascón, J. (2005). Didactic restrictions on the
teacher’s practice: the case of limits of functions in Spanish high schools. Educational
Studies in Mathematics, 59(1), 235–268.
Beer, F., Johnston, E. R., DeWolf, T., & Mazurek, D. F. (2012). Mechanics of materials (7th
edition). New York, USA: McGraw-Hill Education.
Bingolbali, E., Monaghan, J., & Roper, T. (2007). Engineering students’ conceptions of the
derivative and some implications for their mathematical education. International Journal
of Mathematical Education in Science and Technology, 38(6), 763–777.

123 [Link]:indrum2018:172539
Castela, C. (2016). When praxeologies move from an institution to another: an
epistemological approach to boundary crossing. In R. Göller, R. Biehler, R. Hochmuth &
H.-G. Rück (Eds.), Proceedings of the KHDM Conference: Didactics of Mathematics in
Higher Education as a Scientific Discipline (pp. 153–161). Kassel, Germany:
Universitätsbibliothek Kassel.
Castela, C. & Romo Vázquez, A. (2011). Des mathématiques à l’automatique: étude des
effets de transposition sur la transformée de Laplace dans la formation des ingénieurs.
Recherches en Didactique des Mathématiques, 31(1), 79–130.
Chevallard, Y. (1999). L’analyse des pratiques enseignantes en théorie anthropologique du
didactique. Recherches en Didactique des Mathématiques, 19(2), 221–266.
Ellis, J., Kelton, M., & Rasmussen, C. (2014). Student perceptions of pedagogy and
associated persistence in calculus. ZDM – The International Journal on Mathematics
Education, 46(4), 661–673.
Fadali, M. S., Johnson, J., Mortensen, J., & McGough, J. (2000). A new on-line testing and
remediation strategy for engineering mathematics. Paper presented at the 30th Annual
Frontiers in Education Conference. Building on A Century of Progress in Engineering
Education. Conference Proceedings (IEEE Cat. No.00CH37135).
Feodosyev, V. (1973). Strength of Materials. Moscow: MIR Publishers.
Flegg, J., Mallet, D., & Lupton, M. (2011). Students' perceptions of the relevance of
mathematics in engineering. International Journal of Mathematical Education in Science
and Technology, 43(6), 717–732.
González-Martín, A. S., & Hernandes Gomes, G. (2017a). How are Calculus notions used in
engineering? An example with integrals and bending moments. In T. Dooley & G.
Gueudet (Eds.), Proceedings of the Tenth Congress of the European Society for Research
in Mathematics Education (pp. 2073-2080). Dublin, Ireland: DCU Institute of Education
and ERME.
González-Martín, A. S., & Hernandes-Gomes, G. (2017b). The use of integrals in
Mechanics of Materials for engineering: the first moment of an area. In B. Kaur, W. K.
Ho, T. L. Toh, & B. H. Choi (Eds.), Proceedings of the 41st Conference of the
International Group for the Psychology of Mathematics Education (Vol. 1, pp. 198).
Singapore: PME.
Hochmuth, R., Biehler, R., Schreiber, S. (2014). Considering mathematical practices in
engineering contexts focusing on signal analysis. In T. Fukawa-Connolly, G. Karakok, K.
Keene & M. Zandieh (Eds.), Proceedings of the 17th Annual Conference on Research in
Undergraduate Mathematics Education (pp. 693-698). Denver, Colorado: RUME
Mesa, V., & Griffiths, B. (2012). Textbook mediation of teaching: an example from tertiary
mathematics instructors. Educational Studies in Mathematics, 79(1), 85-107. doi:
10.1007/s10649-011-9339-9
Stewart, J. (2012). Calculus (7th edition). Belmont, CA: Cengage Learning.

124 [Link]:indrum2018:172539
 Predictors of performance in engineering mathematics
Julia Gradwohl1 and Andreas Eichler²
1
University of Kassel, Germany; ²University of Kassel, Germany,
eichler@[Link]

Mathematics in university courses was identified as a main obstacle for engineering

students in the beginning of their study. Since difficulties with mathematics could
lead to a dropout, our research aims to analyse students’ profiles referring to
individual characteristics that allow identifying possible risks for students’
achievement or success in the first year of study. As a first step to identify possibly
risky profiles, we started to find possible predictors of students’ performance. For
this, we give a short overview of the research state and our derived research
interest. We discuss theoretical constructs that are possibly crucial characteristics of
students with respect to encountering mathematics as an obstacle. Further, we
describe the method for measuring different variables of 182 engineering students.
Finally, we present results referring to predictors of performance in engineering
mathematics and discuss further steps of our research.
Keywords: Teaching and learning of specific topics in university mathematics,
Mathematics for engineers, motivational variables, students’ profiles, students’
achievement.

INTRODUCTION
Besides the technical disciplines, mathematics is a crucial part of higher engineering
education (SEFI, 2013). Especially in the first year, mathematics is usually taught
without considering practical applications. In lectures and tutorials mathematical
basics are provided for subsequent technical courses. However, engineering students
“encounter epistemological/ cognitive, sociological/ cultural and didactical
obstacles” (Gómez-Chacón et al., 2015, p. 2117) with mathematics struggling with
the transition from school mathematics to university mathematics (Gueudet, 2008).
Considering mathematical school skills, students show remarkable deficits at the
beginning of their study (e.g. Knospe, 2012; Thomas et al., 2012). Empirical studies
show the importance of cognitive variables, since school grades and domain-specific
previous knowledge are identified as important predictors of academic achievement
(e.g. Hailikari et al., 2008). Moreover, there is also some evidence that mathematics
plays a crucial role when dropouts from engineering studies are regarded. Heublein
(2014) stated the highest dropout rates for mathematics laden studies that were partly
caused by a low motivation and partly by excessive demands in the first part of the
studies. Also, an international review study mentioned mathematical competencies as
part of reasons for dropping out at universities (Søgaard Larsen, 2013).

125 [Link]:indrum2018:174410
Apart from cognitive aspects, further individual characteristics are relevant in the
context of learning and study success. The importance of these different aspects, e.g.
socio-demographic, motivational, emotional and social aspects, are explained by
different theoretical models (e.g. utilization of learning opportunities model:
Schrader & Helmke, 2015; models of dropout: Heublein et al., 2010) and proved by
empirical findings. The meta-analysis by Hattie (2009) summarizes the results of
over 800 studies and provides an overview of factors influencing learning success in
school. Moreover, there is evidence for the impact of self-efficacy beliefs (e.g.
Fellenberg & Hannover, 2006), academic self-concept (e.g. Hattie, 2009) and interest
(e.g. Schiefele et al., 1993a) on performance.

RESEARCH QUESTION
In view of the research state, empirical findings show insufficient mathematical
skills, high dropout rates and difficulties with mathematics at the beginning of
engineering study. Research concerning higher engineering education mostly deals
with the improvement of mathematics teaching through developing and evaluating
interventions (e.g. through integrating mathematical and technical disciplines: Rooch
et al., 2013). According to the utilization of learning opportunities model, learning
success does not only depend on the teaching offer but also on its utilization by
students. Therefore, in this project engineering students should be explored in more
detail, especially with respect to mathematics and individual characteristics.
Moreover, the study of Fellenberg & Hannover (2006) gives hints that a domain-
specific investigation is also empirically meaningful. Concerning the time frame, our
project focuses on the first year of engineering study because the secondary-tertiary
transition and dropout surveys indicate serious problems at the beginning of study. In
particular, more students decide to abandon one’s studies within the first two
semesters (Heublein et al. 2010).
Empirical basis for the relevance of individual characteristics in learning processes
exists. However, most of the studies were conducted in the context of school. Since
the subject matter and learning environment changes with the transition from school
to university, these results cannot be transferred immediately. In contrast, studies of
higher education with a special view of mathematics are rather rare. In particular,
most studies concentrate on single aspects and not on an overview of different
impacting variables (e.g. Schiefele et al. 2003). Therefore, our project draws on
previous findings to explore a multitude of impacts of individual characteristics in
learning processes of higher education with the main aim of developing mathematics
related profiles of engineering students. As a student’s profile we understand the
characteristic of different individual variables, e.g. performance, motivational
variables like interest (Gómez-Chacón et al., 2015), or engagement (Rach & Heinze,
2013), and further the relationships among these variables. Students’ profiles should
allow identifying possible risks for study success in the first year of the study. The

126 [Link]:indrum2018:174410
identification of students’ profiles is useful because it allows the development of
goal-oriented and adequate support services for students who encounter difficulties.
Since insufficient motivation and excessive demands for achievement are primary
reasons for dropping out, we focus on motivational and cognitive variables as a first
step to developing students’ profiles. Moreover, students leaving their course of
study are difficult to access, so we focus on available data like students’ performance
and individual characteristics with respect to the following research question:
Which domain-specific predictors of study success or failure can be
determined in the first year of engineering mathematics?

THEORETICAL CONSTRUCTS
As a main construct that impacts on students’ achievement, we refer to the construct
of learning motivation as an umbrella term for different motivational variables
(Spinath, 2011). Firstly, we choose all of the motivational variables that are
summarized by the term learning motivation in order to identify the crucial impacting
factors. Furthermore, all of the constructs are connected to the subject matter, so they
might play a crucial role in the transition from school to university mathematics, a
situation characterised by a changing subject matter and learning environment.
A first and main part of motivation is an individual’s goal orientation (Dweck, 1986).
This dispositional variable involves individual’s beliefs about appropriate goals as a
trait referring to specific and, thus, context-related tasks (Elliot et al., 1999). A
further dispositional and motivational variable is interest which is differentiated into
three components: feeling-related valences, value-related valences and an intrinsic
character. Interest could be understood as an individual’s development of an
appreciation for a specific subject like mathematics (Wild & Möller, 2009). This
definition involves the necessity to regard interest context-specific. One aspect of the
construct of interest, i.e. the feeling-related valences is also a part of the expectancy-
value-theory of Wigfield and Eccles (1992). They derive an individual’s motivation
for doing a task from the individual’s expectancy of the success on a specific task
and the incentive value of this task. Besides the intrinsic value, which is similar to
the feeling-related valences of interest, further variables, i.e. attainment value, utility
value and costs are part of achievement-related values. The expectancy of a success
referring to specific tasks could be understood as individuals’ self-efficacy beliefs
(Wigfield & Eccles, 1992) that are close to the construct of self-concept (Shavelson
et al., 1982).
Learning strategies is a further umbrella term that includes variables which also
could have an impact on students’ achievement (Wild, 2005). Learning strategies
include cognitive learning strategies like strategies for elaborating a specific issue,
meta-cognitive strategies like planning or monitoring the process of learning, or
strategies to use resources like a specific learning environment. Finally, although
students’ achievement or success is hardly to define (Heublein, 2014), it could be

127 [Link]:indrum2018:174410
understood as the achievement in exams referring to a specific subject like
mathematics or to proceed in a field of study like engineering despite encountered
difficulties.

METHODOLOGY
Sample
Our first research step involved 182 engineering students at the University of Kassel
enrolled in a calculus course in the summer semester 2017. Among the participants
were 158 men and 24 women. Most of them started their second semester (67 %),
though a small group of beginners is integrated (14 %). In the beginning of the
semester, students of the calculus course were given a questionnaire concerning
sociodemographic factors and motivational orientation towards mathematics. To
achieve a high response rate, the students had two weeks to answer the questionnaire
and received additional points for their permission to the final exam that could be
achieved by solving weekly exercises. Students were also assured of the anonymity
of their responses.
exam questionnaire
(exam)

linear algebra course calculus course

winter semester 2016/17 semester break summer semester 2017 semester break
Figure 1: Time of data collection
Instruments
The so called SELLMO instrument (Spinath et al., 2012) was used to measure
students’ goal orientation towards mathematics. Twenty out of thirty-one items were
chosen and especially referred to mathematics courses at university. Goal orientation
is divided in four sub-scales with each five items concerning approaching and
avoiding achievement goals, work avoidance and learning goals. One example for an
item referring to learning goals is: “My aim for mathematics courses at university is
to gain a deep understanding for the content.”
For interest, we used a scale with nine items of Schiefele et al. (1993b) and adapted it
to mathematics. One example for a negative formulated item of this scale is: “To be
honest, I less care for mathematics.” Referring to the expectancy-value-theory, we
measured values with a scale involving six items that we developed according to
Wigfield and Eccles (1992). One example for an item of this scale is: “Mathematical
skills will be crucial for my future professional career.” Further, we adapted each
three items from the PISA study (Kunter et al., 2002) to measure students’ self-
concept and self-efficacy with respect to mathematics.
To measure the students’ learning strategies, we focussed in the first step of our
research on students’ self-reports about the use of resources like lectures, tutorials

128 [Link]:indrum2018:174410
and special exercises. Whereas in lectures different mathematics topics are taught,
students practice in tutorials by means of additional exercises. In the special
exercises they discuss their homework and resolve open questions. Finally, students’
success was measured by the self-reported grade in the final exam of the linear
algebra course (see Fig. 1). By contrast, further grades of final exams in the
abovementioned calculus course was directly given but is not analysed yet. Finally,
we collected sociodemographic variables, e.g. the type of matriculation standard,
according to a questionnaire used for dropout studies (Heublein et al., 2010).

RESULTS
Results in the first step of our research firstly refer to an evaluation of the
instruments concerning the quality of scales. We further proved the predicting power
of different variables on the students’ performance measured by the self-reported
grade in the final exam of the linear algebra course that students have taken the
previous semester.
Evaluation of the instruments
In the first analyses, Cronbach’s alpha estimates of reliability were determined for
the scales from each instrument (see Tab. 1). Measures are adequately reliable, with
values ranging from .552 to .844. Most of the values are appropriate, the lowest
value (.552) was found for a scale with only three items.
Construct Number of Items Cronbach’s Alpha
Mathematics Interest 9 .844
Goal orientation:
Approach achievement goals 5 .722
Avoidance achievement goals 5 .827
Word avoidance 5 .727
Learning goals 5 .739
Expectancy-value-theory
Mathematics self-concept 3 .704
Mathematics self-efficacy 3 .552
Value of mathematics 6 .690

Table 1: Sample constructs and Cronbach’s Alphas

Possible predictors of performance in engineering: correlations
In each of the following analyses, we defined students’ achievement as the self-
reported exam grade of the linear algebra course that students have taken the
previous semester. We firstly assessed the relationship between students’ individual

129 [Link]:indrum2018:174410
characteristics referring to grades achieved in school. As seen in Table 2, the school
grades and exam grades of the linear algebra course are significantly and positively
correlated.
Math school grade Final school grade
exam grade .321** .393**

Table 2: Pearson’s correlation coefficients between the exam grade and students’
achievement in school (*p < 0.05; **p < 0.01)
We further proved the correlations between the exam grade and variables
constituting the expectancy-value-theory. Except for interest, the correlations
between the motivational variables and the students’ achievement are significant.
Particularly, there is a considerable relationship between the mathematics self-
concept and the students’ achievement.
self-concept self-efficacy values interest
exam grade .554** .363** .385** .166

Table 3: Pearson’s correlation coefficients between the exam grade and students’
individual characteristics (expectancy-value-theory)
By contrast, the correlations between the students’ achievement and the students’
individual characteristics referring to the construct of goal orientation are weak and
except of the working avoidance, not significant.
AAG1 AGG2 WA LG
exam grade .000 .177 .207* .109

Table 4: Pearson’s correlation coefficients between the exam grade and students’
individual characteristics referring to AAG1 (Approach achievement goals), AAG2
(Avoidance achievement goals), WA (Work Avoidance), LG (Learning goals)
Using correlation analysis, we finally assessed the relationships between students’
achievement and the students’ engagement referring to external resources given by
the attendance rate of lectures, tutorials and special exercises. However, the
attendance rates seem to be independent of the students’ achievement.
Lecture Tutorials Special exercise
exam grade .133 .074 .120

Table 5: Pearson’s correlation coefficients between the exam grade and students’
engagement

130 [Link]:indrum2018:174410
Possible predictors of performance in engineering: group comparison
To identify other possible predictors of the performance in engineering mathematics,
we compared between distinct groups using t-tests. Firstly, we compared students
that were enrolled in an advanced math course and students that were enrolled in
usual math courses. Students in advanced math courses get more math lessons in a
week and, thus, examine mathematics in a greater extent than students of usual math
courses. As expected, students attending a math advanced course have significantly
better grades in the exam of the linear algebra course (see Tab. 6). Further, we
compared the group of students who were at a technical secondary school in which
the extent of mathematics is lesser than in usual secondary schools. As expected, on
average, students who had attended a technical secondary school obtained in the
exam of the linear algebra course a grade of 4.0, whereas the corresponding grade for
students who had not attended a technical secondary school was 3.5. Thus, students
from a technical secondary school significantly perform worse in the linear algebra
course than those who attended another type of school (see Tab. 6).
Technical secondary school Math advanced course
1 0 1 0
Exam grade M 4.0 3,5 3.3 3.9
SD 1,0 1,0 1,2 1,0
Sig. .016 .019

Table 6: Exam grade of the linear algebra course depending on different subgroups
(1 = attended; 0 = not attended)

DISCUSSION AND CONCLUSION

The development of engineering students’ profiles referring to cognitive and
motivational variables could potentially result in identifying students’ risks for an
undesirable low success or a dropout. For identifying possibly risky profiles, we
started to find possible predictors of the students’ achievement. Final school grades,
maths self-beliefs, values of maths, the type of matriculation standard and the choice
of advanced courses in school are meaningful predictors of performance in
engineering mathematics. Hence, the results show that several variables determine
the maths performance which should be considered seriously for the development of
support services. However, the results reveal new questions leading to further
possible research steps.
In conformity with the research state, the mathematical achievement as well as the
amount of mathematics in school seem to be a predictor of the success in a final
exam of the first semester. As a subsequent issue it is a crucial question if the impact
of the former school time on the students’ success in mathematics courses at

131 [Link]:indrum2018:174410
university decreases or disappears. Moreover, it is interesting whether previous
knowledge measured by a skills test has a greater effect compared to the school
grades. Perhaps, domain-specific skills can be recognised as special predictors of
mathematics courses in engineering. In addition, the results show that the highest
correlation exists between the final school grade and students’ performance. This
implies further influencing variables developed in the students’ school time that
impact on the students’ achievement at university, especially learning activities and
strategies. The fact that the attendance rates seem to be independent of the students’
achievement strengthens this perspective.
Concerning the motivational variables, it is interesting that not all of them have an
impact on the students’ achievement. The students’ mathematics self-concept that is
also developed in the time of learning mathematics in school shows the highest
correlation to the students’ achievement. Therefore, support services should not only
focus on the deficits in mathematics skills but also on the assistance of students’ self-
beliefs. Mathematics interest shows no impact on the mathematics performance. This
result could approve findings like Eilerts (2009). Since engineering students do not
choose mathematics voluntarily, mathematics interest might have no predicting
power in this context. In this respect, it could be also interesting to differentiate in
further analyses specific groups of students, e.g. concerning gender, the school form
or other variables and to investigate if different groups show different relationships
between motivational variables and the students’ achievement.
Regarding the method, proven scales have been used and adapted to mathematics.
Only scales with a low number of items can be extended to improve the reliability.
All analyses base on simple correlations. Results can be improved and deepened by
using further methods like regression analyses or structural equation models, so that
an investigation of indirect effects is facilitated.
To conclude, in further steps of our research, the observation of motivational
variables of engineering beginners should be continued and extended to the
investigation of their development. Additionally, engineering students should be
surveyed in respect of learning activities and strategies that we involved in this first
study only by collecting data to the use of attendance rates (external resources). A
detailed investigation of students’ motives for non-attendance would give more
information about engineering students’ learning behaviour. Thus, building upon the
first results of our research, we expect to deepen the desirable insight into students’
profiles in the next steps of our research.

REFERENCES
Dweck, C. S. (1986). Motivational processes affecting learning. American
Psychologist, 41(10), 1040–1048. [Link]
Eilerts, K. (2009). Kompetenzorientierung in der Mathematik- Lehrerausbildung:
Empirische Untersuchung zu ihrer Implementierung. Zürich, Münster: LIT.

132 [Link]:indrum2018:174410
Elliot, A. J., McGregor, H. A., & Gable, S. (1999). Achievement goals, study
strategies, and exam performance: A mediational analysis. Journal of Educational
Psychology, 91(3), 549–563. [Link]
Fellenberg, F. & Hannover, B. (2006). Kaum begonnen, schon zeronnen?
Psychologische Ursachenfaktoren für die Neigung von Studienanfängern, das
Studium abzubrechen oder das Fach zu wechseln. Empirische Pädagogik, 20(4),
381-399.
Gómez-Chacón, I. M., Griese, B., Rösken-Winter, B., & Gónzalez-Guillén, C.
(2015). Engineering students in Spain and Germany – varying and uniform
learning strategies. In K. Krainer & N. Vondrova (Eds.), CERME9. Proceedings of
the Ninth Congress of the European Society for Research in Mathematics
Education (pp. 2117–2126). Prague: Charles University.
Gueudet, G. (2008). Investigating the secondary–tertiary transition. Educational
Studies in Mathematics, 67(3), 237–254. [Link]
9100-6
Hailikari, T., Nevgi, A. & Komulanien, E. (2008). Academic self-beliefs and prior
knowledge of student achievement in mathematics: a structural model.
Educational Psychology, 28(1), 59-71.
Hattie, J. (2009). Visible learning. A synthesis of over 800 meta-analyses relating to
achievement. Abingdon: Routledge.
Heublein, U. (2014). Student Drop-out from German Higher Education Institutions.
European Journal of Education, 49(4), 497–513.
[Link]
Heublein, U., Hutzsch, C., & Schreiber, J. (2010). Ursachen des Studienabbruchs in
Bachelor- und in herkömmlichen Studiengängen. Forum Hochschule. Hannover:
Deutsches Zentrum für Hochschul- und Wissenschaftsforschung.
Knospe, H. (2012). Zehn Jahre Eingangstest Mathematik an Fachhochschulen in
Nordrhein-Westfalen. Proceedings zum 10. Workshop Mathematik in
ingenieurwissenschaftlichen Studiengängen (pp. 19-24). Mülheim an der Ruhr:
Hochschule Ruhr-West.
Kunter, M., Schümer, G., Artelt, C., Baumert, J., Klieme, E., Neubrand, M., Prenzel,
M., Schiefele, U., Schneider, W., Stanat, P., Tillmann, K.-J. & Weiß, M. (Eds.).
(2002). PISA 2000: Dokumentation der Erhebungsinstrumente. Berlin: Max-
Planck-Institut für Bildungsforschung.
Rach, S., & Heinze, A. (2013). Welche Studierenden sind im ersten Semester
erfolgreich? Journal für Mathematik-Didaktik, 34(1), 121–147.
[Link]
Rooch, A., Kiss, C. & Härterich, J. (2013). Brauchen Ingenieure Mathematik? – Wie
Praxisbezug die Ansichten über das Pflichtfach Mathematik verändert. In I.
Bausch, R, Biehler, R. Bruder, P. R. Fischer, R. Hochmuth, W. Koepf, S.

133 [Link]:indrum2018:174410
 Schreiber & T. Wassong (Eds.), Mathematische Vor- und Brückenkurse. Konzepte,
Probleme und Perspektiven. (pp. 398-409). Wiesbaden: Springer Spektrum.
Schiefele, U., Krapp, A. & Schreyer, I. (1993a). Metaanalyse des Zusammenhangs
von Interesse und schulischer Leistung. Zeitschrift für Entwicklungspsychologie
und Pädagogische Psychologie. 25(2), 120-148.
Schiefele, U., Krapp, A., Wild, K.-P., & Winteler, A. (1993b). Der Fragebogen zum
Studieninteresse (FSI). Diagnostica, 39(4), 335–351.
Schiefele, U., Streblow, L., Ermgassen, U. & Moschner, B. (2003). The influence of
learning motivation and learning strategies on collage achievement: Results of a
longitudinal analysis. German Journal of Educational Psychology, 17(3/4), 185-
198.
Schrader, F. W. & Helmke, A. (2015). School achievement: motivational
determinants and processes. International Encyclopedia of the Social &
Behavioural Science, 21, 48-54.
SEFI. (2013). A framework for mathematics curricula in engineering education.
Brussels. Brussels: European Society for Engineering Education (SEFI).
Shavelson, R. J., Hubner, J. J., & Stanton, G. C. (1982). Self-Concept: Validation of
Construct Interpretations. Journal of Educational Psychology, 74(1), 3–17.
[Link]
Søgaard Larsen, M. (2013). Dropout phenomena at universities: what is dropout?
Why does dropout occur? What can be done by the universities to prevent or
reduce it? A systematic review. Clearinghouse - research series: [Link] Danish
Clearinghouse for Educational Research.
Spinath, B., Stiensmeier-Pelster, J., Schöne, C., & Dickhäuser, O. (2012). SELLMO.
Skalen zur Erfassung der Lern- und Leistungsmotivation. Göttingen: Hogrefe.
Spinath, B. (2011). Lernmotivation. In H. Reinders, H. Ditton, C. Gräsel, & B.
Gniewosz (Eds.), Empirische Bildungsforschung (pp. 45–55). Wiesbaden: VS
Verlag für Sozialwissenschaften. [Link]
Thomas, M., de Freitas Druck, I., Huillet, D., Ju, M.-K., Nardi, E., Rasmussen, C. &
Xie, J. (2012). Survey team 4: key mathematical concepts in the transition from
secondary to university. ICME 12, Seoul, Korea.
Wigfield, A., & Eccles, J. S. (1992). The development of achievement task values: A
theoretical analysis. Developmental Review, 12(3), 265–310.
[Link]
Wild, E., & Möller, J. (2009). Pädagogische Psychologie. Springer-Lehrbuch.
Berlin, Heidelberg: Springer-Verlag Berlin Heidelberg. Retrieved from
[Link]
Wild, K.-P. (2005). Individuelle Lernstrategien von Studierenden. Konsequenzen für
die Hochschuldidaktik und die Hochschullehre. Beiträge zur Lehrerinnen- und
Lehrerbildung, 23(2), 191–206.

134 [Link]:indrum2018:174410
 “Making connections” in the mathematics courses for engineers: the
example of online resources for trigonometry
Ghislaine Gueudet1 and Pierre-Vincent Quéré2
1
CREAD, ESPE de Bretagne, University of Brest, France, [Link]@espe-
[Link]; ²CREAD, Université de Bretagne Loire, France, pierre-
[Link]@[Link]
Abstract: This paper concerns the teaching of mathematics for future engineers,
focusing on the theme of trigonometry. We claim that the use of trigonometry in
engineering courses requires different kinds of connections: connecting different
domains, different concepts, frames and registers. We use here the concept of
connectivity, developed in the frame of e-textbooks analysis, to analyse online
courses for future engineers in France. We evidence that these courses propose some
connections; but their connectivity is not developed enough to meet the requirements
of engineering courses.
Keywords: Connectivity, Teaching and learning of specific topics in university
mathematics, The role of digital and other resources in university mathematics
education, Teaching and learning of mathematics in other fields, Mathematics for
engineers.

INTRODUCTION
How can a teaching of mathematics answer the needs of engineering courses, i.e. can
provide the mathematics needed to understand the course and solve the problems
proposed? Which should be the features of such a teaching, and do existing courses
present such features? This is the general theme of the work presented here. Previous
works addressing this theme observe a gap between the mathematics taught in
“mathematics courses” and the use of them to solve problems in engineering courses
(e.g. Redish 2005; Biehler, Kortemeyer & Schaper 2015). Winsløw, Gueudet,
Hochmut and Nardi (2018) note that several works presented at CERME conferences
identify “a lack of connectedness of curricula integrating mathematics and other
disciplines”. Interviewing French engineers about their mathematical needs in the
workplace (Quéré 2017), and how the mathematics courses they followed as students
answered or not these needs, we noticed that several of them declared that the
mathematics courses “did not make enough connections”. These connections,
according to the engineers, can be of different kinds: links between mathematics and
the real world, between different mathematical contents, between different
representations etc. We consider that this is an important issue for understanding
mathematics applied to engineering and their teaching. Which kinds of connections
should propose a teaching of mathematics for engineering?
Moreover, in the frame of another starting study, we are interested by the possible
design of innovative curriculum resources for the teaching of mathematics for

135 [Link]:indrum2018:171131
STEMSS (Science, Technology, Engineering, Medicine and Social Sciences)
courses. Before starting the design of such resources, we have investigated existing
resources, trying in particular to observe whether they tried to build the kind of
connections evoked by the engineers.
This investigation led us to choose a focus on trigonometry. Indeed, trigonometry
appeared as extensively used in different kinds of engineering courses, and offering
many possibilities of connections of all the kinds evoked above. In what follows, we
firstly introduce our theoretical tools and research questions; then we present related
works on the teaching of trigonometry. To exemplify the mathematical needs, we
analyse the use of trigonometry in an electricity course for first year students; then
we consider the trigonometry content in two online mathematics courses for
engineering students.

THEORETICAL PERSPECTIVE AND RESEARCH QUESTIONS

The overarching perspective guiding our work is an institutional perspective
(Chevallard, 2006). We consider that the mathematics taught are shaped by the
institution where they are taught. Engineering studies at university or in engineers’
schools constitute an institution, different from mathematics courses for maths
majors. Within engineering studies, engineering courses also constitute an institution
different from mathematics courses. Our aim here is to compare how these two
institutions shape the mathematics contents, more precisely the trigonometry
contents.
Our central focus is on “making connections”. Students’ understanding in terms of
mathematical reasoning and problem-solving has been linked by several authors to
“making connections” and “connectivity” (e.g. Hiebert & Carpenter 1992). Drawing
on these works, we have chosen to look for:
“connections in, between, and across individuals’ cognitive/learning tasks and
activities, and how e-textbooks may support those (micro level); as well as for
‘connected learning’ between and across groups of individuals, teachers or students
(macro level).” (Gueudet, Pepin, Sabra, Restrepo & Trouche to appear)
We have therefore proposed a concept of “connectivity” to analyse e-textbooks
(encompassing various kinds of digital curriculum), with the intention to evaluate
their potential for the building of connections for the students.
Thus, connectivity has two components: “macro-level connectivity”, which considers
the e-textbook as a whole; and “micro-level connectivity”, where the focus is on a
particular mathematical topic (trigonometry here). In this paper we only use “micro-
level connectivity”; more precisely, we observe in curriculum resources available
online the presence of:
- Connections between different topic areas or frames;

136 [Link]:indrum2018:171131
 - Connections between different semiotic representations (e.g. text, figures,
static and dynamic); […]
- Connections between different concepts […]” (Gueudet et al. to appear).
Here we want to compare the connections concerning trigonometry when it is used in
engineering courses and the connections concerning trigonometry in mathematics
courses for engineers, more precisely the connectivity of online courses for
engineers. Our purpose is not to discuss whether trigonometry should be introduced
in mathematics courses or engineering courses, but to compare how it is
introduced/used in these institutions. Hence the research questions we study here can
be formulated as:
- Which connections concerning trigonometry appear in non-mathematical
engineering courses?
- Which connectivity, concerning trigonometry, can be observed in online
resources for mathematics courses for engineers, and how does this compare
with the connections in non-mathematical courses?
In terms of methods, we have searched for curriculum resources and online courses
on three major websites used in France: Unisciel1, meaning “online science
university”, gathering many online courses; IUTenligne 2, a website for technicians
institutes within universities; and FUN3, meaning France Digital University, the
national platform offering MOOCs. We have selected all the resources
corresponding to mathematical courses for engineers or technicians on trigonometry,
and have looked at the same time for resources on non-mathematical subjects using
trigonometry. We have eventually chosen the theme of electrical engineering,
because we have identified in it specific mathematical needs in trigonometry. Before
analysing these resources, we now consider some works about trigonometry in
mathematics education, and how they enlighten the connections issue.

CONTEXT AND RELATED WORKS

In France, trigonometry is firstly introduced in grade 8 or grade 9 through the
definitions of cos, sin and tan as quotients of lengths in a right-angled triangle. The
angles are measured in degrees. In grade 10, the unit circle and the radian are
introduced, together with a new frame for cos and sin, which are now the coordinates
of a point on the unit circle. In grade 11, cos and sin are studied as functions. Hence
connections between frames and registers are extensively present in this teaching,
and these connections can raise difficulties for the students (Berté et al. 2004). For

1
[Link]
2
[Link]
3
[Link]

137 [Link]:indrum2018:171131
example, grade 9 students already use for solving geometry exercises the “cos-1” key
on their calculator; nevertheless, this key refers to a reciprocal trigonometric
function, which is only presented at university, and moreover belongs to the
functional frame (Bueno-Ravel & Gueudet 2010).
The international research on the teaching and learning of trigonometry acknowledge
the existence of all these different registers and representations and investigate their
consequences. Kendal and Stacey (1997) compare two teachings in grade 10 in
Australia, using respectively ratios and the unit circle to introduce sin, cos and tan;
they conclude that the ratio method appears as a better choice. Also at the university
level, trigonometry remains a difficult subject for the students (Weber 2005). Mesa
and Goldstein (2017), studying the presentation of trigonometry in college textbooks,
have evidenced that these textbooks propose different conceptions of angles,
trigonometric and inverse trigonometric functions; depending on these conceptions,
some problems can be delicate to tackle for the students. The textbooks do not try to
link different conceptions, and do not highlight which one is more relevant for a
given problem.
Trigonometry clearly requires many connections between frames and registers. It is
moreover linked with many different mathematical subjects (geometry, functions, but
also complex numbers); and is extensively used in physics. Several researchers have
also studied trigonometry within physics courses. Chiu (2016) studied the impact of
a new curriculum in Taiwan, where contents of physics requiring trigonometry are
taught before the corresponding mathematics courses. She observes that, while
curriculum designers are positive on the possible consequences of a teaching of
trigonometry by the physics and then by the mathematics teachers, the students and
the teachers are mostly negative about this experience. A teaching of trigonometry in
mathematics courses seems necessary before using it in physics.
In his comparative study between France and Vietnam, Nguyen Thi (2013) shows
that in both countries, trigonometry is present in physics courses with mathematical
models for periodical phenomena, under two forms: uniform circular motion
(represented by a point moving on a circle, in a graphic or algebraic register) and
harmonic oscillation (represented by functions in a graphic, algebraic or vectorial
register). Nevertheless only a few exercises propose modelling activities (in physics
as well as in mathematics); and the two models are almost never connected.

TRIGONOMETRY IN ELECTRICAL ENGINEERING COURSES

In this section we draw on the content of several courses for future engineers or
technicians, available on French websites (e.g. Piou 2014). In electrical engineering,
one of the major subjects is the study of the “alternating sinusoidal regime”. In
circuits with such a regime, the different signals (current and voltage in particular)
are of the form: s(t) = A2sin(t+).

138 [Link]:indrum2018:171131
In a circuit where reactive loads are present (like capacitors or inductors) energy
storage in the loads results in a phase difference between the voltage (u) and the
current (i) waveforms. This difference is firstly introduced in the context of
functions, with formulas like: u(t)=U2sin(t); i(t)=I2sin(t-), and associated
with a graphical representation as two curves on the time axis (Figure 1, left part).
A connection is immediately established with a geometrical representation of these
signals, through “Fresnel vectors”. A signal defined by s(t) = A2sin(t+) (where A
is positive) can be represented by a vector of length A, and a direction forming an
angle  with the horizontal direction. Hence u(t) can be represented as a horizontal
vector, and i(t) as a vector forming an angle (-) with it (figure 1, right part). Some
courses also propose an interpretation in terms of complex numbers; for the sake of
brevity, we do not present it here.

Figure 1. Signals in the alternating sinusoidal regime and phase difference. On the left:
functional frame; on the right, Fresnel vectors in the vectorial frame.
We argue that the students in this case have to master connections between these two
representations of signals: as two curves with a gap of  on the horizontal axis; or as
two vectors forming an angle . Some courses propose animated pictures or
exercises to work explicitly on this connection (see figure 2 for an example extracted
from a teacher’s website, [Link]

139 [Link]:indrum2018:171131
Figure 2. Connection between a rotating vector and the corresponding curve, extract
of an animated picture. The blue points on the right appear when the vector rotates.
In terms of micro-level connectivity, electrical engineering courses naturally connect
trigonometry with electrical engineering; they also connect different concepts: in the
example we considered here, functions and vectors, and in other courses also
complex numbers. Since vectors are represented as arrows, and functions represented
by their graphs, these two kinds of representations are also connected in the text of
the course.

TRIGONOMETRY IN MATHEMATICS ONLINE COURSES FOR

ENGINEERS
A MOOC presenting the basics of mathematics for future engineers
The MOOC “Basics of mathematics”, available on the FUN platform (freely
available after inscription) presents its objective on its first page: “This MOOC aims
at revising the basic notions of mathematics, needed to start engineering studies”.
The MOOC lasts 12 weeks, corresponding to 12 chapters. Chapter 2 is entitled
“trigonometry”. It comprises 7 course videos (from 5 to 12 minutes), interactive
multiple-choice questions, and a final assessment. Most of the notions presented in
chapter 2 of the MOOC are taught in France at secondary school; nevertheless the
reciprocal functions (like sin-1) are only presented during the first year of university.
The course proposes almost no connection with engineering activities. The first
video says that “Trigonometry appears in many domains, like drawing plans,
navigating or mechanics”; the last video mentions the task of “triangulation”,
without definition. Three exercises are associated with each video. Most of these
exercises are situated in the geometrical frame; some of them are in the frame of
trigonometric equations. There is only one exercise in the whole chapter with a
concrete context: computing the length of a cable joining the top of a pole to the
ground, over a mountain with slope of 15%. The final evaluation is composed of two
problems; the first one has a concrete context, measuring lengths, and the second one
concerns trigonometric equations.
The definitions of cos, sin, tan are introduced in the geometrical frame in the right
triangle (video 1). In video 2, a software (Maple reader) is used to establish a
connection between cos, sin and tan in the unit circle and their graphs as functions
(figure 3). The animated picture supports the discourse of the teacher explaining this
link. This picture associates in fact the unit circle with the graphs of the functions,
but also with the triangle. In fact the radius of the circle can be changed; it is not
only the unit circle, but any circle; and the words “Adjacent” and “Hypotenuse” on
the left of the screen link the circle and the other geometrical view as a triangle.

140 [Link]:indrum2018:171131
 Figure 3. A dynamic representation making connections in the Mooc, case of cos.
This dynamic image is clearly used to connect the geometrical frames (triangle and
unit circle) and the analytical frame; nevertheless, it might be very difficult to
understand for participants who do not remember their school courses.
As a summary, we retain that micro-level connectivity in this MOOC comprises
some connections between concepts and representations, including dynamic
representations; but almost no connections with other domains or engineering
contexts. Moreover the connections with dynamic representations can remain unclear
for students because of a lack of explanations.
“Mathematical tools for physics”, an online course
“Mathematical tools for physics”4 is an online course addressed to first year students
in physics, freely accessible (without inscription). It belongs to a complete first year
course, which is always organised in three sections: “learning” (course) “practice”
(exercises) and “self-assessment”. It comprises 11 chapters; chapter 6 is entitled
“Circular trigonometry – Hyperbolic trigonometry”. This chapter comprises 8
subsections; here we are only interested in the 4 subsections concerning circular
trigonometry.
After recalling the definition of an angle, the first subsection defines sin, cos and tan
in the frame of the unit circle. Nevertheless, an animated picture proposes a link with
the frame of the right-angled triangle. A subsection about formulas associates a
functional frame: specific values of sin, cos and tan and register of the unit circle.
Then two sections are dedicated to the properties of the “direct circular functions”
and of the “reciprocal circular functions”, and only mention the functional frame.
This course makes no link with physics or any real-life context.

4
[Link]

141 [Link]:indrum2018:171131
There is only one problem, divided in three questions, in the “practice” section.
Interestingly, it is a problem of physics: “the Compton effect” (scattering of a photon
by an electron). To solve this problem, the students must master contents of physics:
the law of conservation of energy and quantity of movement (and the associated
formula). The initial model is in the frame of vectors; these vectors are projected on
the two axes, and the students have then to use trigonometric formula, and finally to
work in the frame of functions. The difference between the wavelengths before and
after the scattering is indeed of the form coswhere is the angle
characterizing the direction of the photon after the scattering; the students must
observe that this function is increasing over [0, ]: a larger angle corresponds to a
larger change in the wavelength.
The “assessment” section comprises five parts: 2 on circular trigonometry, 2 on
hyperbolic trigonometry and one entitled “composition of vibratory motions”. There
is here again a connection between trigonometry and physics. Nevertheless its
mathematical part remains entirely in the functions’ frame.
Finally, concerning the micro-level connectivity of this online course, we retain that
it proposes some connections with physics in the exercise and assessment part (but
no such connection in the course part). The connections between concepts and
between registers are restricted to the case of a single problem.

CONCLUSION
Trigonometry is a domain of mathematics where many different concepts (angles,
vectors in geometry; functions) and semiotic registers (triangles, arrows, circle in
geometrical register; curves in a graphical register; equations etc.) can be connected.
It is recognised as a difficult domain for students. Nevertheless, within mathematics
some exercises are limited to a single register: the study of trigonometrical functions
for example does not always require thinking in terms of angles. Using trigonometry
in engineering courses, on the opposite, always requires such connections. The
students must be able to associate an expression like s(t) = A2sin(t+) both with a
function and its graph; and a vector represented by an arrow. Engineering courses
have a high degree of micro-level connectivity, for trigonometry. Our analyses of
two online courses of trigonometry for engineers in France lead us to observe that
they have a reduced level of micro-level connectivity: limited to connections
between concepts and semiotic registers for the first one, while the second one on the
opposite offers more connections between trigonometry and physics, but reduced
connections between concepts and semiotic registers.
Our exploratory work leads to formulate recommendations for mathematics courses
for future engineers, concerning trigonometry (and possibly other topics). Our study
of engineering courses confirms that developing the ability of these students to make
connections is an important aim. Connections between frames, between registers, but
also connections with engineering are possible, as evidenced by Nguyen Thi (2013).

142 [Link]:indrum2018:171131
Online resources could propose such connections, and moreover could draw on the
possibilities offered by dynamic representations and various kinds of software.
Contributing to the development of such resources is an important aim for
mathematics education research, to address the need for students to make a relevant
use of the mathematics they learn at university in and for their engineering courses.

Acknowledgments
The authors warmly thank Alain Jameau and Damien Grenier for their contribution
on electrical engineering.

REFERENCES
Berté, A., Chagneau, J., Desnavres, C. , Lafourcade, J., & Sageaux, C. (2004). Aide
apportée aux enseignants par la recherche en didactique. Un exemple : le cosinus
en quatrième. Petit x 65, 9-35.
Biehler, R., Kortemeyer, J. & Schaper, N (2015). Conceptualizing and Studying
Students' Processes of solving typical problems in introductory engineering
courses requiring mathematical competences. In K. Krainer & N. Vondrova (Eds)
Proceedings of the Ninth Congress of the European Mathematical Society for
Research in Mathematics Education (pp. 2060-2066). Prague, Czech Republic:
Charles University in Prague, Faculty of Education and ERME.
Bueno-Ravel, L., & Gueudet, G. (2009). Online resources in mathematics: Teachers’
geneses and didactical techniques. International Journal of Computers for
Mathematical Learning, 14(1), 1-20.
Chevallard, Y. (2006). Steps towards a new epistemology in mathematics education.
In Bosch, M. (Ed.) Proceedings of the Fourth Congress of the European Society
for Research in Mathematics Education. (pp.22-30). Sant Feliu de Guíxols, Spain:
FUNDEMI IQS – Universitat Ramon Llull and ERME.
Chiu, M.-S. (2016). The Challenge of Learning Physics Before Mathematics: A Case
Study of Curriculum Change in Taiwan. Research in Science Education 46(6),
767-786.
Gueudet, G., Pepin, B., Sabra, H., Restrepo, A. & Trouche, L. (2017). E-textbooks
and connectivity : proposing an analytical framework. International Journal for
Mathematics and Science Education, doi:10.1007/s10763-016-9782-2, online.
Hiebert, J., & Carpenter, T. (1992). Learning and teaching with understanding. In A.
Grouws (Ed.), Handbook of Research on Mathematics Teaching and Learning
(pp. 65-97). New York, NY: Macmillan.
Kendal, M. & Stacey, K. (1997). Teaching trigonometry. Vinculum 34 (1). 4-8.

143 [Link]:indrum2018:171131
Mesa, V. & Goldstein, B. (2017). Conception of Angles, trigonometric functions,
and inverses trigonometric functions in college textbooks. International Journal
of Research in Undergraduate Mathematics Education 3 (2), 338-354
Nguyen Thi, N. (2013). Fonctions trigonométriques et phénomènes périodiques. Un
accès à la modélisation dans l’enseignement secondaire ? Petit x 91, 27-48.
Piou, M. (2014). Electricité. Analyse des circuits et des signaux électriques.
Retrieved september 2017,
[Link]
n/pdf/[Link]
Quéré, P.-V. (2017). French engineers' training and their mathematical needs in the
workplace: interlinking tools and reasoning. In Dooley, T. & Gueudet, G. (Eds.)
Proceedings of the Tenth Conference of the European Society for Research in
Mathematics Education (CERME 10, February 1 – 5, 2017). Dublin, Ireland:
DCU Institute of Education and ERME
Redish, E. F., (2005). Problem solving and the use of math in physics courses. In
Proceedings of the Conference, World View on Physics Education in 2005:
Focusing on Change, Delhi, August 21-26, 2005, Retrieved from:
[Link]
Weber, K. (2005). Students’ understanding of trigonometric functions. Mathematics
Education Research Journal 17(3), 94-115.
Winsløw, C., Gueudet, G., Hochmut, R. and Nardi, E. (2018). Research on
University Mathematics Education. In:T. Dreyfus, M. Artigue, D. Potari, S.
Prediger & K. Ruthven (Eds.), Developing Research in Mathematics Education:
Twenty Years of Communication, Cooperation and Collaboration in Europe.
London: Routledge.

144 [Link]:indrum2018:171131
Engineering students’ engagement with resources in an online learning
environment
Shaista Kanwal
University of Agder, Norway, [Link]@[Link]
In this paper, we investigate how undergraduate engineering students interact with
an online learning environment provided to them in a Calculus course. The
constituent resources of this environment include tutorial videos, textbook and
MyMathLab – an online interactive system for mathematics. A qualitative case study
involving a small group of students has been conducted. We investigated which
resources these students used and the manner in which they incorporated these
resources in their online mathematical work.
Keywords: Students’ interactions with resources, the role of digital and other
resources in university mathematics education, mathematics for engineers.

INTRODUCTION
In recent years, digital resources are increasingly used for teaching and learning of
mathematics (Borba et al., 2016; Pepin, Choppin, Ruthven, & Sinclair, 2017). The
presence of wide range of digital resources in terms of their functionalities allows
various possibilities of creating digital environments for students to learn
mathematics. Each digital environment might afford unique interactive and learning
opportunities; therefore, empirical research closely looking at students’ engagement
and the opportunities for their learning in such environments is well needed. This
study deals with one digital learning environment provided to undergraduate
engineering students for practicing mathematics. The aim is to explore students’
interactions with the constituent resources of this environment to elucidate the
learning opportunities in this environment.
Adler (2000) introduced the term resource to embrace several agents such as
physical, human and cultural tools and aids intervening in a teacher’s activity. In this
paper, however, we distinguish between digital and classical resources and focus on
students’ work with resources. The use of digital resources is relevant in the context
of engineering mathematics in the sense that engineers during their professional
activities rely on technology for solving mathematical tasks (van der Wal, Bakker, &
Drijvers, 2017). The framework for mathematics curricula in engineering education
(Alpers et al., 2013) recommends the use of technology aimed at fostering
engineering students’ mathematical competencies. In the next section, we present the
theoretical framework, and the subsequent section contains introduction to the
constituent resources of the online learning environment.

145 [Link]:indrum2018:174742
THEORETICAL PERSPECTIVE
In order to study students’ interactions with the resources, we employ the
documentational approach to didactics (Gueudet, Pepin, & Trouche, 2012; Gueudet
& Trouche, 2009) which is grounded on Rabardel’s work (Rabardel, 2002) and
enlarges the instrumental approach (Trouche, 2004) in mathematics education. One
important distinction between the two approaches lies in the extension of the concept
of artefact, in the former approach, to resource which allows considering wider set
of materials intervening in the teachers’ and students’ activities. A resource can be
conceptualised “as both noun and verb, as both object and action that we draw on in
our various practices (Adler, 2000, p. 207)”. Thus, the approach has the potential to
take in consideration material, human and cultural resources such as language, time,
mathematics teachers, etc. Moreover, a resource is never isolated but belongs to the
wider set of resources (Gueudet & Trouche, 2009).
While one focus of this approach is on the teacher’s work with the resources, the
study of students’ use of resources can provide the overview of their actual use
(Gueudet & Pepin, 2016). Also, this approach has the potential to provide rich
analyses if used to evaluate students’ work in terms of interactions with different
resource systems (Trouche & Pepin, 2014) or with a particular resource (Aldon,
2010). We will employ this approach to analyse how students interact with available
resources.
In particular, we analyse students’ techniques when working digitally in mathematics
(Artigue, 2002). A technique is perceived as “a manner of solving a task (Artigue,
2002, p. 248)”. While students work on mathematical tasks in a digital environment,
they might adopt paper and pencil based techniques or instrumented techniques. The
obvious and easily observable objective of each technique is to reach the goal of the
activity i.e. to produce the results whereas the contribution of a technique to the
learning of involved mathematical concepts might not be easily recognisable. The
former corresponds to pragmatic value while the latter corresponds to
epistemological value liked to each technique.
We seek to explore the kind of techniques implemented by the students in the digital
environment to make sense of how students interact with this environment while
working on mathematical tasks. Furthermore, realisation of the values attached to the
students’ instrumented techniques will also help to understand the role of digital
resources in their learning (Guin, Ruthven, & Trouche, 2005). There are several
resources involved in present situation, therefore, we confine to the general features
of corresponding techniques in the present paper. By this, we mean to consider
students’ general organisation of digital work with several resources related to all
contents in a Calculus course. We ask the following question: How do engineering
students incorporate resources during their work in an online learning environment?

146 [Link]:indrum2018:174742
THE SETTING
This study took place in a Norwegian public university during the spring of 2017.
Undergraduate students enrolled in electronics engineering program participated in
this study. In their Calculus course, students were offered an online learning
environment such that they could work remotely by interacting with the provided
resources. These resources were made available to them electronically to work and
proceed through the course. There were no mandatory lectures, and they could access
the lecturer in the case they needed additional support. The final examination was
also in digital format where the students were allowed the access to tools and aids.
The resource system comprised MyMathLab environment, tutorial videos coupled
with the notes, and the textbook. The students’ homework and the formative
assessments were administered online through MyMathLab system. MyMathLab is
an interactive learning system for practicing mathematics online (figure 1). While
this system provides an online platform for homework and assessments, it also
facilitates students in solving the tasks by providing help and feedback. Students can
seek help through utilising “help me solve this” or “view an example” functions in
the system. The former lets the student solve a similar task by guiding on each step
whereas the latter shows a similar worked-example. The interactive nature of
MyMathLab system allows considering it as a resource which can potentially
influence students’ activity in this course.

Figure 1. Interface of MyMathLab environment.

The tutorial videos are created by the lecturer, and recorded by using a document
camera. Each video deals with a specific section in the book and is named

147 [Link]:indrum2018:174742
accordingly. In these videos, the lecturer explained the topics in the book and
worked through the relevant examples occasionally. The notes pertaining to the
video tutorials were available online. The length of these videos varies depending on
nature of the concerned topics. The tutorial videos replaced lectures and it was
expected that students would watch the videos to learn mathematical topics. The
textbook served as the central resource in the sense that MyMathLab and tutorial
videos were based on contents in the book.
In this course, a compulsory task was the group project in which students were
required to prepare a question bank related to integration. That question bank was
needed to be programmed in the STACK environment, a computer aided assessment
platform. Maxima is the programming language used in the STACK, thus they were
required to learn Maxima to complete the project. The intention was to make
students familiar with programming language and its use in mathematics.

RESEARCH DESIGN AND METHODS

The case study research design (Yin, 2013) has been followed in this study. A group
of three students has been observed over the semester. The methods used to generate
data include group observations, semi-structured interviews, individual weekly
journals and field notes. Using multiple methods for data collection contributed to
triangulation of data.
In order to be able to observe participants’ activity, we requested them to work at
campus each week for which they agreed. During these sessions, they worked on
their routine work including homework and assessments. Video recordings of their
group work accompanied with the screen recordings to follow the activity on their
computer screens have been collected. Screen recordings of their individual work
external to these group sessions have also been collected. Furthermore, weekly
journals containing self-reports about their use of resources were included to get the
detailed overview. The journal was provided to participants in tabular format which
they filled and submitted electronically each week. In the journal, they were asked to
specify the resources they used and state how the use of a particular resource helped
them in their work each week. The semi-structured interviews were held occasionally
to understand the emerging patterns in their use of resources. During the group work
sessions, participants communicated in their native language whereas the interviews
were held in English. Both the group sessions and the interviews were transcribed.
We analyse participants’ weekly journals, a semi-structured interview in the middle
of the semester, screen recordings, and the field notes for reporting on students’ use
of resources in their work. This interview is being counted on because the
participants were inquired about the general manner in which they used the
resources. The observations, screen recordings and the field notes are being counted
on while identifying participants’ techniques during their work.

148 [Link]:indrum2018:174742
ANALYSIS
Participants’ weekly self-reports about use of resources
Table 1 presents the overview of participants’ use of several resources as they
reported in their journals. The manner in which they used them in their work and
their evaluations of resources have been extracted from their journal inscriptions.

Resource used How they incorporated resources Comments about resources (if any)
in their work
Tutorial videos Watched to get information to Easy to understand through videos
complete homework
MatRIC videos Skimmed through the video at
amplified speed
Own note Used the already solved similar
problems in the notes, to recall the
problems (methods for solution)
Textbook Read through the book, found
formulas to work on homework, got
questions from book (during project)
Maxima Programmed tasks in Maxima for the Programming in Maxima is hard but
project, used while doing homework, when it is done, all the problems are
solved tasks using Maxima easy to solve
WolframAlpha Used as a shortcut to get answers, Easier to use than Maxima, Faster
compared answers obtained from than using calculator, useful when the
Maxima, got help with solving answer is in the form of expression
difficult tasks instead of numbers
MyMathLab Worked on homework, learnt Powerful tool, easier to get help and
specific topic, solved some questions information online
with higher difficulty
Internet
Lecturer’s notes Tailored” for the tasks at hand, the
most relevant piece of information
Youtube vidoes Watched Maxima tutorials
Mathway and Solved questions Severely increase the probability to
other online get the correct answer, and therefore
calulators the overall score.
STACKS Made some questions in STACKS
Table 1: Overview of participants’ use of resources.

149 [Link]:indrum2018:174742
The three participants, Tor, Per and Jan, used MyMathLab almost every week
because homework and assessments were required to be done in this system. As
regards the textbook, Tor did not report the textbook in the journals rather he used
the lecturer’s notes. While in Per and Jan’s weekly reports, they pointed out few
ways in which they used the textbook on different occasions. The textbook served as
a source of getting questions, checking answers to those questions, getting help with
formulas, and going through examples in the book. During their project work, they
consulted the book to take questions and subsequently checked the answers for those
questions.
The tutorial videos were reported to be used by Jan and Per during their work. Jan
occasionally watched the videos and when specifying about the kind of help, he used
the word understand linked with this resource such as “to try to understand how to
calculate…” and “to understand the calculation behind the math”. Per has also
mentioned the use of videos and commented, “I easily understand it when someone
explains me the way of solving a problem”. Tor did not mention any tutorial video
provided by the lecturer, however he watched few videos on other platforms,
MatRIC TV (an online resource containing videos aiming to support students in their
transition from high school to university) and YouTube, once for getting introduction
to partial integration and at another occasion to learn Maxima – the programming
language.
It can be seen that participants used some other resources in their work such as
online calculators, WolframAlpha, Maxima and internet (cf. Table 1). Tor named
several online calculators including Mathway ([Link] and
WolframAlpha ([Link] to solve the tasks and to compare
the answers they got in Maxima while working on the project. He mentioned that he
used online calculators for saving time, however, he wrote, “I did not learn anything
doing this, but it severely increases the probability to get the correct answer, and
therefore the overall score”. Wolfram Alpha has also been used by Per and Jan in
order to verify whether the answers they got were correct. While working on the
project, they picked some questions from the book and programmed in Maxima. To
check the answers to those questions, they used WolframAlpha.
After completing the group project that involved learning Maxima, this programming
language became an important resource for them to solve tasks in homework and
assessments. Both Per and Jan began making programs for solving each task to
liberate themselves from calculations. Per inscribed in a weekly journal, “(I) used
Maxima to make a program to solve the problems in an easy way. This is hard to
make, but when it is done, all the problems are easy to solve”. Tor did not seem to
use Maxima a lot, he spent some time on learning how to use Maxima for solving
tasks in one week, and then spending some more time in the next week, he rather
chose to focus on MyMathLab. He inscribed that, “it’s (MyMathLab) a more
powerful tool and it’s easier to attain help and information online”.

150 [Link]:indrum2018:174742
In response to a question about using videos in a semi-structured interview, Per
explained his way of working on homework using the provided resources.
Per: These topics I think are quite hard to learn all by yourself. When I get a new
topic, I first try to solve it myself, if I can’t do that I try to look at the
examples in MyMathLab… and if I don’t completely understand the
examples I take a look at Olav’s (lecturer) video…mainly the examples’
videos because then I get to see the practical kind of way to do..to solve
questions.
Int: How would you rank the provided resources? Which one do you first consult
with?
Per: First, I will try to do it myself because then I think I… remember and learn it
the best because then I have to think and ….and if I can’t do it that
way…then I will try to look at example just to get a few hints. If that does
not work then I watch the videos because I can’t look at the notes (provided
by the lecturer)…I have to get explanation of what he is doing step by step.
Tor’s response was somewhat similar as he replied:
Int: Did you use any video while working on last week’s homework?
Tor: No, I think MyMathLab seemed sufficient so far.
Int: Ok. So which resource did you use for getting introduction to the new topic?
Tor: I tried first MyMathLab but it went fine so I just carried on. …I check the
notes and watch the videos if I get stuck..
Int: So, you turn to the videos when you get stuck.
Tor: When it is a new topic, then I just skim through his notes, but since we have
integration from a couple of weeks now, I am pretty confident and go
straight with it.
While Jan responded to the same question as follows.
Jan: I did not watch that many videos. I mostly use MyMathLab and just see the
examples…and if I can’t get it from there then I go to…to the book because
it is faster… and eventually go to the videos if I do not get constructive help
from there.
The participants preferred MyMathLab during their work for being the source of
quick and most relevant help in comparison to the other available resources. This
approach of working on the tasks saved them time and effort to search for the
required piece of information from other resources such as the videos and the
textbook. However, the use of MyMathLab can be considered more pragmatic as
both Per and Jan mentioned that the kind of help they get from MyMathLab is in the
form of examples which contributes more towards producing the results.

151 [Link]:indrum2018:174742
Another approach was to watch the videos when the help from MyMathLab was not
sufficient as evident through participants’ responses in the interview. The use of
videos has not been preferred much but participants reported that they consulted the
videos when they needed to understand something. As discussed earlier, the help and
feedback in MyMathLab concern the task only as it offers the formula and solution-
steps for the task. They might have needed to consult the videos to learn the concepts
involved in those tasks in case when just knowing the solution steps in a question did
not work. In the journal data, Jan and Per wrote that they used the videos to
understand thus it indicates the epistemic value linked to usage of videos.
Observing participants’ activity helped in finding that the use of different resources
affected their manner of working on tasks i.e. techniques. We seek to categorise the
participants’ techniques pertaining to different resources they used, and by
considering their motives behind use of each resource helped in recognising the
pragmatic and epistemic value of their techniques. It is found that they increasingly
used the digital tools to solve the tasks in MyMathLab environment with the
progression in the course. This led to the use of more instrumented techniques
instead of paper and pencil techniques promoted in the lecturer’s videos and through
MyMathLab. For instance, Tor mentioned in his weekly journals and it is observed
in the screen recordings of his individual work that he used several calculators to
work on homework as well as assessments. The participants themselves perceived
this technique of using online calculators to solve the task as pragmatic.
Two of the participants used Maxima in their work as evident from journals and
could be seen through the screen recordings of their work. They wanted to be
pragmatic in order to make their future work easier. Making programs for each task
for the first time can not be considered as merely pragmatic as Per mentioned that he
found it hard. The difficulty in making programs may be linked to their knowledge of
programming in order to code mathematical tasks. However, the extent to which it
contributes epistemically in learning mathematics is not covered in present paper.

DISCUSSION AND CONCLUSION

In this study, we observed how a small group of three students interacted with the
resources when provided with an online learning environment in their Calculus
course. The environment allowed self-regulated learning and students could work
remotely on their homework and assessments. To make sense of the opportunities for
students’ learning with resources in this environment, we explored their manner of
incorporating the resources in general organisation of their digital work.
Furthermore, we discussed the epistemic and pragmatic potential of participants’
techniques.
In terms of resource usage and the corresponding techniques, participants opted for
the resources and the techniques which were pragmatic in terms of producing results
for the assigned tasks. Pragmatic techniques involved the use of online calculators,

152 [Link]:indrum2018:174742
using help in the MyMathLab to produce the results for tasks. Watching videos for
learning mathematical concepts seemed to be time consuming and hence not
preferred much. Participants appropriated the programming language to work on the
tasks with the motive to be more pragmatic and produce results easily in their work.
An important factor which is likely to cause the preference for more pragmatic
instrumented techniques was the online final examination where they could use the
resources. As for students, it is quite important to prepare according to the
examination to be able to score better.
This case study provides an example of a self-regulated learning environment created
for students to work independently. Our findings suggest some general prospects
which are worth paying attention when assigning online homework to students.
Combination of an online homework with online examination is likely to cause
students to use unexpected use of resources and techniques, for instance, online
calculators and solution tools in the present case. This observation also relates to the
nature of tasks posed in an online homework environment. Variety in the nature of
tasks, such as open-ended tasks, may lead students to interact with resources
epistemically.

REFERENCES
Adler, J. (2000). Conceptualising resources as a theme for teacher education. Journal
of Mathematics Teacher Education, 3(3), 205-224.
Aldon, G. (2010). Handheld calculators between instrument and document. ZDM,
42(7), 733-745.
Artigue, M. (2002). Learning mathematics in a CAS environment: The genesis of a
reflection about instrumentation and the dialectics between technical and
conceptual work. International Journal of Computers for mathematical
learning, 7(3), 245-274.
Borba, M. C., Askar, P., Engelbrecht, J., Gadanidis, G., Llinares, S., & Aguilar, M. S.
(2016). Blended learning, e-learning and mobile learning in mathematics
education. ZDM, 48(5), 589-610.
Gueudet, G., & Pepin, B. (2016). Students' work in mathematics and resources
mediation at university. In E. Nardi, C. Winsløw, & T. Hausberger (Eds.),
Proceedings of the First Conference of the International Network for Didactic
Research in University Mathematics (INDRUM 2016, 31 March-2 April 2016)
(pp. 444-453). Montpellier, France: University of Montpellier and INDRUM.
Gueudet, G., Pepin, B., & Trouche, L. (Eds.). (2012). From text to ‘lived' resources:
Mathematics curriculum material and teacher development. New York:
Springer
Gueudet, G., & Trouche, L. (2009). Towards new documentation systems for
mathematics teachers? Educational Studies in Mathematics, 71(3), 199-218.

153 [Link]:indrum2018:174742
Guin, D., Ruthven, K., & Trouche, L. (Eds.). (2005). The didactical challenge of
symbolic calculators: Turning a computational device into a mathematical
instrument. New York: Springer
Pepin, B., Choppin, J., Ruthven, K., & Sinclair, N. (2017). Digital curriculum
resources in mathematics education: foundations for change. ZDM, 49(5), 645-
661.
Rabardel, P. (2002). People and Technology: A cognitive approach to contemporary
instruments (translation of Les Hommes et les Technologies). Retrieved from
[Link]
[Link]/file/index/docid/1020705/filename/people_and_technology.pdf
Trouche, L. (2004). Managing the complexity of human/machine interactions in
computerized learning environments: Guiding students’ command process
through instrumental orchestrations. International Journal of Computers for
mathematical learning, 9(3), 281-307.
Trouche, L., & Pepin, B. (2014). From instrumental to documentational approach:
towards a holistic perspective of teachers' resource systems in higher
education. Research in Mathematics Education, 16(2), 156-160.

154 [Link]:indrum2018:174742
 Central dialectics for mathematical modelling in the experience of a
study and research path at university level
Nacho Monreal Galán1, Noemí Ruiz-Munzón1, Berta Barquero2
1
Pompeu Fabra University. Escola Superior de Ciències Socials i de l'Empresa –
Tecnocampus, Mataró, Barcelona, Spain, nruiz@[Link],
jimonreal@[Link]; 2University of Barcelona. Faculty of Education,
Barcelona, Spain, bbarquero@[Link];

This paper presents the a posteriori analysis of a study and research path (SRP) on
comparing reality and forecasts of the number of users of certain social networks,
which appears as a teaching and learning proposal for mathematical modelling. We
analyse the main elements of the SRP that have been experienced with a first-year
course at university in management sciences degrees in two consecutive courses,
2015/16 and 2016/17. We focus our analysis on two essential dialectics for
mathematical modelling to be developed: the questions-answers and the media-
milieu dialectics. In particular, we take empirical results from the two successive
implementation of the SRP to outline through which mechanism these two dialectics
could be prompted.
Keywords: Mathematical modelling, study and research paths, dialectics, questions-
answers, media-milieu.

INTRODUCTION: THE SRP AS TEACHING PROPOSAL FOR

MATHEMATICAL MODELLING
The starting point of the research is the problem of inquiring into the conditions that
can help, and constraints that hinder, that mathematical modelling can be integrated
and developed in the teaching and learning of mathematics into current educational
systems, in particular, at university level. Researchers and practitioners agree that
teaching should not be focused only on the formal transmission of knowledge, but
also should provide students of the tools for enquiring into the study of real
phenomena and integrate mathematics as an essential modelling tool. This change
requires moving from a more traditional pedagogical paradigm of transmission of
knowledge, which mostly focuses on introducing students to already built
mathematical knowledge, to a paradigm of inquiry where the solving of problematic
questions leads learning processes and motivates the study of new knowledge.
In the particular case of the research on modelling and their applications and on
inquiry-based approaches some big steps have been made showing how, under
certain suitable conditions in different educational levels and curricular frames,
modelling activities may be successfully put into practice (Artigue & Blomhøj,
2013). However, although school institutions and researchers agreed that modelling

155 [Link]:indrum2018:174841
should play an important role for a change towards a new pedagogical paradigm, the
real situation in school and university is not satisfactory (Stillman et al. 2013) and
the dissemination and long-term survival of these teaching proposals based on
modelling follows as a big challenge for mathematics education (Galbraith 2007,
Burkhardt 2006).
In the case of applications and modelling a shared excitement unites many who have
enthused about early experiences in the field, for example when students unleash latent
power that for whatever reason had remained fettered in their previous mathematical life.
However this very exhilaration can work against further progress, both individually, and
particularly at a system level, by creating a sense of adequate achievement that obscures
the reality that there is so much more to do.
In our research, developed in the framework of the anthropological theory of the
didactic, we focus on the use of the study and research paths (SRP) as
epistemological and didactic model (Chevallard, 2015; Winslow et al., 2013; Serrano
et al., 2013) to face the problem of moving towards a functional teaching of
mathematics and, particularly, where mathematics are conceived as a modelling tool
for the study of problematic questions. According to Barquero and Bosch (2015), the
starting point of an SRP should be a lively question of real interest for the
community of study (students and teacher/s). The study of Q0, called the generating
question, evolves and opens many other derived questions Q1, Q2,…, Qn. The
continuous looking for answers to Q0 (and to its derivative questions) is the main
purpose of the study and an end in itself. As a result, the study of Q0 and its derived
questions Qi leads to successive temporary answers Ai that can be helpful in
elaborating a final response R♥ to Q0. These first characteristics can be associated to
the first level of analysis of the SRP that we here consider, it consists in the
dialectics establishing between the questions posed and the likely answers appearing
(questions-answers dialectic) which also provide the basic structure of an SRP to be
implemented and to be enriched after each implementation. This first layer refers to
the evolution of questions to be faced and the necessary knowledge to be used.
Another central dimension for an SRP is the media-milieu dialectics, which
constitutes the second level of analysis. As described in the aforementioned
investigations, the implementation of an SRP can only be carried out if the students
have some pre-established responses accessible through the different means of
communication and diffusion (that is, the media), to elaborate the consecutive
provisional answers Ai. These media are any source of information, such as:
textbooks, treatises, research articles, class notes, or the teacher acting as main
media. However, the answers provided are constructions that have been elaborated to
provide answers to questions that are different to the ones that may be put forward
throughout the mathematical modelling process. Thus they have to be re-constructed
according to the new needs. Other types of milieus will therefore be necessary to test
the validity and appropriateness of these answers. This second level of analysis put
attention to the evolution of the students’ milieu.

156 [Link]:indrum2018:174841
With this aim, we present an analysis of a particular SRP about the evolution of users
of certain social networks that we will analyse in term of these two central dialectics
and, more concretely, focusing on two critical questions:
(1) How to enhance dialectics between posing questions and looking for answer as
engine of the modelling process? How to transfer to students the responsibilities of
posing questions and looking for answers? (2) What milieu is necessary for students
to facilitate a rich development of modelling? How a richer media-milieu dialectics
can be developed?

DIDACTIC ANALYSIS OF A MODELLING PROCESS: THE CASE OF AN

SRP ABOUT THE EVOLUTION OF THE NUMBER OF FACEBOOK USERS
We focus on analysing the case of an SRP on Comparing forecasts against reality in
the case of Facebook users’ evolution. The first time it was experienced was during
the winter term of the academic year 2015-16 with first-year students of Business
Administration Degree and of Innovation Management (BAIM), all from the ‘Escola
Superior de Ciències Socials i de l’Empresa-Tecnocampus’, Pompeu Fabra
University (see Barquero, Monreal, Ruíz-Munzón & Serrano, 2017). During the
academic year 2016-17 it has been implemented again in the same university degree.
The SRP has run in a modelling workshop that was optional activity for students
during these last two academic years. In this paper we analyse and compare both
implementations by using two central dialectics: the questions-answers and the
media-milieu dialectics.

The initial situation starts from real news about a research performed by Princeton in
2014, in which it was predicted that Facebook would lose the 80% of its users before
2017. Hence, the generating question Q0 presented to students is about: Can these
forecasts be true? How can we model and fit real data about Facebook users’
evolution to provide our forecast the short- and long-term evolution of the social
network? How can we validate the conclusions of Princeton? The experimentation
was structured in three interconnected phases linked to the generating question Q0,
building up the a priori design of the SRP, then reflected in the design of the c-book
unit. A first phase that focuses on the open research of real data about Facebook
users, a second one focused on finding mathematical models (mainly based on
elementary functions) that may provide a good fitting to real data, and a third one
about the use of these models to forecast the behaviour of the social network in short,
medium- and long-term in terms of number of users and about how to decide about
best and most reliable model.
Previously, during the first term (4 ECTS of the subject) students had been getting
familiar with the main properties of some groups of functions (polynomial, rational,
irrational, exponential and logarithmic functions) as well as with basic topics on
differential calculus and its applications to the study of the monotony and
optimization of one real variable functions. Actually, before starting the first session

157 [Link]:indrum2018:174841
we asked the students to answer a test on some of the mathematical tools that mainly
make the workshop up, as indentifying some types of elementary functions or the
concept of fitting model in certain scatter plots.
In the first experimentation 27 students, working in ‘consultant teams’ of 3-4 people,
got the order from MS2 Consulting (‘Mathematical Solutions Squared’) previously
described as Q0 and they were asked to deliver a final report by the end of their work
as an oral presentation as response to the MS2 request. The implementation
combined face-to-face sessions in the teaching device called ‘Math modelling
workshop’ (in a total of six 90-minutes weekly sessions) for the miss-in-common of
the junior consultant teams’ partial reports, with work out of the classroom. For the
second experimentation 12 students (18 students started the workshop, but they left it
in the second session due to external matters) worked also in teams of 3-5 people.
This time we opened a Moodle virtual classroom to provide the students the teaching
aid of the workshop, as well as some communication and collaborative tools (forum,
a different wiki for each phase, etc.) to write their progress and pose their new
questions. The generating question Q0 was presented in a small dossier, next to the
initial subquestions of each of the SRP phases (Q1, Q2 and Q3). The workshop run
over seven face-to-face 90-minutes sessions before the final session, in which
students should present their conclusions in an oral presentation in front of an
external committee with representatives from MS2 Consulting.
Next we sketch in the case of the two implementations how the different dialectics
were prompted by both: (a) the design of the unit (by its initial design but also by the
different changes introduced according to students’ requirements: new questions and
answers not envisioned, new media required, etc.) and (b) the didactic gestures and
devices to manage its implementation.
Integrating the dialectics of questions-answers as engine of the SRP
The a priori design of the SRP was basically the same in both implementations,
structured in three interconnected phases linked to Q0, which guided the design of
the workshop throughout its implementation. A first phase focuses on the open
search of data about Facebook users; a second one focused on mathematical models
(mainly based on elementary functions) that might provide a good fit to Facebook
users data; and a third part focused on the use of these models to provide short-,
medium- and long-term forecasts about the number of users of Facebook and on how
to decide on the best and most reliable model. Figure 1 (and the explanation below)
shows the link between different questions (Qi) that were planned as likely to appear
in the real implementation of the SRP and some expected answers (Ai) from the
working teams. The only difference of the second design with respect to the first one
was motivated for the context in which Q0 was presented originally: the predictions
made by Princeton were supposed to happen in 2017, and this year was present tense
for the students of the second experimentation. Hence, we decided to make the same

158 [Link]:indrum2018:174841
questions, but giving freedom to students of focusing in any other social network
students were interested in.

Figure 1: Tree of questions and answers of the different phases of the SRP

Q1: Which data sets about the users of the social network are better to consider in
our research?  A1: Each group look for the data to be used and shared; the whole
community agree on the terminology (year, period, units, etc.) and on the
dependent and independent variables to take into account.
Q1.1: Which time intervals may be considered? Q1.2: How can data be well-organized? Q1.3: How to
organise and visualise data? Q1.4: What can we say about the growth tendency of the data analysed?
Q2: Which mathematical models provide the best fitting of data about the network
users?  A2: Each consultant group is asked to propose and justify three
mathematical models fitting real data.
Q2.1: Which models (based on elementary functions: linear, parabolic, exponential, etc.) may fit the
data? Q2.2: How can the coefficients of the model be determined?
Q3: How can we decide about the ‘best’ fitting model? Can we use this model to
predict the future evolution of users?  A3: The teams need to create tools to
justify why a mathematical model/s is/are the ‘best’ with respect to: (a) fitting data
and (b) forecasting the evolution of users.
Q3.1: How can we compare the error committed between reality and forecasts provided by models?
Q3.2: Can be the same model used for the short- and long-term forecasts?
Let us now comment the main features of the a posteriori analysis of the
experimentations, referring here to the questions-answers dialectics level.
Regarding the first phase, we should remark the ease with which the students found
real data about the evolution of the social net. The students mainly found the
information by means of a graphical representation. This fact strongly determined
their analysis, since they mainly focused in the graphical analysis growth tendency of
the data, but not in their numerical versant, making Q1.4 being treated before the

159 [Link]:indrum2018:174841
other ones (it was considered that students would have data in table format before
having graphs). With respect to the first experimentation, the fact that many groups
found the same data triggered an intense debate and interchange of ideas among
them, which took us to consider a brainstorming session about the previous
hypothesis in the classroom, and as a consequence, the duration of the first phase
was extended from 3 to 4 sessions. Due to the wealth of answers collected during the
brainstorming we asked the students to deliver a first report in a poster format,
synthesizing their findings, conclusions and new questions made by them. In the
second experimentation the fact that students could choose a social network implied
a disruption with the usual topos of the students in the process of study, since they
were responsible on the delimitation of the field of study. They noticed about the
difficulty of finding reliable data of some of their choices (Snapchat, Instagram,
Twitter…), so finally only Facebook and Instagram were object of study, and not
only the number of users with respect to the time, but also other variables that could
have a relation. Another question that raised here was the role of the intervals of time
of the data obtained, and how to work when data are not regularly spaced in time.
These questions enriched the a priori design of the SRP. The presentation of the first
phase was done on the third session, and there had not been interchange of ideas with
other groups during the first phase. Furthermore, we asked students to present their
plan of work: the questions that they wanted to deal with, when and how. This
showed that each group had planned the next steps in many different ways and with
many different variables. Nevertheless, the lack of time and our interest in the study
of one real variable function made us proposed the students to use only the variables
“Time” and “Users”.
Let us focus now on the second phase. In both experimentations the analysis of the
different proposals made arise a non-expected aspect: the use of piecewise functions.
Then the expected answer to Q2 about the consideration of models based elementary
functions (linear, quadratic, exponential, etc.) was extended. In the case of the first
experimentation, since many groups worked finally with very similar data on the
worldwide evolution of FB users, we took two new decisions: (a) give each team a
second set of different data, corresponding to different geographical areas, in order to
contrast their hypothesis and extend their study; and (b) ask for more than one fitting
model for each data set. This was not necessary in the second experimentation, since
each group had different data sets. Besides, in both workshops new questions and
answers appeared at this stage with respect to the change of tendency of the fitting
models, in accordance to a particular action or to decisions of the corresponding
social network (IPO, new rival social nets, purchases of the company, new
developments, etc.), which determined the moments of change of tendency.
Furthermore, in the second experimentation we let the students choose a software for
representing their data and the functions. This made question Q1.2 emerge again,
since they needed to adapt their data to the different software used. Just one group
decided to use Geogebra, so they were provided the applets we used in the previous

160 [Link]:indrum2018:174841
experimentations (Barquero et al. 2017). Instead, another one decided to make
interpolation in order to find functions fitting their data, so they used Symbolab and
added some questions about how to solve non-linear systems of equations. The third
group worked with Excel for representing a scatter plot, and used linear and non-
linear regression. This motivated a big change in the SRP, since question Q3.1
emerged naturally in the exposition of their findings at the beginning of the third
phase (since they have used the R-squared of their model given by the software).
This gave birth to an interesting discussion on different ways of measuring the error,
and the professors had to present this question as a central matter.
Concerning the third and last phase, in both experimentations we only had two face-
to-face sessions of the workshop, but were not enough for a rich development of Q3.
Although this time constraint, in the first implementation of the SRP there were some
applets designed and made available for students to help on the simulation of models
and its contrast to real data. It helped students to delve into Q3, but not many new
questions appeared from this work. With respect to question Q3.2 only one group
dared with long term forecasts to give a date for the moment in which the users of the
chosen social net would start decreasing. Both implementations finished with a final
presentation of their modelling work and conclusions to an external committee.
Before finishing, we should remark that in both implementations the common
discussions, presentations and brainstorming session became the main device for
students to formulate and organize new questions, debate answers and contrast them.
The progressive enrichment of the milieu: the media-milieu dialectics
Since we have the first layer of analysis of the SRP in terms of the arborescence of
the questions-answers, it is important to ask when, where and how questions can
arise and answers can be developed. It is at this new level when there may appear the
different elements taking part of the milieu, composed of varied elements: questions,
temporary answers, pre-existing answers in or out school, means to validate answers,
experimental data, etc., accessible through different kind of media (textbook,
lectures, website resources, etc.). The relation among these elements can be analysed
through the media-milieu dialectics. The constant dialectics between the search for
data (for instance, real data about users of social networks, or about the company
changes) and pre-existing answers (ways to organise data, common models to fit
population evaluation, elementary functions, tools to control error, etc.) that exist in
different media available for students (web resources, contents of Mathematics
course, answers from lecturers from other courses...) and the creation of the
appropriate means (milieu) to integrate (or refuse) them has been central in our SRP.
Let us stress the importance of some of them.
In the first phase of the SRP, it was important to some groups the topics worked in
another course called ‘Introduction to digital communities’ (running in parallel to the
workshop) who helped on providing a general sense and functionality to Q0 and to

161 [Link]:indrum2018:174841
show how the students could look for real data and some techniques to organise
them. All these elements took part of the media accessible to students, at the time it
enriches students’ milieu mainly composed at this stage of the data sets that each
team chose to work with (shared and debated early with the whole class in the first
implementation, and before the second phase in the second implementation; even
strongly in this case, since the variety of data found was higher and let them make
comparisons between different social networks). All these elements helped them to
prepare a first report with the first temporary answer A1 (a poster format given in the
first implementation, and a face-to-face presentation in the second). Here we should
remark the importance of making their plans explicit (especially in face-to-face
sessions) to construct a common frame to be the source of new questions, as well as
and to integrate in their milieu new concepts about modelling, and ideas of other
groups that could help them. It is in the second phase in which we find more
differences. In the first experimentation, the a priori design contained some
Geogebra applets proposed to help students to explore different models based on
elementary functions (Q2). These applets provided the main media for students to
visualize data jointly with model simulation, and also took part of their milieu as
main tools for contrasting, comparing and deciding on the ‘best’ models to choose.
Nevertheless, there were other tools not planned in the a priori design (as piecewise
functions, or Gaussian functions, most of them part of their milieu, since they had
been introduced in previous courses) but provided by designing new applets. In the
second experimentation only one group used these applets, so their path followed
was more similar to the first ones; but two groups decided to use other software
mentioned above (that they could know from Statistics or other subjects), which
made the main difference with the first implementation: meanwhile the first applet
seemed to drive students to apply only a trial-error method, tools like interpolation or
regression made students arise an earlier answer. Here again the common forum
stated as a face-to-face session motivated an enrichment of the student’s milieu.
Regarding the third phase, there were several important questions that were not
addressed properly, such as Q3.1 about the way to measure the differences between
data and forecasts, but here there is a main difference between both implementations.
In the first one the students assumed and uncritically used the milieu made available
through the design of an applet, a sort of black box to get immediate answers.
Instead, in the second one students had to construct their own tool for measuring the
error, and one group made it with Excel. Just one group could answer Q3.1 but the
answer was totally produced by them, so they could communicate it to the rest of the
class, extending the appropriate milieu to other groups.

FINAL REFLECTIONS
First of all, we should mention that students are not in general motivated to validate
their results after a work of research, since a lecturer will finally do it. In this
workshop students were responsible to validate or justify every decision they made

162 [Link]:indrum2018:174841
by the end of each phase. And this is the main reason why other questions arise and
contribute to enrich the a priori design of the SRP.
In this paper we focus on the case of an SRP on comparing forecasts against reality
in the case of the evolution of the number of users of certain social networks to show
the use of two dialectics: the one of the questions-answers and of the media-milieu,
corresponding to two of the three complementary level of didactic analysis of
teaching and learning processes (Chevallard, 2008). Besides their analytic use, they
suppose a productive framework to enrich teaching and learning practices, in
particular, on modelling.
In what concerns to the questions-answers dialectics, the generating question Q0
about the controversy of the article by Princeton was adopted by the students with a
great interest from the very beginning and, up to the end of the process, was kept
alive. From the two presented implementations we can underline very important
conditions that were created. First, the flexibility of the lecturers and designers team
that were opened to readjust the schedule according to students’ team work.
Furthermore, they were very attentive to integrate in their presentations all new
questions and means that the students asked for. Second, students were very active
on the sessions to share their proposals, making derived questions emerge naturally,
some of them planned in the a priori design, some others that extended the initial
proposal. Regarding the media-milieu dialectics, in the case of this SRP, we took
several decisions along the implementation of transforming the media offered to
students to help them in the modelling process and also to observe the impact new
media had on students’ milieu. Nevertheless, giving students the chance of using
their own ICT tools, as was decided for the second experimentation, enriched the
media-milieu dialectics, since it helped to arise other different answers that had not
happened during the first experimentation. We may insist again on the role played by
very important contributions, such as collaboration with other subjects, focusing
some workshop sessions on discussing external answers that students brought, the
creation of applets to foster students’ experimental work, among other interesting
aspects.

ACKNOWLEDGMENTS
The research leading to these results has received funding from the Spanish R & D project: EDU2015-69865-C3-1-R
and EDU2015-64646-P.

REFERENCES
Artigue, M., & Blomhøj, M. (2013). Conceptualizing inquiry-based education in
mathematics. ZDM Mathematics Education, 45, 797–810.
Barquero, B., & Bosch, M. (2015). Didactic engineering as a research methodology:
from fundamental situations to study and research paths. In A. Watson & M.

163 [Link]:indrum2018:174841
 Ohtani (Eds.), Task Design in Mathematics Education. An ICMI study 22 (pp.
249-272). Dordrecht: Springer.
Barquero, B., Monreal, N., Ruiz-Munzón, N., & Serrano, L. (2017). Linking
transmission with inquiry at university level through study and research paths: The
case of forecasting Facebook user growth. International Journal of Research in
Undergraduate Mathematics Education. [Link]/10.1007/s40753-017-0067-0
Burkhardt, H. (2008). Making mathematical literacy a reality in classrooms. In D.
Pitta-Pantazi & G. Pilippou (Eds.), Proceedings of the Fifth Congress of the
European Society for Research in Mathematics Education (pp. 2090-2100).
Larnaca: University of Cyprus
Chevallard, Y. (2008). Un concept en émergence : la dialectique des médias et des
milieux. In G. Gueudet & Y. Matheron (Eds.), Actes du séminaire national de
didactique des mathématiques 2007 (pp. 344-366). Paris: IREM de Paris 7 et
ARDM. Brest, France.
Chevallard, Y. (2015). Teaching Mathematics in Tomorrow’s Society: A Case for an
Oncoming Counter Paradigm. In S.J. Cho (Ed.), Proceedings of the 12th
International Congress on Mathematical Education (pp. 173-187). Dordrecht:
Springer.
Galbraith, P. (2007). Beyond the low hanging fruit. In Blum W., Galbraith P. L.,
Henn H., Niss M. (Eds.), Modelling and applications in mathematics education:
The 14th ICMI study (pp. 79-88). NY: Springer.
Serrano, L., Bosch, M., & Gascón, J. (2013). Recorridos de estudio e investigación
en la enseñanza universitaria de ciencias económicas y empresariales. UNO:
Revista de didáctica de las matemáticas, 62, 39-48.
Stillman, G., Kaiser, G., Blum, W., & Brown, J. (Eds, 2013). Teaching Mathematical
Modelling: Connecting to Research and Practice. New York: Springer.
Winsløw, C., Matheron, Y., & Mercier, A. (2013). Study and research courses as an
epistemological model for didactics. Educational Studies in Mathematics, 83(2)
267-284.

164 [Link]:indrum2018:174841
 Task design for Engineering Mathematics:
process, principles and products
Karsten Schmidt1 and Carl Winsløw²
1
Technical University of Denmark, DTU Compute, Denmark, ksch@[Link];
²University of Copenhagen, IND, Faculty of Science, Denmark, winslow@[Link]
We present and analyse principles and process employed at the Danish Technical
University to use authentic problems from engineering (APE) in a first year
mathematics course, along with some of the products (actual student assignments).
Keywords: Mathematics for Engineers, Task Design
INTRODUCTION
Mathematics is as important in most branches of engineering, as engineering is to the
prosperity and development of contemporary society. Thus, it is of great importance
to investigate exactly what mathematics is needed by (future) engineers, and how it
could be effectively taught to them; such research is only emerging (see e.g. Winsløw
et al., to appear, sec. 2.5). In the common case where mathematics is taught in
separate “service” courses which cater to several different study programmes, these
questions may be considered in entire separation: a syllabus for the mathematics
course is decided based on needs in the different study programmes of engineering
which include the course, and subsequently the syllabus is delivered by mathematics
faculty. This amounts to a complete separation of external and internal didactic
transposition, in the sense of Chevallard (1992), where the selection of mathematical
contents to be taught may be based on needs and priorities from the engineering
disciplines, while the actual teaching is carried out according to generic standards and
methods for teaching mathematics. The aims of the overall study programme (in
engineering) are only considered in the external transposition of the mathematical
knowledge (see Fig. 1). We can call this model a parallel model for teaching
mathematics to engineers, as the internal didactic transposition runs in parallel to the
rest of the programme and does not interact with it (while it is certainly intended that
the students’ learning serves in other courses, later on).

Scholarly Mathematics Mathematics

knowledge to be knowledge actually
Mathematics taught (syllabus) taught
Int. Did.
Coordinated Transpo.
External Didactic
Transpositions Int. Did.
Transpo.
Engineering Engineering
Scholarly
knowledge to be knowledge actually
Engineering taught (syllabus) taught

Figure 1: The parallel model for didactic transposition in engineering education

165 [Link]:indrum2018:172127
In the literature on university mathematics education, it is widely agreed that the
parallel model has drawbacks:
- students may experience the mathematics teaching as unmotivated and
difficult, which is reflected in relatively high failure and attrition rates for some
engineering programs (e.g. Baillie & Fitzgerald, 2010).
- the knowledge they acquire in the mathematics course may not transfer readily
to engineering contexts, in the sense that students are able to invest the
knowledge acquired in mathematics courses when they need to do so in other
courses of the programme (e.g. Britton et. al., 2005).
Motivated by these well-known problems, the model in Fig. 1 has been modified, in
many universities, by various attempts to relate the internal didactic transposition of
mathematics more closely to the rest of the engineering programs (e.g. Kumar &
Jalkio, 1999).
One of the most common ideas to further such an interaction is that to include, in the
mathematics course, more or less simple examples and student assignments where
engineering problems are solved based on mathematical methods and theoretical
notions (see, for instance, Härterich et al., 2012). A main challenge here is that
university mathematics teachers usually have no in-depth knowledge of any
engineering discipline, let alone of all the specialties which the course they teach
caters to. Of course they may then ask engineering specialists for help to identify
authentic problems from Engineering which can be solved using the mathematics to
be taught in their course (we abbreviate this kind of problems as APE). In that way,
“Scholarly Engineering” may exercise a more direct influence on the internal didactic
transposition of mathematical knowledge (cf. Fig. 1). In this paper, we investigate
some general questions related to the implementation of this (quite common) idea at
the level of the internal didactic transposition:
RQ1. How could the identification and transposition of APEs be organised, given
the academic and institutional separation of university mathematics teachers from
their colleagues in engineering?
RQ2. What didactic variables (cf. eg. Gravesen, Grønbæk and Winsløw, 2017) are
relevant to the construction of assignments based on APEs?
It is clear that answers to these questions will depend on institutional conditions and
that even when such conditions are given, one will at most obtain very partial
answers in the sense of reasonably validated examples of organisations (RQ1) and
didactic variables (RQ2). As always in education, transfer of “answers” from one
context to another will require some adaptation and interpretation, but this limitation
may in fact be less important for the above questions, given the relatively high
similarity of mathematics courses, the engineering programs they serve and the
institution which deliver them. For these reasons, and given the importance of the
matter already argued, it appears worthwhile to present such locally and partially
validated answers. Concretely we will present and analyse the process, principles and

166 [Link]:indrum2018:172127
products of APE design done at the Technical University of Denmark since 2000.
Founded by H. C. Ørsted in 1829, this is one of the most prestigious Schools of
Engineering in Europe, and by far the largest in Denmark.
A TASK DESIGN PROCESS
Mathematics 1 (hereafter Mat1) is the basic mathematics course for 17 different
[Link].-programmes at the university, catering to about 1100 students per year.
The course occupies 1/3 of the students’ time during the first year, and covers
complex numbers, basics of Linear Algebra, Ordinary Differential Equations
including linear systems, and multivariate and vector Calculus up to Gauss’ theorem.
Besides ensuring a technical foundation for later work, the university also considers a
common course on mathematics as important to the formation of an engineer identity.
Most of the course is quite traditional, however with intense use of the computer
algebra system Maple. Exercises with easy applications to engineering occur.
However, during the last four weeks of the course, the students work on a “project”.
This is an assignment containing about 20-30 more or less challenging tasks, related
to a mathematical model related to an APE. The model is usually given in the
assignment, and while some new mathematics may be introduced, the starting point is
Mat1. Each project assignment is presented in a text of varying extent (ranging from
4 to 29 pages, averaging 11); it is those texts which we aim to analyse in this paper.
The students do the projects in groups, hand in a report of about 20-50 pages, and
defend their work during an oral exam, which accounts for 25% of their grade.
The groups can choose their assignment from a list of 4-5 projects, in part depending
on the study programme, with titles like those shown shown in Fig. 4. As the titles
suggest, the project problems come from many different areas of Engineering. Every
year, new projects are added and some are dropped; and the details of retained
projects are updated based on teachers’ experiences. The elaboration of new projects
is a particularly delicate undertaking. When the first projects were done from 2000-
2006, a systematic effort was deployed to engage researchers from the university –
both applied mathematicians and researchers from Engineering at large - to propose
project topics; they were then, mostly, drafted or adjusted by the course responsible.
Some are still used in revised form.
It is the task of the course responsible to organise the production and revision of
projects. The initiative can come from teachers at the course or other mathematicians,
who identify a more or less classical APE which can form material for a project; then,
the motivation is often that some specific parts of Mat1 can be worked on in new
ways. But the initiative also frequently come from colleagues from other
departments. In some cases, their motivation is personal fascination with mathematics
in a more or less current APE, and possibly ongoing collaboration with mathematics
colleagues in this relation (reflecting an interaction between Scholarly Mathematics
and Engineering, which could be added to Fig. 1). In other cases, they propose an
attractive current APE for a Mat1 project, in order to attract students to their specialty

167 [Link]:indrum2018:172127
later on – these colleagues then, sometimes participate for free as supervisors on the
students’ project work.
Summing up what the process involved in Mat1 could contribute to RQ1, at least two
venues can be identified in relation to Fig. 1:
- Scholarly (applied) Mathematics and other basic sciences such as chemistry
and physics, where the main source of motivation is mathematical contents
related to Mat1; but work done here can still involve or lead to APE. Project
proposals from this source are typically mathematically “rich” but are not often
related to current research.
- Scholarly Engineering, often with current APE’s; the elaboration of a project
typically necessitates considerable adaption to fit Mat1, and is often tailored to
the interest of students from a small range of study programmes.
Finally, the genesis of a project may involve a mixture of both sources, when the
APE is identified by scholars with a deep involvement in both areas.
DIDACTIC VARIABLES AND PRINCIPLES
To present and analyse the project assignments which have appeared over the last 15
years, we have defined 10 didactic variables (DV) which are relevant to classify
them according to the aims which have, explicitly or implicitly, been pursued (Fig.
2). Each variable has, in principle, a non-numerical range, but can be determined with
relatively high objectivity for each assignment, based on the text. The detailed
presentation of any project in terms of the variables will, naturally, be difficult to
compare with others when given in this form. So when considering all projects it
appears useful to assign indicatory numerical values to the DVs on a scale from 0 to
2: for instance, to assess the breadth of Mat1 contents which a given project requires
the students to work with, 0 indicates that only one topic (such as systems of linear
equations) is involved, 1 that a few topics from both Calculus and Linear Algebra are
involved, and 2 that the project combines many topics. Naturally, this “grading” is
not absolute but relative to other projects (cf. also Fig.2). In the next section, we
outline a concrete assignment and explain, at the same time, how the numerical
values of the other DV’s are set. The variables were initially formulated in by the
authors (based on the first authors’ many years of involvement in the design) and
subsequently validated and adapted during the actual analysis of assignments. The
variables thus constitute a concrete answer to RQ2, which is of course a partial
answer based on experiences from context we described. In the rest of this paper, we
provide more explanation on how the variables can be used to analyse concrete
projects and, potentially, to direct and systematize the design of student assignments.
For each DV, Fig. 2 also includes a brief description of the more or less explicit aims
which have been pursued in the construction of projects over the past 17 years. The
brevity required in the Table format does not allow for much nuance. We note that
what is ideal use of Maple is not subject to complete consensus among the teachers of
the course, or in relation to the rest of the institution. On the one hand, some course

168 [Link]:indrum2018:172127
teachers consider that students should use Maple whenever it is useful; while others,
including colleagues from other departments, often insist on the value of students’
mastery of basic manual computation (as reflected in the aims for DV3 given in Fig.
2). The variables DV4-10 all describe aspects of the relationship between the internal
transposition represented by the assignment, and Scholarly Engineering (cf. Fig. 1).
Their values are thus of specific importance to go beyond the parallel transposition.
Didactic variable (DV): Aim of designers:
DV1 What breadth of content areas from As many as possible, preferably
Mat1 are needed to solve the involving new combinations. Depth
assignment? What depth of use? beyond “standard tasks” required.
DV2 What new mathematical contents Contents in continuation of Mat1, not
are introduced? excessive for students to cope with
DV3 How must/can Maple be used? Maple should mostly be used to:
- DV3a How essential is the Maple use? avoid tedious
- DV3b What types of Maple functions (numerical, computations,
symbolic, graphical…) are relevant? and for tasks
- DV3c Are the relevant use known or new to students? which the
- DV3d Is there black box use of Maple ? students could not
- DV3e What parts of the Maple use are prescribed? handle otherwise
DV4. What is the ”theme” and source of Origin in APE, if possible source in
the problem the project attacks paper or ongoing research in engineering
DV5. Breadth of engineering problem – Ideally more than one branch of
are more disciplines involved? engineering involved
DV6. How is the mathematical model Ok if model is given in the assignment,
established and worked with? but the students should work with its
details and structure
DV7. How realistic is the model? As much as possible for the students
DV8. How are data used? Data from the source, used as there
DV9. Should the students look up This is not a main aim, except students
information outside assignment? should use Mat1 course material
DV10. How complete answers does the Clear and definite answers/points, to
model give to the main problem? give students a satisfying experience
Figure 2. Didactic variables for the analysis of project assignments.
PRODUCTS
A total of 37 projects have been proposed during the past 10 years. Not all projects
are used every year, and all are revised before use, in the light of past experience,
new needs in the course, and in a few cases, updates to the APE and its solution from

169 [Link]:indrum2018:172127
Scholarly Engineering. We first give a relatively detailed presentation and analysis
(based on the DVs) of one project; at the same time, we describe how each of the 10
DVs is assigned a value as described above, for a rough analysis of the projects. Then
we present an overview and rough analysis of the whole inventory of projects.
In-depth presentation of one project
We now take a closer look at one of the projects, entitled: Heat flow in a house –
simulation and dimensioning. The assignment is relatively long, 18 pages, including
about 5 pages of data. The first paragraph outlines the underlying APE:
The building sector accounts for about 40% of the total energy consumption in Denmark.
It is a common assumption that there is a large unrealized potential for reducing the
consumption (…) in a financially sound way. This requires knowledge of the physical
processes which affect the energy consumption of buildings, the financial aspects of the
construction and maintenance of buildings, as well as the mathematical methods used to
compute these.
It turns out that the energy flow in the building is modelled as an analogy of currents
in electric circuits (cf. Fig. 3). The project is based on a genuine APE, and
bibliography of the assignment includes a reference to Nielsen (2005) which is the
essential source (DV4 = 2), along with a “pricelist” from the construction industry,
and the last part of the assignment draws on a simple model of investments and
interest. Relative to other projects, this assignment involves a relatively broad area of
Engineering fields, and DV5 is set to 2. The introduction acknowledges that the
model proposed in assignment is “a bit simplified”, but in fact it still gives similar
results; we assign DV7 to 2, in spite of some problems (see end of this section).

Figure 3. A figure from the project assignment “Heat flow in house”.

The central model, illustrated in Fig. 3, concerns a house with three rooms, called
“climate zones”. Here 𝑄𝑄𝑘𝑘′ are the internal and external heat sources (heaters and
sunlight), while 𝐶𝐶𝑘𝑘 are the heat capacities of the rooms and 𝐾𝐾∗ are the heat
transmission coefficient of the walls of the house, reflecting that these walls involve a
variety of layers. Before introducing the final model house, the students’ work with
the simpler case of a one-room house, and an external temperature 𝑇𝑇𝑒𝑒 which is a

170 [Link]:indrum2018:172127
given sine function. Based on further assumptions, this leads to the model for the
internal temperature 𝑇𝑇𝑖𝑖 , as a function of time 𝑡𝑡:
𝑑𝑑𝑇𝑇𝑖𝑖
(*) 𝐶𝐶 = 𝐾𝐾(𝑇𝑇𝑒𝑒 − 𝑇𝑇𝑖𝑖 ) + 𝑃𝑃(𝑇𝑇𝑠𝑠𝑠𝑠𝑠𝑠 − 𝑇𝑇𝑖𝑖 )
𝑑𝑑𝑑𝑑
where 𝑇𝑇𝑠𝑠𝑠𝑠𝑠𝑠 is the desired internal temperature (constant), and 𝑃𝑃 is the performance of
the internal heat source. While (*) is just a first order ODE, it still gives rise to
interesting Engineering tasks: the investigation of stationary solutions, the
performance needed to ensure an average temperature of 19.80C, and the thickness
required to respect given limits on the oscillation of 𝑇𝑇𝑖𝑖 . The full model consists of a
system of three differential equations which are similar to (*) but with an added
complexity due to the heat contribution from sunlight which, moreover, is
investigated with two different models. The students must also take into account a
model of the walls involving layers of materials to be computed using authentic data.
Finally, to take into account the cost of construction, the students are given a simple
mathematical model for the total economy involving investment, interest, and
operation costs; the mathematics is very simple but still gives rise to interesting
questions regarding how to optimize, for instance, insulation thickness. Throughout,
the students use real data (DV8=2) but these are all given, so DV9=0. Throughout the
models are given to the students, and while students are given full and extensive
explanations, they are not really asked to do more than apply them; thus DV6 = 1.
The project draws on a broad range of Mat1-topics: harmonic oscillations and
complex exponential function, single and coupled differential equations, solved using
advanced matrix algebra, involving both eigenvalue problems and quadratic forms.
Thus DV1=2, while DV2=0 as almost no new mathematics is introduced (the
exception being the argument required to justify the stationary solution to the system
of differential equations, which involves an extended eigenvalue problem).
While most of the tasks can in principle be solved manually, the visualizations of
temperature variations corresponding to different parameter values decisively require
a tool like Maple. The assignment moreover invites to numeric experimentations,
possibly based on graphical representations, and standard use for tedious operations
like inversion of matrices, make the overall potential of Maple-use relatively average
for projects; we thus assign DV3=1, even if the realized use by some students in
some cases goes beyond a mere use of techniques known from the rest of the course.
In real practice, the project also suffers from some flaws. Some of the questions lead
to less interesting results (like tedious computations leading to a requirement of a
four-doubling of the wall thickness in order to reduce an already negligible
oscillation of 0.120C for 𝑇𝑇𝑖𝑖 ). More serious is the breakdown of the model when taking
into account the contribution of sunshine at low temperatures such as 80C, where the
stationary answer cannot be found. It can be argued that such problems often arise
with simplified models, but it still leaves a negative impression on many students,
which might be avoided by future revisions. Altogether, we consider DV10=1.

171 [Link]:indrum2018:172127
Inventory of project assignments
Fig. 4 lists the inventory of projects used in the last 10 years, analysed using the DVs.
DV (cf. Fig. 2)
Project name (shortened in a few cases) 1 2 3 4 5 6 7 8 9 10
Oscillations in Axle-bearing Systems 2 0 1 2 1 1 2 2 1 2
Micro/Nano Cantilever Based Mass Sensor 1 2 1 2 1 1 2 1 1 1
Enzymatic Hydrolysis of Cellulose 1 1 1 0 1 2 1 1 0 1
Modelling 2D Halbach permanent Magnets 1 2 2 2 1 2 2 1 0 2
Factorization of Integers 0 2 2 0 1 0 1 0 0 1
Heat flow in house – simulation, dimensioning 2 0 2 2 2 1 2 2 0 1
Quantum Mechanics in a Nutshell 2 2 2 2 2 1 2 2 0 1
Red Blood Cells – Optimization in Nature 1 2 2 1 1 2 1 0 0 2
Utilization of the Waste Product Whey 1 0 1 2 1 2 1 1 0 2
Forced Pendulum 1 2 2 1 1 2 1 0 0 2
Stability in Chilled Tank Reactor 1 2 1 1 1 1 1 1 0 2
Optimization of Work Cycles 2 1 2 2 1 1 1 0 0 2
GPS and Geometry 1 2 2 2 2 1 1 1 2 1
Oscillations in Grid Constructions 2 2 1 2 1 2 1 0 1 1
Groundwater Flow in the Forest Vestskoven 1 2 2 2 1 1 2 2 0 1
Internet Hit lists 1 2 2 2 1 1 2 2 1 2
Short Circuit in Electric Networks 1 2 2 1 1 1 1 0 0 2
Simulation of Stretch Reflex 1 1 1 2 1 1 1 0 0 1
Parking Orbits of Satellites 2 1 0 1 1 2 1 0 1 2
Solar Energy Absorption in Curved Glass houses 2 0 2 1 1 1 1 0 0 1
Flow in Chemical Reactors 2 0 1 1 1 2 1 0 1 2
Finite elements in One Dimension 1 2 2 2 1 2 1 0 1 1
Geodesic Curves 1 2 2 0 0 2 2 0 0 2
The Brains Glycose Metabolism 1 2 2 2 1 1 2 1 1 1
Resistors and Markov Chains 1 2 2 1 1 1 1 0 0 2
Dosage of Anaesthesia 2 1 1 2 1 2 1 1 0 2
Anthrax – Attack, Escape and Rescue 1 2 2 2 1 2 1 1 2 2

172 [Link]:indrum2018:172127
Decomposition of PCE 1 1 1 2 1 1 1 2 1 0
Modelling Concrete Moulding 1 2 2 2 2 1 2 2 2 1
Soap Membranes 1 2 2 1 1 1 2 0 0 1
Distribution of Electrons in Semiconductors 1 2 2 1 1 1 2 0 0 1
Methane Concentration Profiles in Soil 1 2 1 1 1 1 1 1 0 2
Train Running in the Alps 2 1 2 1 2 2 1 0 1 1
Proteins’ 3 Dimensional Structure 0 2 2 2 1 1 1 2 1 1
Reaction Kinetics 1 2 2 1 1 1 2 1 0 2
Error Correcting Codes 0 2 2 1 1 1 1 0 0 2
Phononic bandgaps 2 0 2 2 2 2 1 0 0 2

Figure 4: Inventory of current projects with values of the didactic variables

A number of interesting tendencies can be identified in the above table, including
apparent dependencies of some variables, potentials which appear relatively
unexplored (like DV9), etc.; some of these are still not fully analysed. We stress that
a simple sum of the values of didactic variables, for a specific project, cannot be
construed as a measure of the “didactic quality” of the assignment. One reason is that
the variables are not of equal importance (in particular, DV1, 4 and 10 are essential).
But more importantly, one cannot always construe the number two as being
objectively “the best possible value” of the DV; the aims listed in Figure 2 are open
to debate and the viewpoint of teachers and designers may differ. A good example is
DV4, where we have given “2” for projects with a clear APE, “1” for projects with an
authentic problem from basic science (e.g. Chemistry) and “0” for projects which are
not based on an APE but on a (prima facie) purely mathematical problem, such as the
project “Geodesic curves”. One can argue that a project of the “0” type can also be of
high quality as a project for engineers, in view of the importance in several branches
of the mathematical problem (in the example, DV2=2 and indeed, geodesic curves
have multiple applications in many branches of engineering, see e.g. Patrikalaksis
and Maekawa, 2010, 265-291). A similar uncertainty must also be pointed out for
other variables such as DV3, where the further graduation suggested in Fig.2 could be
useful to provide a more nuanced picture than in the analysis in Fig. 4, where “2”
merely means that Maple is indispensable for large parts of the project.
SUMMARY AND OUTLOOK
We have presented the principles, process and products of a relatively longstanding
effort to integrate elements of scholarly Engineering (APE’s) in the internal didactic
transposition of basic mathematics in a course catering to a wide range of
Engineering programmes, going well beyond isolated “applications” of a Calculus or
Linear Algebra. We have emphasised the multiple dimensions which such an effort

173 [Link]:indrum2018:172127
needs to consider, in order to maintain the link with the mathematical knowledge to
be taught within the module in question, and to establish non-trivial links with
Scholarly Engineering (cf. Figure 1). Certainly, the concrete inventory of variables
can be developed and adapted further, and we believe it can eventually become a
valuable explicit basis for the discussion of aims (right column in Fig. 3) of projects
in our and other similar contexts. More importantly, considering such explicit
variables could be an important tool for systematizing the design process, both as a
check list for constructing new projects and (in combination with the analyses behind
Figure 4) to identify potentials for enriching existing projects. We expect that the
variables will also become useful guidelines for investigating the effects of the
project work in this course as a means to facilitate the transition to later courses
where mathematics is so fully integrated into the Engineering knowledge to be taught
that the latter is in practice as inseparable from mathematics as music is from sound.
REFERENCES
Baillie, C. and Fitzgerald, G. (2000). Motivation and attrition in engineering students,
European Journal of Engineering Education 25(2), 145-155.
Britton, S., New, P., Sharma, M. and Yardley, D. (2005) A case study of the transfer
of mathematics skills by university students. International Journal of
Mathematical Education in Science and Technology 36(1), 1-13.
Chevallard, Y. (1991). La Transposition didactique: du savoir savant au savoir
enseigné (2nd edition). Grenoble: La Pensée Sauvage.
Gravesen, K., Grønbæk, N. and Winsløw, C. (2017). Task design for students’ work
with basic theory in analysis: the cases of multidimensional differentiability and
curve integrals. International Journal for Research in Undergraduate Mathematics
Education 3, 9-33.
Härterich, J., Kiss, C., Rooch, A., Mönnigmann, M., Darup, M. and Span, R. (2012).
MathePraxis - connecting first-year mathematics with engineering applications.
European Journal of Engineering Education 37(3), 255-266.
Kumar, S. and Jalkio, J. (1999). Teaching mathematics from an application
perspective. Journal of Engineering Education July 1999, 275-279.
Nielsen, T. (2005). Simple tool to evaluate energy demand and indoor environment in
the early stages of building design. Solar Energy 78(1), 73-78.
Patrikalaksis, N. and Maekawa, T. (2010). Shape Interrogation for Computer Aided
Design and Manufacturing. Berlin, Heidelberg: Springer.
Winsløw, C., Gueudet, G., Hochmut, R. and Nardi, E. (to appear). Research on
University Mathematics Education. In: T. Dreyfus, M. Artigue, D. Potari, S.
Prediger & K. Ruthven (Eds.), Developing Research in Mathematics Education:
Twenty Years of Communication, Cooperation and Collaboration in Europe.
London: Routledge.

174 [Link]:indrum2018:172127
 Mathematical modelling and activation –
a study on a large class, a project-based task and students’ flow
Thomas Gjesteland, Pauline Vos, and Margrethe Wold
University of Agder, Norway, [Link]@[Link]

We studied how engineering students in a large class (n=346) can be activated by a

project-based task, in which they have to model mathematically the motion of an
object. The students had to throw an object, use (1) their smart phones for filming,
and (2) tracker software for capturing the motion. Through a poster, they had to
report their video analysis. We framed activation through the concept of flow, which
is a state of being fully absorbed by an activity. We administered a web-based
questionnaire (response rate 69%). The results show that such a project-based task is
feasible with >300 students and activated them: three out of five experienced flow.
Also, we validated the theory that for experiencing flow, a task must be perceived as
challenging and that one’s skills should match that challenge.
Keywords: flow, large class, mathematical modelling, mathematics for engineers,
novel approaches to teaching, project-based tasks.

INTRODUCTION
Harris et al. (2015) studied engineering students’ values regarding mathematics
finding that not many first-year engineering students have a positive stance towards
mathematics. The students see mathematics as a hurdle in their studies, and they are
disappointed by the mathematical demands in the first year of their studies. Some
even indicate that they wouldn’t have chosen the engineering direction if they had
known about the mathematics demands before. Nevertheless, mathematics needs to
be part of engineering studies, because alumni from engineering studies, such as
engineers, managers, researchers, etc., need mathematical modelling competencies to
describe, analyse, and predict phenomena to solve problems at the workplace (Alpers
et al., 2013). This means, that in particular mathematical modelling needs to be
included in engineering studies. It can be integral part of the mathematics curriculum,
but the learning of mathematical modelling can also take place in other disciplines,
such as physics, where mathematical models are used to describe and analyse
physical phenomena. The study described in this paper centres on a mathematical
modelling task situated within kinematics (the physics of movement).
In university first-year studies, engineering students often attend large-scale lectures
and have tutorial sessions to practice examination-like exercises. However, research
has demonstrated the advantages of activating, inquiry-based tasks over these
traditional instruction methods (De Jong, Linn, & Zacharia, 2013; Freeman, et al.
2014). This means that we need research in engineering education into what
mathematical modelling tasks can be activating, how these can be organised, how

175 [Link]:indrum2018:173961
students experience these tasks, what task characteristics create challenges, etc.
Moreover, studies with large groups are scarce; the review by Freeman et al. (2014)
shows that most studies on students’ activation are carried out with small or medium
size classes (up to 110 students). With more than 300 first-year students, we can add
to the research on how engineering students in large classes can be activated.
Sullivan et al. (2011) describe challenging tasks as requiring students to: plan their
approach, especially sequencing more than one step; process multiple pieces of
information, with an expectation that they make connections between those pieces,
and see concepts in new ways; choose their own strategies, goals, and level of
accessing the task; spend time on the task and record their reasoning; explain their
strategies and justify their thinking to the teacher and other students. We used a task
format that fits this description: a project-based task, which is a task that cannot be
completed within limited time, which has a clear, but not straight-forward goal, there
are various approaches to tackle it, and results must be presented through a product,
such as a written report or an oral presentation (Blomhøj & Kjeldsen, 2006).

THE TRACKER PROJECT TASK

Domínguez et al. (2015) did research with a group of 20 engineering students and
asked them: a child is throwing a candy to another; make a mathematical model of
this movement. This modelling task is an open-ended task with characteristics of ‘a
challenging’ task (Sullivan et al., 2011): students need to sequence more than one
step; process multiple pieces of information and connect the throwing and the model;
choose their own strategies, goals, and level of accessing the task; spend time on the
task. We adapted the task in the following way: (1) students could choose whatever
movement of whatever object: throwing a ball, jumping their skate board, etc.; (2)
students were asked to use their smart phones for filming, as nearly all students
nowadays have smart phones with high quality cameras; (3) students were asked to
download tracker software ( [Link] ), which captures motion in
videos based on contrasts and yields a table of time and position coordinates
(measurements). We made a tutorial video on the use of Tracker. The measurements
were to be mathematically modelled (i.e. create a formula that approximates the
movement). The required, final product was a poster, in which students presented
their reasoning – another characteristic of a ‘challenging task’ (Sullivan et al., 2011).
The poster had to contain the video analysis, including a discussion of the accuracy of
their mathematical model in comparison to the measurements. The task had to be
done in groups of two or three. Collaboration was convenient, because one student
alone cannot throw and film simultaneously. In our communication with the students
we indicated the task as the Tracker Project.
It was our first time to implement a project-based mathematical modelling task with
such a large group. Unlike earlier studies (e.g. Domínguez et al., 2015) we did neither
have a group of 20 students, nor uniform equipment, nor sufficient staff. We couldn’t
learn from earlier experiences, as – to our knowledge – there are no reports of similar

176 [Link]:indrum2018:173961
studies carried out with more than 300 students. The few studies on the activation of
students in large classes centre on using clickers in lectures (Freeman et al., 2014).
Thus, we didn’t immediately want to focus on students’ learning, but instead, first
study the feasibility of such a task with such a large group, with the variation of
cameras, and with students who have little experience with open-ended tasks. We felt
that we – as lecturers – should first take the opportunity to learn how it worked in
practice, whether students liked the task and how they engaged with it.
In this paper, we report on our research into the extent to which students’ were
activated by the modelling task. By activation, we mean – for the time being – that
the task grasped them and that they liked working on it. Thus, our study is on
students’ attitudes, which is an aspect of their affect. Based on Harris et al. (2015), we
expected the engineering students to have preconceived beliefs about mathematics,
and we wanted to avoid that our research would be contaminated by their biases.
Therefore, we undertook our research by limiting the use of the word mathematics in
our communication with students. Abundant use of the term mathematics could
trigger memories and bias of traditional mathematics education, which could interfere
with their evaluation of the Tracker Project Task.

THEORETICAL FRAME
Recent research in the field of mathematics education and affect conceptualize the
latter in terms of complex, dynamic systems and participatory environments (Pepin &
Roesken-Winter, 2015). However, while distinguishing between aspects of affect
(values, emotions, beliefs, attitudes, etc.), these researchers don’t differentiate
between aspects of mathematics education. Yet, mathematics education contains
many aspects, such as instruction formats, teacher attitudes, tasks, etc. These become
invisible when researchers address mathematics holistically and ask students to mark
their (dis-)agreement to statements such as: ‘mathematics is my favourite subject’.
(Dis-)agreement to such a statement gives little room for nuances and contexts. A
student partly agreeing with this item might rather have said: “mathematics with this
particular teacher is my favourite subject, but last year it was the opposite” or
“mathematics could be favourite, if it had relevance for my future”.
We wanted to study students’ affect through an activity that differed from standard
activities within traditional mathematics education. Thus, we sought an activity-based
conceptualisation of affect. An activity-based perspective in mathematics education
aligns with a socio-cultural perspective. One of its promoters, Lerman (2000),
describes mathematics as a socio-cultural practice embedded within a community.
Within a school institution, mathematics is a practice embedded in a community of a
teacher and a group of students, its rules, language, etc. The activities consist, among
others, of explanations by the teacher, and work on tasks by students. This practice
differs markedly from mathematics as a practice within a research community,
whereby the actors organize mathematical patterns, solve creatively a non-routine
problem by using mathematics, and actors may reach different answers. Describing

177 [Link]:indrum2018:173961
mathematics socio-culturally as a practice embedded within a community entails
focusing on the activities undertaken by the actors, which are mediated by language,
tools, etc. Using an activity-based conceptualisation of mathematics enabled us to
relate affect to distinct activities and not to mathematics holistically, whereby we
could distinguish mathematical activities as having different contexts. In our study
the activity was guided by the Tracker Project Task and students had to use
mathematics within a kinematics context.
We sought an activity-based conceptualisation of students’ affect with respect to
them being activated. Activation is an aspect of attitude, just like boredom or anxiety
(Pepin & Roesken-Winter, 2015). For this conceptualisation, we turned to a concept,
which describes “a state in which people are so involved in an activity that nothing
else seems to matter; the experience is so enjoyable that people will continue to do it
even at great cost, for the sheer sake of doing it” (Csíkszentmihályi, 1990, p.4).
Nakamura and Csikszentmihalyi (2009) describe how they observed rock climbers,
gamers, painters and researchers during their challenge, and how these people got
absorbed in their activities, felt happiness, forgot about time and basic needs (eating,
resting), and were intrinsically motivated (motivated by the activity itself, not by an
external incentive). They coined this state: flow.
Flow is an activity-based concept: without activity, there cannot be an experience of
flow. Flow is an experience of an individual, yet, the activity is culturally embedded
(e.g. gamers play a game created by others, painters expose their work). In fact, social
activities can intensify flow through group cohesion (group flow). We will use
students’ self-reported experience of flow as an operationalisation of their activation
through the Tracker Project task. Flow has also been studied in mathematics
education (a.o. Armstrong, 2008; Drakes, 2012; Liljedahl, 2016), observing that
many students in traditional mathematics classes is to not experience flow at all.
Figure 1 (left) illustrates how flow depends on the perceived challenge of a task and
perceived skills of a person engaging in the task (Nakamura & Csikszentmihalyi,
2009). If the activity is too challenging for the skills, then the task may cause anxiety.
If the activity is too easy for the skills, then the task may cause boredom. When
challenge and skills match, a person engaging in a task may experience flow. In later
work, Csíkszentmihályi and colleagues adapted the diagram, adding more affective
states, and stating that flow can be only experienced when a participant perceives the
task as more than averagely challenging, and that he/she thinks to have the skills that
match this challenge, see Figure 1 (right). The older diagram still appears in recent
studies (e.g. Liljedahl, 2016). Therefore, we opted to use our study to empirically
validate the old versus the new theory and see whether flow occurs only when the
actor perceives a more than average challenge. Our research was guided by two
questions. The first was empirical: To what extent did the Tracker Project Task make
students experience flow? The second was about the choice of flow diagram: Can

178 [Link]:indrum2018:173961
one of the Csíkszentmihályi diagrams of flow be confirmed by plotting skills,
challenge, and flow into one diagram?

Figure 1: Flow and other affective states related to task challenge and a person’s skills
(adapted from Nakamura & Csikszentmihalyi, 2009)

METHODS
In the Spring of 2017 we offered the Tracker Project task to all first-year students in
engineering at our university (Mechatronics, Electrical Eng., Data Eng., Renewable
Energy, ICT). There were 346 students for whom the task was mandatory.
The research design for studying students’ activation in terms of flow was a survey.
We collected data through a digital questionnaire. Participation in the survey was
voluntary, but encouraged with prizes of NOK 500 (approx $50) for three randomly
drawn participants. After removing seven participants (four had constantly chosen a 3
as answer, three were 2nd-year students for whom the task wasn’t mandatory), we had
n=239 students. This response rate of 69% is very high (Bryman, 2015).
Based on instruments from earlier research (Armstrong, 2008; Egbert, 2004), we
developed 15 items in alignment with the task. Each item consisted of a statement,
asking students for their (dis-)agreement on a 5-point Likert scale, from 1 (strongly
disagree) to 5 (strongly agree), see the Appendix. Five items were designed to
measure students’ perception of flow. For this, they could indicate, for example,
whether they forgot about the time, and whether they even would do the task if it
wasn’t obligatory. By having several items related to flow, a participant’s score is
indicator of the extent to which he/she had experienced flow. Five other items were
designed to measure students’ self-perceived skills (e.g. ‘the Tracker technology was
easy to use’ or (inverted) ‘It was complicated to find the right formula of the model’).
And a further five items were designed to measure students’ perception of the task’s
challenge (e.g. ‘during this task I started thinking about other movements (what if..?)’
and (inverted) ‘this task was more for secondary schools’).
We make a difference between flow as a concept (written in italics), and the scale of
Flow (with a capital letter). The concept of flow is a psychological state of a person,
and therefore it cannot be measured. However, we assume that it can be approximated
by a score on the scale of Flow. A student’s score on this scale results from his/her

179 [Link]:indrum2018:173961
answers to the five questions in our questionnaire. The score on the Flow scale is
calculated by adding the scores on the five questions. As the score on one question
ranges from 1–5, the score on the Flow scale ranges from 5–25. Likewise for
respectively, challenge and the Challenge scale, and skills and the Skills scale. To
increase reliability, within each scale one or two questions were inversely posed, and
the scoring was inverted, too. As measure of reliability (internal consistency), we
calculated Cronbach’s Alpha: the Skill scale yielded 0.55, the Challenge scale yielded
0.73, and the Flow scale yielded 0.63. A scale is considered unreliable if Cronbach’s
Alpha is less than 0.5 (Bryman, 2015). Thus, the three scales can be considered as
being reliable.

RESULTS
We observed students everywhere on campus, flying paper helicopters, riding
skateboards, or throwing apples, cats or balls. We received more than 100 posters in
our Virtual Learning System. As explained before, in this study we didn’t want to
focus on students’ performance (the precision of their measurements, their
understanding of modelling, the depth of their analysis, etc.). Instead, we focused on
the feasibility of an activating tasks for massive students groups, which would show
in their activation in terms of flow as measured through the questionnaire. Second, we
aimed at seeing whether the measurement reproduced one of the two flow diagrams.
The Appendix shows mean scores on all items.
The mean score on four items in the Flow scale is higher than 3.5, being well on the
positive side. This indicates that a majority of the students experienced a state of flow
to quite an extent, in particular with respect to losing track of time, and not being
easily distracted. Only item 14 was answered below the middle range. This item
focuses on doing the task even if at some costs (Csíkszentmihályi, 1990), which
translates in our study to: one out of four would even do the task voluntarily.

Scale mean (std dev)

Flow (5 items) 17.0 (3.1)
Challenge (5 items) 15.5 (3.3)
Skills (5 items) 18.8 (2.7)

Figure 2: Scores to Skills, Challenge, and Flow scales (n=239)

When adding the students’ scores on the five questions, we obtain their score on the
scale Flow. See Figure 2 (right) for a bar graph. This graph shows the frequencies of
scores (number of students with certain scores). The green bars of the Flow scale
show a skewed distribution. On this Flow scale, 31 students (13%) scored 13 points
or lower, 67 students (28%) scored in the middle range of 14–16 points, and 141

180 [Link]:indrum2018:173961
students (59%) scored 17 points or higher. When we take 17 points as a threshold,
then three out of five students experienced flow to quite an extent. The table in Figure
2 presents mean scores on the scales for Skills, Challenge and Flow (minimal score =
5, middle score range = 14–16, maximal score = 25). The scores on Skills are highest:
generally, students perceived themselves as highly skilled; the low standard deviation
indicates a high agreement among students. The scores on Challenge are around the
middle range; these scores are most “normal” (making a Gauss curve).
To validate the Csíkszentmihályi diagrams (Figure 1), we created a scatter diagram.
Each student was represented by a dot defined by his/her Skills score on the x-axis
and his/her Challenge score on the y-axis, see Figure 2. The resulting diagram shows
a scattered distribution, which means that there is no correlation at all between the
scales Challenge and Skills (r = 0.097). In this diagram, we added the third scale, the
one for Flow, by colouring the dots depending on the student’s Flow scores. These
scores range from red to orange (13 or lower), via yellow (middle range, 14–16) to
green (17 or higher). Roughly, one can discern overlapping red, yellow and green
areas. The red area is more visible at the bottom showing the students who
experienced little flow (13% of the students). These students indicated that the task
posed little challenge, independently of their perceived skills. The yellow area runs
from bottom right to the centre showing the students who experienced medium flow
(28% of the students). These students either indicated low challenge and high skills,
or medium challenge and medium skills. The green area is the largest with the
majority of students (59%). It is in the top-right, fading towards the centre, showing
the students who experienced flow to quite an extent. These students indicated that
they perceived the task as challenging, and they perceived themselves skilled.

Figure 3: Flow score indicated by color, as depending on Skills and Challenge

This colour distribution of Flow does not confirm the earlier Csíkszentmihályi
diagram (Fig 1, left), as the green dots do not centre on the diagonal. Instead, the

181 [Link]:indrum2018:173961
green dots are more to be found in the area where the later Csíkszentmihályi diagram
(Fig 1, right) situates flow: the task must be perceived as quite challenging, and this
challenge must match one’s skills. A majority of the students in our study indicate
that they perceived the Tracker Project Task as such.
In light of the different regions in the later Csíkszentmihályi diagram, we also see
many students who may fit into the affective states of ‘control’, ‘arousal’ and
‘relaxation’. Only few students may fit the more negative affective states of ‘apathy’,
‘boredom’, ‘anxiety’ and ‘worry’.

CONCLUSION, DISCUSSION, RECOMMENDATIONS

We studied whether a project-based task was feasible with a class of more than 300
students, that is: whether the task activated individual students. The first research
question asked: to what extent does the Tracker Project Task make students
experience flow? The results from the survey showed that a majority of the students
(59%) experienced flow to quite an extent, forgot about time and wanted more of
such activities. This result was confirmed by anecdotal evidence of their boasting
stories in the tutorials of them throwing objects, and the high response rate to the
survey. This means that the Tracker Project Task activated a majority of the students
and that they had positive attitudes towards it. Thus, an activating mathematics task
can be feasible with a large class of engineering students, even if they are known to
have a negative stance towards mathematics (Harris et al., 2015).
The Tracker Project Task was designed to be challenging with characteristics such as:
expecting students to process multiple pieces of information, that they make
connections between those pieces, choose their own strategies, and explain their
strategies to others (Blomhøj & Kjeldsen, 2006; Sullivan et al., 2001). We observed
other characteristics in the Tracker Project Task that activated the students. First, the
task had a clear goal, which was understandable to all students. We observed this
through the few questions that we got from the students on how to carry out the task.
Thus, the task was easily accessible, also known as having a low floor. Second, the
better students were able to challenge themselves further, allowing for a high ceiling.
Third, the use of readily-available technology (cameras in smart phones, tracker
software) may have captivated the engineering students, who are known to be
technology minded. Fourth, the task was a mathematical modelling task embedded in
engineering practices, whereby mathematics served non-mathematical purposes; this
showed students the relevance of mathematics to their studies, and contrasted with
bare mathematics tasks that alienate and demotivate students.
Our second research question pertained to the theory of flow and how it can be
conceptualized in a diagram (Nakamura & Csíkszentmihályi, 2009). Our data reject
the earlier theory that flow depends on the alignment of skills and challenge. Instead,
our data support the later theory that flow occurs when the participants perceive the
task more than average challenging, and that their skills should match this challenge.

182 [Link]:indrum2018:173961
Furthermore, we take from our study that the concept of flow proved useful for
activity-based research on affect in mathematics education.

REFERENCES
Alpers, B. A., et al. (2013). A Framework for Mathematics Curricula in Engineering
Education. Brussels: European Society for Engineering Education (SEFI).
Armstrong, A. (2008). The fragility of group flow. Journal of Mathematical
Behavior, 27(2), 101–115.
Blomhøj, M., & Kjeldsen, T. H. (2006). Teaching mathematical modelling through
project work. ZDM Mathematics Education, 38(2), 163–177.
Bryman, A. (2015). Social research methods. Oxford, UK: Oxford University Press.
Csíkszentmihályi, M. (1990). Flow: The psychology of optimal experience. New
York: Harper & Row.
De Jong, T., Linn, M. C., & Zacharia, Z. C. (2013). Physical and virtual laboratories
in science and engineering education. Science, 340(6130), 305–308.
Domínguez, A., de la Garza, J., & Zavala, G. (2015). Models and modelling in an
integrated physics and mathematics course. In Mathematical Modelling in
Education Research and Practice (pp. 513–522). New York: Springer.
Drakes, C. I. (2012). Mathematical Modelling: from Novice to Expert. Unpublished
doctoral dissertation. Burnaby, Canada: Simon Fraser University.
Egbert, J. (2004). A study of flow theory in the foreign language classroom.
Canadian Modern Language Review, 60(5), 549–586.
Freeman, S., Eddy, S. L., McDonough, M., Smith, M. K., Okoroafor, N., Jordt, H., &
Wenderoth, M. P. (2014). Active learning increases student performance in
science, engineering, and mathematics. Proceedings of the National Academy of
Sciences, 111(23), 8410–8415.
Harris, D., Black, L., Hernandez-Martinez, P., Pepin, B., & Williams, J. (2015).
Mathematics and its value for engineering students: what are the implications for
teaching? International Journal of Mathematical Education in Science and
Technology, 46(3), 321–336.
Lerman, S. (2000). The social turn in mathematics education research. In J. Boaler
(Ed.), Multiple perspectives on mathematics teaching and learning (pp. 19–44).
Westport: Greenwood Publishing Group.
Liljedahl, P. (2016). Flow: A framework for discussing teaching. In Proceedings of
the 40th Conference of the International Group for the Psychology of Mathematics
Education, Vol. 3 (pp. 203–210). Szeged, Hungary: PME.

183 [Link]:indrum2018:173961
Nakamura, J., & Csikszentmihalyi, M. (2009). Flow theory and research. In S. J.
Lopez & C.R. Snyder (Eds.), Oxford Handbook of Positive Psychology (pp. 195-
206). Oxford, UK: Oxford University Press.
Sullivan, P. et al. (2011). Challenging mathematics tasks: What they are and how to
use them. In L. Bragg (Ed.), Maths is multi-dimensional (pp. 33–46). Melbourne:
Mathematical Association of Victoria.

APPENDIX
Mean scores on all items (1=lowest, 3=middle, 5=highest).
Flow questions mean (std dev)
(Inv) This Tracker task took too much of my time 3.67 (0.88)
Time was flying when we worked in this task. 3.40 (0.92)
(Inv) I was easily distracted when we worked on this task. 3.55 (0.91)
I would do this task even if it wasn’t obligatory. 2.60 (1.13)
I would like to have more of such practical tasks. 3.70 (1.02)

Skills questions mean (std dev)

The Tracker technology was easy to use. 3.89 (0.88)
(Inv) It was complicated to find the right formula of the model. 3.38 (0.92)
The aims of the task were clear to me. 3.96 (0.91)
During this task I had full control over what we did. 3.77 (1.13)
Filming the movement of an object was easy. 3.76 (1.02)

Challenge questions mean (std dev)

This “Modelling med Tracker” task made me curious. 3.61 (0.75)
Making a poster made me feel like a “real scientist”. 2.52 (1.03)
(Inv) This task is more suitable for secondary schools. 2.58 (0.95)
This task helped me to better understand the theory. 3.39 (0.88)
During this task I started thinking about modelling other 3.31 (1.12)
movements (what if..?).

184 [Link]:indrum2018:173961
 Using schematic representation of resource systems to examine how
first year engineering students use resources in their studies of
mathematics
Eivind Hillesund
University of Agder, Norway, [Link]@[Link]
Very little research has been done on how students use resources when studying
mathematics. My project aims to examine this both quantitatively and qualitatively.
The qualitative data collection includes hierarchical focus interviews, with
schematic representation of resource systems as a supplement. The intent is to
reduce the degree of co-producing answers and imposing terminology on the
students.
Keywords: Students’ practices at university level, the role of digital and other
resources in mathematics education, mathematics for engineers, documentational
approach, hierarchical focus interviews, schematic representation of resource
systems.

INTRODUCTION
The focus of my PhD-project is engineering students’ use of resources for learning
mathematics. Data collection takes place at three Norwegian universities. I focus on
which resources they use, to what extent and in what situations, as well as their
rationale for how they use resources.

THEORETICAL FRAMEWORK
I use the documentational approach (Gueudet & Trouche, 2009) to examine those
questions. Relevant to the approach is the term ‘document’ which is the joint entity
of a set of resources and schemes for utilizing them in certain situations to achieve
certain goals. Also relevant is the term ‘resource system’ for the set of all resources
an individual is using, structurally organized. For instance, what resources to use in
what situation can be part of the structure. The framework is designed to examine
teachers’ practices and professional development, but can be adapted to examine
students’ use of resources as well. One of the focuses of the framework is how
students’ documents develop (called ‘documentational genesis’). I look at students’
practices during their first year of university, when I expected a lot of development
to occur.

METHODOLOGY
I use both qualitative and quantitative means for data collection. I do so because I
want to study resource with some depth, while also getting an indication of the
variety of uses. Here, I will focus on qualitative data collection. I used hierarchical
focus interviews (Tomlinson, 1989). These contain strategies to reduce the degree to

185 [Link]:indrum2018:173373
which the interviewer co-produces the answers. Prior to the first interview (of three),
I asked each student to draw a ‘mind map’ about their use of resources in
mathematics. I used their mind maps to create schematic representations of their
resource systems (SRRS), inspired by Pepin, Xu, Trouche and Wang (2016). From a
pilot interview in spring 2017, I theorized that the construction of a mind map helped
students structure their thoughts about using resources prior to the interview.

RESULTS
Qualitative data collection spanned the fall semester of 2017, with nine students from
three different universities. All students created a mind map during the first
interview. In the other 18 interviews, they made changes to their mind maps a total of
eight times. The students seemed comfortable talking about their use of resources
after constructing a mind map. Their descriptions of their mind maps also yielded
interesting insight into how they perceived their use of resources.
The students structured their mind map in several ways. They all contained
resources, but five students also had categories in their mind maps, three had
situations, five had what purpose they used certain resources for and one had features
they appreciated about certain resources. The resources that the most students put in
their mind maps were the textbook (nine), fellow students (seven), lectures (six),
exercises (six), lecturer (five), pencil and paper (five), Wolfram Alpha (four),
calculator (three) and google (three).

CONCLUSION
It is difficult to discern whether differences in students’ mind map structure only
represent stylistic choices, or meaningful differences in how they perceive their use
of resources. While a student’s mind map can say much about their documents,
analysing the mind map in a vacuum is insufficient. However, when mind map
construction is combined with an interview, the two forms of data may shed some
light on one another. The interviews may also benefit from the construction of the
mind map, as it gives students some time to consider their use of resources prior to
the interview.

REFERENCES
Gueudet, G., & Trouche, L. (2009). Towards new documentation systems for
mathematics teachers?. Educational Studies in Mathematics, 71(3), 199-218.
Pepin, B., Xu, B., Trouche, L., & Wang, C. (2016). Developing a deeper
understanding of mathematics teaching expertise: an examination of three Chinese
mathematics teachers’ resource systems as windows into their work and
expertise. Educational Studies in Mathematics, 1-18.
Tomlinson, P. (1989). Having it both ways: hierarchical focusing as research
interview method. British Educational Research Journal, 15(2), 155-176.

186 [Link]:indrum2018:173373
 Mathematics Teaching for Economics Students, But How?
Ida Landgärds
University of Agder, Faculty of Engineering and Science, Norway,
[Link]@[Link]
The research presented in this poster addresses the poor performance of many
economics students in their first mathematics course at university level. Two different
universities are involved in the research, trying to answer the question of “how to
structure teaching in mathematics to economics students at university level to
strengthen their mathematical competences?”
Keywords: Teaching and learning of mathematics in other fields, curricular and
institutional issues concerning the teaching of mathematics at university level.
How to structure teaching to improve economics students’ mathematical
competences? Research papers within economics are often heavily mathematical.
Mathematics is an indispensable tool in studying economics and in the economist’s
working day. The Norwegian Association of Higher Education Institutions (2011)
therefore stipulates that mathematics should be a useful tool for students in the
learning of other subject areas within economics.
Research has proved that mathematical knowledge in algebra and arithmetic is a good
indicator for performance in introductory economics courses (e.g. Ballard & Johnson,
2004). However, many universities are struggling to find satisfactory formats for
teaching mathematics as a service discipline, particularly in their economics
undergraduate degree courses.
At the University of Agder there is a first year, first semester course in mathematics
for economics students. The proportion of students failing this course, has for several
years been about 40%. This is an alarming issue and I have experienced similar
problems at Åbo Akademi University in Finland, the university where I studied for
my master’s degree. Currently these two universities have taken opposite directions
regarding the teaching of mathematics to economics students.
Both universities face the issue of students’ diverse mathematical backgrounds. At
the Åbo Akademi University the course in mathematics for economics students has
been moved to the first semester, to more naturally be a continuation of school
mathematics. At the University of Agder, a diagnostic test, compulsory for all
economics students, is being implemented. The test is followed by a preparatory
mathematics module; although not compulsory it is recommended that students with
weak mathematical background take it, prior to the main course in mathematics. The
preparatory module will consist of online, self-study material with a clear structure,
implemented in the Canvas digital learning environment with 3-4 workshops.
The overall goal of the research is to improve the teaching of mathematics for
economics students and to optimize their learning. This poster is about the first stage

187 [Link]:indrum2018:174025
in the research process and considers productive opportunities for research on the
diagnostic test and the online preparatory mathematics module. Students studying and
participating in activities online produce a huge amount of data which gives the
opportunity to use Learning Analytics (LA) as a research method to better understand
the mathematical needs of economics students and their use of mathematical
resources. e a-Ayala (2017, p. 6) writes: “LA in the context of higher education is
an appropriate tool for reflecting the learning behaviour of students and provide
suitable assistance from teachers or tutors.” I thus want to find out what factors in the
preparatory mathematics module contribute or do not contribute to the students’
progress in learning mathematics. Because of the preparatory module being optional
there is also the possibility to find out amongst students recommended to make use of
the preparatory mathematics module, did those who took part in the module perform
better in the main mathematics course examination than those who did not take part?
At this stage, there are no concrete results to report. Data will be collected and
analysed in autumn 2018. The theoretical framework to be used is under
consideration. The theoretical framework will underpin the formulation of research
questions. Clow (2013) argues for Learning Analytics being atheoretical. There are
limited studies linking learning theory to learning analytics, but for example
Macfadyen and Dawson (2010) acknowledge the importance of effective student
centred learning and mention, as an example of the social aspect for learning, the
possibility of learner-to-learner communication within digital learning environments,
and thus they look into the socio-constructivist paradigm.
At the time of the conference I hope to be more knowledgeable about what
theoretical framework could guide the proposed research. It will be a valuable
opportunity to discuss the theoretical framework with more experienced scholars in
the field during the conference.
As e a-Ayala (2017, p. 68) writes: “LA aims at developing models, methods, and
tools that can be widely used, whose deliverables are reliable and valid at a scale
beyond a course or cohort to provide benefits for learners and educators without
distracting or misleading them” I hope the proposed research will provide new
knowledge about teaching of mathematics to economics students.
REFERENCES
Ballard, C. L., & Johnson, M. F. (2004). Basic Math Skills and Performance in an Introductory Economics
Class. The Journal of Economic Education, 35(1), 3-23. doi:10.3200/JECE.35.1.3-23
Clow, D. (2013). An overview of learning analytics. Teaching in Higher Education, 18(6), 683-695.
doi:10.1080/13562517.2013.827653
Macfadyen, L. ., & Dawson, S. (2010). Mining LMS data to develop an “early warning system” for
educators: A proof of concept. Computers & Education, 54(2), 588-599.
doi:[Link]
e a-Ayala, A. (2017). Learning analytics : fundaments, applications, and trends : a view of the current
state of the art of enhance e-Learning. In Studies in Systems, Decision and Control, Vol. Volume 94.
The Norwegian Association of Higher Education Institutions. (2011). Plan for the Bachelor's Degree in
Business Administration. In. Retrieved from
[Link]

188 [Link]:indrum2018:174025
Non-standard Problems in an Ordinary Differential Equations Course
Svitlana Rogovchenko1, Stephanie Treffert-Thomas2, and Yuriy Rogovchenko1.
1
University of Agder, Faculty of Engineering and Science, Norway,
[Link]@[Link]; 2Loughborough University, School of Science, UK.

We report first results from a teaching intervention in an ordinary differential

equations (ODEs) course for engineering students. Our aim is to challenge traditional
approaches to teaching of Existence and Uniqueness Theorems (EUTs) through the
design of problems that students cannot solve by applying well-rehearsed techniques
or familiar methods. We analyse how the use of non-standard problems contributes to
the development of students’ conceptual understanding of EUTs and ODEs.
Keywords: existence and uniqueness theorems, design research, non-standard
problems, commognitive theory, mathematical discourse.
INTRODUCTION AND BACKGROUND TO THE STUDY
Although ODEs are an important topic in the engineering curriculum, students
experience difficulties with mastering ODEs and with the very concept of a differential
equation (Arslan, 2010). In our study, the lecturer, a mathematician, devised a set of
non-standard problems (see below, Problem 1 of 6) to challenge students’ conceptual
understanding of the EUTs. These problems formed an assessed piece of coursework.

Figure 1. One of the problems in the study

We analysed how solutions changed and developed as students worked on the
problems. Students' discussions in small groups were audio-recorded, transcribed, and
then analysed using constructs from commognitive theory (Sfard, 2008). We are
currently in the initial phase of the data analysis aimed at answering the following
question: How do non-standard problems contribute to the development of students’
mathematical discourse and further their conceptual understanding of fundamental
notions and results in an ODE course?
RESULTS
For Problem 1a (P1a), students could use one of two solution methods: M1
(substitution) and M2 (integration). Working on the problems, several students
changed their approach. In the final script, only one student produced a correct and

189 [Link]:indrum2018:174855
complete solution (M2) while 14 (of 19) students used M1 verifying that a given
function is a solution (which is sufficient for the particular solution), but failed to
explain why this solution is the general one (hence incomplete M1). We conducted
similar analyses for Problem 1b (P1b).
We present one extract (for P1b) as an example of our analyses of students' group
discussions using commognitive constructs - narrative, routine, ritual, substantiation.
S12. The first idea was just to try to solve for C and I got the same constant, so that’s
OK. And I checked for asymptotes and I got one on x=0, so I noted that the equation
is split to get two curves, at least, according to calculator we got it split about zero.
S11. So it’s undefined at zero.
S12. Undefined at zero, so we get two different curves and both solutions work. We do
not have a continuous curve which happens to intersect at these two points […]
S14. It’s not continuous for 𝑥 = 0?
S12. No. So if we take an interval from -3 to 1, it’s discontinuous in this interval, so it’s
not a curve that happens to just hit these two points, it is two individual curves that
have the same solution. So it’s correct in just a tiny area.
S11. That was my argument as well. As the theorem states, there is a continuous
interval but here it is split into two which contain two different t0’s.
S13. The theorem says, that there is a unique solution for every interval where the
function is continuous. Since there are two intervals and there are two solutions, it
does not conflict with the theorem. […]

Note that S12 is using two different visual realizations of solutions, first the algebraic
representation and then the graph plotted by calculator. He shows that the realizations
are not equivalent, they do not produce the same result. We see how the student
demonstrates the ability to solve the problem by developing the realization tree and
employing the mathematical object of “continuous solution” (discursive object). S11
is not so sure at the beginning, he is guided by S12 (considered “more experienced”
interlocutor) and adopts the narrative offered by S12. S13 concludes by reformulating
the expression “does not violate the theorem” as “does not conflict with the theorem”.
We see how students worked to substantiate the narrative. This routine can be
characterized as the exploration. Students gradually improved their abilities in
developing and endorsing the EUTs narratives while working on all six tasks during
the group discussions.
REFERENCES
Arslan, S. (2010). Do students really understand what an ordinary differential equation
is? International Journal of Mathematical Education in Science and Technology, 41,
873-888.
Sfard, A. (2008). Thinking as communicating: human development, the growth of
discourses, and mathematizing. Cambridge: Cambridge University Press.


190 [Link]:indrum2018:174855
 Students’ understanding of mathematical models in three Norwegian
biology courses
Floridona Tetaj
University of Agder, Faculty of Engineering and Science, Kristiansand
[Link]@[Link]
Keywords: mathematical models, mathematical understanding, biology students
INTRODUCTION
Development in Biology over the last three decades have greatly increased the
need for using mathematics in this field. This has influenced biology students
which need to develop mathematical knowledge to be able to work with
contemporary biology models, including frameworks that are applicable in
analyzing the overwhelming flow of biological data (i.e. Labov, Reid &
Yamamoto, 2010). Although a quantitative approach is often used in university
biology courses, yet they remain largely qualitative and descriptive (Nelson, et. al.,
2009). In Norwegian context, where this study will be conducted, there is an
increased awareness towards mathematics from the Departments of Biology.
However, there is a lack of information on students’ understanding and usage of
mathematics within this context. This poster addresses this issue with a particular
focus on the use of mathematical models (MM) as dynamic tools that allows us to
observe various aspects of students’ understanding.
UNDERSTANDING OF MATHEMATICAL MODELS
Mathematical understanding has been the scope of many researchers. Sfard (1994)
distinguish between operational and structural way of understanding. While the
operational understanding includes highly manipulative skills and use them as
principal means in their quest after meaning, the structuralist is more capable of
direct-grasp understanding. She defines as reification the transition from an
operational to a structural way of thinking, and states that this transition is a basic
phenomenon in the formation of a mathematical object.
According to Niss (2012), a mathematical model can be defined as mapping
(translation), f, from a mathematical domain, D, to a mathematical realm, M. In this
context, D and M, represent not only sets of objects but also collections of
relationships, phenomena or questions, while f operates on objects and the
relationships, phenomena or questions. It is important to point out the distinction

191 [Link]:indrum2018:174802
between mathematical model and mathematical modeling. While mathematical
model is a tool for facilitating quantification, analyses, predictions and gaining
insight for a real-world situation, mathematical modeling is the process of the
creation of such tool (GAIMME, 2016). In this study, I focus only on the use of
mathematical models as a tool that allows one to observe different shades of
students’ mathematical understanding, and not on the process of mathematical
modeling. Observing biology students while engaging in biology problems that use
mathematical models as their representation, I aim to describe their way of
thinking and reasoning – as operational or structural thinking.
METHODOLOGY
This study is a descriptive case-study in a naturalistic paradigm. For the pilot phase
of this study, have been selected three different biology courses from a Norwegian
university (two courses of bachelor level and one in master level). These courses
have been selected considering their use of mathematical models. In these courses,
mathematics is implicitly presented using mathematical models (i.e., population
dynamic models). All students taking these courses have taken previously at least
one mathematical course (mostly, Calculus course or Statistic course).
For collection of the data, I plan to video-record sessions in these courses when
students are engaged in group-work. I will use semi-structured interviews with
some of the students after some sessions, and personal written notes.

Bibliography
Applications, C. f. (2016). GAIMME Guidelines for Assessment & instruction in
mathematical modeling education. Bedford, MA 01730 USA: COMAP.
Labov, J. B., Reid, A. H., & Yamamoto, K. R. (2010). Integrated biology and
undergraduate science education: A new biology education for twenty-first
century? CBE- Life Sciences Education, 9, 10-16.
Nelson, K., Marbach-Ad, G., Schneider, K., Thompson, K., Shields, P., & Fagan,
W. (2009). MatheBench Biology Modules, Web-Based for All Biology
Undergraduates. Journal of College Science Teaching, 34-39.
Niss, M. (2012). Models and Modelling in Mathematics Education. EMS
Newletter, pp. 49-52.
Sfard, A. (1994). Reification as the Birth of Metaphor. For the Learning of
Mathematics, 14(1), 44-55.

192 [Link]:indrum2018:174802
TWG 3: Number, Algebra, Logic
Contemporary Research Practices in Discrete Mathematics - A way to
enrich the understanding of Discrete Mathematics at University Level
Elise Abdallah
University of Reims Champagne-Ardenne (URCA) - CEREP- EA 4692- France
Lebanese University- EDLS- Lebanon
elizeabdallah@[Link]
This paper is part of a thesis about discrete mathematics and its teaching in higher
education. The literature on the didactics of discrete mathematics questions this
branch at different levels: its integration in teaching, the particularity of its affective
dimension, and its epistemological specificities especially in the fields of proof and
modeling. We seek to epistemologically define this field and to characterize its
corresponding mathematical activity by studying the processes of knowledge
construction, the types of problems, the specificity of concepts and proofs, and also
the existing links between discrete mathematics and other disciplines. This
epistemological study has a didactic purpose of defining and analyzing the teaching
of discrete mathematics in higher education.
Keywords: teaching and learning of number theory and discrete mathematics,
teaching and learning of logic and proof, higher education, functional definition,
epistemology.
INTRODUCTION AND CONTEXT
Research on discrete mathematics has rapidly developed in its methodologies, in the
way it is viewed by mathematicians, and in its range of applications. Discrete
mathematics has been described by TSG-17 Teaching and Learning of Discrete
Mathematics at the ICME-13 (2016) as a comparatively young branch of
mathematics with no agreed-on definition but having old roots and emblematic
problems. Moreover, it is a robust field with applications to a variety of real world
situations, and of on growing importance to contemporary society (Hart & Sandefur,
in press). Over the past several decades, discrete mathematics has proved to be an
important part of the recommended program for students of computer science
(Maurer, 1997; DeBellis & Rosenstein, 2004; Grenier & Payan, 1998; Borwein,
2009; Epp, 2016; Rosenstein, 2016). Epp (2016) points out the strong necessity for
engaging students in abstract thinking for the course of discrete mathematics and its
applications in computer science. Discrete mathematics also seems to be a very
important tool for research in biology and chemistry. On the other hand, discrete
mathematics has been influenced by a variety of mathematical results, methods, and
representations (group theory, number theory, geometry, algebraic combinatorics,
graph theory, and cryptography). Their integration and combination in a profound
theory is essential for research in discrete mathematics (Heinze, Anderson, & Reiss,
2004). A recent publication that looks into the future of mathematics, The
Mathematical Sciences in 2025 (Committee on the Mathematical Sciences in 2025,

194 [Link]:indrum2018:174690
2013) identifies two new drivers of mathematics: computation and big data. For both,
it describes how discrete mathematics plays an important role like discrete
mathematics algorithms for mathematical processing, dynamical systems in ecology,
networks in industry and the humanities, and discrete optimization (p.77).
The growing importance of discrete mathematics leads us to define this field for an
educational purpose. We seek to develop a “functional definition” [1] of discrete
mathematics, in order to use it to analyze and design didactical situations.
Specifically, we are concerned with the university level. We first present the state of
art in the teaching and the learning of discrete mathematics mainly at secondary level
pointing out some of its epistemological aspects. We then state our research questions
and describe the methodology aimed at developing a “functional definition” of
discrete mathematics. Our research is inscribed in a “contemporary epistemology” [2]
that draws on interviews with mathematicians. Our working hypothesis is that such
interviews can update and enrich our functional definition. Finally, we discuss some
preliminary results of the interviews with mathematicians and close with some
concluding remarks.
TEACHING AND LEARNING OF DISCRETE MATHEMATICS-
SUMMARY OF A STATE OF ART
This young field of mathematics with numerous interconnections has no agreed-on
definition shared by mathematicians (Maurer, 1997; Hart & Martin, 2016) and has
blossomed in several directions. There exist different attempts to define discrete
mathematics, by mathematicians (like in the United States) and by mathematics
educators (like in France). These attempts depend on the epistemological posture of
the authors and on the intended function of the definition (e.g. to enable
mathematicians to define a field, mathematics educators to characterize a domain,
and teachers to present a topic of mathematics…etc.). For example, in an attempt to
define discrete mathematics, a mathematician proposed two standard approaches
toward this definition (Maurer, 1997): by specifying properties or by lists of topics.
The defining lists are too many (courses aiming for computer science majors,
algorithm-oriented course, finite mathematics course for social science and business
majors, high school course) (Maurer, 1997). Mathematics educators have also
proposed a definition of discrete mathematics such as the following:
“The main idea is that discrete mathematics is the study of mathematical structures that
are “discrete” in contrast with “continuous” ones. Discrete structures are configurations
that can be characterized with a finite or countable set of relations” (Ouvrier-Buffet,
2014, p. 181).
Moreover, discrete mathematics acquires particular objects and methods (Grenier &
Payan, 1998). However, these attempts to define discrete mathematics are not all-
inclusive as they overlook many characteristics of the concepts and proofs involved
in this field. In our opinion, what is also important for the didactics of mathematics is
to uncover the specificities of this field of mathematics in comparison to others.

195 [Link]:indrum2018:174690
Recent research in discrete mathematics, computer science, and mathematics
education has led to a serious discussion of the principles of proof, the teaching and
learning of proof, the validity of computer-based proofs or of visual proofs etc. The
distinction between the terms reasoning, proving, augmenting, demonstrating, and the
complex relationship between argumentation and demonstration (argumentation
considered as an epistemological obstacle for the learning of proof), calls for a debate
and an in-depth analysis (Balacheff, 1987; 1999; Reid & Knipping, 2010). In their
book, Proofs in Mathematics Education, Reid and Knipping included several
examples in their discussions, and there might be a reason that a large number of
these examples come from discrete mathematics (Reid & Knipping, 2010). In the
special issue of ZDM (2004) and in the ICME 13 monograph (2016), discrete
mathematics continues to be promoted as the essential mathematics in a 21st century
school curriculum. Its power lies in the opportunity it provides for supporting
reasoning, problem solving, modeling, and systematic thinking in the school
curriculum. Besides, recursion and recursive thinking seem to be powerful modeling
and problem solving strategies throughout mathematics in general and in the teaching
and learning of discrete mathematics in particular. The latter has been highlighted in
the studies of part III of ICME 13 monograph entitled recursion and recursive
thinking. They describe the integration of recursive thinking with iterative as well as
algebraic thinking, and they present the benefits of this integration as means to
deepen the students understanding of each of the geometry of transformations and
covariation of variables.
Some epistemological aspects of discrete mathematics pointed out in didactics
Researchers in didactics of discrete mathematics have proposed several
characteristics, of epistemological nature, of discrete mathematics. These
characteristics are the result of their research aiming at investigating the place and
role of discrete mathematics in education, analyzing the teaching and learning
situations, integrating new content into the curricula, studying the place and role of
proof in the curricula, and examining the mathematical expression (symbolic and
visual) and the use of language. Accordingly, several aspects have revealed such as:
problems in discrete mathematics encourage the development of heuristic and
affective processes (Goldin, 2016), there exists a specific relationship between
discrete mathematics and proof-existence of different situations that provide different
views on proof (Grenier & Payan, 1998), there exist different models in discrete
mathematics which necessitates the work on modeling (Grenier & Payan, 1998),
discrete objects and situations are easily accessible (Grenier & Payan, 1998; Maurer,
1997; DeBellis & Rosenstein, 2004), there exist different definitions of different
natures for discrete objects (Grenier & Payan, 1998; Maurer, 1997; Ouvrier-Buffet,
2011; 2006; Balacheff, 1987), and the fact that examples from discrete mathematics
enhance the semantic development of mathematical concepts and proving skills
(Alcock, 2009). Discrete mathematics provide the opportunity to develop students
reasoning ability, communication skills, problem solving ability, and modeling skills,

196 [Link]:indrum2018:174690
as well as mathematical habits of the mind that are specifically cultivated by studying
discrete mathematics such as algorithmic problem solving, combinatorial reasoning,
and recursive thinking. In short, as Hart & Martin (2016) say, discrete mathematics is
empirically powerful as a tool to enhance modeling and solving fundamental
contemporary problems, and it is pedagogically powerful in that it can be used in the
curriculum to simultaneously address content, process, and affect goals of
mathematics education.
RESEARCH AIM AND RESEARCH QUESTIONS
The importance of discrete mathematics in both research and in education has been
highly marked and extensively studied in the literature. However, the inclusion of
discrete mathematics in school curricula faces challenges worldwide. There are
countries like Hungary and Germany in which discrete mathematics has been taught
since a long time and as early as primary years of school. In France, the recent
introduction of graph theory for grade 12 classes of specialty “ES” (economy and
social) represents an official entry of discrete mathematics into the classrooms, yet
this integration is still far from that of other European countries. In the United States,
since 2000 discrete mathematics had been integrated into the curricula such as
“combinatorics, iteration, and recursion, and vertex-edge graphs…” as mathematical
topics at school level (K-12) (NCTM, 2000, p. 31). Yet, the new Common Core State
Standards for mathematics that were developed in 2009 and adopted soon afterwards
by most of the states in the United States excluded discrete mathematics (Rosenstein,
2016). Rosenstein explains in his paper that the reasons for this exclusion are: (1) the
shift in focus from college-readiness to calculus-readiness, (2) the desire to expand
the STEM pipeline by ensuring that students take more calculus at secondary level,
and (3) the concerns for international assessments. He calls out the international
mathematical education community to have an active role in introducing discrete
mathematics into the curricula of their countries’ schools by developing their own
curriculum material to promote a broader curriculum. However, although discrete
mathematics is taught in a shy manner in some countries, this does not mark the
existence of didactics of discrete mathematics, as a well-structured branch of
mathematics in the same way there exists the didactics of algebra, calculus or
geometry. Discrete mathematics exists at the frontiers with other fields like computer
science. Hence, the teaching of discrete mathematics constitutes a challenge (a
complex choice of topics with a high demand for instruction). We believe that proof
processes of discrete mathematics are abundant, diverse, and particular, and we aim
at exploring this aspect and its connection with other mathematical domains.
The literature led to the following research questions: how can we define
“functionally” discrete mathematics (that is how can we describe its epistemological
aspects, the links between discrete mathematics and other domains, and what are the
most recurrent types of problems that arise), and how can we describe the teaching of
discrete mathematics at university level. Their treatment, based on the “contemporary
epistemology”, will contribute to the delimitation of the field of discrete mathematics,

197 [Link]:indrum2018:174690
hence an objective of our study. In particular, this treatment will update and enrich
the conceptions [3] of mathematics educators about discrete mathematics and lead to
the development of teaching and learning situations.
Towards a Functional Framework
Therefore, our research aims at further investigating the above questions and
exploring the reality of the teaching and learning of discrete mathematics. Our
objective is to develop a “functional framework” for discrete mathematics in order to
conduct didactical studies of discrete mathematics. In this way, a “functional
definition” of discrete mathematics will have two main functions: (1) to delimit the
mathematical domain of discrete mathematics (epistemological level) and (2) to open
new horizons for the integration of this field in teaching (didactic level).
The epistemological aspects of this framework are a very important asset and often
not taken into consideration explicitly by university teachers. Indeed, Artigue (2016)
claims the existence of a disconnection between the mathematician’s experience as
researchers and their experience as teachers. This might be caused by the absence of
the epistemological dimension in their work as educators. The importance of
developing these epistemological aspects is linked to the following characteristics as
stated by Radford (2016), quoting Artigue (2016): (1) epistemology allows the
reflection on the manner in which objects of knowledge appear in the school practice,
(2) epistemology offers means through which we understand the formation of
knowledge (historical production and social production), and (3) epistemology allows
the reflection on the notion of epistemological obstacle. Accordingly, this first
function of our “functional framework” concerns the delimitation of the field of
discrete mathematics, by its contents, its types of problems, and to highlight the
specificity of the work on proof in relation to other mathematical domains. The place
and role of modeling in discrete mathematics will also be investigated. As discrete
mathematics interacts with other mathematical fields, we will also need to
characterize the links between discrete mathematics and arithmetic, number theory,
algebra among others. Moreover, since the epistemological definition of discrete
mathematics is linked to that of computer science (via the problems of counting and
combinatorics among others), we will be specifying the links and interactions
between these two scientific domains, explicitly relying on the “contemporary
epistemology”, i.e. the current problems and interactions between discrete
mathematics and computer science. Finally, we will integrate into our definition a
strong didactical perspective by studying the place and the role of discrete
mathematics in the articulation between secondary and university education
(particularly between university education and teacher training). We are also
interested in investigating the process of evaluation conducted at the university level
of the concepts and procedures proper to discrete mathematics. Ultimately, our
purpose is to be able to make coherent epistemological propositions for the teaching
of discrete mathematics at a given level.

198 [Link]:indrum2018:174690
RESEARCH METHODOLOGY
Our research methodology to address our first research question, which is “how to
characterize discrete mathematics at the epistemological level?” is based on a
contemporary epistemology relying on the experiences of researchers in the field of
discrete mathematics who are also instructors of discrete mathematics at the
university level. Our approach is inspired by several previous work relying on
interviews with mathematicians and mathematics educators such as Nardi (2008). We
will also base our work on the notion of praxeology of Chevallard, particularly
sequences of praxeologies, for the elaboration of our framework in order to describe,
analyze and structure specific contents at the heart of the teaching and learning
process. The work of Hausberger (2017) on structuralist praxeology in Abstract
Algebra could be an inspiring example. He uses a historical and epistemological
study of structuralist thinking and practices combined with a study of few textbooks
to develop his notion (Hausberger, 2017). In our study, we will be considering the
choice of particular emblematical textbooks of discrete mathematics at university
level along with the interviews to study the teaching practices.
Our study is an exploratory one in which we will conduct interviews with the
researchers aiming at reinterpreting the literature findings, investigating the
coherence between the literature and teacher practices, and identifying other
epistemological aspects. We have conducted interviews with instructors of discrete
mathematics at each of the Lebanese University and the Mathematical Society in
France. In accordance with the literature findings, we have developed a questionnaire
that included open-ended questions concerning the definition of discrete
mathematics, types of problems, particularly proofs, in discrete mathematics, and the
utility of discrete mathematics at university level (teaching and learning). We have
noted important aspects of discrete mathematics, which will enrich our “functional
definition”, and they will be presented as soon as we complete the rest of the
interviews. At the methodological level, Table 1 represents our first approach to
analysis. However, to better frame the conceptions of the researchers, we will be
developing in parallel other analyses methods. This will be done using two
complementary approaches: the first based on the praxeologies of Chevallard and the
second relying on a theoretical model regarding “conceptions” (Balacheff, 2013).
Axis Criteria for analysis
Conception on the definition of discrete Identify different points of view for
mathematics researchers (since the definition is not
(in teaching and in research) agreed-on)

Topics from discrete mathematics Identify and categorize topics

Conception on proofs in discrete Identify types of problems, types of
mathematics (in teaching and in research) reasoning, characteristics of concepts,
place and role of modeling

199 [Link]:indrum2018:174690
Links between discrete mathematics and Categorize links; are they being
other disciplines (in teaching and research) worked in class?
Learning of discrete mathematics Identify objectives, learning outcomes,
learning difficulties, student behavior

Table 1-Criteria for analyzing the interviews with the researchers in discrete
mathematics
To test our questionnaire, we have conducted two pilot interviews with two graph
theorists, one in Lebanon and the other in France. The interviews were recorded and
transcribed. We have selected some instances from the pilot study, and they will be
presented in this paper in the following section.
PRELIMINARY RESULTS
Researchers’ conceptions about the definition of discrete mathematics
In order to analyze the conception of the interviewees about the definition of discrete
mathematics, we tried to elicit some epistemological aspects of discrete mathematics.
The pilot interviews showed that for the two interviewees Michel (researcher and
instructor of graph theory in Lebanon) and Bertrand (researcher and instructor on
graph theory in France) discrete mathematics is difficult to define, and sometimes it
is easier to define what is not discrete. Both interviewees used the term “separable” to
describe discrete objects:
Bertrand: […] so basically one could say that discrete mathematics concerns objects
that can be separated […] (our translation)
Michel: […] the elements can be manipulated separately […] (our translation)
Interesting examples illustrating this important aspect of the definition discrete
mathematics (“separable”) will be presented for discussion during the conference.
However, the interviewees had different opinions regarding the teaching strategies
and the origin of student difficulties. Michel focuses on the teaching of concepts
whereas Bertrand puts more emphasis on the methods and strategies (through games
and experimentations).
Michel: […] in the courses, I try to convey the basic ideas like in graph theory:
definition of graphs, adjacency matrices, standard objects such as […] (our
translation)
Bertrand: […] in fact, it is to train for reasoning skills ... and by the extrapolation to
critical thinking [...] that is by working on the problems I have proposed
like […] (our translation)
It is this discrepancy between teachers' perceptions of discrete mathematics and their
corresponding teaching practices at university level that we intend to further explore
in our study.

200 [Link]:indrum2018:174690
Researchers’ conceptions on proofs in discrete mathematics
In order to characterize proofs in discrete mathematics in general, we made an
attempt at identifying the types of proofs used in discrete mathematics. We find that
the proofs by contradiction, by induction, and by recurrence are the most used. The
exhaustive proofs are also used frequently, and according to the interviewees, this is
due to the fact that oftentimes problems require very complex strategies, which
compels the students to perform case-by-case analysis. Apparently, this exploratory
phase of problems is a remarkable requisite of topics in discrete mathematics more
than in other branches of mathematics.
Moreover, heuristic processes show in the students’ development of methods, to find
approximate solutions instead of exact solutions to problems. According to Bertrand,
it is widely used in the experimentations for proof and in the mathematical
investigation processes. For Bertrand, in discrete mathematics, heuristics consist of
taking particular cases (like combinatorial optimization problems), extirpating to
arrive at a clear solution (questions of tiling and stacking), and modeling illustrations
especially in difficult problems.
We have also noticed that the “proof” activity in mathematics has a different status
than the “demonstration” activity. It is affirmed by the interviewees that there is a
difficulty for students in writing proofs:
Michel: […] they feel at ease, they understand everything that is explained but they
feel unable to reproduce […] (our translation)
Bertrand: […] we think it’s clear and that we are convinced; however when asked to
write, to formalize, we do stupid things …] (our translation)
Therefore, we notice that the place of proof in discrete mathematics is not well
defined and needs more investigation especially when it comes to its characteristics
and the distinctions between the terms proof, demonstration writing, argumentation,
etc.
CONCLUDING REMARKS
Currently, we limit our work to researchers of discrete mathematics particularly
graph theory. For the rest of our work, we plan to complete the analyses of interviews
with the researchers to further develop the state of the art (to further explore proof,
modeling and their particularities in relation to discrete mathematics). At the
conference, we will present some more refined results of these interviews along with
to the questionnaire used. This mapping along with the review of literature will allow
us to better develop the criteria that would ultimately lead to a functional definition of
discrete mathematics. We are also interested in exploring the teaching practices of
researchers in order to make informed suggestion on the training of instructors at the
university. An extension to this work might possibly be in interviewing researchers in
contiguous disciplines like computer science, algebra, or number theory.

201 [Link]:indrum2018:174690
[1] We aim at constructing a representation of discrete mathematics that presents the concepts,
types of problems, proof processes and strategies, reasoning skills and other particularities of the
field of discrete mathematics.
[2] The adjective “contemporary” indicates that our research focuses on the researchers’ practices in
statu nascendi. We have conducted interviews with mathematicians to this end.
[3] In this paper, we use the word “conception” in the common sense, not yet in any specific
theoretical sense.
REFERENCES
Alcock, L. (2009). Teaching proof to undergraduates: semantic and syntactic
approaches. Proceedings of the ICMI 19 conference: Proof and Proving in
Mathematics Education 1 (pp. 29-34). Taipei: The Department of Mathematics,
National Taiwan Normal University.
Artigue, M. (2016). Mathematics education research at university level: achievements
and challenges. Proceedings of Indrum 2016, (pp. 11-27). Montpellier.
Balacheff, N. (2013). cK¢, a model to reason on learners' conceptions. In M.
Martinez, & A. Castro Superfine (Ed.), Proceedings of the 35th annual meeting of
the PME-NA (pp. 2-15). Chicago, IL: University of Illinois at Chicago.
Balacheff, N. (1999). L'argumentation est-elle un obstacle? Invitation à un débat...
International Newsletter on the Teaching and Learning of Mathematical Proof , 1-
5.
Balacheff, N. (1987). Processus de preuve et situations de validation. Educational
Studies in Mathematics , 18, 147-176.
Borwein, J. M. (2009). Digitally assisted discovery and proof. Proceedings of the
ICMI study 19 conference: Proof and Proving in Mathematics Education. 1, pp. 3-
11. Taipei: The Department of Mathematics, National Taiwan Normal University.
Committee on the Mathematical Sciences in 2025. (2013). The mathematical sciences
in 2025. Board on mathematical Sciences and Their Applications- Division on
Engineering and Physical Sciences- National Research Council. Washington, DC:
The National Academies Press.
DeBellis, V. A., & Rosenstein, J. G. (2004). Discrete mathematics in primary and
secondary schools in the united states. ZDM , 36 (2).
Epp, S. S. (2016). Discrete mathematics for computer science. 13th International
Congress on Mathematical Education- TSG 17, (pp. 1-8). Hamburg.
Goldin, G. (2016). Discrete mathematics and the affective dimension of mathematical
learning and engagement. Proceedings of the 13th International Congress on
Mathematical Education. Hamburg: Springer.
Grenier, D., & Payan, C. (1998). Specifités de la preuve et de la modélisation en
mathématiques discrètes. Grenoble: Département de Mathématiques Discrètes,
Laboratoire LEIBNIZ, Université Joseph Fourier.

202 [Link]:indrum2018:174690
Hart, E. W., & Martin, W. G. (2016). Discrete mathematics is essential mathematics
in a 21st century school curriculum. 13th International Congress on Mathematical
Education. Hamburg: Springer.
Hart, E. W., & Sandefur, J. (in press). Teaching and learning liscrete mathematics in
the school curriculum worldwide, an ICME-13 monograph. (E. W. Hart, & J.
Sandefur, Eds.) Hamburg: Springer.
Hausberger, T. (2017). Structuralist praxeologies as a research program on the
teaching and learning of abstract algebra. International Journal of Research in
Undergraduate Mathematics Education .
Heinze, A., Anderson, I., & Reiss, K. (2004). Discrete mathematics and proof in the
high school, introduction. ZDM , 36 (2), 44-45.
Maurer, S. (1997). What is discrete mathematics? The many answers. In J. G.
Rosenstein, D. S. Flanzblau, & F. S. Roberts (Eds.), DIMACS Series in Discrete
Mathematics and Theoretical Computer Science (Vol. 36, pp. 121-132). American
Mathematical Society.
Nardi, E. (2008). Amongst mathematicians teaching and learning mathematics at
university level. USA: Springer.
NCTM. (2000). Principles and standards of school mathematics. (N. C.
Mathematics, Ed.) Reston, VA: NCTM.
Ouvrier-Buffet, C. (2011). A mathematical experience involving defining processes:
in-action definitions and zero-definitions. Educational Studies in Mathematics , 76
(2), 165-182.
Ouvrier-Buffet, C. (2014). Discrete mathematics teaching and learning. In S. Lerman
(Ed.), Encyclopedia of Mathematics Education. Dordrecht: Springer.
Ouvrier-Buffet, C. (2006). Exploring mathematical definition construction processes.
Educational Studies in Mathematics , 63 (3), 259-282.
Radford, L. (2016). Epistemology as a research category in mathematics teaching and
learning. (B. Hodgson, A. Kuzniak, & J. Lagrange, Eds.) The didactics of
mathematics: Approaches and Issues , 31-41.
Reid, D. A., & Knipping, C. (2010). Proof in mathematics education. Rotterdam:
Sense Publishers.
Rosenstein, J. G. (2016). The absence of discrete mathematics from primary and
secondary education in the united states... and why that is counter positive.
Proceedings of the 13th Internatio0nal Conference on Mathematical Education.
Hamburg: Springer.

203 [Link]:indrum2018:174690
 Analysing regressive reasoning at university level

Marta Barbero1,2 and Inés M. Gómez-Chacón2

1
Università di Torino ([Link]@[Link]), 2Universidad Complutense de Madrid

ABSTRACT:
This paper focuses on the epistemic and cognitive characterization of regressive
reasoning in resolving strategic games. It explores the use of the Finer Logic of
Inquiry Model as a tool for the analysis of the regressive reasoning. It reports the
results of a study carried out on 32 undergraduate students who are studying a
Mathematics Degree in a university of Spain.
Key words: Teaching and learning of specific topics at university level mathematics,
Teaching and learning of logic, reasoning and proof, Regressive reasoning, Logic of
Inquiry, Strategy games.

1. INTRODUCTION
The method of analysis has proved to be extremely stimulating in various fields, and
has played a crucial role in the emergence of the modern world-view. The combination
of the two branches of analysis and synthesis has been applied to several fields of
artificial intelligence, theoretical computer science, and in programming methodology
(Peckhaus, 2000; Grosholz, Breger, 2000). For many engineering students and
mathematics undergraduate students, learning the method of analysis in tertiary
education mathematics is a critical issue. They have the challenge of incorporating it
in different disciplines related to the design and production of products and services,
such as, Project Management, Systems Engineering and Design Science. They have no
theoretical and methodical basis (Koskela and Kagioglou, 2006). A conscious
integration of regressive reasoning in mathematics university learning raises the need
for articulation between epistemological and cognitive aspects. Regressive reasoning
is not completely logically determined, but has elements of contingency, creativity and
intuition. The purpose of this text is to highlight the potential of Finer Logic of Inquiry
Model (Arzarello 2014) as a tool for the didactical analysis of the regressive
reasoning. This model has been used at secondary level education, not being used at
tertiary level so far.
Here we will report the results of a study carried out on 32 undergraduate students
studying a Mathematics Degree at a Spanish University, using strategy games in order
to promote the regressive reasoning. The choice of strategy games is justified by
antecedents to this study in which they have been shown to be a key tool for teaching
problem solving and regressive reasoning (Gómez-Chacón, 1992).
The present research is primarily exploratory for two reasons: 1) Regressive reasoning
has been scantly analysed in mathematics and educational psychology; 2) the use of

204 [Link]:indrum2018:172311
the Finer Logic of Inquiry Model methodology to analyse data from mathematical
thought at tertiary education is a new development. The theoretical background and
empirical studies related to regressive reasoning needs to be developed.

2. REGRESSIVE REASONING
In mathematics, progressive reasoning alone is not exhaustive to fulfil the tasks of
solving problems. Great mathematicians like Pappus, Descartes, Leibniz, in their
discussions about analysis and synthesis, emphasize this fact (Peckhaus, 2000).
Regressive reasoning is known by different denominations: regressive analysis,
backward solution, method of analysis, etc. This process includes different ways of
proceeding in problem solving: backward strategy, strategy of assuming the problem
solved, Reductio ad Absurdum, beginning at the end of the problem, etc.
Pappus was the mathematician who has contributed substantially to the clarification
and exemplification of the method. In the seventh book of his Collection he deals with
the topic of Heuristics (methods to solve the problems).Where he exemplifies the
method of analysis as the method of synthesis, therefore making the development of
this reasoning clearer. Pappus defines the method of analysis as follows: “In analysis,
we start from what is required, we take it for granted; and we draw correspondence
(ακολουθον) from it and correspondence from the correspondence, till we reach a
point that we can use as a starting point in synthesis. That is to say, in analysis we
assume what is sought as already found (what we have to prove as true).” (elaboration
by Polya, 1965 and by Hintikka and Remes, 1974). Subsequently he points out: “This
procedure we call analysis, or solution backward, or regressive reasoning.” (Hintikka
and Remes, 1974) And on the Method of Synthesis: “In synthesis, on the other hand,
we suppose that which was reached last in analysis to be already done, and arranging
in their natural order as consequents the former antecedents and linking them one with
another, we in the end arrive at the construction of the thing sought. This procedure we
call synthesis, or constructive solution, or progressive reasoning.”(Hintikka and
Remes, 1974)
In summary, the following was considered backward reasoning: the practice that
involves the making of a number of arguments from the bottom of the problem and
proceeds through logical correspondences which allow to obtain something known or
to be reached through other paths. The analytical method consists of a procedure that
starts with the formulation of the problem and ends with the determination of the
conditions for its solution.
3. FINER LOGIC OF INQUIRY MODEL (FLIM)
Trying to overcome the static approach of habitual logical mathematical reasoning,
Hintikka (1996, 1999) developed what he calls Logic of Inquiry. The idea, already
elaborated by ancient Greek philosophers, is building knowledge through a
questioning process, implicit or explicit. The knowledge is the result of research

205 [Link]:indrum2018:172311
generated by a specific question. The philosopher introduces it as the “logic of
question and answer”.
In his approach he considers Game Theory and game semantics to support formal
epistemic logic. Hintikka overcomes the limitations and excessive abstractions of
Tarski's Definitions of Truth (Sher, 1999), which leave the process used to reach the
truth unexplained. He introduces a top-down definition of truth (Hintikka, 1995)
unlike the classical and tarskian bottom-up view, highlighting the regressive way of
proceeding in problem solving from an epistemological point of view. Hintikka (1995)
retakes the idea of Wittgenstein's language games and some aspects of Game Theory,
elaborating on a theory where the centre is “a path towards the formulation of a truth
that, instead of proceeding recursively from atomic to complex formulas, reverses the
approach and proceeds from the more complex ones to their simplest constituents”. In
this research, the study of games will try to explain this interlacing between game
theory and strategic rules that allows the student to win.
The FLIM elaborated on by Arzarello (2014) sought to propose a concretion of
Hintikka`s proposal to be used in the Didactics of Mathematics. More specifically, he
explained the elements needed to analyse the interactions between strategic and
deductive components of students’ resolution protocols. This model allows for the
structuring of the resolution in two components: Inquiry Component (IC) and
Deductive Component (DC).
In the Inquiry Component the subject alternates a series of questions, answers and
explorations, according to Hintikka's Logic of Inquiry. Its purpose is to meet the aim of
the problem, solving conjectures that gradually rise from results of two explorations:
• Exploration: in order to analyse and understand the situation in which the
subject is involved
• Control: in order to verify the ideas or conjectures that came out during the
development of the activity.
In the above Component, the cognitive dimension of reasoning is necessary. From a
cognitive point of view, the progressive-regressive reasoning movement has been
highlighted by studies such as those of Saada-Robert (1989). The psychological model
for solving mathematical problems focuses on the distinction between two phases of
the resolution: investigate why things are like this (backwards, until reach a plausible
hypothesis -abduction- or a known fact) and verify this investigation (forward,
codified by the classical logic). Based on Saada-Robert's model, Arzarello (2014) and
Soldano (2017) characterized this cognitive dimension through the sequence of
actions in three different modalities: ascending, neutral and descending.
Ascending modality (A) refers to the path towards the formation of ideas and
conjectures after a phase of exploration. Descending modality (D) characterizes the
transition from a conjecture to an investigation. The purpose of descending modality is

206 [Link]:indrum2018:172311
to find an equivalence between the object of thought (the conjecture, the idea) and the
object of work (the problem and its resolution). Neutral modality (N) marks the change
between the ascendant and the descendent; it is the moment in which a conjecture is
formulated. Observable actions in the subjects are: formulations (of questions, of
resolutions plan, of conjectures), affirmations, explorations and controls.
In the Deductive Component the subject is not directly involved in the investigation
and verification of conjectures and uses a language with a logical nature to formally
formulate the truth. Three specific modalities are added: detached modality, logical
control and deductive modality (Arzarello, 2014; Soldano, 2017). Detached modality
is the moment in which a conjecture, which has not arisen immediately after an
exploration, is formulated. Logical control is the time when an exploration-control is
done without using instruments. It is characterised by the use of formal language.
Deductive modality characterises control phases where instruments are involved.
Deductive Steps and Logical Chains are added to the Inquiry Component actions.
Inquiry and Deductive components are not often well differentiated during problem
resolution where the subject passes from one component to another, even more than
once. We can say that the typical components structure is nested in this way: (IC ~ (DC
~ (IC ...))) with “~” that expresses the passage from one component to the other.
Observable actions
Modalities
General Specific
Question
Affirmation Ascendant
Verbal
Conjecture Neutral
Handwritten
Exploration Descendant
Gestures
Control Detached
Others (gaze, …)
Plan formulation Logical Control
Silent
Deductive step Deductive
Logical chain

Table 1
Table 1 summarizes some observable actions and their modalities according to the
definitions given and that will be considered in the analysis.

4. AIM AND METHODOLOGY

Aim
The aim of this paper is to show an evaluation tool for examining how regressive
reasoning develops in university students. In particular, how the FLIM can be a valid

207 [Link]:indrum2018:172311
tool to analyse the interplay between cognitive and epistemic in the regressive
reasoning.
Participants and instrument
Data were collected in 2014 from 32 (19 women and 13 men, aged between 21 and 23)
Caucasian undergraduates working toward a BSc. in mathematics. All of the
participants were in their last year of academic studies. They were following advanced
courses in several areas of geometry, algebra, probability and analysis. With regard to
solving problems, the students had been introduced to the problem solving heuristics.
They had not received any special training about backtracking heuristics.
The work dynamic started with individuals being given paper and pencil with which
they need to resolve two games, each lasting one and a half hours. Figure 1 shows the
problem which we will analyse in the results section. Strategy games allow for the
natural development of regressive reasoning. These games are disconnected from the
mathematical content which forces the student to use their mathematical knowledge
acquired in their university degree.
The Triangular Solitaire (Gómez-Chacón, 1992) is a game for a single person that
requires a board with 15 boxes as the figure shows.

These are the rules:

1. Place the pegs in all boxes, except in the one marked in black.
2. The player can move as many pegs as they like as long as they are
able to jump over an adjacent peg and onto an empty space (along the
line). At the same time, he "eats" the peg that was jumped over and that
peg gets taken out of the game. All pegs move in the same way. Pegs
can move around the table in any direction.
Objective: The player wins when there is only one peg on the table.

Figure 1
Students were given the game and asked to describe their approaches to solving the
problem on protocols including: thought processes in the resolution, explanations of
the difficulties they might face, and strategies they would use in order to solve with
paper and pencil. A qualitative analysis was chosen to examine the resolution
protocols of the students through the “Finer Logic of Inquiry Model” (Arzarello 2014).
A general analysis of 32 students took place before a case study was carried out. In this
paper we describe an individual student case in order to show a deep understanding of
the tendencies of the behaviour related to the sequences of actions and movement
between modalities of reasoning. The protocols analysis, at a macroscopic level of this
case, provides the identification of reasoning difficulties and way of using backward
reasoning that determined success or failure in the resolution. It’s worth noting that
Student M (see section 5) is a key informant of the group because he belongs to 60% of

208 [Link]:indrum2018:172311
students that use the backward strategy and incorporates graphical representations to
achieve the transition between modalities.

5. REGRESSIVE REASONING USE (CASE STUDY)

Regressive reasoning use varies among the group of students. Let us examine a case
study. A student (Student-M) has combined regressive reasoning with different
strategies and auxiliary constructions: drawings, graphical representations. Student-M
indicates difficulties in creating the solution because of the actions which are needed
for discovering the solution and because of the recognition of representational
equivalences. The visualisation and representations which are used help during the
resolution process; Student-M performs continuous control over its own resolution
process. She is able to slightly modify the strategy or even change it completely to
reach the solution. For analysis purposes, Student-M’s protocol has been divided into
the following phases: familiarisation, exploring and carrying out the strategy, results
verification. According to the Finer Logic of Inquiry Model, this student’s protocol is
mainly characterised by the inquiry component. This begins with the first part of the
protocol, corresponding to the familiarization phase. The entire protocol has been
translated highlighting the parts where student M uses backward reasoning (in Italics).
Student M protocol
1 To accomplish the exercise, I’m going to number the
holes on the board in order to leave a trace of the
movements I'm doing. At the beginning, all the holes are
filled except number 5.
Figure 3
2 I observe that you can only start with two movements 14-9-5 or 12-8-5.
3 Since this is an equilateral triangle, I think it does not matter what the starting
movement is because they should lead to "symmetrical" solutions.
4 I’ll start to do it roughly.
5 The steps I’ll take are: 14-9-5; 7-8-9; 12-13-14; 2-4-7; 11-7-4; 10-9-8; 3-6-10.
6 At this point, I note that the only way to eliminate 1 would be to move 8-5-3.

7 Here I notice that [with these

movements] the game cannot be
solved because the 4 cannot be
eliminated and the remaining pegs
cannot eliminate each other.

Figure 4
8 I realise that I can try to go backwards, that is, starting with just one peg in one
position and undo the jumps trying to fill the table with the exception of a hole.

209 [Link]:indrum2018:172311
9 Looking at the board, I think that maybe the fact that the last piece stays on the board
(the peg from which I start to move backwards), in a position that you can come up
with many jumps, facilitates the strategy. These places are positions 4, 6 and 13
because you can get to them with 4 jumps.

10 To fill up the game table I will have to do 13 moves, because there are 15 holes, an
initial peg and an empty final hole.
11 Let's start only with peg 13.

Figure 5
12 Here I already notice that I do not reach the solution because I will never fill the top
corner due to the absence of a peg in the 3rd row; I should do 11-7-4 leaving corner
11 without a peg [so that the top corner will be filled].
13 Let's start with the reason for the various steps:
13-14-15: I want to start filling the corners as soon as possible because these holes
are the hardest to fill up (the peg is in hole 15 and I will not move it anymore).
14-13-12: Random movement.
12-8-5: I want to leave hole 12 free to get to the next step at corner 11.
8-9-10: I want to leave hole 8 free to retrieve peg 12 (to fill 13 and 14) in the next step,
so I can complete it later [the row].
12-13-14: I want to complete the row below.
5-8-12: I want to complete the row below.
14 I think trying to fill the centre was not a good strategy...
15 … so now I'm going to try to fill the outside of the triangle, that is, [I'll try to] undo the
jumps to the corners and sides. (Playing normally would involve jumping to the centre
avoiding corners and sides if possible.).
16 I also get stuck [on the fact] that by eating pegs or
undoing the jumps, the movements that are made are
triangular. So I will try to fill the smaller triangles
contained in the big triangle. Figure 6
17 First, I will fill the lower right triangle.
18 Now I’m going to fill the upper triangle; to do so (Since i do not want to remove the
peg I placed in position 1), I have to get some pegs in the 4th row that, undoing the
jump fills the 2nd and 3rd row. I undo the jump with the 9.
19 Now you have to fill the lower left triangle.

Figure 7

210 [Link]:indrum2018:172311
 20 Now I just have to write the jumps in the correct order

Figure 8
The following table shows actions and cognitive modalities associated with each
protocol line and figure; a check ( ) indicates the lines where the regressive reasoning
is used. The last column of the table shows different strategies involved.
Familiarization phase
Protocol parts Action Modality R.R Strategy
Lines 1-4 y Fig. 3 Exploration Descendant
Line 5 Affirmation Neutral
L. 6-7 and Fig. 4 Exploration Ascendant
Explore and carry out the strategy
Line 8 Plan Neutral Backward
Line 9 Exploration Ascendant Begin from the end
Line 10 Affirmation Descendant
Line 11 y Fig. 5 Exploration Descendant
Line 12 Affirmation Ascendant
Line 13 Exploration Ascend/Descen
Line 14 Affirmation Ascendant
L. 15-16 and Fig. 6 Plan Neutral Auxiliary construction
Lines 17 y Fig. 7 Exploration Descendant
Line 18 Exploration Ascendant
Line 19 y Fig. 7 Exploration Ascendant
Results verification
Line 20 y Fig. 8 Control Detached

Table 2

211 [Link]:indrum2018:172311
This cognitive analysis shows that the first two resolution phases are characterised by
a continuous alternation of explorations and plan formulations together with an
alternation of descending and ascending modalities. The second resolution phase
involves the continuous use of the going backward strategy. Subdivision of the board
into rows and then into triangles is fundamental to reach the solution. Student-M
modifies the strategy slightly by adding new elements in the resolution (board
subdivision into rows and triangles) typical of problem solving using regressive
reasoning. Crucial points of backward reasoning are reached in the ascending
modality (see in Table 2) where ideations occur. A routine that can be established
regarding the use of modalities is A~N~D~(A~N~D~(A~…)). The neutral modality
marks the transition between A and D and it is characterised by the incorporation of
auxiliary constructions as generating tools of new knowledge (epistemic transaction).
In the third phase of the resolution, by writing and graphically representing the steps
taken to reach the solution, Student-M (in detached modality) checks the result
obtained by going backwards.

6. CONCLUSION
Analysis with the FLIM model allows to model student’s cognitive movement in a
logical concatenated way. The strategic aspects are more dominant in the ascending
and descending modality, while the epistemic ones are prevailing in the neutral
modality. Our study confirms results obtained by Soldano (2017) (with upper
secondary school students in geometry): the ascending modality characterises the
backward way of thinking, while descending is the cognitive modality that
characterises the progressive way of reasoning. However, most likely, abductive
reasoning has been used in the formulation of conjectures in ascending modality, but
we cannot be sure of it by only analysing the protocol, we need to complete this
information by interview. This is an open question for further research.
At a phenomenological level, this method allows us to analyse the development of
strategic aspects within the cognitive modality movement to reach the solution. But it
mainly focuses on cognitive modalities while it doesn’t distinguish between the
strategic principles that are used. Through this tool it’s possible to emphasise that
regressive reasoning involves auxiliary intuition elements that are necessary to
achieve the solution; these aspects are developed by looking at the consequence and
looking for the premises. A larger sample size with two different tasks, find the
winning strategy and mathematically solving the game, would allow us to advance in
the development of the tools for evaluating regressive reasoning.

7. ACKNOWLEDGMENT
Supported by NILS Program (ES07) (007-ABEL-IMI-2013), Project “Mathematical
Working Space” (UCM-CmdeGuzman-2015-01), University of Torino doctoral
scholarship and by special action IMI-INVEDUMAT_uni.

212 [Link]:indrum2018:172311
REFERENCES
Arzarello, F. (2014). Notes for a seminar “A new structural approach to argumentation
in mathematics: from Toulmin model to Hintikka Logic of Inquiry”, in Workshop:
Mathematics education as a transversal discipline. Università di Torino.
Gómez-Chacón, I. Mª (1992). Los juegos de estrategias en el curriculum de
matemáticas. Narcea, Madrid.
Grosholz, E., Breger, H. (2000). The Growth of Mathematical Knowledge. Kluwer:
Dordrecht/Boston/London.
Hintikka, J. (1984). Rules, Utilities, and Strategies in Dialogical Games in Vaina, L.,
& Hintikka, J. Cognitive constraints on communication: Representations and
processes. Boston: Reidel.
Hintikka, J. (1996). The Principles of mathematics revisited. Cambridge Univ. Press.
Hintikka, J. (1995). The Games of Logic and the Games of Inquiry. Dialectica. vol. 49,
pp. 229–249.
Hintikka, J. (1999). Inquiry as Inquiry: A logic of Scientific Discovery. Springer
Science+Business Media Dordrecht.
Hintikka, J., Remes, U. (1974). The Method of Analysis: Its Geometrical Origin and
Its General Significance. Reidel, Dordrecht.
Koskela, L. and Kagioglou, M. (2006). The proto-theory of design: the method of
analysis of ancient geometers. 9th International Design Conference. Dubrovnik.
Peckhaus, V. (2000). Regressive Analysis. University Essen. Forschungsbericht.
Polya, G. (1945). How to solve it? USA: Princeton University Press.
Saada-Robert, M. (1989), La microgénèse de la representation d’un problem.
Psychologie Française. 34, 2/3.
Sher, G. (1999). What is Tarski’s Theory of Truth? Topoi 18. p. 149-166.
Soldano, C., Arzarello, F. (2016). Learning with touchscreen devices: game strategies
to improve geometric thinking. Math. Education Research Journal. 28(1), 9-30.
Soldano, C. (2017). Learning with the logic of inquiry. A game-approach within DGE
to improve geometric thinking. PhD Dissertation. Università di Torino.

213 [Link]:indrum2018:172311
 A Study of Students’ Reasoning About “There exists no …”
Stacy Brown
California State Polytechnic University Pomona, brown@[Link]
In this paper, we report findings from two studies of students’ engagement in
metatheoretical tasks drawn from a model of the reasoning requirements of a proof
by contradiction. The studies aimed to explore students’ engagement in the tasks, the
extent to which they were successful, and the similarities and/or differences between
students’ and mathematicians’ approaches. Findings indicate students tend towards
syntactic, logical theory approaches while mathematicians gravitate towards
semantic, mathematical theory approaches. Drawing on interview data, it is shown
that students may use symbols to avoid employing fragile content knowledge, yet
encounter further difficulties by viewing quantifiers as appended symbols.
Keywords: metatheoretical reasoning, proof by contradiction.

METATHEORETICAL DIFFICULTIES: AN OVERVIEW

Mariotti, Bartolini Bussi, Boero, Ferri & Garuti (1997) argue, “what characterises a
mathematical theorem is the system of statement, proof and theory” (p. 183). By this
they mean a statement and its proof are situated within a theory from which one
draws not only axioms, definitions, and theorems (i.e., a mathematical theory) but
also rules of inference (i.e., a logical theory). The fact that proofs are contingent on
both a mathematical and a logical theory is best illustrated by the fact that there are
statements that are valid in some mathematical theories (e.g., Euclidean geometry)
that are not valid in others (e.g., Hyperbolic geometry). Thus, their validity depends
on the mathematical theory referenced. Moreover, if one does not employ standard
logic, further shifts in the status of theorems may occur; a point illustrated by
Brouwer’s rejection of his Fixed-Point Theorem, after adopting intuitionistic logic.

Building on this model of mathematical theorems and the theory of Cognitive Unity,
Antonini and Mariotti (2008) demonstrated that students’ difficulties with indirect
proofs may occur within the mathematical theory or the logical theory. To
demonstrate the latter, a compelling example is given where a university student,
Fabio, describes difficulties accepting a proof by contraposition because of the
movement back and forth between the statement-to-prove and the contrapositive.
Fabio: Yes, there are two gaps, an initial gap and a final gap. Neither does the initial gap
is comfortable: why do I have to start from something that is not? [...] However, the final
gap is the worst, [...] it is a logical gap, an act of faith that I must do, a sacrifice (Antonini
and Mariotti, 2008, p. 407, sic).
Indeed, Fabio speaks both of his acceptance of the proof of the contrapositive and his
difficulties accepting the contrapositive proof as a proof of the statement-to-prove.
Antonini and Mariotti refer to these difficulties as metatheoretical, for they are at the
level of the logical theory that is applied to the mathematical theory. Their work is of

214 [Link]:indrum2018:174100
interest, for it raises many questions: To what extent are novices successful when
engaging in metatheoretical tasks? What approaches do they employ? Do their
approaches differ from mathematicians’? The purpose of this paper is to take a
preliminary step towards answering these questions. Specifically, we report on two
studies of participants’ responses to metatheoretical tasks drawn from a model of the
reasoning requirements of a proof by contradiction, which is described below.

PROOF BY CONTRADICTION AND ITS REASONING REQUIREMENTS

In this section, our aim is to model the reasoning requirements of a proof by
contradiction of a universally quantified conditional statement. To aid our discussion,
we consider a specific example: Theorem 5. For all positive integers n, if n mod(3) º
2 then n is not a perfect square. To prove the theorem by contradiction, one must
correctly negate the universally quantified conditional statement and take the
resulting statement as one’s primary assumption. Such actions require one accept (at
least at an intuitive level) that for a conditional statement to be true universally, it
must not be the case that there is some element in the universe of discourse for which
the premise is true (has a truth-value of true) and the conclusion is false (has a truth-
value of false). In the case of our example, we assume “There exists a positive integer
n, such that n mod(3) º 2 and n is a perfect square.” As shown by Wu Yu, Lee, &
Lin (2003), this task is far from trivia for students 17-20 years of age. Moreover, as
Antonini and Mariotti (2008) note, the validity of the work is determined by theorems
that reside within the logical theory (i.e., the metatheory).
Having assumed the negation of the statement-to-prove, one must now explore the
consequences of this assumption and identify a contradiction. Three aspects of this
work are important. First, to carry out this work one must move back to the
mathematical theory, for it is here that the contradiction will reside. Second, one’s
goal is open-ended, for one does not know in advance where one will find the
contradiction. In fact, there may be many. Third, one must know one’s commitments
with regard to the mathematical theory. Otherwise, one will not have the means to
recognize a contradiction. This point was made by Sierpinska (2007) who argued,
“sensitivity to contradictions in mathematics requires theoretical thinking …
(thinking) concerned with internal coherence of conceptual systems” (p. 1-54). Once
the contradiction is identified, one’s work is not done. One must make sense of it.

In our example, we claimed that an integer existed but having produced a

contradiction we now know that such a number cannot exist. Hence, one must
conclude,i there exists no integer n, such that n mod(3) º 2 and n is a perfect square.
And it is at this point that one is faced with the very requirement that Antonini and

215 [Link]:indrum2018:174100
Mariotti’s (2008) Fabio rejected; namely, seeing the proof of this statement as a proof
of the statement-to-prove. In other words, having shown S*: There exists no integer
n, such that n mod(3) º 2 and n is a perfect square, one must recognize (from a
logical standpoint) that one has proven S: For all positive integers n, if n mod(3) º 2
then n is not a perfect square; that is, we must recognize S* implies S since S* is a
non-identical but logically equivalent form of S. As Antonini and Mariotti (2008)
note, this work relies on theorems in the logical theory rather than the mathematical
theory; that is, it is metatheoretical. Thus, a proof by contradiction imposes two
unique metatheoretical requirements. First, at the beginning, when one must produce
the negation of a statement. Second, at the conclusion, when one must recognize that
S* implies S. And, it is the latter requirement that is the focus of the reported studies.

AN OVERVIEW OF THE STUDIES

The reported studies examined students’ engagement and extent of success in the
metatheoretical reasoning requirements that arise at the conclusion of a proof by
contradiction. All studies were conducted at a minority-serving university, where the
majority of students qualify for need-based financial assistance and are first-
generation university students. Study 1 explored the extent to which novices (i.e.,
students without prior logical training or who have limited training) are successful
evaluating claims of the form S*implies S. Study 2 explored students’ and
mathematicians’ approaches to and success with metatheoretical tasks. The aim of the
combined studies was to explore the reasoning practices that may inhibit or support
students’ metatheoretical work and consequently, play a role in the extent to which
students reach or fail to achieve cognitive unity in relation to indirect proofs.

Study 1 Methods and Findings

To explore novices’ success with metatheoretical reasoning tasks prior to instruction,
46 university students were surveyed. The surveys were administered on the first day
of a “Basic Set Theory and Logic” course that served as the universities’ first logic
course and their “Introduction to Proof” course. Prior to the course, students would
have been enrolled in computation-focused courses on calculus and differential
equations. Included on the survey were two tasks that asked students to compare a
pair of statements and determine, “Can you prove Statement A by proving Statement
B?” (Figure 1.) Task 1 involved a universally quantified statement and an incorrect
alternative. Task 2 involved the same Statement A and a correct alternative.
Task 1.
Statement A. For every integer n, if n is a perfect square, then n has an even number of factors.
Statement B. There exists no integer n such that n has an even number of factors and n is not a perfect square.
☐ Yes, you can prove Statement A by proving Statement B.
☐ No, you cannot prove Statement A by proving Statement B.
Task 2.
Statement A. For every integer n, if n is a perfect square, then n has an even number of factors.
Statement B. There exists no integer n such that n is a perfect square and n has an odd number of factors.
☐ Yes, you can prove Statement A by proving Statement B.

216 [Link]:indrum2018:174100
 ☐ No, you cannot prove Statement A by proving Statement B.

Figure 1. Study 1’s Task 1 and Task 2

Survey results indicated that of the 46 students surveyed, 50% were successful at
Task 1, 47.8% were successful at Task 2, and 24% were successful at both tasks.

Study 1 Discussion
The findings of Study 1 demonstrate that most of the students did not enter the Basic
Set Theory and Logic course reasoning in ways aligned with the metatheoretical
requirements of indirect proofs, as the rates were at or below guessing and less than a
quarter successfully answered both questions. While the findings are not startling,
they provide a warrant for further research. Indeed, prior to Study 1 there were no
studies of novices’ responses to such tasks prior to instruction. Thus, the findings
warrant the following questions: Do novices’ difficulties persist after instruction? Do
students’ and mathematicians’ approaches differ?

Study 2 Methods
Study 2 aimed to explore university students’ and mathematicians’ extent of success
and approaches to the metatheoretical task in Figure 2. Participants were 21 students
drawn from the same student population as Study 1 and 6 mathematicians. However,
the Study 2 students had completed the Basic Set Theory and Logic course. As the
course focused on set theory and logic in the service of proof writing, the instruction
on set theory and logic was limited to basic properties, terms, and definitions, as well
as symbolizing practices, and then on specific proof techniques and/or strategies. All
participants took part in video-recorded interviews during which the task was
presented on a large piece of paper. The participants were given as much time as
requested and then asked to explain their answer to the stated question.
Question: Can you prove Theorem 5 by proving Statement A?
Theorem 5. For all positive integers n, if n mod(3) º 2 then n is not a perfect square.
Statement A. There exists no positive integer n such that n mod(3) º 2 and n is a perfect square.

Figure 2. Study 2, Interview Task

To identify approaches the analysis focused on which theory (mathematical or
logical) the participant worked in and how they engaged in that theory. Responses
were considered mathematical theory approaches (MTA) if the participant was
observed: (1) explicitly exploring mathematical statements, definitions and/or terms;
and/or (2) constructing a proof of either statement. Responses were considered
logical theory approaches (LTA) if the participant was observed: (1) posing explicit
questions of equivalence; (2) constructing truth-tables and/or working with symbolic
logic; and/or (3) citing logical theorems or practices. In addition to the approach,
participants’ responses were analysed for the form of engagement. Specifically,
coding noted participants’ use of syntactic and/or semantic reasoning, with semantic
referring to reasoning that employs meanings and multiple representational systems

217 [Link]:indrum2018:174100
and syntactic referring to rule-based reasoning within a representational [Link]

Study 2 Findings
In Table 2, we report the percentage of correct responses. The reader will notice that
among the 21 students five types of responses were observed: yes, yes-no-yes, no-yes-
no, no, and don’t know. Yes refers to students who, after a period of exploration,
decided without hesitation that one can prove Theorem 5 by proving Statement A.
Yes-no-yes refers to students who repeatedly switched answers, expressed hesitation
and doubt, but ultimately choose “yes.” No-yes-no were similar to yes-no-yes but
were students who repeatedly switched answers and ultimately choose no. No refers
to students who reached, with evident certainty, the decision you cannot prove
Theorem 5 by proving Statement A. Uncertain refers to students who, after
deliberation, responded to the prompt by remarking they “didn’t know.”
Prompt: Can you prove Theorem 5 by proving Statement A?
Student Responses N %
Yes 6 28.6%
Yes-no-yes 5 23.8%
No-yes-no 4 19.0 %
No 5 23.8%
Uncertain (Don’t Know) 1 4.8%
Mathematician Responses
Yes 6 100%

Table 2: Student and Faculty Response by Category

As seen in Table 2, less than one-third of the students (28.6%) who had completed
the Basic Set Theory and Logic course stated with certainty, yes one could prove
Theorem 5 by proving Statement A. And, nearly as many (23.8%) reached this
conclusion with significant hesitation (Yes-no-yes). Furthermore, 42.8% argued
either with certainty (No) or with hesitation (No-yes-no) that you cannot prove
Theorem 5 by proving Statement A. These findings indicate instruction had little
impact on the students’ success with the metatheoretical requirements of a proof by
contradiction. In contrast, (without surprise) all of the mathematicians replied yes.
Since findings that indicate the prevalence of difficulties are of little use without
information on the nature of students’ engagement, we turn to the analysis of
participants’ approaches. This analysis focused on the question of which theories the
participant engaged with and their form of engagement (see Table 3).
Student Response By Type and Form
(n = 21) Mathematical Theory Approach (MTA) Logical Theory Approach (LTA) NE*
Response Type Semantic Syntactic Semantic Syntactic
Yes 1 3 1
Yes-no-yes 3 3
No-yes-no 4
No 1 1 3
Don’t know 1

Table 3: Student and Faculty Response by Type and Form *(NE is no evidence)

218 [Link]:indrum2018:174100
Looking at Table 3, the reader will notice that the majority of the students (18 of 21;
85.7%) engaged in a logical theory approach (LTA). For many this work occurred
symbolically, with 15 of the 18 (LTA) students replacing the open sentences with the
symbols (e.g., P, Q, ~P or ~Q or P(n), Q(n), etc.) and the phrases for all and there
exists no with " and ∄, [Link] Indeed, except for one LTA-semantic (Don’t
know) and two LTA-syntactic (No-yes-no), the students worked symbolically. When
asked about the use of symbols many students noted their discomfort with the
content, “mod is really rough in my memory right now”, and that “it’s easier to work
with symbols.” Thus, the symbolic approaches enabled the students to avoid content
for which they lacked confidence in their mathematical understandings.
For 3 of the LTA students their symbolic approach led to a quick and definitive yes,
as shown Figure 3. The reader will notice the student initially focuses on the
relationship between the quantifiers and then on how translating from ∄ to "
requires one to act on the open sentences by negating a sentence of the form (PÙ ~Q).

Student A: Yes, you can prove Theorem 5 by proving Statement A … because when
you say there exists no that implies ... well that’s a for all statement and
then you have to negate the umm … the umm … (writes PÙ ~Q).

Figure 3: Student A’s (Correct) Syntactic-LTA Response

In contrast to those who readily replied yes, nearly half of the LTA students
experienced a significant amount of hesitancy and doubt (Yes-no-yes; No-yes-no).
Many of these students articulated difficulties with the phrase “there exists no” while,
at the same time expressing certainty regarding the logical relationship between for
all and there exists (i.e., they asserted the negation of one quantifier produced the
other). Among these students, it was not uncommon for them to argue that there
exists no means nothing and that, “nothing is the opposite of everything,” a point
which left many confused having already noted for all and there exist were
“opposites” in logic. For nearly a third of the students (6 total) recognizing there
exists no as the opposite of for all and the open sentence “n mod(3) º 2 and n is a
perfect square” as the negation of “if n mod(3) º 2 then n is not a perfect square” led
to the conclusion Statement A is the negation of Theorem 5, as illustrated in Figure 4.
Two aspects of this approach are important to note. First, the student compares the
quantifiers (for all and none) and then compares the open sentences. Hence, the
quantifiers are not seen as variable-binding operators that act on open sentences but

219 [Link]:indrum2018:174100
rather as appended symbols. Such reasoning enables the student to translate ∄ into
its “opposite” " independently of translating PÙ ~Q into its “opposite” PÞQ.
Second, such reasoning relies on the student incorrectly viewing ∄ (PÙ ~Q) as being
of the form (~$) (PÙ~Q) rather than as of the form ~[($) (PÙ~Q)].

Student B: they’re opposites […] this (Statement A) is the negation of Theorem 5

…it’s saying for all of them, it’s saying none of them [...] Yeah, (writes
"(PÞQ)) and (writes ~ symbol before "(PÞQ))) is (writes ∄ (PÙ ~Q)).

Figure 4: Student B’s (Incorrect) Syntactic-LTA Response

In addition to the LTA responses, two MTA responses were observed. In the MTA-
semantic response, the student spent his time considering numbers that satisfy n(mod
3)º2 and trying to understand the structure of a number that would disprove Theorem
5 or Statement A. Eventually, this student decided Statement A was false and,
therefore, could not be used to prove Theorem 5. In the MTA-syntactic response, the
student immediately remarked, “it’s by contradiction.” The student then proceeded to
determine if Statement A provided the needed claims for such a proof:
Student C: by contradiction […] he’s claiming that there is no positive integer n, …
such that (points to Statement A’s open sentences) […] so, he’s saying there
is no positive integer n here so you can use that argument (points to open
sentences again) and … so, yeah, you can put those together and prove it.
As seen in Table 4, the mathematicians’ responses were quite different, with all but
one engaging in an MTA. Though not shown, it is important to note that in three of
the five MTA-semantic responses, the mathematician spent the majority of the time
proving (or considering how they would prove) Theorem 5.
Faculty Response Category
(n = 6) Mathematical Theory Approach (MTA) Logical Theory Approach (LTA)
Response Type Semantic Syntactic Semantic Syntactic
Yes 5 1

Table 5: Faculty Response by Type and Form

This work lead all three to realized they would use a proof by contradiction and in so
doing, prove Statement A to prove Theorem 5. In the other two MTA-semantic
responses, the mathematicians repeatedly rephrased the statements, while explicitly

220 [Link]:indrum2018:174100
noting the everyday meanings of the words, until they had convinced themselves that
the statements were “essentially the same.’ This work was often well-situated in the
mathematical theory, as seen in the transcript below where the mathematician speaks
of “turning around” Statement A and “running through” sets of numbers.
Mathematician a: I tend to take statements like that [Statement A] and try to rephrase
them, so … for me, I would say, what does that actually say? It says that,
umm, whenever, umm, a positive integer n is congruent to, umm, is
congruent to 2 mod 3 then n cannot be a perfect square … like I … I try to
turn it around … I’m sitting here almost hesitant about whether or not I’ve
even done it correctly. But let me think … so, umm, let’s see, so there exists
no positive integer n such that these two things are true … so that’s … what
is that the same thing as saying, it’s saying that, umm, if you ran over the
positive integers n which were congruent to 2 modulo 3 you are never going
to hit a perfect square but then that’s what this is saying (point to Theorem
5), umm, if I think of for all positive integers n and this part is true, that n is
congruent to two modulo three, then I am never going to hit a perfect
square. So, … umm, actually, I think, umm, I would almost rephrase these
things as being equivalent but I am feeling a little bit hesitant about that.
Here, it is important to note that in addition to Mathematician a, three other
mathematicians expressed hesitancy with regard to their own reasoning; e.g., Dr. b
remarked “just doubting myself for some reason.” In each case, the mathematician
was asked “Do you have some doubts about your answer?” and all responded “No.”
Thus, the participants appeared to be applying inferences with a high degree of
(perhaps intuitive) certainty, while also doubting their own judgements of those
inferences. Finally, two other observations are of note. First, like the students, one
mathematician translated the statements into symbols. However, they immediately
pushed the paper away saying, “I am not going to do that.” Second, in the case of the
LTA-semantic approach, the mathematician translated both statements into Venn
diagrams (Figure 5) and then, by comparing the diagrams, reasoned through the task.

Dr. b: I’m going to draw some sets. … Statement A says to me the sets of n
mod(3) … congruent to 2 and perfect squares … (long pause)… are
disjoint. Right. There exists no positive integer … so this says for all
positive integers if n mod(3) is 2 expressed then … this will be expressed as
a containment and ….it’s not a perfect square … perfect square is on the
outside …….. and, umm, let’s see, if-then means that ….(long pause) that

221 [Link]:indrum2018:174100
 that set is inside that set or is it the other way … that implies that …(long
pause). Yeah, it looks like it (laughs quietly).

Figure 4: A Mathematician’s (Correct) Semantic-LTA Response

Study 2 Discussion
Study 2 aimed to explore students’ and mathematicians’ success with and approaches
to a metatheoretical task. The data demonstrate that post-instruction, novices
continued to struggle with the metatheoretical requirements of proof by contradiction
and gravitated towards syntactic-LTA approaches, while the mathematicians tended
towards semantic-MTA approaches. Furthermore, students’ remarks indicated their
use of syntactic-LTA approaches enabled them to avoid perceived content knowledge
weaknesses, whereas the mathematicians drew heavily on this knowledge to produce
proofs and explore concepts. The study also revealed a tendency among students who
struggled with the tasks; namely, a tendency to view quantifiers as appended symbols
rather than as variable-binding operators that act on open-sentences. Though far from
providing definitive evidence, the study contributes to the literature by highlighting
the logical complexities novices may encounter when producing or comprehending
proofs by contradiction, given the approaches they gravitate towards.

CONCLUDING REMARKS
One question raised by the studies is, why didn’t the students’ reasoning progress,
even after completing the Basic Set Theory and Logic course? Certainly, the lack of
progress may be due to poor instruction, an insufficient curricular treatment, or the
cognitive demand of the tasks. Turning to the curricular materials used, Chartrand,
Polimeni, and Zhang’s (2008) Mathematical Proofs: A Transition to Advanced
Mathematics, one finds little in the ways of support for the metatheoretical tasks
studied. This text includes an introductory chapter on logic with two subsections on
quantifiers. In these subsections, the quantifiers for all and there exists are defined
and discussed with regard to the variations of these phrases used in mathematics (e.g.,
for some; at least one, etc.). Neither are quantifiers discussed as variable-binding
operators nor is the phrase “there exists no” or the symbol ∄ mentioned. The same is
true in a latter chapter focused on proof by contradiction, where emphasis is placed
on moving from “for all” to “there exists” when proving by contradiction without
mention of what one must do once one determines something “does not exist.” Thus,
the lack of progress may be tied to an insufficient curricular treatment of the topic.
Turning to the mathematicians’ responses an alternative rationale for students’
persistent difficulties becomes evident. As discussed, most of the mathematicians in
Study 2 expressed a lack of confidence in their own reasoning, while none wished to
change their answer due to an intuition (i.e., “gut feeling”). Hence, it seems
reasonable to conclude that the task was cognitively demanding. Consequently, even
with instruction, we might expect low success rates among undergraduates, who are
at the early stages of the education and lack the content knowledge experts employed.

222 [Link]:indrum2018:174100
Lastly, in an interesting study of effective proof comprehension strategies, Weber
(2015) found mathematicians preferred students, “rephrase theorems in their own
words” and that students not use the strategy, “rewrite the theorem in first-order
logic.” These views reflect the practices of the mathematicians in Study 2, for none
used the latter strategy, while nearly all used the former. However, their rephrasing of
the statements relied on their extensive content knowledge; namely, as a tool for
inferring meanings. Hence, the findings raise questions regarding whether or not the
mathematicians would use these approaches if they were working with unfamiliar (or
difficult) content. Indeed, it seems that we must be careful inferring instructional
recommendations from the mathematicians’ practices. Many appeared to generate
inferences automatically – a practice that seemed to inhibit them from rationalizing
their judgements; as illustrated by the mathematicians’ repeated expressions of
hesitancy.

REFERENCES
Antonini, S., & Mariotti, M. A. (2008). Indirect Proof: What is specific to this way
of proving? ZDM- The International Journal on Mathematics Education, 40,
401-412.
Mariotti, M. A., Bartolini Bussi, M. G., Boero, P., Ferri, F., & Garuti, R. (1997).
Approaching geometry theorem in contexts: from history and epistemology to
cognition. In E. Pekhonen (Ed.), Proceedings of the 21st PME International
Conference, Vol. 1, 180-195.
Sierpinska, A. (2007). I need the teacher to tell me if I am right or wrong. In J. H.
Woo, H. C. Lew, K. S. Park, & D. Y. Seo (Eds.), Proceedings of the 31st
conference of the international group for the psychology of mathematics
education (Vol. 1) (pp. 45–64).

Wu Yu, J., Lin, F., Lee, Y. (2003). Students’ understanding of proof by
contradiction. In N.A. Pateman, B.J. Dougherty, & J. Zilliox (Eds.),
Proceedings of the 2003 joint Meeting of PME and PMENA, (Volume 4, pp.
443-449). Honolulu, Hawai’i: USA.
Weber, K. (2015). Effective proof reading strategies for comprehending
mathematical proofs. International Journal of Research in Undergraduate
Mathematics Education, 1(3), 289-314.
Weber, K. & Alcock, L. (2009) Proof in advanced mathematics classes: Semantic
and syntactic reasoning in the representation system of proof. In D. Stylianou,
M. Banton, and E. Knuth (Eds.) Teaching and Learning Proof Across the
Grades: A K-16 Perspective. Routledge: New York, NY.

i
An alternative conclusion is the mathematical theory is inconsistent. This progression, therefore,
assumes a stance where inconsistencies are rejected. This issue is beyond the scope of the paper.
ii
The proposed definitions of syntactic and semantic are heavily influence by those of Weber and
Alcock (2009) but are not identical for they are not tied to the representational system of proof.
iii
In American mathematics texts ~ symbolizes “not.” The symbol ¬ may be more familiar to some.

223 [Link]:indrum2018:174100
 Students’ problems in the identification of subspaces
in Linear Algebra
Yael Fleischmann1 and Rolf Biehler1
1
Universität Paderborn, Institut für Mathematik
[Link]@[Link], biehler@[Link]
The goal of the study presented in this paper is the investigation of students’
problems with exercises concerning central topics of linear algebra courses at
university level. We present the results of our analysis of students’ work on an
exercise about subspaces of . We evaluated the written solutions of the task
as well as transcripts based on videos taken of student groups working on the
problem. We identified and classified descriptions of vector spaces and sub-
spaces that varied widely and demonstrated highly different skills in working
with geometric or formal algebraic objects. We analyzed how far students could
progress in a complex reasoning process, and identified those steps in the rea-
soning process on which students needed support to continue.
Keywords: Linear algebra, vector space, subspace, proof, tutorial groups.
PURPOSE AND BACKGROUND
Problems in teaching and learning of linear algebra have a long history in many
countries (Dorier & Sierpinska, 2001). Frequently, the abstract character and the
formalism of mathematics that students have not been exposed to in school be-
fore is named as a central obstacle (a variety of studies are outlined and evalu-
ated in Dorier, Robert, Robinet & Rogalsiu, 2000). Since vector spaces are a
central part and moreover of special importance for almost all disciplines related
to mathematics at university, special attention has been paid to them (Dorier,
2000, Stewart, 2017). Generally, students do not develop a clear concept of vec-
tors at school level (Mai, Feudel & Biehler, 2017), and the more abstract ap-
proach to this subject taught at universities is described as being “out of reach”
by some students (Stewart, 2017). Wawro, Sweeney and Rabin (2011) investi-
gated concept images of subspaces in interviews with students and identified
recurring concept images, distinguishing between a subspace as a part of a
whole, a geometric object, and an algebraic object. The introduction of first
concepts in tutorial meetings in linear algebra, with a special focus on the behav-
ior and influence by the tutor, has been studied by Grenier-Boley (2014).
CONTEXT AND DESIGN OF THE STUDY
In this study, we investigated the problems of students shortly after their first
encounter with vector spaces and subspaces at university level. The participants
of our study were students with major mathematics or computer science, en-
rolled for bachelor of science or bachelor of education (for secondary school,
“Gymnasium”), most of them in their first semester. In our study, we collected
data from students working on tasks in groups during their tutorial group meet-

224 [Link]:indrum2018:174624
ings (1.5 hours), where the tutors were advised to answer questions but to only
intervene when the students had substantial problems to continue. The students
worked on exercises about the content of a recent lecture under the supervision
of a tutor. In this context, we assigned special tasks that we developed ourselves
together with the lecturer and his assistant, but we did not influence the course
design otherwise. We will report only on one of them in this paper. The course
can be considered to be typical for a beginners’ lecture in linear algebra, which
normally is rather abstract, and was given by an experienced lecturer. During
their tutor meetings, the students worked on our exercises on separate sheets that
we collected, scanned for later analysis and gave back to the students in the next
meeting without any grading or corrections. We gathered between 78 and 130
written works on each exercise. Moreover, we also took video recordings of
groups of 2–4 volunteering students working on these exercises. They worked
on the exercise under the supervision of a student tutor who was part of the re-
search team and familiar with our a priori analysis of the task. The experienced
tutor was advised to help the students if they struggle with the exercise in the
same way as she would do in an ordinary tutor group meeting. We were inter-
ested in identifying important didactic variables. The results obtained by analyz-
ing the first implementation of the exercise about vector spaces are currently be-
ing used for designing a second implementation in the course Linear Algebra I.
The task for students in our study and preliminary research questions
In this paper, we concentrate on an exercise about subspaces and vector spaces
(see figure 1) that was part of the exercise sheet during week 7 of the course,
immediately after the notion of subspaces had been introduced. Students are
taught analytical geometry and linear algebra at school level, where vectors are
introduced as tuples (or classes of arrows), but they do not as a rule have a clear
concept of a vector (cf. Mai, Feudel & Biehler, 2017). Students know equations
of planes and lines in and , without considering them as subspaces, be-
cause this notion is not taught at school level.
The following exercise was designed in order to provide two different kinds of
learning potentials (as described in Gravesen, Grønbæk and Winsløw, 2016):
1. Linkage potential: In part a) to e), our intension was to motivate the students
to activate their school knowledge concerning the description of geometric
objects using equations; we hoped that they would recognize the sets as de-
scriptions of lines, points, parabolas etc., and connect this knowledge with
the new concepts of vector spaces and subspaces.
2. Research potential: Part f) of the exercise was created in order to engage the
students in a research-like activity. Even if achieving a complete solution
seemed unrealistic for most of them, we were interested in how the students
would approach this open question. They had to formulate a hypothesis and
use abstraction to identify and construct subspaces. The exercise can be seen

225 [Link]:indrum2018:174624
 as a “mini research project” that differs in type from standard exercises.

Figure 1: Exercise on subspaces (translated from German)

The parts a) to e) can be solved by a formal check whether the properties of sub-
spaces are satisfied by the provided sets. As this was learned in the previous lec-
ture, this is a standard task. Geometric ideas are not necessary, but we hoped
that students may do geometric interpretations of the sets to develop a geometric
meaning of subspaces and non-subspaces of Task f) is different, because this
is the first time that this type of question is asked. Students may use the results
from a) to e), that have provided examples and counterexamples of subspaces, to
find the zero space, all lines through the origin, and the whole as subspaces
and give reasons why they are subspaces on some level. The challenging ques-
tion is whether or why these are all subspaces of . Research questions con-
cerning f) are: How many students identify the zero space and the entire as
subspaces? Are all lines through the origin identified as subspaces? Which ar-
guments do students provide for considering a line through the origin as a sub-
space of ? How do they reason, when exploring, whether there are more sub-
spaces in or whether they have already found all?
We were also interested in the sources of knowledge students used, such as their
results on a) to e), parts of the lecture, or geometric interpretations related to
their school knowledge, and concerning the videographed tutorial sessions,
which kind of support by the tutor they can use in their reasoning process.
METHODOLOGY AND DATA COLLECTION
For the analysis of the written work of the students, we followed the method of
Biehler, Kortemeyer and Schaper (2015), by comparing each solution with the
so-called student expert solution (SES), which is a sample solution based on the
idealized actual knowledge of the students at this point of the lecture. Moreover,
the student expert solution contains additional meta-information about the solu-
tion, for example, several alternative opportunities for solutions and explicitly
written-out learning objectives. In relation to Brousseau’s theory of didactic sit-
uations, this method can be seen as a special approach to the development of an
a priori analysis. We evaluated the written work in a two-step procedure: In a

226 [Link]:indrum2018:174624
first step, we categorized the solutions by correctness and collected peculiarities
and mistakes. Based on this and the SES, we developed a detailed coding system
for deeper analysis. The recorded videos have been transcribed in order to allow
a detailed qualitative analysis.
A PRIORI ANALYSIS OF THE TASK
In the lecture, the definition of vector spaces was given in a typical traditional,
abstract way. The zero space and the vector spaces (trivial vector space over
) and ℂ (the latter together with component-wise addition and multiplication)
had been presented by the lecturer as first examples. Apart from this introduc-
tion, the students had only seen the following (relatively abstract) non-trivial
examples for vector spaces in the lecture: (VS1): , the “standard vector
space”, where is any field, with n , including the definition of addition and
the scalar multiplication (component-wise), (VS2): , the vector space of se-
quences over the field , with the component-wise operations. Subsequently,
subspaces of vector spaces had been defined to be subsets of vectors spaces that
are vector spaces themselves with respect to the same operations. Following
this, they had learned that a sufficient criterion for proving that a nonempty sub-
set of a vector space over the field is a subspace is to prove that
ly and secondly .
As examples of subspaces, the trivial subspaces and were nominated
without proof. Moreover, for both vector spaces (VS1) and (VS2), there was an
abstract example for a subspace given, and we state the first one of them here
since it will be of use for our later analysis:
(S1) The set is
the solution space of a homogeneous linear equation system . It
was shown that this set is indeed closed with respect to addition and scalar mul-
tiplication and is a subspace. Note that this example can be applied to , if we
choose . The subspaces in are the lines through the origin expressed by
linear equations. This interpretation could be done by students on the basis of
school knowledge. The lecturer did not provide this specialization himself.
For our later analysis, the following distinction is central. All provided examples
have in common that sets are characterized by equations (subspaces defined by
relations). In contrast, 1-dimensional subspaces could also be defined by explicit
construction: for instance for any : . The lat-
ter way of defining subspaces was not yet a topic of the lecture, which will turn
out to be an obstacle for some students. Moreover, the students had not seen any
geometric interpretation or visualizations of vector spaces or subspaces, in par-
ticular no (concrete) examples of subspaces in In the following, we will give
an overview about possible approaches and steps to part f).
Step 1: Find some subspaces. With the knowledge from the lecture, the trivial
subspaces (the zero space and can be named. To find nontrivial subspaces,

227 [Link]:indrum2018:174624
one can identify again the set of the previous part a) of the exercise as a sub-
space. Starting with this set, one could generalize from numbers (like 2 and 1, as
chosen in part a)) to a general form with coefficients, and give the set
with not both being zero. Instead, if
one abstracts from the mathematical language used in the exercise before, these
sets could also be expressed constructively as
Supported by Dorier, Robert, Robinet and Rogalsiu (2000), we expected dif-
ficulties to translate the relational representation into the constructive represen-
tation and vice versa. Alternatively, with the knowledge from the lecture, one
could apply the example (S1) given in the lecture to the space , and describe
the subspaces in terms of the solutions of homogeneous linear equation systems.
This reasoning can be done just algebraically. It could also happen that students
use geometrical terminology concerning lines through the origin.
Step 2: Verification of the subspace properties. In order to reason why the sub-
sets given in step 1 are subspaces, one could either refer to the solution of part a)
of the exercise or (for the trivial subspaces and in case of the use of the solution
spaces of homogeneous linear equation systems) to the lecture. In case of a ge-
ometric description (“lines through the origin”), either geometric or algebraic
arguments have to be provided to verify the subspace properties.
Step 3: Why are these all subspaces? The final challenge is to reason if and why
all subspaces of have been found. This can be done algebraically, but we did
not expect our students to complete this reasoning process in the given time,
since it requires a development of several successive algebraic arguments. Based
on their school knowledge, the students could recognize the descriptions of ge-
ometric objects by equations in part a) to e) and abstract from the previous re-
sults, leading to the conclusion that lines through the origin are subspaces, but
no other lines, single points or other collections of points. At this point, a suc-
cessful reasoning based on school knowledge could be done constructively,
based on geometric arguments. Trying to construct “bigger” subspaces than just
the lines through the origin, a student could build the union of two different lines
and check whether this set is a subspace. Alternatively, he or she could try to
find the minimal subspace that includes one line through the origin and an
additional point not lying on this line. He or she could come to the conclusion
that this has to be the whole . A formal argumentation here is that every point
can be represented as a linear combination of a point from the line
and , but even if the student does not come to this conclusion at this point,
he or she could have the idea to consider the line through the new point to-
gether with the original line, and therefore check this new set for the subspace
conditions. He or she could check the closure of addition or come to the idea
that further points have to be added to the union in order to get a subspace. Since
this type of reasoning seemed to us more likely to be achieved with the previous
(including school) knowledge of the students, the tutors in the normal tutor
group meetings as well as the tutor in the video study were advised to guide the

228 [Link]:indrum2018:174624
students along this reasoning process if they struggled in approaching the prob-
lem. Based on this sample solution process description, we tried to answer the
following research questions in our analysis:
1. How far in this three-step process would the students come when they work
on this exercise? Would they even be aware of the need to do step 2 and 3?
2. Would they favor one of the described approaches to the problem (geomet-
ric, algebraic), and would they use the constructive or the relational way to
describe the 1-dimensional subspaces? Would they approach step 3 in a con-
structive way, building up subspaces starting with just one point, as de-
scribed above, or would they find other ways (purely algebraic?)?
3. Finally: Would they recognize that parts of exercise f) could be solved by an
application of the example (S1) given in the lecture?
Since we posed the question in part f) in a relatively weak phrasing, we could
not expect the majority of students to give a fully structured, formal reasoning in
this exercise, in particular for the steps 2 and 3. But we were interested if the
exercise itself would stimulate the students to give reasons for their answers and,
in particular, how they would argue in this case.
RESULTS
To find answers to our questions, we analyzed the written works as well as the
video recordings of the students working on part f).
Work on part f): Written exercises
From the written works of 116 students on this exercise, just 48 handed in solu-
tions for part f). This is most likely due to the fact that the time was very limited,
so many students just did not come to part f). We analyzed their work with re-
spect to the three steps of the solution as described in the a priori analysis.

Trivial 33
subspaces
Zero Space 32

1- Solution with any description of the 1-dimensional subspaces (some students 33

dimensional used more than one description)
subspaces
- Relational description: 24

- Constructive description: 4

- Geometric descriptions: “line containing zero”, “line through origin” 12

Table 1: Frequency of the nominations and descriptions of the subspaces

Step 1: Which subspaces do they find? How do they describe them? Do they use
previous parts of the exercise or name the set considered in part a)?
The results in table 1 were collected by counting how often the three types of
subspaces were mentioned in the solutions. Hereby, each notion of a subspace

229 [Link]:indrum2018:174624
counted, as long as it was clear enough to denote the required set. How did the
students describe the 1-dimensional subspaces? We distinguished between “ge-
ometric” descriptions, using expressions like “line containing zero”, “line
through the origin”, relational descriptions using a set like
or something mathematically equivalent (see a priori analy-
sis for a definition of this category) or constructive descriptions like
. Some students used more than one description in their
solution. Apart from this, it was interesting to see that only 8 students did men-
tion any part (mostly a)) of the previous exercise in part f). It is not clear if those
who could not give any (nontrivial) subspace actually never recognized that the
set M1 from part a) is a subspace (since the word “subspace” was not used in part
a)), or if they just forgot about it before they started with part f). Moreover, it is
interesting that the trivial subspaces, which we expected the easiest to find, were
not nominated more often than the 1-dimensional subspaces. We were also sur-
prised to see that only 2 of the students did refer to example (S1) (see a priori
analysis) from the lecture, concerning the solution spaces of homogenous sys-
tems of linear equations.
No rea- Incorrect reasoning/ Partial rea- Complete reasoning
soning unclear approach soning

Step 2 34 5 7 2

Step 3 35 6 6 1

Table 2: Frequency of reasoning in part f)

Step 2: Do they show that the given sets are subspaces? How do they argue?
Most students did not give reasons (see table 2 for results), but within those who
did, we distinguished between approaches that did not go in the right direction
(for instance students just answered by listing all properties of a subspace with-
out proving them or claimed that it was “clear” that the spaces are subspaces),
students who did give a correct approach or a partial proof (they mentioned that
closure must be proved, but did not, or just checked the addition or the scalar
multiplication, or just checked an example etc.) and complete solutions with full
reasoning (using example (S1) from the lecture in both cases).
Step 3: How do they reason that they found all subspaces of ?
Within the 13 solutions that had some kind of reasoning (see table 2 for results),
we distinguished again between unclear or vague approaches to reason the com-
pleteness of the given list of subspaces (for instance the statement “there are no
other possible, because one cannot multiply vectors”), promising but incomplete
approaches (some students gave reasons why lines not going through the origin
cannot be subspaces, but did not consider other subsets, or just discussed the
closure of one of the operations) and complete reasoning (just one case, again
applying example (S1) from the lecture).

230 [Link]:indrum2018:174624
Work on part f): Video recordings
We give a summary about three groups of students that we recorded during their
solution process on part f) under the perspective of our research questions. Due
to limitations of space, we cannot document the method used to analyze the
transcripts and the students’ interaction in more detail.
Group 1: The first group tried to find subspaces by systematically going through
the list of properties, and found the zero space to fulfill them. Then, they re-
membered the set proven to be a vector space in part a) and generalized it to a
set of the form after a discussion
whether the coefficients are arbitrarily exchangeable without harming the sub-
space conditions. They discussed the closure of the vector space operations in
this set, but referring to the proof they had given in part a), they convinced
themselves quickly that there was nothing else to prove. After this, they also
identified the full space since there is no claim for a subspace to be a proper
subset. At this point, they were asked by the tutor if and why they found all sub-
spaces now. They had the idea to consider the set and, referring to their
knowledge about groups, discussed the closure of operations on this set before
they could finally rule it out to be a subspace by the fact that the scalar multipli-
cation with elements from is not closed on . The tutor then asked them to
consider the set geometrically. They start to consider the tuples of coeffi-
cients in the plane instead of the equation . With another
hint from the tutor, they found out that the set whose elements are de-
scribed by the equation denotes lines in the plane, and discussed
the closure of the operations for these lines. The students did not develop an idea
themselves to give arguments why they had found all subspaces. However, the
students were able to follow the geometric constructive reasoning of the tutor
(see a priori analysis).
Group 2: The second group came up with the idea to apply example (S1) from
the lecture. After some discussion and a bit help from the tutor, they found that
the subspaces defined there are the solutions of one or more linear equations,
each having two coefficients. The central difficulty for them was to see that the
number of coefficients is fixed to 2, but there could be an arbitrary number of
equations in a system of linear equations that is still defined in It was a real
discovery later that provided descriptions as are provided in . Up to
this point, they did not consider the trivial subspaces at all. They struggled a bit
to write down the concrete subspaces they could find this way in terms of alge-
braic expressions, but managed it with some help from the tutor. Asked whether
they found all subspaces, they did not develop the idea to consider the spaces as
lines in the plane on their own, but after the tutor came up with this idea, they
were able to work with this concept after a short phase of orientation in which
they convinced themselves that the geometric objects stand for the same sub-
spaces they worked with before. Just at this point, they identified the trivial sub-

231 [Link]:indrum2018:174624
spaces too. Step 3 (to reason that all subspaces were found) was only solved
with tutorial support (similar to group 1).
Group 3: The third group used the previous parts a) to e) in their reasoning and
started with the subspace found in a), but immediately identified this set to de-
scribe a “line through the origin”, which gave birth to a generalization to all
lines through the origin. They continued to orally communicate in geometric
terms, but decided to write down the set using the relational algebraic expression
. They named the trivial subspaces
without further comments. The proof given in part a) sufficed for them for a rea-
soning of step 2, and they started to discuss step 3 quickly. They used references
to part b) to e) to rule out other types of possible subspaces. The group thought
they had finished at this point. It was the tutor who pointed out that step 3 was
not yet satisfactorily answered. Different from the other groups, they took up the
tutors input to construct other subspaces geometrically and in order to find out
that such subspaces have to be equal to . With some minor help from the tu-
tor, they finished this step quickly, needing much less time for the full task than
all other groups.
CONCLUSION AND DISCUSSION
As a result, we can state that students at this point in their studies were able to
find and describe (using varying descriptions) subspaces of , but the question
to find all subspaces was a serious obstacle for the students. Moreover, the step
to translate the algebraic description of the subspaces into a geometric view,
where reasoning could be done with less formality, was a further obstacle for
them, since they seemed not to connect or apply their geometric knowledge
from school to the new problem.
It seems like a geometric approach to this kind of problems is not a natural, au-
tomatic behavior of students at this point of their education. This result resonates
with the observations from Wawro et al. (2011), who stated that intuitive geo-
metric notions can be the preferred approach of first year students to the con-
cepts of subspaces, but also cause problems if they their geometric intuitions are
inconsistent with the formal definition. It is worth pointing out that our students
did not, in opposition to the results of Wawro et al., automatically identify (often
mistakenly, if there was no respect to a necessary embedding) the as a sub-
space of the . A possible explanation for this result is the fact that our results
were obtained shortly after the introduction of subspaces in the lecture, where
Wawro et al. interviewed their students when they already have had more time
to develop a concept image of subspaces, including some misconceptions.
Most students did not connect the different parts of the exercise, appearing in
different “languages” (like the sets in the parts a) to e) and the open question in
part f)) to solve the problem in f). With some help, especially with the request to
consider the sets geometrically, they were able at least to understand reasoning
on this basis, and some students actually could even give proofs or approaches

232 [Link]:indrum2018:174624
to proofs on their own. We came to the result that the students needed more
guidance and preparation to solve this problem, and in particular support that
helps them to deal with each step and even sub-step of the solution of the prob-
lem in part f). In our subsequent study in winter term 2017/2018, we are investi-
gating if explicit indications in a) to e) to consider the sets geometrically and a
rephrasing of part f), splitting it up into more explicitly described steps, have a
decisive influence on the students’ ability to solve task f).
REFERENCES
Biehler, R., Kortemeyer, J. & Schaper, N. (2015). Conceptualizing and studying
students’ processes of solving typical problems in introductory engineering
courses requiring mathematical competences. In K. Krainer & Nad’a
Vondrová (Eds.), Proceedings of the CERME 9 (pp. 2060-2066). Prag:
Charles University in Prague, Faculty of Education and ERME.
Dorier, J.-L. (2000). Epistemological Analysis of the Genesis of the Theory of
Vector Spaces. In: Dorier, J.-L. (Ed.). On the Teaching of Linear Algebra.
(pp. 1-81). Dordrecht: Kluwer.
Dorier J.-L., Robert A., Robinet J., Rogalsiu M. (2000). The Obstacle of For-
malism in Linear Algebra. In: Dorier, J.-L. (Ed.). On the Teaching of Linear
Algebra (pp. 85-124). Dordrecht: Kluwer.
Dorier, J.-L., & Sierpinska, A. (2001). Research into the teaching and learning
of linear algebra. In: D. Holton (Ed.), The teaching and learning at university
level - an ICMI study (pp. 255-273). Dordrecht: Kluwer.
Frovin Gravesen, K., Grønbæk, N. & Winsløw, C. (2016). Task Design for Stu-
dents’ Work with Basic Theory in Analysis: the Cases of Multidimensional
Differentiability and Curve Integrals. International Journal of Research in
Undergraduate Mathematics Education, 3, pp. 9-33.
Grenier-Boley, N. (2014). Some issues about the introduction of first concepts in
linear algebra during tutorial sessions at the beginning of university. Educa-
tional Studies in Mathematics, 87, pp. 439-461.
Mai T., Feudel F. & Biehler R (2017). A vector is a line segment between two
points? - Students’ concept definitions of a vector during the transition from
school to university. Paper presented at the CERME 10, Dublin.
Stewart, S. (2017). School Algebra to Linear Algebra: Advancing Through the
Worlds of Mathematical Thinking. In: Stewart, S. (Ed.), And the Rest is Just
Algebra (pp. 219-233). Springer.
Wawro, M., Sweeney, G. & Rabin, J. (2011). Subspace in linear algebra: inves-
tigating students' concept images and interactions with the formal definition.
Educational Studies in Mathematics, 78, pp. 1-19.

233 [Link]:indrum2018:174624
A TDS analytical framework to study students’ mathematical activity
An example: linear transformations at University
Marc Lalaude-Labayle1, Patrick Gibel², Isabelle Bloch² and Laurent Lévi 1
1
Laboratory for Mathematics and its Applications (UMR 5142 CNRS), France,
[Link]-labayle@[Link];
²Lab-E3D Laboratory of Epistemology and Didactics of disciplines,
[Link]@[Link]; [Link]@[Link], University of
Bordeaux
Our research focuses on the teaching of linear transformations in “Classes
Préparatoires aux Grandes Écoles”. The theory of didactical situations, jointly with
Peirce's semiotics, constitute the main theoretical framework of our works and allow
us to analyse student's reasoning in situations of oral evaluation. We illustrate the
use and utility of this framework with the study of student’s mathematical activity
when they are faced to situations involving complex concepts such as linear
transformations in polynomial spaces.
Keywords: linear algebra, semiotics, theory of didactical situations, transition,
tertiary level

INTRODUCTION
Our work deals with a double object: linear transformations as a structuring concept
of the teaching of linear algebra and a particular institution at the undergraduate
level, the Classes Préparatoires aux Grandes Écoles. Grounded on didactical
motivations, our epistemological analysis allows us to exemplify the crucial role of
linear transformations for the emergence of linear algebra concepts. This
epistemological part of our research, mainly based on the works of Dorier and
Moore, leads us to the use of Peirce’s semiotic and to enrich the analytical model of
Bloch and Gibel (2011) which is rooted in the Theory of Didactical Situations
(TDS). The purpose of this paper is twofold:
 give some responses to the following question: which reasoning forms are
actually produced by a student during the different stages of a situation of oral
evaluation?
 show the utility of our framework to analyse the signs and arguments
produced and thus take part to the development and enrichment of this model.
Within the French didactic tradition, we remind briefly the theoretical tools used in
the elaboration of the framework. Then, we expose our model, using the terminology
of Bloch and Gibel. Equipped with these tools, we use the model to analyse the
arguments produced by students. But, at first, we succinctly introduce the institution
of Classes Préparatoires aux Grandes Écoles by highlighting the differences with the
University, especially regarding the transition phenomenon with secondary level.

234 [Link]:indrum2018:174109
I. THE CLASSES PRÉPARATOIRES AUX GRANDES ÉCOLES: A FRENCH
POST-SECONDARY LEVEL INSTITUTION
The Classes Préparatoires aux Grandes Écoles, which we can translate as Higher
School Preparatory Classes, are part of the French tertiary education system for over
two centuries. They consist in two really intensive years which act as a preparatory
course to train undergraduate students for their further enrolment in one of the
French graduate schools called Grandes Écoles, such as École Polytechnique, École
Normale Supérieure, École des Hautes Études Commerciales, also known as HEC
School of Management … The enrolment in one of these Grandes Écoles depends on
the rating obtained in national competitive and demanding examinations.
We summarize the main differences between the Classes Préparatoires and the
University in the following table. Thus, we highlight some facts that Winslow (2007)
showed to think about the study of the transition phenomena from the secondary to
the post-secondary level.
Classes Préparatoires University
Full-time teacher Part-time teacher, part-time researcher
One teacher by class Several teachers for one class
One class per teacher Several classes for one teacher
Non adoption of semesters Adoption of semesters
Common national curriculum Local curriculum
Non degree course Degree course
In High School At University
Class councils No class council
Report cards No report card
Selection of students No selection of students

Table 1: Comparison of CPGE and University

As Winslow (2007), Castela (2011) and more recently Farah (2015) wrote, these
differences have deep didactic implications, relative to the theoretical knowledge and
praxeologies, to the problem solving approaches, to the evaluations and to the
personal homework just to name a few. In our experimental work, we studied some
arguments produced by second year students from one of these Classes Préparatoires
during an oral evaluation on linear algebra.
As noted by Bloch and Gibel (2016) for calculus, in order to recognize and analyse
the reasoning forms actually produced by a student,
it is necessary to classify the objects, signs and reasoning processes they have to cope
with during resolution of calculus problems. (Bloch & Gibel, 2016, p. 44)

235 [Link]:indrum2018:174109
II. A MODEL TO ANALYSE STUDENTS’ REASONING
We then need a tool to modelize the students’ productions, more precisely to
identify, while in a mathematical situation, which is the knowledge they rely on
during their activity; we want to identify the structure and functions of students’
reasoning processes in this situation. Briefly, this model should allow us to seize the
complexity of the reasoning processes a student has to cope with during the
resolution of mathematical problems. By reasoning process, we mean valid or
erroneous ones according to the work of Gibel (2015) about reasoning.
This model takes its origins in the Theory of Didactical Situations (Gonzalez-Martin,
Bloch, Durand-Guerrier, and Maschietto 2014) and in the semiotics of Peirce. We
briefly recall the main theoretical elements of TDS and semiotic used in the
elaboration of the model.
The TDS theory
TDS relies on a two basic premises concerning the mathematical activity and the
learning of mathematical knowledge. For TDS, the mathematical activity consists of
distinct stages: a situation of action, followed by a situation of formulation and then
a situation of validation phase. To take the learning activity into account, TDS adds
two more stages: the phase of devolution and the phase of institutionalization. TDS
defines three fields to construct and analyse such situations. The theoretical field
which the domain of elaboration of fundamental situations relative to a knowledge.
The a priori experimental field which envisages a situation at a specific level of
teaching, taking into account the didactic repertory as defined by Gibel (Gibel,
2004). The a priori analysis of the situation, which checks if the conditions of
devolvement are fullfilled, takes place at this second level. The third field is the field
of eventuality where the situation is actually implemented. In short, TDS is a
didactical framework which tries to implement situations with adidactical parts and
offers tools to analyse the teaching and learning activities. These adidactical stages
allow students to face a heuristic phase of research and then, through a confrontation
to the elements of an adequate milieu, to test, validate or invalidate their conjectures.
The notion of milieu appears to have a central role. TDS organizes situations with up
to seven logically successive phases, but in our work we will mainly work with three
of them: a heuristic one, grounded in a problem, and a formulation and validation
one, composing the adidactical moments of the situation, and then the
institutionalization by the teacher or with his/her help, which is the didactical
moment. The dynamic of the nesting of the situations with the paired levels of milieu
illustrates the dynamic of the learning processes involved. The following chart
(Bloch 2006, Bloch and Gibel 2016) sums up the levels of milieu paired to the
different phases of situations corresponding to the experimental situation
M1 Didactical milieu E1: reflexive subject P1: P. planner S1: situation of
project
M0 Learning milieu : E0: generic student P0: professor S0: Didactical
institutionalization teaching situation

236 [Link]:indrum2018:174109
M-1 Reference milieu: E-1: The subject as P-1: Professor S-1:Learning
Formulation and validation learner Regulator situation
M-2 Heuristic milieu: E-2: The subject as P-2: P devolves S-2: Situation of
action, research an actor and observes reference
M-3 E-3: epistemological S-3: Objective
Material milieu subject situation

Table 2: Structuration of the didactical milieu

The levels M-2 and M-1 are the ones that will allow us to identify, describe and
analyze the elements (signs, processes …) associated to the emergence of an
argumentation within the proof process. More precisely, as Bloch and Gibel (2016)
write
The place where we hope to see the expected reasoning processes appear and develop is
located at the articulation between the heuristic milieu and the reference milieu. (ibidem,
p.46)
Thus, the TDS theoretical framework allows us to consider not only the reasoning
processes produced, but their functions within the situation and the levels of
argumentation they rely on: it already gives us a glimpse of the multidimensionality
of our model.
The semiotic tools
To take into account the semantic dimension, i.e. the meaning of the signs and
arguments produced, with certain accuracy, we need semiotic tools. The signs a
student produce during an adidactical situation, whether formal or informal, are the
only observable phenomena that can sustain our semiotic analysis: roughly speaking,
the signs produced (syntactic aspect) are in relation to an object (semantic aspect)
creating a instantiated sign (pragmatic aspect) relative to the milieu, the didactical
repertoire and the repertoire of representations as defined by Gibel (2015). This
triadic relation, linking the sign, the object and the student’s instantiated sign, led us
to use Peirce’s semiotics. In this paper, we only analyse the relation between a sign
and its object, its content. But in his thesis, one the authors (Lalaude-Labayle 2016)
conducted a semiotic study, relying on the full Peirce’s triadic classification of signs
as ‘put in algebra’ by Marty (1990). Applying the universal categories to the relation
between a sign and its object refers to the notions of icon, index and symbol,
describing the abstraction level of this relation. An icon is a sign which stands for an
object because of its physical resemblance: a drawing, a triangular matrix
represented with triangles … An index is a sign physically connected to its object:
the columns of the pivots in a reduced echelon form of a matrix M is an index that
points to the basis of Im(f), where M is canonically associated to f … A symbol is a
sign that refers to its object by virtue of a law: ker is a symbol whose object is the
concept of kernel of a linear transformation.

237 [Link]:indrum2018:174109
The logical inferences
In our work we understand the term ‘reasoning’ in it broadest sense. More precisely,
by reasoning we mean a sequence of representations, following some intern and
potentially explainable rules that lead to reach some explicit goal. Postulating, as it is
the case within the TDS theory, some rationality of the students, we need to define
these rules or inferences. Peirce saw mathematics as the science of drawing
necessary conclusions, studying what is and what is not logically possible. But, since
one does not think about logical propositions but about and with signs, Peirce
broadened the notion of inference. He then distinguishes three kinds of rational
inferences: abduction, induction and deduction. Deduction, or necessary reasoning,
deduces a proposition B from a proposition A, where B is a formal logical
consequence of A. Induction goes from the particular to the general; it allows
inferring B from A. Abduction allows inferring A as a probable explanation of B. So,
deduction proves that something must be, induction shows how something
effectively operates and abduction suggests that something could be.
The framework to analyse students’ productions
To analyse students’ processes of reasoning, Bloch and Gibel (2011) develop a
multidimensional model. They focus their didactical analysis on three main axes.
The first axis is related to the level of milieu and so to the phase of the situation in
which the student produces his/her reasoning (cf. Table 2). The second axis of the
model is linked to the notions of didactical repertoire, of organizational system and
of a repertoire of representations. It studies the functions of the reasoning produced
and is in close relation to the first axis. As Bloch and Gibel (2016) state it, they aim
at
linking these two axes, showing how the reasoning functions are linked specifically to the
levels of milieu and how these functions also manifest these levels of milieu. (ibidem,
p.47)
Semiotic analysis of observable signs constitutes the third axis of the model. Marc
Lalaude-Labayle enriched the model by adding a fourth axis about the forms of
inference applied. This logical axis links the second and third axis by setting out and
clarifying the organization of the reasoning signs and their functions. This fourth
axis helps to ‘make visible’ the organization within the system of representations and
its actualization. We sum up this model in the following table
Milieu M-2 Milieu M-1 Milieu M0
Heuristic level Formulation, validation Institutionalization
R1.1 SEM R1.2 SYNT/SEM R1.3 SYNT
Nature and - Decision of a working - Generic calculations - organization of the
functions of frame (DOO) - Formulation of signs
reasoning - Decision of underpinned - formalization and
transformation (semiotic conjectures (right or certification of
register) wrong) validations

238 [Link]:indrum2018:174109
 - Decision of calculation - Decision on a - Formalization of
- Heuristic tools; errors mathematical object proofs within the
- Exhibition of an mathematics involved
example /a counter ex. theory
- pattern
research/identification
Level of use R2.1 SEM R2.2 SYNT/SEM R2.3 SYNT
of symbols Icons or indices Local or more generic Formal and specific
depending on the context arguments: indices, arguments related to
(schemas, intuitions…) calculations the chosen frame
Actualisation R3.1 SYNT/SEM R3.2 SYNT/SEM R3.3 SYNT
of the - Ancient knowledge Enrichment at the - Formalized proofs
repertoire - Enrichment at the argumental level: - Signs within the
heuristic level(patterns, - statements relevant theory
praxeologies ...) - organizational system - theoretical elements
Forms of R4.1 R4.2 R4.3
reasoning - deductive - deductive - deductive
- inductive - inductive
- abductive
Table 3 – A model to analyse situations
SEM signifies that the formulations are made on a semantic mode whereas SYNT is
for syntactic mode. This model emphasizes the fact that the mathematical activity,
with its reasoning processes, appears in the heuristic and reference milieu (cf. Table
2). These two milieus, and the articulation between them, will thus be of particular
interest for our work, even if the situation is an ordinary one.
Let us insist on the fact that the use of this model relies on a precise a priori
mathematical analysis of the situation and of its components, e.g. the problem to be
solved. Within the TDS this step appears to be necessary to clarify the didactical a
priori analysis.

III. A PRACTICAL USE OF THE MODEL IN A LINEAR ALGEBRA ORAL

EXAMINATION
We analyse the productions of a second year student of Classes Préparatoires in the
context of an oral examination. The students of the Classes Préparatoires, by group
of three, pass such an oral exam once every two weeks. It lasts one hour during
which the teacher asks to the three students to solve different mathematical
problems. They work simultaneously and individually on a large blackboard on the
problem they just discovered, in front of the teacher. The teacher helps the students
by giving advices or clues. Taking into account the writings noted on the chalkboard,
she/he can ask for some explanation or clarification. This can give rise to an oral or
written answer and possibly to some discussion to deepen and enrich the repertoire

239 [Link]:indrum2018:174109
of the representations of the student. The oral exam we analyse deals with linear
algebra, and more specifically linear transformations and matrices.
The student is confronted to the following problem:
Let n be a integer, greater than or equal to 2 and let φ defined on Rn[X] by: for all
polynomial P from Rn[X], φ(P)=P(X+1)-P(X).
1. Show that φ is an endomorphism of Rn[X].
2. Is φ injective? Surjective? Bijective?
In France, Rn[X] symbolises the vector space of real polynomials of degree at most
n.
As briefly explained earlier, our model provides a framework for investigating
mathematical and didactic activities in terms of milieu, focusing on the reasoning
processes, signs and their dynamics and on the conditions that enable their
development during the situation, be it ordinary or not. As is done within TDS, our
didactic analysis is divided in several stages: a detailed and strucured a priori
mathematical and didactic analysis, enriched with a specific a priori analysis of the
reasoning involved; follow then an a posteriori analysis organized in our model.
A priori analysis
From the mathematical point of view, showing that φ is an endomorphism of Rn[X]
can be approached in different ways, engaging several frames: indeed the stability of
Rn[X] under φ can be done in a purely algebraic frame using the degree and the
i
composition rule, in a functional frame using the decomposition P= ∑ ai X or in an
algebraic frame using the linearity and showing that forall integer i between 0 and n,
φ(Xi)∈Rn[X]. To study the injectivity and surjectivity of φ, the student can again
choose between several frames and several registers of semiotic representation
(Duval, 2017): she/he can use an algebraic frame with an example φ ( 1 )=⃗0 and then
applying the theorem linking rank and kernel dimension; but she/he can also try to
find precisely Ker φ, that is find a basis. To do this, she/he can use the functional
decomposition of polynomials, she/he can solve a linear system or she/he can
determine the matrix of φ in the canonical basis of Rn[X].
From a didactic point of view, the situation studied here is said to contain an
adidactic dimension. Most of the actions, of the frames and of registers of semiotic
representation are devolved to the student. So, a first difficulty that occurs for the
student can be the control she/he has on the objects involved: a circular application
of the definition of linearity of φ to prove its linearity, the complexity of formulas to
write down φ(P) for a general P, the non operability of φ with wrong calculations of
φ(1), φ(X), φ(X²) for example. For the injectivity of φ, the student can “forget” the
structure of the space Rn[X] he is working on and try to check whether φ(P)=φ(Q)
implies that P=Q. All the reasoning processes and objects involved are part of the
didactic repertoire of the class the student belongs to.

240 [Link]:indrum2018:174109
A posteriori analysis: analysis of a student’s production
In the following, we extract some translated excerpts of the student’s answer, the
whole solution can be found in Lalaude-Labayle (2016).
To solve the first part of the problem, which is an ordinary situation in undergraduate
level, the student is mainly confronted to the milieu of reference, articulating objects
and processes involved in his repertoire of representations. Here the heuristic milieu
is not really requested. He starts by showing the stability Rn[X] under φ and that
writes changing X with X+1 doesn’t change the degree of P. He then proves the
linearity in an algebraic frame. Doing so, he makes a formalization of proofs within
the required theory and thus reaches the level R1.3. The semiotic analysis shows
that he uses generic arguments (R2.2) and more formal one (R2.3). These arguments
and signs don’t give any hint to how φ operates on Rn[X]. Its argumentation
validates its use of the didactic repertoire, and reveals some implicit assumptions:
φ(P) is a polynomial is here implicit, as is its use linking degree and composition of
polynomials. He uses mainly hypothetical-deductive inferences. But, as an
introduction to his argumentation, the student asks himself whether deg(φ(P)≤n: he
formalizes here the start of an abductive reasoning.
To study the injectivity of φ, the student applies a transformation of register of
semiotic representations to formalize the link with ker φ, starts within an algebraic
frame then uses the decomposition of P to study ker φ, but without success: he
cannot make φ operate on P and says not to have any clue to study the kernel. During
this phase, the student tried to use some deductive reasoning involving objects from
the reference milieu: the lack of articulation between the heuristic milieu and the
reference milieu confirm the difficulties encountered in R1.3 with an aimless
organisation of the signs. The semiotic analysis underlines the lack of quantifiers
which leads to a incomplete apprehension of the objects (R2.2) and reinforces the
feeling of lack of goal in the reasoning. The teacher asks then the student to consider
the tools he has got in his repertoire to “calculate” objects in linear algebra. With no
answer, he asks the student to consider the matrix of φ in the canonical basis. Doing
so, he tries to force the student to face the heuristic milieu and tries to maintain an
adidactic dimension in the situation. The calculations produced confirm the fact that
the student doesn’t know how to compute φ: he obtains the identity matrix. This
matrix doesn’t appear to be an index to control the reasoning (articulation R1.2 and
R1.3). With a new oral intervention of the teacher, the student writes the right matrix.
Some misinterpretations are following: Rn instead of Rn[X], surjectivity is meant
instead of injectivity (R3.1), φ(1)=0 is used as an symbol for ker φ=vect(1) instead
of a simple index of it (R2.1). The student uses the theorem of the rank (R3.2),
relying then on a deductive form of reasoning. But some of his deductions rely on
the preceding explicit calculations (articulation R4.2 R4.3).

241 [Link]:indrum2018:174109
CONCLUSION
As stated in the introduction, the purpose of this paper is twofold: try to determine
the reasoning produced by a student in a specific mathematical situation and show
the utility of our framework to analyse the signs and arguments produced in this
situation.
Regarding the first question, our analysis allows us to say that this particular student
has difficulties to reach fully the institutionalization milieu (level R3): the reasoning
and the articulation of the objects involved do not ease his control over his
arguments and eventually lead him to aimless computations. Moreover he seems to
get stuck in the reference milieu (level R2) and to hypothetical-deductive inferences.
The student does not rely on reasoning made in a heuristic milieu (level R1) that
would be appropriate to linear transformations and polynomials. The problem we
analyze in this work contains an adidactic dimension but fails in asking the student
to make effectively operate φ on Rn[X]. In semiotic terminology, we can postulate
that the at least incomplete pragmatic dimension in the reasoning leads to some
confusion and lack of pertinent association between the syntactic and semantic
dimensions (Bloch and Gibel, 2016).
Regarding the second question, our work seems to confirm that the model used
within TDS constitutes an efficient framework, as stated in Bloch and Gibel (2011,
2016). It helps specifying the reasoning and signs on which it relies both for the a
priori and a posteriori analysis and highlights the obstacles contained within the
mathematical notions. Our future work should be working more specifically on the
abductive reasoning and on situations encouraging students to adopt a heuristic
approach.

REFERENCES
Bloch, I. (2003). Teaching functions in a graphic milieu: what forms of knowledge
enable students to conjecture and prove? Educational Studies in Mathematics, 52,
3-28.
Bloch, I. (2006). Quelques apports de la théorie des situations a la didactique des
mathématiques dans l’enseignement secondaire et supérieur. Note de synthèse
présentée en vue de l'Habilitation à Diriger des Recherches, Université Paris 7.
Bloch, I., Gibel, P. (2011). Un modèle d'analyse des raisonnements dans les
situations didactiques: étude des niveaux de preuves dans une situation
d’enseignement de la notion de limite, A model to analyse reasoning processes in
didactic situations: a study about levels of proofs in a situation concerning the
teaching of limits Recherches en Didactique des Mathématiques, 31-2, 191-227.

242 [Link]:indrum2018:174109
Bloch, I., Gibel, P. (2016). A model to analyse the complexity of calculus knowledge
at the beginning of University course. Two examples: parametric curves and
differential equations. Proceedings of INDRUM, Montpellier, Nardi, Winslow
&Hausberger, ed. ISSN: 2496-1027 (online) p. 43-52.
Castela C. (2011). Des mathématiques à leurs utilisations, contribution à l'étude de
la productivité praxéologique des institutions et de leurs sujets / Le travail
personnel au cœur du développement praxéologique des élèves en tant
qu'utilisateurs de mathématiques, Note de synthèse présentée en vue de
l'Habilitation à Diriger des Recherches, Université Paris Diderot. Paris : Irem 7
Duval, R. (2017). Understanding the Mathematical Way of Thinking – The Registers
of Semiotic Representations, Springer International Publishing
Farah L. (2015), Étude et mise à l'étude des mathématiques en classes préparatoires
économiques et commerciales : point de vue des étudiants, point de vue des
professeurs, Thèse de doctorat, Université Paris Diderot (Paris 7)
Gibel, P. (2004). Fonctions et statuts des différentes formes de raisonnement dans la
relation didactique en classe de mathématiques, Doctorat de l’Université de
Bordeaux 2.
Gibel, P. (2015). Mise en œuvre d’un modèle d’analyse des raisonnements en classe
de mathématiques à l’école primaire. A model to analyze reasoning processes at
primary school. Éducation et Didactique, 9-2, 51-72, PUR.
González-Martín, A., Bloch, I., Durand-Guerrier, V., Maschietto, M. (2014).
Didactic Situations and Didactical Engineering in University mathematics: cases
from the study of Calculus and proof. International Journal of Research in
Mathematics Education 16(2), 117-134.
Lalaude-Labayle, M. (2016). L'enseignement de l'algèbre linéaire au niveau
universitaire : Analyse didactique et épistémologique. The teaching of linear
algebra at tertiary level: A didactic and epistemological analysis]. Université de
Pau et des Pays de l’Adour.
Marty R. (1990), L'Algèbre des signes, Essai de sémiotique scientifique d'après C.
S. Peirce, Amsterdam, John Benjamins Publishing (Foundations of Semiotics Series
24)
Winsløw C. (2007), Les problèmes de transition dans l'enseignement de l'analyse et
la complémentarité des approches diverses de la didactique, Annales de Didactique
et de Sciences Cognitives, Vol. 12 , 189-204

243 [Link]:indrum2018:174109
 Leveraging Specific Contexts and Outcomes to Generalize in
Combinatorial Settings
Elise Lockwood and Zackery Reed
Oregon State University, [Link]@[Link]
Generalization is a fundamental aspect of mathematics, and it is a practice with
which undergraduate students should engage and gain fluency. It is important for
students in combinatorial settings to be able to generalize, but combinatorics lends
itself to engagement with specific examples, concrete outcomes, and particular
contexts. In this paper, we seek to inform the nature of generalization in
combinatorial settings by demonstrating ways in which students leverage specific,
concrete settings to engage in generalizing activity in combinatorics. We provide
two data examples that highlight ways in which concrete and specific ideas can be
leveraged to help students develop generalizations in combinatorial settings.
Keywords: Combinatorics, Generalization, Examples, Discrete Mathematics.

INTRODUCTION AND MOTIVATION

Generalization is a foundational mathematical activity, a mathematical practice that
both researchers and policy-makers value (Amit & Neria, 2008; Ellis, 2007). At the
undergraduate level, given the nature of abstract, advanced mathematics, it is
important to learn how to facilitate generalizing activity for students. We have
recently conducted a study designed to investigate undergraduate students’
generalizing activity, and we explored the students’ generalizing activity in the
context of combinatorial problems. In this study, we aimed to examine ways in
which to foster students’ engagement in generalizing activity. In combinatorics,
however, it is often important and even necessary to focus on specific contexts or to
consider particular, concrete outcomes. Indeed, in our prior work (e.g., Lockwood
2013, 2014) and in this current study, we have found that concrete, specific
instantiations of problems, outcomes, and examples are particularly important for
students’ combinatorial thinking and activity. We believe that in the domain of
combinatorics in particular, such specific instantiations are especially important for
developing combinatorial thinking. Given that we want our mathematics students to
be able to reason generally in combinatorial settings, we examine the interplay
between the natural need for specific contexts and outcomes in combinatorics and
the desire to have students engage in meaningful generalization. In this paper, we
seek to answer the following research question: In what ways can specific examples,
concrete outcomes, and particular contexts be leveraged to foster generalizing
activity in a combinatorial setting?

244 [Link]:indrum2018:174870
LITERATURE REVIEW AND THEORETICAL PERSPECTIVE
A Piagetian perspective on generalization and generalizing activity. As a broad
theoretical perspective, we adhere to a constructivist view of learning, asserting that
students construct their knowledge of a given situation based on their mathematical
experiences. We fundamentally view generalization as being related to Piaget’s
notions of reflective abstraction, and we emphasize the importance of having
students engage with and reflect upon their prior activity as they engage in
generalization. Many researchers have studied generalization in a variety of contexts
involving both school-aged children (Amit & Neria, 2008; Ellis, 2007; Rivera, 2010)
and undergraduate students in a variety of areas (e.g., Dubinsky, 1991; Harel & Tall,
1991). This report contributes to the growing body of literature by examining the
nature of generalization in an undergraduate combinatorial setting.
We follow Ellis (2011) and take generalization to mean engaging in “at least one of
three actions: (a) identifying commonality across cases, (b) extending one’s
reasoning beyond the range in which it originated, or (c) deriving broader results
from particular cases” (p. 311). To describe students’ activity as they generalize, we
adopt Ellis’ (2007) taxonomy of generalizing activity, Ellis describes three main
categories of generalizing actions: relating, searching and extending. In this paper,
we focus especially on relating, which occurs when “a student creates a relation or
makes a connection between two (or more) situations, problems, ideas, or objects”
(p.235). In this paper, the term “generalization” need not involve a formal, final
statement of a general rule or property, but rather it may refer to the results of a
students’ generalizing activity, even if that activity is incomplete or not normatively
correct.
Combinatorial thinking and activity.
Combinatorial enumeration problems, or “counting problems,” are easy to state, but
they can be surprisingly challenging for students to solve. This is due in large part to
the fact that counting problems are not reliably solved using prescribed, fool proof
algorithms (e.g., Kapur, 1970). Solving counting problems thus provides
opportunities for students at all levels to engage in rich mathematical thinking. There
is ample evidence that students struggle with solving counting problems (e.g.,
Batanero, Navarro-Pelayo, & Godino, 1997). Although researchers have taken
strides in identifying productive strategies and ways of thinking that might help
address such difficulties, there remains much to learn about how we might
effectively help students to count successfully.
In this paper, we examine the role of generalizing in students’ counting, and we
explore how to frame generalizing activity in terms of Lockwood’s (2013) model of
students’ combinatorial thinking. Lockwood (2013) suggested that there are three
key components to students’ combinatorial thinking (Figure 1) and that solidifying
the relationships between these components is an important aspect of successful

245 [Link]:indrum2018:174870
counting. Formulas/expressions refer to numbers and/or variables that represent the
answer to a counting problem. Counting processes refer to the enumeration process
in which a counter engages as they solve a problem. Sets of outcomes are the
collection of (encoded) objects that are being counted.

Figure 1: Lockwood’s (2013) model of students’ combinatorial thinking

To exemplify the model, we discuss the Horse Race Problem, which is a problem
that we discuss in the Results Section. The problem states: There are 10 horses in a
race. In how many different ways can the horses finish in first, second, and third
place? Note that one counting process to solve this problem is to consider options
for which horse could be first, second, or third place. We can argue that there are 10
options for which horse is first, and for any choice of which horse is first there are 9
options for which horse is second, and then for any of those there are 8 choices for
which horse is third. This counting process yields an expression of 10*9*8, which is
720. This process would organize the set of outcomes lexicographically, grouped
according to which horse finished first, then second, then third.
Lockwood has emphasized the importance of sets of outcomes in a number of
studies. In particular, she has advocated for a set-oriented perspective toward
counting (Lockwood, 2014), in which the act of counting is viewed as inherently
involving structuring and enumerating the set of outcomes. In addition, she has made
a case for the value of listing outcomes, demonstrating that listing outcomes was
positively correlated with solving problems correctly for novice undergraduate
students (Lockwood & Gibson, 2016). In this way, Lockwood has emphasized the
importance of considering concrete outcomes as students solve counting problems.
The point of the set-oriented perspective is that students should think carefully about
what they should be solving in a given problem.
On one hand, then, this prior research suggests that it is useful for students to
consider the concrete, specific outcomes that they should count. Further, when a
student solves a counting problem such outcomes are necessarily tied to that problem
and context. We want students to be able to think about what constitutes an outcome
in a particular combinatorial situation. On the other hand, though, we want to foster
generalization for students and to encourage them to engaging in generalizing
activity, even in combinatorial situations. We want students to be able to develop and
apply general formulas, or to be able to make general arguments about their counting
processes. In this paper, we describe specific ways in which students use concrete
settings to leverage general thinking and activity in the domain of combinatorics.

246 [Link]:indrum2018:174870
METHODS
We report on data from a study designed to study students’ generalizing activity in
the context of combinatorics. We report on two data sources. First, we report on a
design experiment (Cobb, Confrey, diSessa, Lehrer, & Schauble, 2003) with four
undergraduate calculus students, and we focus on one particular student Carson
(student names in this paper are pseudonyms). The students were chosen based on a
selection interview; they had not taken a discrete mathematics course in the
university and were novice counters who could explain their thinking. The students
were interviewed together as a group of four during nine 90-minute sessions. The
interviews were audio and video recorded. During this time, the students worked
both individually and together on combinatorial activities, and the interviewer often
asked probing questions or asked the students to explain their work. These tasks
included solving basic counting problems, coming up with general formulas for
counting problems, and solving problems related to combinatorial proof.
Second, we report on an individual interview with a calculus student, Tyler, who had
similarly not taken a discrete mathematics course in the university. The interview
was individual and 60 minutes long. We gave Tyler tasks involving determining the
number of 3, 4, 5, and eventually n-length passwords using As and Bs, and then
passwords consisting of the characters As, Bs, and the number 1. We had him write
tables in which he recorded the number of passwords with a certain number of As,
and ultimately the tasks could yield the binomial theorem (which we do not discuss
in this paper). These tasks were broadly designed to target students’ generalizing
activity in combinatorial tasks specifically, and we sought both to learn about
students’ combinatorial reasoning and about their combinatorial generalization.
The design experiment sessions and the interviews were transcribed, and we created
enhanced transcripts in which we inserted relevant images and descriptions of
activity into the transcripts. For the purposes of this paper, we identified two cases of
Carson and Tyler as students who leveraged particular problems and situations in
order to engage in generalizing activity. We focused on these students’ data and
identified relevant episodes that shed light on this phenomenon. We reviewed the
transcripts and the videos and discussed these cases with the research team.

RESULTS AND DATA EXAMPLES

In our results, we seek to demonstrate instances in which students leveraged and use
specific, concrete examples in combinatorics in order to engage in generalizing
activity. These are meant to shed light on the interplay between particular situations
and contexts that are important in combinatorial settings and the broader practice of
generalization. We argue that specific examples, concrete outcomes, and particular
contexts remain a fundamental aspect of combinatorial reasoning that can help to
facilitate generalization. We offer two specific examples of how this phenomenon
occurs, shedding light on the nature of generalization in combinatorial settings.

247 [Link]:indrum2018:174870
Students leverage activity on particular problems to generalize counting
formulas and principles.
In this case, a student in the design experiment, Carson, repeatedly referred back to
his prior work on a particular problem that stood out to him as being important. We
view this as an example in which work on a particular problem can be leveraged to
help students engage in generalizing activity. During initial problem solving in the
first session, Carson had solved the Horse Race Problem, described previously. We
will demonstrate that as he proceeded to solve other tasks and solve other problems,
he repeatedly referred back to his prior activity on this problem, related it to other
situations, and used it to generalize a counting formula. Carson solved the problem
in a different way than we had described above, arriving at a correct expression of
10!/7!. Note, this is equivalent to the expression 10*9*8, but, as he explained below,
Carson used a different counting process. He had a particular way of reasoning about
this solution, leveraging the notion of division and equivalence to explain why the
division by 7! makes sense combinatorially.
Carson: …So, there’s 10 factorial total outcomes, and then we know for any given
first 3 there’s gonna be 7 factorial, because that’s saying we know the first 3
horses have finished – how can the last 7 horses finish, so that’s gonna be 7
factorial. But all we care about is how many given first 3s there are. So, if
we divide the total number of outcomes by the number of potential of
outcomes for the last 7 horses that will give us the potential number of
outcomes for the first 3. If that makes sense?
Carson argued that for any particular arrangement of all ten horses, since all that
matters is how the first three horses finish, we can divide by the number of identical
arrangements of the last seven horses. This is a valuable way to think about these
problems, and understanding and articulating this counting process seemed to be an
important moment for Carson. As we proceeded to consider more problems, Carson
repeatedly returned to this Horse Race Problem. We will demonstrate ways in which
Carson has engaged in the generalizing activity of relating (Ellis, 2007) by using this
problem as he approached additional tasks. In this way, this problem served as a
generic example (Mason & Pimm, 1984), a way in which he could make general
arguments and connect his reasoning to other problems. We now describe several of
the ways that Carson leveraged this particular problem.
First, we see that Carson engaged in relating by connecting back to the Horse Race
Problem as he solved other problems. For example, in solving a problem of arranging
4 of 7 books in a row on a shelf, Carson arrived at the correct answer of 7!/3!. The
interviewer asked him how he was thinking about the problem, and his response
below shows the connection he made to the Horse Race Problem.
Carson: Yes, kind of similar to the horse problem. You can say they’re all in a race,
you wanna see how many ways the first 4 books could finish in the race.

248 [Link]:indrum2018:174870
We later had the four students categorize problems they had solved, and from that
exercise we asked them to generalize formulas. One of these formulas was the
number of ways of arranging some number of objects from a larger set of objects.
Carson had indicated that he saw several problems as being the same, and in the
excerpt below, he explained why he viewed the problems as being the same. Again,
he referred to the podium and the division that he had articulated on the Horse Race
Problem as being a distinguishing feature of all of these problems.
Carson: So, essentially all of them are asking for a ranking of a given set of objects
and asking how many arrangements there are for a given number of places,
right? So, the cats are racing to get the collars you could say or the
restaurants are racing to get the top five rankings in the town or the horses
are racing in a race. Then each of the rankings or the collars are a ranking in
the race. Yeah, then you can just divide by the duplicates for leftover ones,
the ones that didn’t make the podium finish or whatever amount of finishes
there are or whatever podium they’re asking for.
His reference to the horses and to the podium suggest to us that this continued to be a
salient aspect of his reasoning. We interpret that Carson was engaging in the
generalizing activity of relating (Ellis, 2007), and, in terms of Lockwood’s (2013)
model, he related the counting process of arranging all of the objects and then
dividing by the ways to arrange the leftover objects. He also seemed to emphasize
the nature of the set of outcomes (arrangements). He recognized that counting
process as similar among the problems he grouped together, and he related each of
those other problems to the ranking and podium language he used in the Horse Race
Problem.
Further, we also asked the students to come up with a general formula for the
problems they had grouped together. They did this for several problems, but we
highlight the formula for the permutation problems. In trying to articulate the kind of
problem they were dealing with, again Carson referred to his activity on the Horse
Race Problem.
Carson: Right. I mean thinking about the method for solving this, it’s the factorial
from above, right? So, we have ten horses in a race. How many ways can
the horses finish, but then how many of those have a unique podium, right?
So, how many times are the first, second, and third place different?
The students then had a conversation about what the formula would be. They came
up with the formula a!/(a-b)! for arranging b objects from a set of a distinct objects,
which another student, Josh, articulated:
Josh: No, I think that it would be something like if you have a objects, you would
have a factorial – that’s the total number of things that you can select – over
a minus b factorial, where b is the number of slots that you have.

249 [Link]:indrum2018:174870
After they agreed upon this formula, Carson explained how he was thinking about
this general formula they had come up with. The excerpt below shows that he
referred back to the imagery of the podium, using that context to make a general
argument.
Carson: … A is your total number of arrangements for the entire thing and then you
want to divide by the number of ways that the places you’re not selecting
can be arranged, right? So, if you’re selecting first, second, and third, then
you have fourth through tenth and those can be arranged in ten minus three
factorial ways, right? So, we can just divide by that number of arrangements
[begins motioning slots with hands] for the back end to get just one for the
front end because that’s what we’re asking for is how many ways can that
podium be arranged?
We contend that in relating back to the Horse Race Problem, Carson was relating
back to different components of Lockwood’s (2013) model, including formulas, sets
of outcomes, and counting process. This exchange suggests that Carson had a well-
developed understanding of the specific problem in terms of the components of the
model, and he related different aspects of the problem in different situations. From
our Piagetian perspective we view Carson and the students as constructing a formula
that makes sense to them, and Carson reflected upon his prior activity in order to
develop a statement of a more general formula.
The Horse Race Problem came up in additional settings for Carson, including during
reasoning about combinatorial proof in a later session. Ultimately, Carson
acknowledged how important this problem was for him. In the final interview, when
we were reflecting on the entire design experiment, Carson shared that he continued
to think about subsequent problems in terms of the Horse Race Problem. We
interpret that his language below means that he felt that he conceptually understood
the ideas in the Horse Race Problem, perhaps in a deep way that he felt confident
about.
Carson: For whatever reason, the horse race problem is the one that’s in my head
forever. And it must have just been where it clicked in the interview
because that’s kind of what I refer to. If somebody says how many ways can
a horse finish in the podium, how many ways can the podium be organized,
things like that. And that’s kind of where I keep going back to. And I don’t
know why that is.
This case serves as an example of a student leveraging activity on a particular
problem for a number of other activities, particularly generalizing activities of
relating (Ellis, 2007). There have other examples of students drawing on prototypical
problem types in combinatorics. Maher and colleagues have talked about students
referring to Pizza Problems or Block Towers problem, demonstrating how students
think about and use powerful particular problems in other (e.g., Maher, Powell, &
Uptegrove, 2011). We build on such work by explicitly drawing attention to the

250 [Link]:indrum2018:174870
generalizing activity that a specific problem fostered for students, highlighting the
affordances that can stem from a student deeply understanding and justifying his or
her activity on a particular counting problem.
Students compare and contrast specific examples to identify general structure.
We briefly describe an additional example in which a calculus student Tyler was
relating two different situations while working on the Passwords Activity. In the
interview Tyler was counting two kinds of passwords – those involving either upper
case As and/or Bs, and those involving As, Bs, and the number 1 (with repetition
allowed). Tyler had initially engaged in systematic listing activity to count the
number of possible 4-character AB passwords. In particular, he created the list of 4-
character AB passwords with exactly two As (Figure 2a), and the table of passwords
according to number of As (Figure 2b).

Figure 2a, b: Tyler’s arrangements of 4-character AB passwords with exactly 2

As and his complete 4-character AB table
Later in the interview, Tyler was in a situation of counting 4-character passwords
consisting of uppercase As, Bs, and 1s. We had asked him to create a table based on
the number of 1s in the password. To complete this table, one can first consider
placing the 1(s) and then filling the remaining positions with either A or B. Notably,
placing As and Bs then reduced to the prior problem Tyler had solved, namely
counting 4-character AB passwords. Tyler realized that there were the same number
of arrangements of two types of characters, and he made a general statement about
this case, recognizing that he will always have six ways of arranging two kinds of
characters. Tyler was able to speak generally about counting arrangements of two
kinds of characters. That is, he recognized that the tables gave him totals for the
number of ways of arranging two characters, not just that they had to be As and Bs or
1s and xs. In the excerpt below he had been working on an extension problem, and
he speaks about two different “things” that are changing, suggesting he had
extrapolated a notion of arranging two characters independently of what the
characters are specifically.
Tyler: Um well these are all the number of combinations I can do, um, with 2
different, 2 things that are changing, and this number of letters.

251 [Link]:indrum2018:174870
Here, we conjecture that reasoning about the particular situations and engaging with
the actual outcomes allowed Tyler to make an important connection between
arranging As and Bs and 1s and xs. The similar nature of the activity when listing in
both cases allowed him to draw attention to the similar counting process in which he
was engaging and the fact that the outcomes he was generating were fundamentally
similar – arrangements of two kinds of characters. Ultimately this allowed him to
make and use a useful generalization, and he understood the values in the rows in the
AB tables as representing the number of arrangements of two kinds of characters.
We infer that Tyler engaged in relating (Ellis, 2007), and that comparing both
situations allowed him to draw out some general commonalities between the two
specific settings. Here, we argue that reasoning carefully about the concrete
examples and actually engaging in concrete listing activity may have helped to
solidify a broader combinatorial process.

CONCLUSION AND DISCUSSION

Prior research (e.g., Lockwood, 2013; 2014; Lockwood & Gibson, 2016) has
emphasized the importance of having students focus on sets of outcomes as they
solve counting problems. Often this focus on outcomes necessarily means that
students reason about very specific contexts and concrete objects, and we view this
as a fundamental aspect of counting and combinatorial activity. However, we also
acknowledge that part of mathematical engagement involves looking beyond
particular situations and contexts, and we have tried to demonstrate certain ways in
which the particular contexts and concrete outcomes can be leveraged to facilitate
meaningful generalizing activity for students.
Specifically, we offer two qualitatively different examples in which students
leveraged the structure of specific combinatorial contexts to establish more general
relationships. In our first example, Carson used his activity on and solution of the
Horse Race Problem as a template for a specific combinatorial process, which he
then used in similar contexts. Carson’s generalizing activity was manifest through
using this template as a means to relate combinatorial processes that he viewed as
similar in some way and to connect the structure of the Horse Race problem to other
cases. This specific generalizing activity demonstrates a powerful manifestation of
relating, wherein Carson leveraged the structure of a known specific example as a
solution for novel and abstract counting processes.
In our second example, Tyler used a connection between two situations to generalize
a concrete arrangement structure. Tyler had initially recognized that his tables
partitioning the 3-length passwords according to the number of As represented
instead a partition of ways to arrange pairs of objects. Tyler then leveraged the
specific process with which he was familiar by arranging 1s and xs, thus
implementing the same specific counting process in a particular generalized setting.

252 [Link]:indrum2018:174870
Both of these examples involve the activity of using concrete situations to form a
general relationship. These cases help to inform the nature of generalization in
combinatorial contexts, offering examples of specific ways in which concrete
outcomes and situations can be leveraged for use in more general settings.

REFERENCES
Amit, M., & Neria, D. (2008). “Rising to the challenge”: Using generalization in
pattern problems to unearth the algebraic skills of talented pre-algebra
students. ZDM Mathematics Education, 40, 111 – 129.
Batanero, C., Navarro-Pelayo, V., & Godino, J. (1997). Effect of the implicit
combinatorial model on combinatorial reasoning in secondary school pupils.
Educational Studies in Mathematics, 32, 181-199.
Cobb, P., Confrey, J., diSessa, A., Lehrer, R., & Schauble, L. (2003). Design
experiments in educational research. Educational Researcher, 32(9), 9-13. DOI:
10.3102/0013189X032001009.
Dubinsky, E. (1991). Constructive aspects of reflective abstraction in advanced
mathematics. In L. P. Steffe (Ed.) Epistemological foundations of
mathematical experience (pp. 160-187). New York: Springer-Verlag.
Ellis, A. B. (2007). A taxonomy for categorizing generalizations: Generalizing
actions and reflection generalizations. The Journal of the Learning Sciences,
16(2), 221-262.
Ellis, A. B. (2011). Generalizing-promoting actions: How classroom collaborations
can support students’ mathematical generalizations. Journal for Research in
Mathematics Education, 42(4), 308-345.
Harel, G. & Tall, D. (1991). The general, the abstract, and the generic. For the
Learning of Mathematics, 11, 38-42.
Kapur, J. N. (1970). Combinatorial analysis and school mathematics. Educational
Studies in Mathematics, 3(1), 111-127.
Lockwood, E. (2013). A model of students’ combinatorial thinking. Journal of
Mathematical Behavior, 32, 251-265. Doi: 10.1016/[Link].2013.02.008.
Lockwood, E. (2014). A set-oriented perspective on solving counting problems. For
the Learning of Mathematics, 34(2), 31-37.
Lockwood, E., & Gibson, B. (2016). Combinatorial tasks and outcome listing:
Examining productive listing among undergraduate students. Educational Studies
in Mathematics, 91(2), 247-270. DOI: 10.1007/s10649-015-9664-5.
Maher, C. A., Powell, A. B., & Uptegrove, E. B. (Eds.). (2011). Combinatorics and
Reasoning: Representing, Justifying, and Building Isomorphisms. New York:
Springer.

253 [Link]:indrum2018:174870
Mason, J. & Pimm, D. (1984). Generic examples: seeing the general in the particular.
Educational Studies in Mathematics, 15(3), 277-289.
Rivera, F. D. (2010). Visual templates in pattern generalization activity. Educational
Studies in Mathematics, 73, 297 – 328.

254 [Link]:indrum2018:174870
 Discrete mathematics at university level
Interfacing mathematics, computer science and arithmetic
Cécile Ouvrier-Buffet1, Antoine Meyer2, and Simon Modeste3
URCA, CEREP, Reims, France, [Link]-buffet@[Link];
1

2
Université Paris Est - Marne-la-Vallée, LIGM, France, [Link]@[Link];
3
IMAG, UMR CNRS. Université de Montpellier, [Link]@[Link]
Discrete mathematics is a recent field linked with Computer Science. We discuss its
place in university mathematics curricula and in the particular case of France, where
it has difficulties to find its place. We make explicit the didactical challenges posed
by discrete mathematics at university level, and present DEMIPS network and its
plans to tackle them. Through two detailed examples we discuss the reasons for
teaching Discrete Mathematics at university level, and illustrate our conclusions.
Keywords: teaching and learning of number theory and discrete mathematics,
teaching and learning of mathematics in other fields, proof, algorithms.

INTRODUCTION
This paper points out the current need for the construction of resources and debates
regarding discrete mathematics at university level. We wish to emphasize the features
of the French context, both from an educator’s and researcher’s point of view, at the
intersection of didactics, mathematics and computer science. Indeed, teaching and
learning discrete mathematics involves mathematical skills and heuristics (e.g.
different kinds of proofs and reasoning, several ways of modelling etc.) 1 and also
develops objects, concepts, methods and tools that are necessary for computer
science. This link with computer science brings new types of questions to
mathematics (for instance, regarding algorithmic complexity). Then, our aims are to
design original situations for schools and at university level, and to construct
appropriate introductory situations for computer science and maths majors.
We propose an overview of discrete mathematics in mathematics education and make
a focus on the interface between discrete mathematics and computer science. Then,
after presenting our research group in France, we analyse two kinds of situations.

DISCRETE MATHEMATICS IN AND FOR MATHEMATICS EDUCATION

How to define discrete mathematics?
Several mathematical topics are often gathered under the blanket term discrete
mathematics. A first step in contributing to a thorough didactical study of discrete
mathematics is to provide a satisfactory definition, or at least delimitation, of what it
refers to. Several definitions exist, which either attempt to provide a general common
trait to the covered topics, such as “the mathematics of discrete sets”, or resort to an
enumeration of objects, concepts or techniques most often associated with discrete
mathematics. Most of these definitions include or are followed by a discussion on

255 [Link]:indrum2018:174091
some typical difficulties, such as the distinction between finite, discrete and
continuous mathematics (e.g. Maurer, 1997). To clarify the distinction between finite
and discrete mathematics, the MAA (1992) places finite mathematics in the pre-
calculus category and discrete mathematics in the same category as calculus.
We advocate that an interesting way to define discrete mathematics both for research
and didactical perspectives (for the design of courses and of didactical engineering at
university level) is to emphasize the features of the modes of reasoning that are
common (or specific) to the various topics usually recognized as discrete
mathematics, and the discrete nature of the structures they involve. Moreover, a
classification of problems is required in order to structure a didactical analysis of the
field of discrete mathematics. Furthermore, as the development of discrete
mathematics has been strongly directed by the needs for computer science, the links
with computer science must be explicitly explored.
In 1974, Knuth, a pioneer in computer science and its teaching made a similar
analysis (Knuth, 1974, p. 329) :
“The most surprising thing to me, in my own experiences with applications of
mathematics to computer science, has been the fact so much of the mathematics has been
of a particular discrete type [...]. Such mathematics was almost entirely absent from my
own training, although I had a reasonably good undergraduate and graduate education in
mathematics. [...] I have naturally been wondering whether or not the traditional
curriculum – the calculus courses, etc. – should be revised to include more of these
discrete mathematical manipulations, or whether computer science is exceptional in its
frequent application of them.”
We consider that these questions are still topical, even at university level, and
deserve a careful didactical analysis.
Where is discrete mathematics? What questions are relevant at university level?
It is often stated that discrete mathematics can be a tool for improving reasoning and
problem-solving skills (see for instance Rosenstein, Franzblau & Roberts (1997)
who advocated an introduction of discrete mathematics in curricula, asked didactical
questions, and made propositions that were taken into account for Principles and
Standards for School Mathematics NCTM, 2000 for instance). Moreover, discrete
mathematics is an active modern branch of contemporary mathematics with a wide
range of applications in society, which is a very legitimate reason to teach it at
school, high school and college. In fact, discrete mathematics courses are relevant to
a wide variety of majors at university level, including mathematics, number theory,
computer science, and engineering: from an epistemological point of view, discrete
mathematics has an interdisciplinary nature and can provide a mathematical
foundation (with specific ways of reasoning and proving, and mathematical
concepts) for computer science and engineering courses. By 1989, an MAA report
(Ralston, 1989) from an ad-hoc committee consisting of mathematicians and
computer scientists recommended that “discrete mathematics should be part of the

256 [Link]:indrum2018:174091
first two years of the standard mathematics curriculum at all colleges and
universities”. This report also emphasizes the notions of proof, recursion, induction,
modelling, and algorithmic thinking, as well as the benefits of using discrete
mathematics in the secondary level to improve problem-solving skills with the
transition to university level in mind (Ralston, 1989). Moreover, Epp (2016) points
out the strong necessity of abstract thinking for the course and its applications in
computer science. She underscores that it is done in the frame of the current
curricular recommendations, prepared by The Joint Task Force on Computing
Curricula (2013) of the ACM and the IEEE Computer Society, which gives discrete
mathematics as one of the two largest components in the “core body of knowledge”
recommended for all computer science students. Besides, discrete mathematics is in
close relationship with other mathematical areas: other fields of mathematics use its
methods and results, and, are useful for solving some discrete mathematics problems.
What is currently the place of discrete mathematics and its links with other scientific
fields at university level? In several countries (Hungary, USA, Germany for
instance), its significance in university programs is well-established and
acknowledged. That is not always the case in France where the status of discrete
mathematics in the first three university years is unclear, at least in mathematical
curricula. However, discrete mathematics appears sporadically in few parts of
mathematics curricula as probability theory (in particular combinatorics for discrete
probability theory) or arithmetic. It sometimes appears in courses dedicated to the
learning of proving, mathematical reasoning and problem solving, but we question
whether its specificity is emphasized. One is more likely to find courses where
discrete mathematics is taught for itself in computer science or mathematics and
computer science curricula, where there exists a kind of common basis shared
between teachers and including classical contents of discrete mathematics as can be
seen abroad. These reports and recommendations coming from academic societies
and the above remarks underscore two key questions for mathematics education at
university level, and more specifically in France:
 What are the place and role of discrete mathematics at university level? How
to design curricula and didactical engineering for the first university years ?
 What links are there between discrete mathematics and other areas (mainly of
mathematics and computer science) and how are they (or should they be)
practised / worked in the first university years?
These questions are particularly crucial for countries where discrete mathematics
does not have a well-established status is the first university years.
What do we know from a didactical point of view?
In mathematics education, various research regarding the teaching and learning of
discrete mathematics exist, focusing mainly on the primary and the secondary levels
(ZDM (2004), Hart & Sandefur (in press) propose overviews). This research meets
general approval, and points out epistemological features of discrete mathematics

257 [Link]:indrum2018:174091
such as: discrete problems bring out different ways of proving (Grenier & Payan,
1998); discrete structures enable work on the construction of mathematical models,
optimization, operational research and experimental mathematics (e.g. Grenier &
Payan, 1998; Maurer, 1997); discrete concepts are accessible and problems are easy
to understand (Grenier & Payan, 1998; De Bellis & Rosenstein, 2004); discrete
concepts have different kinds of definitions and representations (Ouvrier-Buffet,
2006, 2011); some discrete problems are real world problems developing and using
techniques from mathematics and computer science (Schuster, 2004), etc. Discrete
mathematics problems are also a frame for developing and teaching algorithms;
conversely, the study of algorithms requires a lot of discrete mathematics, and
studying algorithms and programming can be a good way to justify the introduction
of discrete mathematics contents (e.g. Modeste, 2012 & 2016). In all this research,
discrete mathematics seems to be a powerful source of problems for teaching and
learning mathematical proofs and processes and engaging students in developing
new ways of thinking (such as recursive thinking), heuristics and problem-solving
skills from primary school to university. Besides, some researchers point out that its
teaching provides opportunities to bypass some of the sources of commonly-
occurring negative affect in students (e.g. Goldin, 2016).
It appears that the features of discrete mathematics clearly represent challenges for
university mathematics, in particular in France.
THE “DEMIPS” NETWORK – A WAY TO FEDERATE DISCRETE
MATHEMATICS EDUCATION
Presentation of the DEMIPS network
In the French framework of mathematics education, there is a need to federate
(isolated) research in university mathematics. Following the INDRUM momentum,
the national network DEMIPS2 supports the development of new research programs.
DEMIPS’s research involves around 40 researchers in mathematics, mathematics
education, physics education, computer science, and epistemology and history of
mathematics and sciences, and is concerned with five main topics: three topics
dealing with mathematical contents (analysis; linear algebra and abstract algebra;
arithmetic, discrete mathematics and algorithmics) at the secondary – post secondary
transition and at university level (the links with physics and computer science are
questioned); a transversal topic (logic, language, reasoning, proofs - from both a
mathematics and computer science point of view); and a specific topic dealing with
the practices of teachers and teachers-researchers at university level (in mathematics,
computer science and physics).
We (the authors of this paper) organize federative research in the fields of arithmetic,
discrete mathematics and algorithms. The members of our group are mathematics
educators (didacticians) with specific skills in teaching and learning at university
level, mathematicians, and researchers in computer science. We choose to study the
parts of mathematics which lie at the intersection of “classical” mathematics and

258 [Link]:indrum2018:174091
theoretical computer science (for instance discrete mathematics, arithmetic, and
algorithms), which interact and complement each other. As theoretical background
we will follow Brousseau’s theory of didactical situations (Brousseau, 1998) for its
notion of didactical engineering, and the notion of scheme (Vergnaud, 1990) in order
to structure our analysis of mathematical concepts. We organize our questions around
key axes regarding the French university level:
 What are the epistemological features of concepts and reasoning in arithmetic,
discrete mathematics and algorithms? How do they interact? (And then, how
can these interactions be used to enrich the way these concepts are taught?)
 What kind of situations can one design in these mathematical areas for the
university level and for pre-service and in-service teacher training? What for?
 What kind of curricula are there for this kind of mathematics at university
level? What can be said about the design of these curricula?
Our research questions try to break down the barriers between scientific disciplines
involving discrete mathematics. They also underline typical situations and questions
common to mathematics and computer science, and try to put to use didactical
analysis techniques to cast a new light on the way these questions are, or could be,
tackled at university level. We develop below two examples to illustrate our work,
and elaborateon the place and role of discrete mathematics at university level.

SITUATIONS AND IMPLEMENTATIONS AT UNIVERSITY LEVEL -

EXAMPLES FROM DEMIPS’ WORKSHOPS
We develop here two examples to illustrate the potentialities of discrete mathematics
to engage students in learning modelling, proving, and mathematical reasoning and
also to underscore the validity and the interest of keeping in mind the algorithmic
point of view and the connections with computer science. These examples emphasize
new perspectives for the teaching and learning of mathematics. The first example
explore the links between mathematics and computer science in a problem-solving
context and the second deals with a classical “divide and conquer”-type algorithm.
Example 1 – Discrete lines
The mathematical object concerned here is the discrete straight line (colouring
squares, or “pixels”, on a regular rectangular grid, in order to give the best possible
visual impression of a straight line). The (real) straight line can act as a referent.
Discrete straight lines are accessible through their representations (e.g. perceptive
and analytical aspects of geometry) and their definitions and properties are non-
institutionalized (a concept is institutionalized if it has a place in a classically taught
content). Computer programmers are familiar with this concept. Professional
researchers in discrete geometry (both mathematicians and computer scientists) use
several definitions, but the proof of the equivalence of these definitions remains

259 [Link]:indrum2018:174091
worth considering. The complexity of the underlying axiomatization of discrete
geometrical concepts is actually an open and interesting problem.
Ouvrier-Buffet (2006) has analysed the evolution potential of zero-definitions (in
Lakatos’ sense, zero-definitions act as working definitions) of the concept “discrete
straight line” in a defining situation implemented with freshmen. She underscores
several approaches dealing with this concept, namely “real straight line” (What is the
“nearest” pixel to a real line? What kind of modelling should be used?), “regularity”
(What are the properties of the sequence of stages (called chaincode string)?), and
“axiomatization” (What about the existence of the intersection of two discrete
straight lines? Is a discrete straight line unique?). Each point of view brings about
several zero-definitions. To engage into an axiomatic perspective carries great
difficulty. This approach deals with both the perceptive aspect of a straight line and
the axiomatic perspective. We are here confronted with two markedly different
defining styles: a local one and a global-theoretical one, the latter mobilizing some
implicit skills and knowledge in students (e.g. building a theory and choosing among
competing definitions). The main results of this experiment underscore the ability of
students to engage in a defining activity with a “neutral” but complex concept.
Students do not assume an axiomatic perspective but mobilize reasoning involving
approximate methods close to those used for real straight lines (and then arithmetic
tools) and also the characterization of the sequence of stages of pixels (how can we
modify a sequence to obtain a better regularity?) that involves recursive arguments.
From a didactical point of view, this research requires the development of a new
theoretical background in order to model the defining process. From a mathematical
point of view, the discrete geometrical objects, and more specifically the discrete
straight lines can be approached in several ways: differential discrete analysis, the
Bresenham algorithm, algorithms involving combinatorial analysis, several
discretizations using algorithms which generate and study errors (Greene & Yao,
Freeman & Pham, Rosenfeld), and the introduction by Reveillès of the arithmetical
definition of a discrete straight line (1991). For instance, the approach to the
discretization of a real straight line by checking linearity conditions is directly
related to number theoretical issues in the approximation of real numbers by rational
numbers. These linearity conditions can be checked incrementally, leading to a
decomposition of arbitrary strings into straight substrings (Wu, 1982). The ongoing
mathematical problems in discrete geometry are intimately related to questions in
other fields of mathematics and computer science. The construction and the
manipulation of algorithms are important for this purpose.
Example 2 – Exponentiation by squaring
A classic algorithmic problem is that of computing for some natural n the n-th power
an of real number a. A naïve solution is, starting with value 1, to multiply n times this
value by a. The final value one obtains is indeed the expected result, which is not
very difficult to establish. The fact that this algorithm terminates is also trivially true
since it contains a single bounded repetition. Finally, the complexity of this

260 [Link]:indrum2018:174091
computation is clearly in Θ(n) (i.e. asymptotically bounded above and below by n),
counting for instance the number of multiplications performed, and assuming that
multiplication by a is an elementary operation.
This algorithm is not very efficient, considering that its running time is actually
exponential in the representation size of n (which is of the order of log(n)). A more
efficient technique relies on the observation that a n=(a2)n/2 if n is even, otherwise
an=a.(a2)(n-1)/2. Written as a recursive Python function, this algorithm reads as follows 3:
def power(a, n):
if n == 0:
return 1
elif n % 2 == 0:
return power(a * a, n // 2)
else:
return a * power(a * a, n // 2)
We will now study a few common questions asked about algorithms, which will
allow us to illustrate examples of mathematical techniques relevant to the analysis of,
and discussion about, algorithms. In the following, typewriter face (as in n) will be
used for formal parameters, and italic (as in n) to denote actual values.
Termination. A first question when it comes to analysing an algorithm is to
determine whether or not it terminates, i.e. whether its execution on any instance of
the problem (i.e. any pair (a, n)) yields a result after a finite number of execution
steps or elementary operations. A standard technique used to prove this kind of result
in non-trivial cases is the following. Assume here that there exist a 0 and n0 such that
power(a0, n0) performs infinitely many recursive calls. Call n i the value of parameter
n on the i-th recursive call. The sequence of naturals (n i)i≥0 is strictly decreasing,
because whenever ni>0, ni+1 is the quotient of ni by 2, rounded down. This contradicts
the fact that infinitely many calls are made, which means that the value of n
eventually has to reach 0 and the function must terminate for all values of a and n.
Correctness. It remains to prove that the result is indeed correct for any instance of
the problem. This is often done using some form of induction due to the intrinsically
discrete and recursive or iterative nature of algorithms. In our case, we will establish
that the value returned by a call to power(a, n) is indeed a n, via a simple recurrence
on the call depth, which is the maximal number, say k, of generated nested calls. The
base case (k=0) is obvious: since there is no recursive call it must mean that n=0 and
the returned is indeed 1 = an. In the inductive case, assume the property holds for call
depth k and consider a call of maximal depth k+1. Necessarily n must be greater than
0. If n is even, n//2 evaluates to n/2, a single nested call power(a*a, n//2) is
performed and the obtained value is returned directly. This call itself has call depth
exactly k, therefore by induction hypothesis its return value is (a 2)n/2=an. Similarly if
n is odd, n//2 evaluates to (n-1)/2, the value returned by power(a*a, n//2) is, by
induction hypothesis, (a2)(n-1)/2=an-1, and the value returned by the main call is

261 [Link]:indrum2018:174091
a*power(a*a, n//2), which evaluates to an. Therefore by the recurrence principle, the
function returns the correct value whatever the initial value of its parameters.
Complexity. In the study of termination, we observed that in a call power(a, n), the
value of n for the next call (if there is one) is divided by two (rounded down). One
may observe the successive values of n more easily when it is written down in
binary. Indeed, the operation of dividing a number by two and rounding down
corresponds, in binary representation, to erasing its rightmost digit. The algorithm
stops when n is 0, and performs one recursive call otherwise, modifying its value as
we just saw. The number of nested calls for some initial value of n is therefore equal
to the length, say k, of its binary representation, in other words its number of digits.
Moreover, when n is even, exactly one multiplication is performed in the current call,
two when it is odd. Therefore, denoting by m the number of digits equal to 1 in the
binary representation of n, the total number of multiplications performed by the
power(a, n) is exactly k+m, which is asymptotically bounded above by log 2(n).
Summary. We chose this example to illustrate, on a simple problem, the type of
questions which can be asked about algorithms and the methods which are likely to
be used to answer them. Note that in this simple case, all three properties could have
been proven simultaneously using a complete recurrence on n. For our purpose, we
chose a more basic and detailed approach. It would have been interesting to show
how these proofs could be rephrased in the context of an iterative function. This
example also tries to advocate the necessity for students in mathematics, computer
science and related topics to have at least a basic understanding of various flavours
of recursion and induction (including basic properties of orderings), to be able to
present rigorous proof arguments (at least informally), and to possess minimal
fluency in arithmetic, in order to be able to envision algorithms as objects worth
studying in their own right. It is moreover often the case that the study of algorithms
provides insight on related mathematical objects (here, the relationship between the
value of a number and the length of its binary representation). Finally, this example
illustrates a typical preoccupation of algorithmics, which is to provide more efficient,
sometimes even optimal, algorithmic solutions to problems.

DISCUSSION AND CONCLUSIONS

Discrete mathematics is now considered as an entire field of mathematics, with many
links to computer science. While it has entered university curricula in many
countries, its status and contour are not always clear, and there are countries (such as
France) where it has difficulties finding a legitimate place. Through the two
examples we have developed (the discrete line and exponentiation by squaring), we
have illustrated that it is legitimate to question the place that discrete mathematics
occupies in university mathematics, for different reasons:
 it allows to develop situations for mathematical reasoning, mathematical
heuristics, and problem solving (by its nature, but also by contrast with
traditional continuous mathematics),

262 [Link]:indrum2018:174091
  many objects and techniques of discrete mathematics are required knowledge
for computer science curricula; these contents must be identified and analysed
from a didactical point of view, to design appropriate activities and situations,
 discrete mathematics involves specific questions and types of problems (such
as complexity questions, combinatorial problems, etc.) that must be studied in
order to understand their place in university curricula.
The DEMIPS network, through the topic group arithmetic, discrete mathematics and
algorithmics, aims at addressing these questions. We pointed out the necessity to
develop a didactical research on the topic of discrete mathematics at university level
and its articulation with other fields of mathematics and other disciplines. This
didactical research must rely on an institutional analysis of the situation in
universities, and most importantly on a thorough epistemological study of discrete
mathematics and its specific branches. It also requires to select and develop
appropriate theoretical frameworks. Such work, started in the DEMIPS topic group,
requires a plurality of viewpoints and interactions between (discrete)
mathematicians, computer scientists, and didacticians of mathematics.

NOTES
1. Problems that can be identified as belonging to discrete mathematics can be found in many books
aiming at developing “mathematical thinking”, such as (Mason, Burton & Kaye, 1985).
2. Didactique et Epistémologie des Mathématiques, et liens Informatique et Physique dans le
Supérieur: Didactics and Epistemology of Mathematics, and links with Computer Science and
Physics in University Mathematics - with the support of CNRS.
3. Here a is assumed to range over floats, and n over positive integers. Note that in Python 3, n//2
computes the quotient of n by 2, whose value is n/2 if n is even and (n-1)/2 otherwise.

REFERENCES
Brousseau, G. (1998). Théorie des Situations Didactiques. La Pensée Sauvage.
DeBellis, V. A., & Rosenstein, J. G. (2004). Discrete Mathematics in Primary and
Secondary Schools in the United States. ZDM, 36 (2), 46-55.
Epp, S.S. (2016). Discrete mathematics for computer science. Proceedings of the
13th International Congress on Mathematical Education - TSG 17. Hamburg.
Goldin, G.A. (2016). Discrete mathematics and the affective dimension of
mathematical learning and engagement. 13th ICME - TSG 17. Hamburg.
Grenier, D., & Payan, C. (1998). Spécificités de la preuve et de la modélisation en
mathématiques discrètes. RDM, 18(1), 59-100.
Hart, E.W & Sandefur, J. (in press). Teaching and Learning Discrete Mathematics in
the School Curriculum Worldwide. An ICME-13 Monograph. Springer.

263 [Link]:indrum2018:174091
Heinze, A., Anderson, I., & Reiss, K. (2004). Discrete Mathematics and Proof in the
High School, Introduction. ZDM, 36 (2), 44-45.
Knuth, D. E. (1974). Computer Science and its Relation to Mathematics. The
American Mathematical Monthly, 81(4), 323-343.
MAA (1992). Statistical abstract of undergraduate programs in the mathematical
sciences and computer science in the United States, 1990-1991, CBSM Survey,
First Two Years, MAA notes n°23, Washington DC.
Mason, J., Burton, L., & Kaye, S. (1985). Thinking Mathematically. Workingham:
Addison-Wesley Publications.
Maurer, S. B. (1997). What is Discrete Mathematics? The many answers. Discrete
Mathematics in the Schools (pp. 121-132). American Mathematical Society.
Modeste, S. (2012). Enseigner l’algorithme pour quoi ? Quelles nouvelles questions
pour les mathématiques ? Quels apports pour l’apprentissage de la preuve  ? (PhD
dissertation). Université de Grenoble.
Modeste S. (2016), Impact of informatics on mathematics and its teaching. On the
importance of epistemological analysis to feed didactical research, in Gadducci, F.,
Tavosanis, M. (Eds.) History and Philosophy of Computing, Series : IFIP
Advances in Information and Communication Technology, Vol.487 (Springer).
Ouvrier-Buffet, C. (2006). Exploring mathematical definition construction processes.
Educational Studies in Mathematics, 63(3), 259-282.
Ouvrier-Buffet, C. (2011). A mathematical experience involving defining processes:
in-action definitions and zero-definitions. Educational Studies in Mathematics,
76(2), 165-182.
Ralston, A. (1989). Discrete Mathematics in the First Two Years. M.A.A. Notes, 15.
Rosenstein J.G, Franzblau D.S., Roberts F.S (Eds) (1997) Discrete Mathematics in
the Schools. DIMACS Series in Discrete Mathematics & Theoretical Computer
Science Vol. 36. American Mathematical Society & NCTM.
Schuster, A. (2004). About Traveling Salesmen and Telephone Networks -
Combinatorial Optimization Problems at High School. ZDM, 36 (2), 77-81.
The Joint Task Force on Computing Curricula, Association for Computing
Machinery (ACM), IEEE Computer Society (2013). Computer Science Curricula
2013, ACM publication.
Discrete mathematics and Proof in the High School (2004). Zentralblatt für Didaktik
der Mathematik, Vol. 36 (2), 44-84 and Vol. 36 (3), 82-116.
Vergnaud, G. (1990). La théorie des champs conceptuels. Recherches en Didactique
des Mathématiques, 10(2.3), 133-170.
Wu, L.D. (1982). On the chaincode of a line. IEEE Trans. on Pattern Anal. Machine
Intell. PAMI 4, 347–353.

264 [Link]:indrum2018:174091
 Tasks for enriching the understanding of the concept of linear span
Evi Papadaki1 and Christos Kourouniotis2
Department of Mathematics and Applied Mathematics, University of Crete, Greece,
1
[Link]@[Link]; 2chrisk@[Link]
The concept of linear span is one of the first abstract notions that students encounter
in a course on Linear Algebra. Using the theoretical construct of concept image and
concept definition (Tall & Vinner, 1981) along with observations about teaching and
learning Linear Algebra, we present two tasks designed to enrich students’ concept
image regarding linear span. These tasks could be included in a problem workshop
of an introductory university course on Linear Algebra. Each task is carefully
created and/or selected so as to foster the ground for potential conflict factors to
arise and be confronted. A preliminary evaluation shows that the tasks are well
received by students and succeed in addressing certain conflicting factors.
Keywords: Teaching and learning of linear and abstract algebra; Teachers’ and
students’ practices at university level; Linear span; Task-design.

INTRODUCTION
Linear Algebra is a subject with many applications in Mathematics and other
sciences, but its teaching and learning proves to be demanding both for lecturers and
students. The difficulties encountered are partly attributed to the way the subject is
usually taught, as well as to students’ lack of familiarity with proofs and limited
knowledge of Logic and Set Theory. (Dorier et al., 2000; Hillel, 2000). Sierpinska
(2000) attributes students’ difficulties in Linear Algebra to their practical rather than
theoretical way of thinking.
The concept of linear span seems to be quite difficult for students. Carlson (1993)
states that difficulties in the notions of subspace, linear span and linear dependence /
independence, if they are not addressed in time, create barriers for students. The
analysis of Stewart and Thomas (2009) showed that students who were taught these
concepts through formal definitions faced significant difficulties in understanding
the concept of span compared to a group who were taught with emphasis on
embodiment (Tall, 2004) and geometry. Moreover, they report that students have
experienced several difficulties in linking the concept of span to the concept of a
base. Finally, Wawro et al. (2012) propose teaching the concept through the solution
of systems of linear equations and present a teaching approach through a series of
realistic mathematical activities.
The main purpose of this paper is to investigate students’ understanding of the
concept of linear span and to use tasks to help resolve conflict factors in the students’
concept image (Tall & Vinner, 1981). Based on a study of first year Mathematics
undergraduates in a Greek university, we identify the misconception many students
have that in a linearly dependent set each vector is in the span of the others. We use a

265 [Link]:indrum2018:174405
set of design principles based on Sierpinska’s (2000) remarks about theoretical
thinking and Harel’s (2000) principles of teaching and learning Linear Algebra, to
create a set of tasks, and we present results of a preliminary evaluation of the tasks
which indicate their potential to address the above misconception.
The work presented in this paper is part of the first writer’s Master’s thesis.

THE SETTING
The course “Geometry and Linear Algebra” is a first year mandatory course for
students following the degrees in Mathematics or in Applied Mathematics at a Greek
University. The course is typically taught through 4 hours of lectures and a two-hour
problem workshop per week. Problem workshops are an important part in the
teaching of the mandatory courses in the department. In the workshops the students
are encouraged to work in groups of 5 or 6 students, on selected problems on the
topics taught that week with guidance from the lecturer and a number of
postgraduate or senior undergraduate students. The role of the latter is to discuss
with students about the problems and the key mathematical ideas that may come up
in the process. Promoting mathematical discussion among the students is a
promindent element of the workshops of this course. During the semester of the
study, the second writer was the lecturer of the course and the first one of the
postgraduate students involved in the workshops.
During the first part of the course, students experiment with the idea of linear span in
Euclidean 2- and 3-space, as an intuitive introduction to the concept. Later on,
students are given a slightly modified version of the formal definition, limited to the
spaces Rn. The notion of linear span is usually described as the “subspace generated
by the set S of vectors in Rn”. In relation to the general goals of the course, students
are expected to familiarize with the concept of linear span in subspaces of Rn, to be
able to identify its geometrical representation in the case of R2 and R3 and to
determine if a vector is in the span of a fixed set of vectors. We note the most
important aspects of the concept. Firstly, linear span is a subspace, hence it is closed
under the operations of a vector space. Secondly, every element in this subspace is a
linear combination of some of the vectors in S. The final aspect is also very
important but sometimes overlooked. In contrast to the concept of basis, there is no
limitation in the choice of the set of generators S.
A starting point for this work was a study of the written answers given by students in
response to a question in the final examination for the “Geometry and Linear
Algebra” course, asking them to determine whether a vector belongs to the subspace
spanned by two other vectors. The findings suggested that some students may have
the misconception that in a linearly dependent set of vectors, every vector can be
expressed as a linear combination of the others (see Papadaki, 2017). This
misconception was found to affect students’ understanding of linear span and to be a
potential conflict factor (Tall & Vinner, 1981). We believe that examining the notion

266 [Link]:indrum2018:174405
of linear span through tasks may offer the opportunity to confront such difficulties in
a meaningful way.

THEORETICAL FRAMEWORK & DESIGN PRINCIPLES

Tall & Vinner’s (1981) cognitive model of concept image and concept definition is
used in the development of the task and to account for students’ responses.
According to them concept image is “the total cognitive structure that is associated
with the concept” (p. 152). For each individual a concept image includes all the
mental pictures (graphs, symbols, formulas etc) generated about the concept,
associated properties and processes. The concept image is unique for each student
and is changing over time when the student meets new stimuli. The term evoked
concept image (Tall & Vinner, 1981) is used to describe the part of a concept image
which is evoked at a specific time. Different parts of the concept image which
contain conflicting aspects are called potential conflict factors (Tall & Vinner, 1981)
and they are not evident to the individual until a stimulus causes the conflicting
images to be evoked simultaneously and create confusion, in which case they are
referred to as conflict factors.
The term concept definition is referring to “the form of words used to specify that
concept” (Tall & Vinner, 1981: 152). The concept definition might be a reflection of
an evoked concept image associated with the definition or a rote memorization of a
given definition with little or no meaning to the student. We adopt Tall & Vinner’s
(1981) differentiation between the personal concept definition, constructed by the
individual, and the formal definition of a concept, the definition accepted by the
mathematical community as a whole. The personal concept definition might contain
aspects not included in the formal definition and/or ignore others. Finally, the
(personal) concept definition creates its own concept image, which is part of the
concept image as a whole, called concept definition image. Tall & Vinner (1981)
argue that potential conflict factors can be an obstacle in understanding the formal
theory, especially the ones that are in contrast with the formal concept definition.
Warwo et al. (2011) investigated students’ concept images of subspace and the links
students create with the formal definition of a linear subspace.
Bingolbali & Monaghan (2008) demonstrated how the construct of concept image –
concept definition can be used in socio-cultural research. They argued that although
concept image is unique to the individual there are aspects that are shared among
students. They link these aspects to teaching and shared experiences in the
department they are studying.
In this paper we adopt the original concept image – concept definition framework
(Tall & Vinner, 1981) along with its more recent developments (Bingolbali &
Monaghan, 2008) to design tasks that can enrich the understanding of linear span of
undergraduate Mathematics students when used in situations which encourage
interaction among students and tutors. We believe that this framework can be easily

267 [Link]:indrum2018:174405
understood and used by mathematicians. Nardi (2006) presents evidence from
discussions with mathematicians which support this idea. Therefore, we find this
framework useful as a means to communicate our design and findings both to
Mathematics lecturers and researchers in Mathematics Education.
In designing the tasks, we take into account Sierpinska’s (2000) remarks about
theoretical thinking. To be more specific, the task should have characteristics that
correspond to theoretical thinking, such as opportunities for conscious reflection,
connections between related concepts or different representations and attention to
contradictory thoughts. Harel (2000) emphasizes the need for curricula tailored to
students’ needs which aid the understanding of abstract concepts in Linear Algebra.
He proposes three principles that we take into account in designing the tasks. That is,
the tasks should include familiar concepts that allow connection with prior
knowledge and language (concreteness principle), they should justify the need of
linear span (necessity principle) and allow generalization of the key ideas
(generalizability principle).
We identify the following principles based on the theoretical framework, the concept
of linear span as thought in the course “Geometry and Linear Algebra” as well as the
needs of our students.
1. Include key aspects of linear span: Closure under the operations of a vector
space; Every vector is a linear combination of the set of generators; No
limitation in the choice of the set of generators
2. Tackle potential conflict factors: The difference between linear combination
and linear dependence; Modes of representation (Hillel, 2000)
3. Promote theoretical thinking (Sierpinska, 2000): Reflection; Connections
between different representations; Attention to contradictory thoughts
4. The three principles of teaching and learning Linear Algebra (Harel, 2000):
Concreteness principle; Necessity principle; Generalizability principle
5. Promote discussion: among the students; between the students and the tutor

METHODOLOGY
The aim of this work is to investigate the conflict factor identified earlier through
tasks that are designed to foster the ground for this conflict to emerge and to be
discussed with the students. We present data collected during a preliminary
evaluation of the tasks through semi-structured interviews with seven students who
had attended the course “Geometry and Linear Algebra” the previous semester. The
analysis of this preliminary evaluation is expected to answer the following questions:
Can the tasks tackle this potential conflict factor? What are the roots of this conflict
factor? Does the discussion around the task help students resolve their
misconceptions? Do students find the tasks interesting and/or useful?

268 [Link]:indrum2018:174405
The following table summarizes the information about the seven participants.
Mathematics Applied Mathematics
st
1 Year 2nd Year rd
3 Year 1st Year 2nd Year
Male 0 1 0 1 0
Female 3 0 1 0 1

Prior to the interviews each student was given a folder including the task and other
necessary information. The students had one week to attempt and review the tasks
before the interviews. All interviews were videotaped. To ensure confidentiality each
student was assigned and referred to with an alias.

ANALYSIS
The first task is based on an exercise from the book “Linear Algebra: Concepts and
Methods” by Antony and Harvey (2012). Its structure was slightly altered to fit that
of the course notes (Kourouniotis, 2014). It aims to create connections with prior
knowledge, known processes and language under the new context and introduce to
students basic ideas linked with the concept through algebraic and geometric
representations of the notion. The task is divided into three interconnected sub-tasks
as a scaffolding strategy to support students.
Task 1: Consider the vectors:
v1= ( –1, 0, 1), v2 = (1, 2, 3), w1 = (–1, 2, 5), w2 = (1, 2, 5)
i) Show that w1 can be expressed as a linear combination of v1 and v2, but w2 cannot be
expressed as a linear combination of v1 and v2.
ii) Explain what subspace of R3 is spanned by v1, v2 and w1. Explain what subspace of
R3 is spanned by v1, v2 and w2. What do you observe?
iii) Show that the vectors v1, v2, w1 and w2 span R3, that is for every u = (x, y, z) there are
a, b, c, d such that:
u = av1 + bv2 + cw1 + dw2
Show also that every vector u ∊ ℝ 3 can be expressed as a linear combination of v1, v2,
w1 and w2 in infinitely many ways.
The first, introductory, sub-task aims to support students’ theoretical thinking in the
following sub-task by limiting its focus on calculations. This task was completed by
all the participants without difficulty prior to the interview. The second sub-task is
expected to enrich students’ image of linear span by making connections between the
algebraic and geometrical representations of the concept in R3. It may also motivate
students to seek a deeper connection between Analytic Geometry and Linear
Algebra. This sub-task was completed by 5 students. Finally, the third sub-task aims
to create a link between the relation of the given vectors and the number of ways
arbitrary vectors can be expressed as a linear combination of the elements in the set.

269 [Link]:indrum2018:174405
Sub-task (iii) proved to be the most difficult for the participants, being completed by
only 2 students before the interviews.
In more detail, the students who did not complete (ii) appeared to have trouble with
methodology. The students are expected to know from the first part of the course
what the geometric representation of a 1-, 2- or 3-dimensional subspace of R3 is,
therefore one has to connect this idea with the notion of linear span and check if the
given vectors are linearly dependent. In both cases the students did not make this
connection beforehand but the problem was quickly resolved through discussion.
Apart from that, six out of the seven students found the question “what do you
observe?” useful. This question was added to the task as an encouragement for
reflection on the effect that different choices of vectors have on the outcome and to
promote discussion. In particular, three of the students indicated that they might not
have given a second thought to their result if it wasn’t for this question. One of the
students found the question stressful, although she had successfully answered it. Her
reaction is significant to us at this point. Clute (1984) found that students with higher
anxiety levels can benefit more from instrumental approaches. Open questions, such
as the above, are not frequent in Greek secondary education. It is therefore
reasonable to assume that some students would have difficulty (and in some cases
anxiety) answering this question in a problem workshop.
While discussing sub-task (ii) an unexpected observation was made by two of the
students. These students interestingly replied that the span of the vectors v 1, v2 and
w1 is the vector space R2. This conflict factor is called by Wawro et al. (2011: p. 13)
the “nested subspaces”. Based on their evidence they hypothesized that this
confusion has roots in students identifying any 2-dimensional subspace of Rn with
R2 and suggested that lecturers must be aware of this as a potential conflict factor.
Their hypothesis was confirmed in these cases too.
In trying to answer sub-task (iii) the biggest pitfall was following the same reasoning
used in subtask (ii). This approach will not help answering the second part which
requires from students to solve a system of linear equations. Despite the instructions
included in the Task, four out of the five students who didn’t complete (iii), tried to
use the same approach as in (ii). Additionally, three of them faced a difficulty
making use of the proposition “for every u = (x, y, z) there are a, b, c, d such that u =
av1 + bv2 + cw1 + dw2” and did not manage to recognize the random vector u = (x, y,
z) as a parameter of the problem. Instead they identified it as another variable. In
each case the task was completed with the help of the interviewer but we find that
subtask (iii) required more guidance from the part of the interviewer compared to
subtask (ii). The fifth student managed to solve the required linear system but she
could not make a connection between the infinite number of solutions and the fact
that the four vectors are more than enough to describe any vector in R3.
The second task was created to address potential conflict factors in relation to the
notions of linear combination and linear dependence in the context of linear span.

270 [Link]:indrum2018:174405
The idea for this task was based on our goal to promote theoretical thinking and
discussion. The conflict is given to the student as a statement - challenge and the
goal is to find an example to support the given proposition. It is expected that
students will first use a trial and error approach by reaching for appropriate vectors
in their example space (Mason & Watson, 2008). This approach will probably fail if
students are not able to identify what are the key relations between v 1, v2 and w in
the prοposition. If one’s concept image includes conflicting ideas about the status of
vectors in a set of generators, it might be difficult to find an example without careful
prompting and discussion. Because of the nature of the problem, we believe that
students would want to cross-examine their findings or get some guidance.
Task 2: Let v1, v2 and w be linearly dependent vectors in R3. It is possible for w not to be
in the space spanned by v1 and v2 although v1, v2 and w are linearly dependent. Give an
example. Why do you think this can happen?
Moving on to the interviews, only one student had found an example of three vectors
fulfilling the requirements of the task before the interview. In four of the seven cases
clear signs of conflicting images emerged. This reinforces our preliminary hypothesis
that students struggle with identifying the difference between the notions of linear
combination and linear dependence. Furthermore, it might be an indication that Task
2 can help potential conflict factors to emerge and be resolved in a controlled
environment. The following quotations capture these observations.
Minos: So, what I thought was that I can have two vectors... which will be linearly
independent that will span a plane in R3. I can of course... I am sure that I can find
another third vector that will not belong in the plane but the relationship to be true... these
three vectors to be linearly dependent.
Minos’ evoked concept image of the linear span is geometric. He thinks of the span
of the two vectors as a plane and he tries to find an example by checking vectors that
are not on that plane. Of course, if the two vectors are linearly independent, adding a
third vector that does not belong in their span will result in a linearly independent
set. It seems that either this fact is not part of his concept image or his evoked
concept image does not include this information because of the phrasing of the task.
In the following two quotations, the conflict can be directly connected to our
preliminary findings in Papadaki (2017). The students seem to struggle with the idea
of three vectors being linearly dependent and at the same time one of them not being
able to be expressed as a linear combination of the others.
Interviewer: Well, so for w not to belong in the span of the two other vectors it could not
be written as a linear combination of them…
Pasiphae: Yes… yes… well… But then how can they be linearly dependent? They are all
together linearly dependent…

271 [Link]:indrum2018:174405
The student thinks of the two notions as equivalent. She later justifies her thinking
by stating that if they are linearly dependent she can solve the algebraic equation
av1 + bv2 + cw = 0 for any of the three vectors. Similarly, Ariadne describes her own
experience with the task. It is worth mentioning that later in the interview Ariadne
successfully refers to the (personal) definitions for both concepts.
Ariadne: To begin with, to me it seemed absurd at first... because… what does it tell me?
It tells me that they are linearly dependent, so if I solve for w, I will find a linear
combination, so based on the theory it belongs to the subspace spanned by v1 and v2.
In Ariadne’s case, it can be assumed that although her concept definition for linear
dependence includes the information that the coefficients a, b and c are not all zero,
in her evoked concept image this statement is replaced by none of them being zero.
The quotations depict two possible roots of students’ difficulties with the task. That
is, thinking of the linear span of two vectors as necessarily a 2-dimensional subspace
or thinking of the algebraic representations of linear dependence and linear
combination as equivalent.
Task 2 was thoroughly discussed with the students using different approaches based
on the line of thinking of the students, but also influenced by the interviewer. The
ideas portrayed in this task were discussed using an algebraic approach with four of
the students and geometrically with two of them. In each interview the final example
was found by the students using an informed trial and error approach. All six
students reported that the discussion was very useful and Task 2 is important for
understanding the concept. Three of them also said that this was the task that made
them the biggest impression and four of the students suggested that it would be better
if this task was presented to them in a problem workshop after a sequence of related
more instrumental tasks.
Concluding, four of the students reported that they understand a notion better
through examples and tasks. The way that students’ concept image is formed through
model examples and experience, is of course well known. What is important is the
fact that the students are aware of this happening. This last observation is an
indication why it is crucial to pay attention to the examples and tasks used in any
course. There are students who are consciously depending on them and expect to
understand the “mysterious” concepts that the lecturer is talking about through them.

RESULTS & DISCUSSION

The analysis of the interviews gave us very important information about how tasks
can be improved and used in a problem workshop for an introductory course on
Linear Algebra. Although all students indicated that they found the tasks useful they
gave us opportunities to reflect upon their design and experiment with different
tactics which can be used by tutors in an attempt to make the most out of these tasks.

272 [Link]:indrum2018:174405
Beginning with the first task, students appeared to have particular difficulty in
subtask (iii). One reason might be that (iii) requires a shift in thinking and cannot be
fully answered by using the same approach as in subtask (ii). In an attempt to resolve
this issue we are also considering a slightly different version of this part of the task
that forces students to begin with the shifted approach as follows:
Show that for every u = (x, y, z) there exist a, b, c, d such that:
u = av1 + bv2 + cw1 + dw2
Conclude that v1, v2, w1 and w2 span R3. Moreover, show that every vector u∊R3 can be
expressed as a linear combination of v1, v2, w1 and w2 in infinitely many ways.
Another observation we made while discussing Task 1 with the students was that of
“nested subspaces”. This is another conflict factor we didn’t take into account at first
and realized it only during the interviews with the students. Our observation is in line
with the hypothesis of Warwo et al. (2011).
Task 2 was fruitful both in terms of meaningful discussion and reflection. Students
found Task 2 important for understanding the concept of span. We also observed
manifestations of cognitive conflict which indicates that the task can be used as a
means to resolve potential conflict factors. Different approaches can be used to
discuss these conflicts with students (algebraically, geometrically or by trial and
error). A useful tactic might be to discuss the conflicting factors using more than one
representation of vectors with the same group of students.
In addition, the indications about the need of examples and tasks made by the
students were of great importance. This fact depicts the necessity of well thought
examples and tasks in order to help students create a coherent concept image.
This paper presents an approach on how lecturers can design tasks inspired by their
observations on students’ misconceptions and taking advantage of the research in
Mathematics Education. The framework could be used as guidelines for tutors that
are interested in developing tasks for a Linear Algebra course based on their students
needs and related research. Finally, the tasks need to be tested in a problem
workshop and be compared to other tasks aiming to familiarize first year
Mathematics undergraduates with the concept of linear span.
References
Anthony, M., & Harvey, M. (2012). Linear algebra: concepts and methods.
Cambridge University Press.
Bingolbali, E., & Monaghan, J. (2008). Concept image revisited. Educational Studies
in Mathematics, 68(1), 19-35.
Clute, P. S. (1984). Mathematics anxiety, instructional method, and achievement in a
survey course in college mathematics. Journal for Research in Mathematics
Education, 50-58.

273 [Link]:indrum2018:174405
Dorier, J. L., Robert, A., Robinet, J., & Rogalski, M. (2000). The obstacle of
formalism in linear algebra. In On the teaching of linear algebra (pp. 85-124).
Springer Netherlands.
Harel, G. (2000). Three principles of learning and teaching mathematics. In On the
teaching of linear algebra (pp. 177-189). Springer Netherlands.
Hillel, J. (2000). Modes of description and the problem of representation in linear
algebra. In On the teaching of linear algebra (pp. 191-207). Springer Netherlands.
Kourouniotis, C. (2014). Geometry and Linear Algebra [lecture notes]. University of
Crete, Heraklion. [Link]
Mason, J., & Watson, A. (2008). Mathematics as a constructive activity: Exploiting
dimensions of possible variation. In M. Carlson & C. Rasmussen (Eds.), Making
the connection: Research and teaching in undergraduate mathematics education
(pp. 191-204). Washington, DC: MAA.
Nardi, E. (2006). Mathematicians and conceptual frameworks in mathematics
education… or: Why do mathematicians’ eyes glint at the sight of concept
image/concept definition. In Retirement as Process and concept; A festschrift for
Eddie Gray and David Tall, (pp. 181-189). Charles University, Prague.
Papadaki, E. (2017) Design of tasks for the understanding of the concept of linear
span (Master’s thesis). University of Crete, Heraklion, Greece.
Sierpinska, A. (2000). On some aspects of students’ thinking in linear algebra. In On
the teaching of linear algebra (pp. 209-246). Springer Netherlands.
Stewart, S., & Thomas, M. O. (2009). A framework for mathematical thinking: The
case of linear algebra. International Journal of Mathematical Education in
Science and Technology, 40(7), 951-961.
Tall, D. (2004). Building theories: The three worlds of mathematics. For the learning
of mathematics, 24(1), 29-32.
Tall, D., & Vinner, S. (1981). Concept image and concept definition in mathematics
with particular reference to limits and continuity. Educational Studies in
Mathematics, 12(2), 151-169.
Wawro, M., Sweeney, G. F., & Rabin, J. M. (2011). Subspace in linear algebra:
investigating students’ concept images and interactions with the formal
definition. Educational Studies in Mathematics, 78(1), 1-19.
Wawro, M., Rasmussen, C., Zandieh, M., Sweeney, G. F., & Larson, C. (2012). An
inquiry-oriented approach to span and linear independence: The case of the magic
carpet ride sequence. PRIMUS, 22(8), 577-599.

274 [Link]:indrum2018:174405
 Delineating Aspects of Understanding Eigentheory through
Assessment Development
Megan Wawro1, Michelle Zandieh², and Kevin Watson1
1
Virginia Polytechnic Institute and State University, Department of Mathematics,
USA, mwawro@[Link]; ²Arizona State University, Mesa, AZ, College of Integrative
Sciences and Arts, USA
In this report, we share insights we have gained from developing an assessment for
documenting students’ understanding of eigentheory. We explain the literature and
theory that influenced the assessment’s development and share question examples.
We frame our results in terms of three eigentheory settings ( ,
and eigenspaces) and four interpretations (numeric, algebraic, geometric, and
verbal). Results from our analysis include students’ reasoning being influenced by
setting, insights into students’ struggle with understanding eigenspaces, and the
importance of making connections between and across various interpretations.
Keywords: Teaching and learning of linear and abstract algebra, teaching and
learning of specific topics in university mathematics.

INTRODUCTION
Linear algebra is particularly useful to science, technology, engineering and
mathematics (STEM) fields and has received increased attention by undergraduate
mathematics education researchers in the past few decades (Dorier, 2000; Artigue,
Batanero, & Kent, 2007; Rasmussen & Wawro, 2017). A specifically useful group of
concepts in linear algebra is eigentheory, or the study of eigenvectors, eigenvalues,
eigenspaces, and other related concepts. Eigentheory is important for many
applications in STEM, such as studying Markov chains and modeling quantum
mechanical systems; however, research specifically focused on the teaching and
learning of eigentheory is a fairly recent endeavor and is far from exhausted.
As part of our ongoing research program analyzing students’ understanding of
eigentheory (e.g., Watson, Wawro, Zandieh, & Kerrigan, 2017; Wawro, Watson, &
Christensen, 2017), we created an assessment instrument focused on the multifaceted
and interconnected nature of eigentheory. The purpose of this paper is to describe
insights have we gained about students’ conceptual understanding of eigentheory as a
result of developing, using, and refining this assessment instrument.

THEORY AND LITERATURE REVIEW

We ground our work in the Emergent Perspective (Cobb & Yackel, 1996), which
assumes that mathematical development is a process of active individual construction
and mathematical enculturation. In this paper we focus on the former by analyzing
mathematical conceptions that individual students bring to bear in their mathematical
work (Rasmussen, Wawro, & Zandieh, 2015). The literature on the teaching and

275 [Link]:indrum2018:174806
learning of eigentheory points to several aspects important to students’ conceptual
understanding. Here we summarize that literature by highlighting what we found to
be important aspects for building a working model for understanding eigentheory.
Literature on student understanding of eigentheory
Thomas and Stewart (2011) found that students struggle to coordinate the two
different mathematical processes (matrix multiplication versus scalar multiplication)
captured in the equation to make sense of equality as “yielding the same
result,” an interpretation that is nontrivial or even novel to students (Henderson,
Rasmussen, Sweeney, Wawro, & Zandieh, 2010). Furthermore, students have to keep
track of multiple mathematical entities (matrices, vectors, and scalars) when working
on eigentheory problems, all of which can be symbolized similarly. For instance, the
zero in refers to the zero vector, whereas the zero in
is the number zero. This complexity of coordinating mathematical entities and their
symbolization is something students have to grapple with when studying eigentheory.
Thomas and Stewart (2011) also posit that this struggle to coordinate may prevent
them from making the needed symbolic progression from to
, which is central to determining the eigenvalues and eigenvectors of . In their
genetic decomposition of eigentheory concepts, Salgado and Trigueros (2015) posit
that students need to interpret the procedure of finding eigenvectors and eigenvalues
of as determining the solution set of the homogeneous system of equations created
by the matrix equation . Harel (2000) posits that the interpretation of
“solution” in this setting, the set of all vectors that make the equation true, entails a
new level of complexity beyond solving equations such as , where , and
are real numbers. When considering the notion of eigenspace in particular, Salgado
and Trigueros (2015) found that students struggle to coordinate the number of
eigenvectors corresponding to a given eigenvalue with the dimension of the space
spanned by the eigenvectors. Thus, understanding eigentheory not only involves
coordinating mathematical processes and entities but also equations and solution sets.
In addition, students have to make sense of instructors’ frequent movement between
geometric, algebraic, and abstract modes of description, but this may be challenging
(Hillel, 2000). In fact, Thomas and Stewart (2011) found that students in their study
primarily thought of eigenvectors and eigenvalues symbolically and were confident
in matrix-oriented algebraic procedures, but the majority had no geometric or
embodied views. In contrast, other researchers have shown how exploration through
dynamic geometry software (Çağlayan, 2015; Gol Tabaghi & Sinclair, 2013; Nyman,
Lapp, St John, & Berry, 2010), geometric interpretations of a linear transformation
(Zandieh, Wawro, & Rasmussen, 2017), or real-world contexts (Salgado &
Trigueros, 2015) can help students develop conceptual understanding of eigentheory.
We similarly agree on the importance of understanding eigentheory concepts in
multiple ways and navigating between various interpretations, and we incorporate
this complexity in our model of student understanding of eigentheory.

276 [Link]:indrum2018:174806
Working Model of Understanding Eigentheory
Regarding what it may mean to have a conceptual understanding of eigentheory, our
current working model is a network of connections within and across three main
settings of how eigentheory is framed. The three sets of relationships that are
pertinent are: (1) relationships indicated by the eigen-equation ; (2)
relationships indicated by the homogeneous form of the eigen-equation
; and (3) relationships indicated by a linear combination of eigenvectors. Within the
first two settings, what is most frequently the focus of inquiry is one particular
eigenvector for either form of the eigen-equation. However, when considering the
collection of all that satisfy either eigen-equation, one arrives at the eigenspace of
associated with . The relationships between vectors in the same eigenspace are the
focus of the third setting. For instance, if and are eigenvectors of with
eigenvalue , then all vectors that are a linear combination of and (i.e.,
for scalars and ) are also eigenvectors of
associated with . Furthermore, reasoning about relationships in this third setting
almost certainly involves reasoning about either the first or second setting as well.
Each of these settings includes entities and relationships between those entities that
may be realized in various ways. We organize this variability in our model according
to four main interpretations: graphical, numeric, symbolic, and verbal.

THE EIGENTHEORY MCE ASSESSMENT

The development of the Multiple Choice Extended (MCE) assessment instrument for
eigentheory grows from our prior work in student understanding of span and linear
independence in which we developed the MCE-style question format (Zandieh et al.
2015); questions begin with a multiple-choice element and then prompt students to
justify their answer by selecting all statements that could support their choice (see
Figure 1). This format was inspired by existing conceptually-oriented assessment
instruments in undergraduate mathematics and physics (e.g., Carlson, Oehrtman, &
Engelke, 2010; Hestenes, Wells, & Swackhamer, 1992; Wilcox & Pollock, 2013).
Development of the Eigentheory MCE involved multiple steps. First, we compiled a
database of questions about eigenvectors, eigenvalues, and related concepts from
research on student understanding of eigenvectors and eigenvalues (e.g., Gol Tabaghi
& Sinclair, 2013; Salgado & Trigueros, 2015; Thomas & Stewart, 2011), online
resources for clicker and classroom voting on linear algebra (Cline & Zullo, 2016),
and previous linear algebra homework assignments, exams, and interview protocols
used by research team members (e.g., Henderson et al., 2010). Second, we used
research results regarding students’ understanding of eigentheory from the literature,
as well as our own teaching experience and theoretical thinking, to determine which
questions seemed to address important aspects of what it may mean to have a
conceptual understanding of eigentheory. Third, the most promising questions were
edited into the MCE format, which involved moulding the problem into a multiple-
choice question and developing six corresponding justification choices that required

277 [Link]:indrum2018:174806
students to reason within and between various eigentheory settings and
interpretations. Fourth, through multiple rounds of administering the assessment to
students, analysing the data, and subsequent refinement, we arrived at the current
Eigentheory MCE. It contains six questions, each with six justification choices; five
questions are in Figure 1 (the sixth is omitted because of space limitations).

Figure 1: Questions 1-5 from the Eigentheory MCE

The MCE questions were created to elicit student thinking about eigentheory within
and across the settings and interpretations within our working model of understanding
eigentheory. For example, the stem of Question 1 is a numeric interpretation; its
given justifications for students to choose as true and relevant, we see that
justification (i) is a symbolic interpretation in the setting, (iii) is a symbolic
interpretation in the setting, and (v) is a geometric interpretation in
the setting (Figure 1). As students choose justifications that support their
answer to the main question, they are prompted to reason about eigenvectors and
eigenvalues within and across multiple settings and interpretations.

METHODS
The Eigentheory MCE was given to two introductory linear algebra classes taught by

278 [Link]:indrum2018:174806
the same instructor at a large, research-intensive public university in the United States
at the end of Spring Semester 2016. The course utilized the Inquiry-Oriented Linear
Algebra ([Link] curricular materials and Lay (2012) as its textbook.
One class (of 29 students) received the MCE with given closed-ended justifications
(see Figure 1), and the other class (of 28 students) received an open-ended version
where students had to write their own justifications for their multiple-choice answer;
we refer to these as Class C and Class O, respectively. Students had 20-25 minutes to
work on the assessment. All student work referred to in this paper is labelled “B#.”
Analysis of the closed-ended MCE consisted of entering the data into spreadsheets
and looking for trends such as: (a) common sets of justifications that students selected
or did not select; (b) how selecting certain justifications may have influenced
students’ multiple choice selection; and (c) instances in which we would have
expected students to select what we viewed as related justifications, but they did not.
We used Grounded Theory (Glaser & Strauss, 1967) to characterize the concepts
students brought to bear in their justifications in the open-ended MCE, coding
independently and discussing our results as a team to find emerging themes. Finally,
we compared students’ responses across questions and across classes to discover
further insight into student understanding of eigentheory.

RESULTS
We include four insights into student understanding of eigentheory discovered from
our MCE data analysis. These selected results are organized by settings (sections 1-2)
and interpretations (section 3) from our working model of understanding eigentheory.

Figure 2: B65’s reasoning within the setting.

Reasoning about relationships within or

We found that as students respond to an MCE question, they seem to situate it within
a particular setting, perhaps the setting they are most familiar or comfortable with,
regardless of the setting in which the question was initially written. Furthermore, a
student’s chosen setting can lead to different ways of reasoning about a problem. We
present two illustrations of this from Class O: B65’s justifications for Q1 and Q2, and
B66’s justifications for Q1 and Q3. First, in Figure 2, B65 seemed to situate both
problems within the setting. On Q1, B65 explained that multiplying the
vector by the matrix resulted in six times that vector. On Q2, B65 explained that

279 [Link]:indrum2018:174806
 needed to be a scalar multiple, and thus the only possible vector would be one
along the same line as the vector , namely the vector . In both cases B65
emphasized that for an eigenvector, multiplying by the matrix yields a multiple of the
original vector, thus working within the setting.
Second, in Figure 3, B66 seemed to situate both problems in the
setting. On Q1, we infer this student first found the matrix , multiplied each
vector from the MCE question by it, and chose the vector that was mapped to the zero
vector. Then, on Q3, although what s/he actually writes is idiosyncratic, we can infer
s/he was still reasoning with the homogeneous equation, imagining the vectors and
being mapped to zero by the matrix , and thus the vector would also map
to zero. In both Q1 and Q3, B66 emphasized an action on the eigenvectors to produce
the zero vector, seemingly invoking the setting.

v is a linear combination of y & z

Since the value 2 already causes y & z to equal 0, adding a multiple to it
will not change that

Figure 3: B66’s reasoning with the setting.

We note that the stems of Q1, Q2, and Q3 are not written so as to elicit student
reasoning within a particular setting. This allows for use of the open-ended
assessment to measure a student’s preferred setting or for the closed-ended
assessment to measure whether students can interpret the problem in either setting
Reasoning about Eigenspaces
The previous section provides examples of the relationships involved in the first two
settings: to . In this section we attend to the eigenspace
setting, which focuses on the relationships involved with scalar multiples or linear
combinations of eigenvectors. An eigenspace, like any vector space, is closed under
addition and scalar multiplication; thus, a linear combination of vectors in an
eigenspace is also an eigenvector with the same eigenvalue as the other vectors in
that eigenspace. When asked about eigenspaces, students may draw on these facts
and/or may work within one of the previous two settings to derive these principles.
For Q3, only six (of 28) students in Class O circled the correct answer (a) that is an
eigenvector with eigenvalue 2, five chose (b) an eigenvalue of 5, and 16 chose (c)
that was not an eigenvector. Approximately half the students in each group used the
phrase “ is a linear combination of” as part of their justification (3 for (a), 3 for (b), 7
for (c)). Sample justifications using “is a linear combination of” are given in Figure 4.

280 [Link]:indrum2018:174806
 Justification given Justification given Justification given Justification given
with choosing (a) with choosing (c) with choosing (b) with choosing (b)
is a linear combination No, because is a is a linear combination of and Both Since it is a linear
of and which have linear combination and are scalar multiples of their combination of the other
the same eigenvalue. of the two vectors. previous form so the resultant vector will eigenvectors, it would also
be an eigenvector as well be an eigenvector.
NOTE: Typed versions are used here to improve readability of students’ handwritten justifications

Figure 4. Example of four students’ open-ended justifications for Question 3

Very few students in Class O gave justifications that brought in the relationship
between eigenvectors and eigenvalues described in the first two settings. One such
student was B66, described in the above section. In Class C, however, 13 (of 29)
students chose justification (iv) (symbolic ), 11 of which correctly selected
answer (a). Because a much higher percentage of Class C circled (a) than in Class O,
it is possible that this justification served as a hint that helped some students choose
the correct answer. On the other hand, this MCE option allowed us to test whether
students recognized the relevance of this set of relationships for the given question.
Although some students who answered (c) used the phrasing “linear combination,”
their arguments focused more on the linear independence of the vectors. The answers
students gave for (c) include: “Eigenvectors must be linearly independent from each
other so if is a linear combination of and then it cannot be an eigenvector,”
[B58], and “Because they all correspond to the same eigenvalue they all must have
unique eigenvectors and is a linear combination of and and therefore not unique
and not an eigenvector of ” [B79]. These answers focus on eigenvectors as
necessarily being linearly independent or unique. This focus may come from students
remembering that eigenvectors of distinct eigenvalues are linearly independent or that
textbook solutions often give an eigenspace basis as the final answer, which may
explain students thinking there are only finitely many eigenvectors for an eigenspace.
In Q5, eigenspaces were represented geometrically, and students who completed it
were rather successful in selecting the correct multiple-choice answer (14/21 in Class
C and 21/26 in Class O). However, many students still focused on finite numbers of
vectors. On Q5, reasons given by some students to support the correct choice (b)
similarly focused on finite numbers of eigenvectors: “Matrix already has 3
th
eigenvectors so there’s no room for a 4 ” [B59], and “ is a linear combination of
and , and there are already 3 eigenvectors for 3 dimensions, so cannot be an
eigenvector of ” [B66]. We conjecture these students may have been conflating the
total number of possible eigenvectors (infinite) for a 3x3 matrix with the number of
linearly independent vectors needed to create the bases for the 1- and 2-dimensional
eigenspaces. Alternatively, B58’s justification for Q5 focuses on dimension: “In a
3x3 matrix there can only be 3 dimensions to the eigenspace. and together span
the entire space of so there cannot be another eigenvector of besides and
.” We conjecture grasping the difference between finiteness of dimensions and
infiniteness of eigenvectors may be particularly vital for understanding eigenspaces.

281 [Link]:indrum2018:174806
Reasoning Across Interpretations
We conclude our results by discussing various aspects of students reasoning across
interpretations and the ways in which the MCE afforded that. In particular, we focus
on symbolic versus geometric interpretations of eigentheory. On Q1, as noted in the
previous section, a majority of students in Class O wrote at least one equation
(symbolic interpretation), but none wrote anything geometric in their justifications.
This could be an indication that students might favour algebraic reasoning over
geometric reasoning when justifying their answers to eigentheory questions, even
though the classes used the IOLA curriculum which specifically introduces
eigenvectors and eigenvalues geometrically. On the other hand, it could be that the
numeric interpretation that Q1 was written in did not elicit geometric interpretations
from students in their open-ended justifications, or that students see symbolic
justifications as more acceptable to the teacher or the broader math community than
geometric ones. In a more direct way of assessing students’ ability to see connections
to the geometric interpretation, the closed-ended MCE gives students the geometric
justification choice (v) on Q1, and roughly half (14/29) of the students in Class C
selected it. Furthermore, over 80% of the total students from both classes answered
the multiple choice stem of Q1 and Q2 correctly (48/57 and 51/57 respectively),
demonstrating some ability to reason both numerically and geometrically about
eigenvectors and eigenvalues. Because the wording of the MCE questions and
justifications makes use of the four different interpretations from our working model,
we are better able to assess students’ understanding of the symbolic, numeric,
geometric, and verbal interpretations in eigentheory, both within and cross settings.

DISCUSSION
Research on student thinking often relies on students’ written work on mathematics
problems as evidence of how students make sense of or reason about particular
content. Our research here is no exception, with student work on the MCE revealing a
variety of ways that students understand aspects of eigentheory. However, the MCE’s
closed-ended justifications extend a written question’s ability to examine connections
between settings and interpretations that students might not have initially considered
or felt the need to include in their justifications. For instance on Q1 in Class O, four
students wrote some form of as part of their justification, ten wrote
some form of , four wrote some form of both equations, and ten students did
not write either equation. In contrast, on Q1 in Class C, 23 students selected both
justifications (i) (symbolic ) and (iii) (symbolic ), and only
one student selected neither. Hence, when students were forced to consider the two
eigentheory settings (i) and (iii), the large majority was able to see both as true and
relevant. As other researchers have pointed out the importance of understanding both
equations in eigentheory, it is significant that the MCE may give new insight in
students’ understanding of connections between these two settings.
We do see some potential limitations of the MCE. First, it is more time consuming to

282 [Link]:indrum2018:174806
take than a simple multiple-choice test, and this affects the number of questions that
can be asked. The MCE can also be cognitively taxing because students must
consider each justification to determine its truth and relevance. Third, scoring MCE
results can be complicated. We hope that further refinement and use of the MCE, as
well as developing possible scoring systems, will continue to broaden and deepen the
mathematical community’s understanding of how students reason about eigentheory.

REFERENCES
Artigue, M., Batanero, C., & Kent, P. (2007). Mathematics thinking and learning at
post-secondary level. In F. K. Lester Jr. (Ed.), Second handbook of research on
mathematics teaching and learning (pp. 1011–1050). Reston, VA: NCTM.
Çağlayan, G. (2015). Making sense of eigenvalue-eigenvector relationships: Math
majors' linear algebra–geometry connections in a dynamic environment. Journal of
Mathematical Behavior, 40, 131–153.
Carlson, M., Oehrtman, M., & Engelke, N. (2010). The precalculus concept
assessment: A tool for assessing students’ reasoning abilities and understandings.
Cognition and Instruction, 28(2), 113–145.
Cline, K., & Zullo, H. (2016). MathQUEST/MathVote. Retrieved from
[Link]
Cobb, P., & Yackel, E. (1996). Constructivist, emergent, and sociocultural
perspectives in the context of developmental research. Educational Psychologist,
31, 175–190.
Dorier, J.-L. (Ed.) (2000). On the teaching of linear algebra. Dordrecht, The
Netherlands: Kluwer Academic.
Gol Tabaghi, S., & Sinclair, N. (2013). Using dynamic geometry software to explore
eigenvectors: The emergence of dynamic-synthetic-geometric thinking.
Technology, Knowledge and Learning, 18(3), 149–164.
Glaser, B., & Strauss, A. (1967). The discovery of grounded theory: Strategies for
qualitative research. Chicago, IL: Aldine Publishing Company.
Harel, G. (2000). Three principles of learning and teaching mathematics. In J.-L.
Dorier (Ed.), On the teaching of linear algebra. Dordrecht, The Netherlands:
Kluwer Academic Publishers.
Henderson, F., Rasmussen, C., Sweeney, G., Wawro, M, & Zandieh, M. (2010).
Symbol sense in linear algebra. Proceedings of the 13th Annual Conference on
Research in Undergraduate Mathematics Education, Raleigh, NC.
Hestenes, D., Wells, M., & Swackhammer, G. (1992). Force concept inventory. The
Physics Teacher, 30, 141–158.
Hillel, J. (2000). Modes of description and the problem of representation in linear
algebra. In J.-L. Dorier (Ed.), On the teaching of linear algebra (pp. 191–207).

283 [Link]:indrum2018:174806
 Dordrecht, Netherlands: Kluwer.
Lay, D. C. (2012). Linear algebra and its applications (Fourth ed.). Boston, MA:
Pearson.
Nyman, M. A., Lapp, D. A., St John, D., & Berry, J. S. (2010). Those do what?
Connecting eigenvectors and eigenvalues to the rest of linear algebra: Using visual
enhancements to help students connect eigenvectors to the rest of linear algebra.
International Journal for Technology in Mathematics Education, 17(1).
Rasmussen, C., & Wawro, M. (2017). Post-calculus research in undergraduate
mathematics education. In J. Cai, (Ed.), The compendium for research in
mathematics education (pp. 551-579). Reston VA: NCTM.
Rasmussen, C., Wawro, M. & Zandieh, M. (2015). Examining individual and
collective level mathematical progress. Education Studies in Mathematics, 88(2),
259-281.
Salgado, H., & Trigueros, M. (2015). Teaching eigenvalues and eigenvectors using
models and APOS Theory. The Journal of Mathematical Behavior, 39, 100–120.
Thomas, M. O. J., & Stewart, S. (2011). Eigenvalues and eigenvectors: Embodied,
symbolic and formal thinking. Mathematics Education Research Journal, 23(3),
275–296.
Watson, K., Wawro, M., Zandieh, M., & Kerrigan, S. (2017). Knowledge about
student understanding of eigentheory: Information gained from multiple choice
extended assessment. In A. Weinberg, C. Rasmussen, J. Rabin, M. Wawro, and S.
Brown (Eds.), Proceedings of the 20th Annual Conference on Research in
Undergraduate Mathematics Education (pp. 311-325), San Diego, CA.
Wawro, M., Watson, K., & Christensen, W. (2017). Meta-representational
competence with linear algebra in quantum mechanics. In A. Weinberg, C.
Rasmussen, J. Rabin, M. Wawro, and S. Brown (Eds.), Proceedings of the 20th
Annual Conference on Research in Undergraduate Mathematics Education (pp.
326-337), San Diego, CA.
Wilcox, B., & Pollock, S. (2013). Multiple choice assessment for upper-division
electricity and magnetism. In P. V. Engelhardt, A. D. Churukian, & D. L. Jones
(Eds.), 2013 PERC Proceedings (pp. 365–368).
Zandieh, M., Plaxco, D., Wawro, M., Rasmussen, C., Milbourne, H., & Czeranko, K.
(2015). Extending multiple choice format to document student thinking. In T.
Fukawa-Connelly, N. Infante, K. Keene, and M. Zandieh (Eds.), Proceedings of
the 18th Annual Conference on Research in Undergraduate Mathematics
Education (pp. 1094-1100), Pittsburgh, PA.
Zandieh, M., Wawro, M., & Rasmussen, C. (2017). An example of inquiry in linear
algebra: The roles of symbolizing and brokering. PRIMUS: Problems, Resources,
and Issues in Mathematics Undergraduate Studies, 27(1), 96–124.

284 [Link]:indrum2018:174806
TWG 4: Students’ practices
 The complexity of knowledge construction in a classroom setting
Tommy Dreyfus1, Chris Rasmussen², Naneh Apkarian², and Michal Tabach1
1
Tel Aviv University, Israel, TommyD@[Link]
² San Diego State University, CA, USA
We study a class of mathematics education MA students in an introductory course on
Chaos and Fractals, as they grapple with the Sierpinksi triangle, and in particular
with the apparent paradox that its area equals 0, while its perimeter is infinitely long.
For this purpose, we network an approach for investigating the construction of
knowledge in small groups with one for examining how ideas and ways of reasoning
function-as-if-shared in a classroom. Our results show complexities: (i) small group
work and whole class discussions mutually influence each other; (ii) ideas may
function-as-if-shared in the whole class even if the majority of students have not
previously constructed them in their groups; (iii) knowledge constructed in the small
groups may or may not later function-as-if-shared in the whole class.
Keywords: Teachers’ and students’ practices at university level, teaching and
learning of analysis and calculus, knowledge construction in classrooms, paradoxes
INTRODUCTION
The research presented here deals with the construction of knowledge in a student
centred, inquiry-based classroom, where small group work (SGW) alternates with
whole class discussions (WCDs). Construction of knowledge is usually investigated
by observing small groups (1 to 4 students) of students. The reason for this is that in
larger groups, the density of information for each student is low and does not allow
the researcher to interpret their utterances or actions. However, intentional learning
more often than not takes place in classrooms with many more than 4 students. We
therefore use different approaches for analysing the SGW and the WCDs. The aim of
our research is to link the two analyses by following ideas from their emergence in
SGW or WCD, via their flow between SGW and WCD settings, until they possibly
function-as-if-shared in the class, even though they may not have been constructed by
all students. We thus aim at tracing and describing the complexity of knowledge
construction across several classroom settings.
THEORETICAL BACKGROUND
The perspective we adopt for analysing the construction of knowledge during SGW is
Abstraction in Context (AiC), a theoretical framework for analysing processes of
constructing abstract mathematical knowledge (Dreyfus, Hershkowitz, & Schwarz,
2015). AiC methodology begins with an a priori task analysis identifying the new (to
the learner) knowledge elements required or useful when solving the task. It then uses
a model of three types of epistemic actions – actions pertaining to the knowing of the
learners – to analyze their learning processes. The model suggests constructing as the
central epistemic action of mathematical abstraction. Constructing consists of
assembling, interweaving and integrating previous constructs to make a new

286 [Link]:indrum2018:171528
construct emerge. It refers to the first time the new construct is expressed or used by
the learner. Hence, while the term constructing refers to the process, the term
construct refers to the outcome of the action.
The perspective we adopt for analysing WCD episodes is documenting collective
activity (DCA). Collective activity of a class refers to the ways of reasoning that
function-as-if-shared (FAIS) as students work together to solve problems, explain
their thinking, represent their ideas, and so on (Rasmussen & Stephan, 2008). These
FAIS ways of reasoning can be used to describe the mathematical activity of a group
and may or may not be appropriate descriptions of the characteristics of each
individual student in the group. The empirical evidence that a way of reasoning is
FAIS is obtained by using Toulmin’s (1958) model of argumentation, the core of
which consists of Data, Claim, and Warrant. Typically, the data consist of facts or
procedures that lead to the claim that is made. To further improve the strength of the
argument, speakers often provide more clarification, which serves as a warrant for
connecting the data to the claim. Backings provide further support for the core of the
argument. For examples, see the data analysis below, e.g., in WCD 9. The following
three criteria are used to determine when a way of reasoning becomes normative: 1)
When the backing and/or warrants for particular claim are initially present but then
drop off; 2) when certain parts of an argument shift position within subsequent
arguments (e.g., a claim shifts to data); or 3) when a particular idea is repeatedly used
as either data or warrant for different claims across multiple arguments.
In earlier studies (Tabach et al., 2014; Hershkowitz et al., 2014), we have shown how
DCA and AiC combine to provide an in-depth analysis of knowledge shifts in the
classroom and of the knowledge agents that initiate these shifts. In Tabach et al.
(2017), we articulate why and how the two approaches are theoretically compatible.
In this paper, we analyse a lesson where students dealt with an apparent paradox
because of its potential to bring to the fore the complex nature of knowledge
constructing processes across social settings in a classroom. Specifically, the paradox
is an infinite perimeter that delimits a shape with no area, a phenomenon occurring in
fractals. While paradoxes are abundant in the study of infinity, we found only two
studies relating to similar ones: Sacristán (2001) examined how the coordination of
visual and numerical representations supported a single student’s resolution of this
apparent paradox. Wijeratne & Zazkis (2015) found that their students were hindered
by contextual considerations when attempting to resolve a similar paradox of a solid
of revolution with finite volume but infinite surface area. Neither of these studies
focused on the construction of knowledge in a classroom community.
METHODOLOGY
The setting for the research was a course on Chaos and Fractals at a US university,
which formed part of the mathematics requirement toward a master’s degree in
mathematics education. Participants were 11 students with an undergraduate degree
in mathematics, the teacher, and an instructor/observer who occasionally intervened.
The teacher and instructor were both part of the research team. Classes took place

287 [Link]:indrum2018:171528
during one semester twice a week for 75-minutes each; typical class periods
alternated between SGW and WCD. During SGW, students worked in four stable
groups; they were invited to use huddle boards - one table sized white board per
group - in order to promote group communication and to facilitate subsequent whole
class presentation of their work. The teacher and instructor went from group to group,
trying to understand student thinking and attempting to focus students’ activity on
what they saw as the main issues; they did this mainly by asking questions but did not
otherwise intervene in the SGW. The four stable groups will be numbered 1 (Carmen,
Jan and Joy); 2 (Kevin, Elise and Mia); 3 (Soo, Kay and Shani); and 4 (Curtis and
Sam). All names are pseudonyms. Groups 1 and 2 were video-recorded during SGW;
the class was video-recorded during WCDs. In WCDs, groups had the opportunity to
use the huddle boards to share their thinking; there were also teacher led discussions
and short lectures whose aim it was to facilitate reflection on issues having been
discussed by some groups.
On Day 9 (out of 24), class work was based on an activity about the Sierpinski
Triangle (ST). As shown in Figure 1, the ST may be produced by a recursive
procedure: Draw an equilateral triangle; connect the midpoints of its sides; remove
the middle triangle to get three equilateral triangles (of side ½ of the original one);
repeat these steps (including the repetition) for each of the three smaller triangles.
The ST is obtained by means of the (infinite) recursion.

Figure 1. The Sierpinski Triangle (shown after 6 iterations)

The activity was based on a three-part worksheet. In Part A, students were asked to
carry out the recursive procedure six times, blackening the removed triangles on the
huddle boards. In Part B, they were asked to imagine continuing the recursion forever
and to discuss the figure they would obtain, in particular its area and its perimeter. In
Part C, they were asked about properties of the resulting figure, especially comparing
it to its parts.
The teacher had planned for the students to come up with enough properties in Part C
to enable a definition of self-similarity, and as a by-product of learning about self-
similarity, to realize that self-similar objects may have finite (or even zero) area and
infinite perimeter. However, as will be seen below, the class had its own emergent
goals, which led us to investigate student reasoning about area (A) and perimeter (P).
Our task analysis yielded 4 knowledge elements for each of A and P: nature, process,
limit, and infinity, denoted An, Ap, Al, A∞, and Pn, Pp, Pl, P∞. Nature refers to the

288 [Link]:indrum2018:171528
nature and properties of the region at step k whose area/perimeter is being considered.
Process refers to the process of removing the triangles and computing the relevant
area/perimeter. Limit refers to the realization that the process is never-ending and the
areas form a sequence converging to zero whereas the perimeters form a diverging
sequence. Infinity refers to the awareness that the actual figure has zero area and an
infinitely long perimeter. The difference between limit and infinity may be construed
as the difference between potential and actual infinity. These knowledge elements
have been formulated in precise language and operational criteria have been fixed to
decide whether a student or group of students has constructed each knowledge
element. As an example, P∞ is defined as follows: Eventually, there actually is a
figure whose perimeter is longer than any finite curve. Operationally, we will say that
a student has constructed P∞ if the student explicitly claims that the eventual shape or
figure or region has an infinitely long perimeter.
The data used in this paper consisted of transcripts from Day 9, images of the groups’
huddle boards, and researcher notes taken in the classroom. SGW was analysed using
AiC. We present only constructing actions here. In most cases, these will be
attributed to the groups rather than individual students; exceptions will be noted. The
WCDs were analysed using DCA; for each argument (numbered as A1, A2, etc.),
Claim, Data and Warrant were identified, so that the criteria for ideas that FAIS could
be applied.
SMALL GROUP WORK AND WHOLE CLASS DISCUSSIONS
The class on Day 9 started by watching and discussing an excerpt from a video about
fractals with real world examples including a cauliflower, mountains, a magnetic
pendulum and the coast of Britain; this took about 25 minutes and included WCD 1,
SGW 2, and WCD 3. Then students were then asked to start working in groups on the
worksheet; they spent about 26 minutes on drawing according to Part A and
discussing the meaning of repeating the repetition (SGW 4, WCD 5). They used
terms such as “infinite loop” and “zooming in”. The focus of this paper is their work
on Part B of the activity during the remaining 24 minutes, split into three episodes of
SGW (6, 8, 10) and three WCDs (7, 9, 11). The class did not reach Part C on Day 9.
SGW 6 (area as process) - The teacher invited the students to develop a conjecture
about area and perimeter. After a brief discussion about the perimeter showing
confusion (Joy: “in one sense it's infinity, because you keep adding a little bit more.
But it should approach a number, right?”), Group 1 focused on area, and attempted to
compute the area after one repetition. A reminder by the teacher to produce a
conjecture led to a seed of the idea of recursion (Carmen: “That’s one-fourth of it, so
each term maybe three-fourths of it”) thus starting the construction of Ap.
Group 2 quickly came up with a formula (Elise: “So it's three fourths to the n of our
A1?”) and spent the reminder of the time discussing what n means and how to denote
things (e.g., A0 for the initial area). We interpret this as having constructed Ap.

289 [Link]:indrum2018:171528
WCD 7 (computations) - The teacher asked whether the groups had come up with a
conjecture, and the students reacted by presenting computational results. Kevin
presented their result as the sequence (3/4)n and Joy added that they had just started in
this direction after computing the area of the first triangle. Student arguments
included only claims and hence were not analysed per the DCA approach.
SGW 8 (perimeter as process) - Group 2 focused on computing the perimeter (Kevin:
“we have an additional a, we have three halves more a”, and later Mia: “So, it's like,
it's going by a scale of three over two, to the n”). The group also made attempts at
seeing what happens in the long run (Mia: “the perimeter is just keeps getting bigger,
and bigger and bigger”; Elise: “Or is there, like, a limit? That it stops?”). While this
points in the direction of Pl, our interpretation is that they have not constructed Pl yet:
In spite of them having identified what an expert might see as a diverging geometric
sequence, they question whether it converges or not. We also note that some students
may be thinking additively rather than multiplicatively.
Group 1 quickly constructed Ap (Carmen: “And then three-fourths of our three-
fourths”; Jan: “It's alright, we got enough… to do the whole”) and somewhat
hesitatingly, Al (Carmen: “Maybe zero?”; Joy: “No no no, because this is like three-
fourths time three-fourths is nine-sixteenths, and after that would be… what?
Twenty-seven over sixty-four?”; Carmen: “Is it approach… zero? I think it does”;
Joy: “Okay, so you are right, it approaches zero”). There is evidence that they also
constructed A∞ (e.g., Joy: “If you keep filling it in, there's not going to be any white
area”). We note that there was no discussion about An. However, they then held a
long discussion about Pn (e.g., Joy: “So what counts as the perimeter?”; Carmen: “Is
it cumulative perimeter?”). Our analysis resulted in the decision that while
constructing Pn was under way, it had not yet been achieved. Next, they mentioned
aspects of Pp (Joy: “So let's say the perimeter of this is three, we would add in… half
of each. So, like, three… times the half”) without completing a constructing process.
They were reminded by the instructor of the area tending to zero, which brought
tension with respect to the perimeter (e.g., Carmen: “if we keep zooming in, there's
no area, so there can be no fence [perimeter]”).
WCD 9 (the controversy) – This discussion in Group 1 prompted the instructor to ask
for the teacher‘s permission to ask Carmen and Joy to present their controversy to the
class. Carmen’s argument (A2) used as data “there’s no area” and claimed “there’d
be no perimeter” with warrant “there’s nothing to… nothing to put a fence around it”.
Joy, on the other hand, argued (A3) for the opposite, using as data “as you zoom in
there’s more and more to fence”, supported by the warrant that one keeps putting in
more fencing material. This brought about a suggestion that when one removed a
triangle (i.e. colours it black), the perimeter of this black triangle is added to the
existing perimeter. In argument (A4) the claim is that “the perimeter of the, the white
is also the same as the perimeter of the, perimeter of the black part” (Curtis), and this
is based on the data “When you shade it in, you’re adding the perimeter of the black”
(Kevin) with the warrant that “the fence is guarding both properties” (Carmen). When

290 [Link]:indrum2018:171528
encouraged by the instructor to explain Carmen and Kevin’s thinking, Soo built
argument A5: Claim: “So you keep adding the numbers, right?”; Data: “So you have
more areas”; Warrant: “You keep zooming in, you’re going to get more triangles
forming”. Next, Mia argued (A6): Claim: “I see the perimeter increasing and then
this, the unshaded area is what’s left over, and that’s constantly decreasing and going
to zero”; Data: “You’re going to have all these shaded triangles, with perimeters”;
Warrant, upon Carmen’s question “Is this a cumulative perimeter or a perimeter at a
point in time?”: “I see it the first way” (Mia). Several more arguments (A7, A8, A9)
in this WCD focused on the area decreasing and tending to zero in an unending
process of creation. The analysis of arguments in this WCD resulted in two ideas that
FAIS:
FAIS A: Perimeter of white is also perimeter of black; this was a claim in argument
A4 but a justification in argument A6; the justification was that the perimeters of the
shaded (black) triangles cumulatively constitute the perimeter of the remaining,
unshaded, white area. Hence this idea satisfies Criterion 2.
FAIS B: The perimeter is cumulative; this was a claim in A5 and a warrant in A6 (we
will see it serving as justification again in A13), and hence also satisfies Criterion 2.
SGW 10 (connections) - Group 1 had a discussion of all four aspects of perimeter,
completing the construction of, at least, Pn and Pp. They built their thinking on the
fact that they used more and more ink at each stage to draw the additionally generated
bits of perimeter, and concluded (Joy): “I thought it went infinitely, because if you
zoom in, there's more fencing to put in. And if you zoom in there's more fence to put
in”. We have no evidence that they constructed Pl and P∞. In fact, this is unlikely
since they only completed constructing Pn and Pp toward the very end of the SGW.
Moreover, Carmen, while admitting that the perimeter tends to infinity, insisted that
intuitively, no area implies no perimeter.
Group 2 attempted to combine what they knew about the unending processes of area
and perimeter. For example, Mia: “There's nothing for the area, but you're still…
you're counting the perimeter of what you're taking out” and Kevin: “…as soon as
you say - as n approached infinity, that means you're going to computation. So, I
think what we want is something general, like - the area is getting smaller, but the
perimeter is getting larger, and just leave it at that general statement”. Our
interpretation is that they may have started constructing A∞ and P∞ but are still far
from completing these constructions.
The constructing processes resulting from the AiC analysis are summarised in Table
1. The table lists only constructs that we have evidence for; in other words, the fact
that, for example, An does not appear does not mean they have not constructed An – it
only means that An has not been discussed during SGW in a manner that lets us as
researchers conclude that An has been constructed.
Group SGW6 SGW8 SGW10

291 [Link]:indrum2018:171528
 1 (Ap) Ap Al A∞ (Pn) (Pp) (P∞) Pn Pp
2 Ap (Pp) (Pl) (A∞) (P∞)
Table 1. The constructs; parentheses mean (under construction)
WCD 11 (linking area and perimeter processes) - The instructor called on group after
group to present their thinking, however tentative. Group 1 used an ink metaphore,
the ink being used to draw the additional perimeter bits. In this way, they explained
how cutting out further triangles reduces the area while using more ink – thus
increasing the perimeter. Carmen, however, added, that for her “that’s not
resonating”. The researchers identified two arguments while this group was reporting
(A10, A11) and one (A12) while another group was reporting. During their report,
Group 2 connected the perimeter process to the area process like Group 1. Elise built
the following argument (A13): Claim: “The area is getting smaller and the perimeter
is getting bigger”. Data: “You’re adding smaller and smaller pieces”. Warrant: “…but
you’re adding those pieces to what you already have”. Backing: “The perimeter is,
like, all of this, combined with all of this, combined with all of this, combined…”.
We note that the argument focuses almost completely on perimeter, although the
claim equally relates to area. Finally, the report of Group 4 included arguments A14
and A15. We only describe A15, produced by Sam: “So the area would be… go to
zero… There would be limited amount of areas, so we're going to have a limited
number of… perimeters. So, we don't have infinite number of perimeters”. Sam’s
claim of a “not infinite perimeter” is based on the data that the area goes to zero and
hence there is a limited amount of area, with the “limited number of perimeters”
serving as warrant. Based on these arguments, we identified two more ideas that
FAIS:
FAIS C: Unending process of creation; this meets Criterion 3: Continued use of an
idea (e.g., keep adding) across multiple arguments to describe the process that is
being analysed. This idea is related to potential infinity.
FAIS D: Area going to zero; this was repeatedly a claim, including in arguments A6,
A7, A8, A9, A13 and became data in argument A15, hence satisfying Criterion 2.
The relationship between SGWs and WCDs
As a preliminary, we note the richness and diversity of students’ ways of reasoning
about area and perimeter, which in a less student-centred classroom might have been
quickly undermined with an infinite geometric sequence that is decreasing (r<1) and
hence tending to zero for area and an infinite geometric sequence that is increasing
(r>1) and hence tending to infinity for perimeter. We note that the term geometric
was once mentioned briefly with respect to area by Sam during WCD 11 (“it's
geometric, so it's going to converge. So, the area would be… go to zero”) but this
was rather toward the end of class and was not picked up by any of the other students.
Generally, students seem to have been satisfied by arguments of the type “the
sequence is infinite and decreases, hence it tends to 0” (as in the discussion of Group

292 [Link]:indrum2018:171528
1 in SGW 8) or “the sequence is infinite and increases” (e.g., Mia and Elise of Group
2 in SGW 8).
The diversity manifests itself, among others, in the metaphors the students used
(fence, ink), in a tendency to use, at least initially, numerical considerations for area
and perimeter, in the attempt to link the area process with the perimeter process, and
the related discussion about the nature of the perimeter as separating the region that
belongs to the ST from the one that doesn’t. Little of this was initiated or suggested
by the teacher or the worksheet.
This last issue appears as Pn in the AiC analysis of the SGWs and as FAIS A in the
DCA analysis of the WCDs. Similarly, there are relationships between the other
FAIS ideas and knowledge elements. Table 2 shows these relationships.
FAIS A B C D
Constructs Pn Pp Al, Pl Al, A∞
Table 2. Relationship between FAIS and constructs
While FAIS idea A (the perimeter of white is also perimeter of black) is related to
construct Pn, the relationship between the constructing process of P n and the
arguments establishing A as FAIS is complex. We don’t have evidence of Pn having
been constructed in either of the two observed groups before the relevant WCD 9;
and the first argument establishing idea A as a claim (A4) was initiated by Curtis but
immediately supported by Kevin and Carmen. Moreover, the second argument, in
which idea A became a justification, A6, was presented by Mia. This may indicate
that the beginning Pn construction we identified in Group 1 was substantial, and
maybe even that the discussion in Group 2, which on the face of it focused on Pp
caused Kevin and Mia to think about Pn. Finally, we ask ourselves to what extent the
constructing process of Pn continued during WCD 9 for Kevin, Carmen and Mia.
FAIS B (the perimeter is cumulative) – is similar to FAIS C (and intimately related to
it from the point of view of the mathematical content). While we hesitated to claim
that Pp has been constructed by Group 2, it is Mia from that group who produced A6
and Elise from that same group who produced A13, the two arguments where the
element switched position to becoming a justification and thus allowed us, according
to DCA to categorize this idea as FAIS.
FAIS C (unending process of creation) exhibits a case in which the two analyses
connect rather smoothly. Group 1 constructed Ap and Pp, and Group 2 may be
assumed to have implicitly constructed Ap and to be progressing in the constructing
process of Pp. The frequent use of this knowledge element in many WCD arguments
may indicate a similar situation in the other two groups. Soo from the Group 3
produced Argument A5; and Sam from the Group 4 produced Argument A9.
Finally, the relationship between FAIS D (area going to zero) and Al seems obvious
and needs little comment. When this idea functions as if shared in the classroom, it is

293 [Link]:indrum2018:171528
possible that some of the students relate to construct A∞ (as shown above for Carmen
and Joy) and others think in terms of Al or even in terms of Ap only. While to the
expert, thinking in terms of Al may be satisfactory at best, and thinking in terms of Ap
may be insufficient, such differences are tended to be glossed over in this classroom
with respect to the rather basic construct of area, and we may speculate that similar
situations pertain to more complex constructs in this and other classrooms.
It becomes obvious that there are many ways, in which SGW and WCD can interact.
The relationship is by no means unidirectional from constructing an idea in SGW to
this idea FAIS in WCDs. Rather, constructing processes may well be continued or
even initiated in WCDs. On the other hand, ideas may FAIS without having been
constructed by all or even by a majority of the students in class. For example, we
have no evidence for Pl having been constructed in either of the two observed groups,
though Group 2 had started this constructing process; but C (unending process of
creation) is anyway FAIS in relation to both area and perimeter. Of course, it could
have been constructed in the groups we have no data on. This raises the question
whether an idea can FAIS if it has not been constructed at least in some group. We
can only say that we have no example for this having happened in these data.
On the other hand, notions can be constructed by some students, or in some groups,
without ever functioning as if shared. In fact, some of the constructs may not have a
chance to come up in any WCD. We have no unequivocal evidence for this
happening but we do know that in the first few minutes of the next lesson (Day 10),
which took place two days later, Kevin referred to the perimeter as an increasing and
hence diverging geometric sequence. We could also point to the fact that Ap has been
constructed by both analysed groups but does not appear in the lower row of Table 2.
However, this argument is weak since Ap appears indirectly as a component of Al.
CONCLUSION
The complexity of knowledge flow in the classroom, even based on this one class
session, is far greater than one might imagine. Inquiry-based instruction features
students’ deep engagement in mathematics and peer to peer interaction. As such
instruction increases at the university level, the field is in need of theoretically
grounded approaches for analysing individual and collective mathematical progress.
This paper makes a contribution in that direction. A strength of AiC is that the
approach allows researchers to gain insight into the ideas that individuals or small
groups of individuals construct, as long as the number of students remains small.
DCA provides a complementary approach that provides researchers insight into the
ideas and ways of reasoning that characterize the collective progress of the classroom
community. In Tabach et al. (2017), we showed how the two approaches combine
theoretically, and the present paper adds to this a detailed analysis of how an AiC
analysis interacts with a DCA analysis to expresses the complexity of knowledge
construction in the classroom. An open question, and one that we are currently
pursuing with this data, is how to coordinate the small group and classroom level
findings with individual interviews conducted shortly after such rich class sessions.

294 [Link]:indrum2018:171528
Finally, the analysis presented here also contributes to what we know about how
students reason about area and perimeter in a paradoxical situation. In contrast to the
findings of Wijeratne and Zazkis (2015), the students in our classroom found ways to
profitably use contextual considerations, metaphors, numerical computations, and
figural reasoning to support their endeavour to understand the apparent paradox.
Acknowledgment
This work has been partially supported by Israel Science Foundation grant 438/15.
REFERENCES
Dreyfus, T., Hershkowitz, R., & Schwarz, B. (2015). The nested epistemic actions
model for abstraction in context - Theory as methodological tool and
methodological tool as theory. In A. Bikner-Ahsbahs, C. Knipping & N. Presmeg
(Eds.), Approaches to qualitative research in mathematics education: Examples of
methodology and methods (pp. 185-217). Dordrecht: Springer, Advances in
Mathematics Education series.
Hershkowitz, R., Tabach, M., Rasmussen, C., & Dreyfus, T. (2014). Knowledge
shifts in a probability classroom – A case study coordinating two methodologies.
ZDM - The International Journal on Mathematics Education, 46, 363-387.
Rasmussen, C., & Stephan, M. (2008). A methodology for documenting collective
activity. In A. E. Kelly, R. A. Lesh, & J. Y. Baek (Eds.), Handbook of innovative
design research in science, technology, engineering, mathematics (STEM)
education (pp. 195 - 215). New York, NY: Routledge.
Sacristán, A. I. (2001). Students' shifting conceptions of the infinite through
computer explorations of fractals and other visual models. In M. van den Heuvel-
Panhuizen (Ed.), Proceedings of the 25th International Conference for the
Psychology of Mathematics Education, Vol. 4 (pp. 129–136). Utrecht, The
Netherland: PME.
Tabach, M., Hershkowitz, R., Rasmussen, C., & Dreyfus, T. (2014). Knowledge
shifts in the classroom – A case study. Journal of Mathematical Behavior, 33, 192-
208.
Tabach, M., Rasmussen, C., Dreyfus, T., & Hershkowitz, R. (2017). Abstraction in
context and documenting collective activity. In T. Dooley & G. Gueudet (Eds.),
Proceedings of the Tenth Conference of the European Society for Research in
Mathematics Education (CERME10) (pp. 2692-2699). Dublin, Ireland: Dublin
City University and ERME.
Toulmin, S. (1958). The uses of argument. Cambridge, UK: Cambridge University
Press.
Wijeratne, C., & Zazkis, R. (2015). On painter’s paradox: Contextual and
mathematical approaches to infinity. International Journal of Research in
Undergraduate Mathematics Education, 1, 163-186.

295 [Link]:indrum2018:171528
 Analysing tasks for the Flipped Classroom from the perspective of
Realistic Mathematical Education
Helge Fredriksen
University of Agder, Norway, [Link]@[Link]
The advent of Flipped Classroom as a framework for organizing the teaching and
learning of mathematics has the potential to revitalize the attention to tasks as a vehicle
for meaningful learning in tertiary education. Flipped Classroom is based on the idea
of student active learning under close guidance of the university teacher. Due to the
possibility to engage the students in meaningful discovery of how mathematics can
relate to real-life situations during in-class sessions, the tasks are seen to have a
central role in a successful implementation of Flipped Classrooms. This paper explores
the realistic mathematics education (RME) as a theoretical framework for task design
and analysis in the context of in-class flipped engineering mathematics classrooms,
and I seek evidence of knowledge construction through analysing students’ work on
modelling the height of a rider in a double Ferris wheel.
Keywords: Flipped Classroom, RME, Mathematics for engineers, the role of digital
and other resources in university mathematics education

INTRODUCTION
One of the fundamental principles of Flipped Classroom (FC) approaches in education
is the opportunity it provides for a more interactive and meaningful use of class time.
(Bergmann & Sams, 2012). The idea is to use valuable classroom time for the more
advanced learning processes like analysing, evaluating, and modelling in a cooperative
environment consisting of peers and tutors. A contrast to this form of in-class student
participation can be the traditional university teaching style where the lecturer is more
or less seeking to “transfer” the mathematics in a monologue fashion (Sfard, 2014).
Moving the lecturing part out of the classroom as a preparatory activity can help
remedy the problem lecturers often feel is present during the delivery of traditional
lectures, the one of disengaged and passive students. Rather, the social setting of the
classroom can be used to let the students discuss, evaluate, try out, and receive
assistance with their ideas and conjectures. However, such learning processes need a
structuring element, which can be the tasks that students work with, preferably in
groups. The purpose of this paper is to advance knowledge on task design and analysis
in connection with FC implementation, specifically considering the theoretical
framework of RME. I attempt to do this through a case study of students’ work with a
modelling task that follows the heuristics of RME.
Previous research
FC is an innovative teaching approach, initially described by Lage, Platt, and Treglia
(2000) that has many implementations. As the popularity of FC teaching has caught

296 [Link]:indrum2018:173684
up, interest has been considerable also for tertiary educational settings (Wasserman,
Quint, Norris, & Carr, 2017).
Various authors have done research on general issues of FC pedagogy (Abeysekera &
Dawson, 2015; Wan, 2015). However, little research has emerged on analysing how
task design should be addressed in connection to FC implementation. In particular, I
was not able to find any research considering RME as a framework for analysing tasks
in connection with FC in-class activities. Several articles seem to call for more research
on providing insights on FC pedagogic heuristics (Song, Jong, Chang, & Chen, 2017;
Wan, 2015). Task design is of particular importance for the in-class FC component,
since the model of FC considers the classroom activities to be the arena for attaining
the highest levels of skills and knowledge (Bergmann & Sams, 2012). The challenge
is to bridge the out-of-class videos with these in-class activities in a meaningful way.
The particular task analysis performed in this paper is based upon students’ group work
with mathematizing the double Ferris wheel movement. A study using a similar type
of problem is described in Sweeney and Rasmussen (2014). They found that the
students’ bodily engagement using gestures and measurements by fingers facilitated a
link between the movement of the rider and the mathematical model.

THEORETICAL FOUNDATIONS OF RME

According to Van den Heuvel-Panhuizen and Drijvers (2014), RME is a domain-
specific instruction theory for mathematics that focuses on rich, “realistic” situations
serving to initiate development of mathematical concepts, tools and procedures. Core
teaching heuristics give these directions:
 The activity principle emphasises students’ direct participation throughout the
learning process. Mathematics is considered to be best learned by doing
mathematics, which is in accordance with student active learning that is
considered a widely accepted principle of FC.
 The reality principle expresses the importance of presenting students with real-
life situations and problems that they can imagine and mathematize upon.
 The level principle highlights the idea that students pass through various
cognitive development phases, from informal context-related descriptions of the
problem towards the use of more formal mathematical language. This process
sets the stage for bridging student understanding that the model of the context-
related situation at hand can become a model for similar kinds of problems.
 The intertwinement principle emphasises that mathematical content domain
should not be considered as isolated fragments, but rather seen as a connected
whole. This principle supports task designs facilitating open problems that
stimulate students’ own thinking and reasoning about which mathematical
solution techniques and mediating artefacts to employ.
 The interactivity principle relates to the idea that learning mathematics is not
purely an individual pursuit, but rather a social activity, where group work and

297 [Link]:indrum2018:173684
 whole-class discussions should be orchestrated by the teacher. This principle
aligns well with socio-cultural learning theories, as also emphasised by Cobb,
Jana, and Visnovska (2008). They consider semiotic mediation and cultural tools
as important means of conveying mathematical meaning in an RME setting.
 The guidance principle refers to the idea of “guided re-invention” of
mathematics. Instructional sequences in task design can involve historical
evolutionary steps in mathematics as inspiration for rich context problems
(Gravemeijer, 1999).
Further, RME makes a distinction between horizontal and vertical mathematization.
According to Van den Heuvel-Panhuizen and Drijvers (2014), students use
mathematics as a tool to understand and organize problems in a real-life context,
typically during task solution. This is called horizontal mathematization. Vertical
mathematization refers to mathematizing one’s own mathematical activity to reach a
higher level of abstraction.
The research question explored in this paper can be framed in the theory of RME:
“To what extent does RME task design facilitate students’ modelling activities and
knowledge construction in a FC context?”

METHODOLOGY
The data for this paper was collected from video filming the work of two groups of
computer engineering students in their first semester of study. These student groups
were part of a larger research project conducted at a university campus in Norway. FC
teaching was performed throughout the whole year of study. I focus on one in-class
session where students were modelling the double Ferris wheel.
The Flipped Classroom setting
Before this in-class session,
the students had watched an
out-of-class session of videos
that introduced the unit circle
and how the y-component
would map to the sine
function as the angle rotated
(Figure 1). In addition to this,
there were videos showing
examples on how to make Figure 1: Mapping of unit circle y-component to the sine-function
mathematical models using this function, and how one could sketch a particular sine
function based on textual information. Lastly, they were showed examples of how to
determine the period, amplitude, baseline and phase shift by considering various graphs
of this function. These videos were provided using English as the spoken and written
language, while students were using Norwegian as working language in class.

298 [Link]:indrum2018:173684
The particular task that I refer to in this paper was given to the students in connection
with the teaching of trigonometric functions during the second week of the semester.
In this task, they were asked to
create a model of the evolution
of the height of a rider on a
double Ferris wheel. The double
Ferris wheel was represented in
an applet available for the
students to watch on their
computer (Figure 2). The Ferris
wheel consisted of two separate
wheels rotating at the same
speed and centred at the end of a
larger rotating bar as seen on Figure 2: The double Ferris wheel simulation
Figure 2. The students had the possibility to study each rotation separately by
starting/stopping each of them. The height measurement of an imagined rider called
Marit (marked with a dot on one of the wheels), was available as a readout at all time.
The interested reader can study the applet at
[Link] The task presented for the
students was the following:
1. Create a representation of your choice that illustrates Marits’ ride on the double-
Ferris wheel
2. Create a sketch/graph of Marits’ height versus time as she ride on the double
Ferris-wheel
3. Create a function for Marits’ height versus time as she travels on the Double-
Ferris wheel.
4. Graph the functions you found in the previous problem using a graphing utility
like geogebra. Does the graph of your function make sense? Is it similar to the
graph you sketched?
The design is seen to follow the principles of RME. Firstly, the purpose of the task is
to engage students in exploring the dynamics of a familiar situation; the Ferris wheel
ride. This meets the demands of the activity and reality principles. Furthermore, the
sequence of the task, asking students to initially make a representation of the
movement, followed by sketching a graph, and finally creating a function, attempts at
making the students move from more informal to formal mathematizing, respecting the
level principle. In setting the stage for employing trigonometric functions for
mathematical modelling, the intertwinement principle is considered. Further, the
interactivity principle is kept through the group work context of the task. The guidance
principle, however, will not apply to this task design. For most themes in engineering
mathematics, a reinvention can be a considerably challenge to embed in task designs.
Students were placed in groups of 3 or 4 when working with the task, and a 90 minutes
session was spent on the task. The author of the paper was the teacher throughout this

299 [Link]:indrum2018:173684
session1. The study was performed under a naturalistic research paradigm to ensure as
realistic settings as possible for the observation of students’ collaboration
(Moschkovich & Brenner, 2000). The students were filmed using a high definition
camera that followed the group work for the whole session. Prior to the filming,
students had acknowledged their participation in the study through signing a letter of
consent. The selected students had watched the videos beforehand. Statistics about
which videos and how long the students had spent watching them were available in the
tool I used for video-distribution to the students. I analysed each recording using
descriptive accounts (Miles & Huberman, 1994), where sessions of data are broken
into separate entities of activity. Each such entity were subsequently analysed for
content informing us on the research question, and I report on a variety of such episodes
in what follows. In the aftermath of the last case, I performed an interview with two
selected students to obtain deeper insights about certain episodes during the work with
the task and general qualities about videos and the session.
Furthermore, the study was performed on the cohort of 2017/2018, a class consisting
of 20 students. The two groups that I report from had highly varied backgrounds, most
of them from a pre-calculus course comprising two years of mathematics from upper
secondary into 1 year of study at the university. They were acquainted with
trigonometric functions through this pre-calculus course before arriving at the
engineering study. However, as one of the students mentioned in an interview
afterwards, he only remembered sparsely what the various mathematical terms was
named by, but claimed to be able to use them anyway. I filmed two of the groups to be
capture some of the variation taking place inside the single classroom environment. I
will refer to these groups as group 1 and group 2.
I specifically looked for traces of content from the videos, that is, the out-of-class
component of FC when analysing students’ discourse. Such content can be word use,
theoretical elaborations like definitions and theorems, visually mediated ideas and
specific examples.

RESULTS
The two groups seemed to develop a rather different focus in terms of how to approach
the task. One group spent much more time on the formal mathematizing part of the
task, compared to the other one that wanted to understand the movement from a more
informal, empirical viewpoint.
Group 1
The students in this group followed the layout of the task in the sense that they
attempted to sketch a graph of the movement before moving on towards formal
mathematical modelling. This seemed to set the scene for a discussion among the group

1
A pilot-study in a very similar setting was performed on the previous cohort of 2016. This was part of a cooperation
project with San Diego State University, where the research fellow Matt Voigt was responsible for the teaching material
in addition to conducting the in-class teaching.

300 [Link]:indrum2018:173684
members on which physical properties were influencing the model. They constantly
argued about where the “x-line/dotted line” should be, meaning the baseline, in
addition to discussing how the various lengths of the rods and rotation times were
influencing the amplitude and period. However, synthesizing the two movements into
one seemed to form the biggest challenge for the group. For instance, the teacher was
consulted on how one could decide the amplitude on a combined sine. After about 15
minutes of trying to comprehend this, the group moved on to a more formal
mathematical phase of modelling. One of the students, let us call him Finnegan,
appeared to get a breakthrough at about 17:40 (timestamp corresponds to video
recording). Figure 3 show this situation where he is pointing to the printed unit circle
that expresses various exact values of sine and cosine, while he states:
“I suggest that we make
two functions that are zero
when… They are not zero,
3𝜋
but at 2𝜋. Then you
2
would have made a
function that is zero when
the period…Well you have
to make two functions…”
After this episode, the
students seemed to prefer Figure 3: Phase shift of the sine mediated through the unit circle
working in two pairs to
discuss further details of the functions. Both pairs of students agreed that the model
should consist of a sum of two different sine functions.
Analysis
We can see that Finnegan struggles with the correct phrasing of what he is trying to
express, but the visual connection between the Ferris wheel circle and the unit circle is
guiding Finnegan towards an understanding of phase shifting. The printout of the unit
circle became an important mediating artefact for Finnegan, although shared with
students earlier for solving trigonometric equations, not for modelling. They spent
much of the time getting accustomed to the mathematical terminology with amplitude,
period and, as one sees, phase shifts. The informal part of the task where students were
asked to elaborate on a sketch of the function seemed to facilitate these discussions,
much in line with the interactivity and level principles of RME.
I was unable to make any clear connection to the videos from the discourse of these
students. During the interview with one of the students in this group, it was pointed out
that using English as the formal language in the videos was difficult to handle.

301 [Link]:indrum2018:173684
Group 2
The other group that was filmed
during this session utilized a variety
of means to depict the movement of
the Ferris wheel. In Figure 4 we can
see two of the students working in
parallel, where one is considering
using a graphing calculator, while
the other is attempting to graph the
model on GeoGebra. On the right Figure 4: Utilizing various digital tools to model
PC the applet with the simulation is
displayed. Another member in this group even mapped the movement from the PC
screen to a sheet of paper, letting the pen trace the movement of a point on the Ferris
wheel, similar to what was reported in Sweeney and Rasmussen (2014). The group also
attempted other practical means of tracing the movement like taking screenshots in
timed intervals and putting the variable height values displayed in the applet in a
function table for later analysis. It was even suggested that they would let GeoGebra
use these data-points to produce an expression for them using interpolation. After
working with various such “empirical” approaches to the problem, mostly on an
individual basis, one student arrived at an expression for the model at 49:16. He gave
this explanation to the others in the group on how he was able to arrive at a summation
of two sinusoidal functions:
“Analyse the small circle, height over time, that became f(x), then we did the same with
the large one, called this g(x), then we made a new one h(x) consisting of the two of them
together”
Analysis
The big surprise in this session was that the one student that had not prepared watching
the videos seemed to be the leading the group discussions. It might seem that RME can
act to remedy group dynamics where certain individuals come unprepared. The reality
principle in RME stresses the importance of using real-life situations which is likely to
provoke engagement, discussions and active participation during group work.
In this particular data set, it seemed hard to find a direct link between not being
prepared by the out-of-class session and poor participation in the group work although
this was a prevalent finding in other analyses of group work (Fredriksen, Hadjerrouit,
Monaghan, & Rensaa, 2017).
An important observation that was made for both these groups was the physical
arrangement of four students placed face to face, which made it harder to collaborate
on the task. This was a setup originally meant to spur more discussion since the students
would face each other. However, it turned out to be a poor design in this particular
setting due to the importance of using the laptops for controlling the visual simulation
in addition to the actual modelling in a digital graphical environment. It became

302 [Link]:indrum2018:173684
apparent that two students would look at one PC, while the adjacent students would
look at an opposing one, effectively hindering much of the discussion across the group
(Figure 3 and Figure 4) The teacher recognized this, trying to convince students to look
at the same PC, but they quickly went back to work in pairs again. One of the students
also confirmed this observation during the post-interview. Thus, it seems that this
configuration interfered with the interactivity principle of RME for the groups.

DISCUSSION
In this study, I investigated to what extent RME task design facilitates students’
modelling activities and knowledge construction in a FC context.
The analysis found students were making active use of computers, calculators,
printouts of the unit circle, internet, and even their mobile phones to support their
exploration of the task. This can be seen to support the intertwinement principle:
students were actively using various tools, in addition to the mathematics introduced
in the videos, to explore the problem.
The second group seemed to have problems progressing towards the formal
mathematizing part of the task. An observation made from studying the group
dynamics was a quite individualistic working attitude. There seemed to be little
collaboration on working towards a common model for the group, effectively violating
the interactivity principle of RME. During the post-interview with one of the students
in the group, I asked him if he shared this impression, which he confirmed, stating that
some kind of internal competition among the members evolved.
The guidance principle of RME refers to students’ re-invention of mathematics.
Although there are examples of RME facilitating such inductive discovery
(Gravemeijer, 1999), we can conjecture that the FC out-of-class preparatory
component turns this table. Through the videos, the students should have attained basic
knowledge of mathematics that is to be articulated through the modelling activities in-
class. Thus, instead of re-inventing the mathematics, they are rather re-employing it at
a given situation and, through this, making conceptual ties towards it.
We can clearly see how the level principle of RME is at play in these observations.
Initially, students are grappling with understanding the basics of the movement, using
gestures, making sketches and talking to each other using informal language. The next
stage of the task prompts students to express a formal mathematical description of the
movement, which all students achieved to some extent. Although the students work
was directed towards making a model of this particular movement, the whole-class
discussion at the end of the session pointed towards similar situations that this could
apply as a model for. This was also done in the videos prior to class.
The reality and activity principle seem to be well safeguarded in this task. This is also
supported by the statements by one of the students during the informal interview
conducted after the session:
“To see the task in a physical real framing was motivating”

303 [Link]:indrum2018:173684
Throughout the session, students used the familiar terms and concepts related to the
sine function. Words like amplitude, phase shift, period and baseline were all referred
to, and the students seemed to be able to connect these to the real-life situation of the
Ferris wheel. If this really stems from watching the videos or originates from previous
experience with the sine function, I have no direct evidence. This of course influences
the capacity to answer the research question. Light on this may be shed by statements
from the other student interviewed:
“It’s very nice that you can see the videos, think about it for some days, get it refreshed in
the lesson and then start working with tasks related to it. This has helped me a lot”.
Although this is a statement about the general nature of the teaching layout, it
nevertheless gives indication about FC as an instructional platform giving learners a
potential for experiencing different motivations for mathematics.

CONCLUSION
When working with task design in mathematics, we employ various theoretical
frameworks that has certain sets of design principles that the designer should adhere to
(Kieran, Doorman, & Ohtani, 2015). We have seen that the Ferris wheel task design
presented in this paper aligns well with the principles of RME for task design.
However, designing such tasks for a FC pedagogic context will need an additional
component, namely the adoption of the video preparation. Clearly, close attention to
the pedagogical structure combining videos and tasks seem to be necessary for
facilitating rich discussions, modelling activity and knowledge construction to take
place in-class. This statement is supported by the cases that were analysed in this paper,
where we saw students engaged in horizontal mathematization utilizing out-of-class
preparation through videos and mathematical modelling of realistic situations.
Acknowledgements
The author is indebted to Matt Voigt and his fellow research team at SDSU, who were
helpful in supporting this research. I also wish to thank MatRIC for funding and
assisting the cooperation between our institutions.

REFERENCES
Abeysekera, L., & Dawson, P. (2015). Motivation and cognitive load in the flipped
classroom: definition, rationale and a call for research. Higher Education
Research & Development, 34(1), 1-14. doi:10.1080/07294360.2014.934336
Bergmann, J., & Sams, A. (2012). Flip Your Classroom, Reach Every Student in Every
Class Every Day. Washington DC: International Society for Technology in
Education.
Cobb, P., Jana, Q. Z., & Visnovska, J. (2008). Learning from and Adapting the Theory
of Realistic Mathematics Education. Education et didactique, 2(1), 105-124.

304 [Link]:indrum2018:173684
Fredriksen, H., Hadjerrouit, S., Monaghan, J., & Rensaa, R. (2017). Exploring
Tensions in a Mathematical Course for Engineers utilizing a Flipped Classroom
Approach. Paper presented at the CERME10, Dublin.
Gravemeijer, K. (1999). How Emergent Models May Foster the Constitution of Formal
Mathematics. Mathematical Thinking and Learning, 1(2), 155-177.
doi:10.1207/s15327833mtl0102_4
Kieran, C., Doorman, M., & Ohtani, M. (2015). Frameworks and Principles for Task
Design. In A. Watson & M. Ohtani (Eds.), Task Design In Mathematics
Education: an ICMI study 22 (pp. 19-81). Cham: Springer International
Publishing.
Lage, M. J., Platt, G. J., & Treglia, M. (2000). Inverting the Classroom: A Gateway to
Creating an Inclusive Learning Environment. The Journal of Economic
Education, 31(1), 30-43. doi:10.1080/00220480009596759
Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded
sourcebook (2 ed.). Thousand Oaks, CA.: Sage Publications, Inc.
Moschkovich, J. N., & Brenner, M. E. (2000). Integrating a Naturalitic Paradigm Into
Research on Mathematics and Science Cognition and Learning. In A. E. Kelly
& R. A. Lesh (Eds.), Handbook of research design in mathematics and science
education (pp. 457-486). New Jersey: LEA Publishers.
Sfard, A. (2014). University mathematics as a discourse – why, how, and what for?
Research in Mathematics Education, 16(2), 199-203.
Song, Y., Jong, M. S. Y., Chang, M., & Chen, W. (2017). "How" to Design, Implement
and Evaluate the Flipped Classroom? - A Syntheis. Educational Technology and
Society, 20(1), 180-183.
Sweeney, G., & Rasmussen, C. (2014). Reconceiving Modeling. An Embodied
Cognition View of Modeling. In L. D. Edwards, D. Moore-Russo, & F. Ferrara
(Eds.), Emerging perspectives on gesture and embodiment in mathematics (pp.
197-226). Charlotte NC, USA: Information Age Publishing.
Van den Heuvel-Panhuizen, M., & Drijvers, P. (2014). Realistic Mathematics
Education. In S. Lerman (Ed.), Encyclopedia of Mathematics Education (pp.
521-525). Dordrecht: Springer Netherlands.
Wan, N. (2015). Technology Integration and the Flipped Classroom. In N. Wan (Ed.),
New Digital Technology in Education (pp. 149-170). Switzerland: Springer
International Publishing.
Wasserman, N. H., Quint, C., Norris, S. A., & Carr, T. (2017). Exploring flipped
classroom instruction in Calculus III. International Journal of Science and
Mathematics Education, 15(3), 545-568.

305 [Link]:indrum2018:173684
 Engaging with feedback: How do students remediate errors on their
weekly quiz
Cillian Copeland1, Emma Howard1, Maria Meehan1, and Andrew Parnell1,2
1
School of Mathematics and Statistics, University College Dublin, Dublin, Ireland
2
Insight Centre for Data Analytics, University College Dublin, Dublin, Ireland
[Link]@[Link]
Maths for Business is a first-year mathematics module for approximately 500 non-
mathematics specialists. It has continuous assessment consisting of ten weekly
quizzes, worth 40% of the final mark. In 2016/17, students who did not receive the
maximum five marks on their weekly quiz were offered the opportunity to resubmit
their quiz, with correction(s) and an explanation of their error(s), for one additional
mark. We refer to this process as ‘remediation’. In this paper, we examine how
students remediate their errors in order to identify features of a ‘good’ remediation.
These features are identification, description, and correction of errors. By analysing
a subset of students (n=31), we observe that a student’s quiz mark, and the cognitive
level of the quiz question may impact the nature of the remediation provided.
Keywords: assessment practices in university mathematics education, feedback,
remediation of errors, students’ practices at university level.

INTRODUCTION
Maths for Business is a core first-year mathematics module for non-mathematics
specialists enrolled on three business programmes in University College Dublin,
Ireland. Topics from one- and two-variable Calculus are covered in the module and
given the cumulative nature of the content, students should (ideally) achieve the
learning outcomes for a topic before proceeding to the next. To encourage mastery of
learning outcomes, the module has a continuous assessment component consisting of
ten weekly quizzes, worth 40% of the final mark. A week after sitting a quiz, graded
quizzes are returned to students with a mark out of five, and tutors provide oral
feedback to each tutorial group highlighting the most common errors made. In
addition, the lecturer provides an online video entitled “Most Common Errors” and
posts a pdf of the quiz solutions online. With our focus on mastery, we believe that
students who do not get full marks on a quiz should engage with the feedback to
identify and remediate their errors in a timely manner. However, Gibbs and Simpson
(2004) discuss how, even if timely and good quality feedback is provided to students,
there is no guarantee they will engage with it. Therefore, to encourage this
engagement, in 2016/17 we offered students who did not receive full marks on a
quiz, one extra mark if they resubmitted their graded quiz one week after it was
returned with error(s) identified and corrected. We refer to this process as
“remediation”.
1

306 [Link]:indrum2018:174756
Handley, Price and Millar (2011) propose a shift from examining feedback
evaluation and attributes to investigating the process of students’ engagement with
feedback. To this end, we first wanted to explore: which students were most likely to
participate in the remediation process; which feedback resources were they most
likely to access; and, whether engagement in the process impacted academic
achievement in the module. The analysis and findings from this part of the study are
described in detail in Howard, Meehan and Parnell (under review). The main
findings were that 70% of students who had the opportunity to remediate did so; the
most accessed feedback resource was the pdf of the quiz solutions which were made
available online; and, students who achieved an average quiz mark of 3-4 (excluding
remediation marks) and who consistently engaged in the remediation process,
exhibited the most learning gains as measured by their performance on the final
examination. Secondly, we wanted to examine how students remediated quizzes in
order to identify aspects of a “good” remediation, and from these findings, refine the
instructions given to students at the start of the module on how to remediate their
quizzes. We also want to identify and explore what factors might influence the nature
of students’ remediations. It is this second part of the study that we wish to focus on
in this paper by addressing the following research questions:
1. What ways, in general, do students remediate their weekly quizzes?
2. What instructions would we give to future students to assist them in
remediating their quizzes?
3. What factors may influence the nature of a student’s remediation?

LITERATURE REVIEW
Assessment and feedback
There have been a number of in-depth reviews in the area of assessment and
feedback (Bennett, 2011; Sadler, 1989) with some focusing specifically on higher
education (Evans, 2013). Assessment is generally discussed under the headings of
formative assessment, where the primary objective is to provide feedback to the
student and evaluation of students’ knowledge is secondary; and summative
assessment, where the primary role is to evaluate students’ knowledge and feedback
is secondary. Ramaprasad (1983, p. 4) describes feedback as “information about the
gap between the actual level and the reference level of a system parameter which is
used to alter the gap in some way”. Building on this description of feedback in terms
of its effect rather than its content, Sadler (1989) argues that the learner must:
(a) possess a concept of the standard (or goal of reference level) being aimed
for, (b) compare the actual (or current) level of performance with the standard,
and (c) engage in appropriate action which leads to some closure of the gap (p.
121, italics in original).

307 [Link]:indrum2018:174756
While most of the major studies on assessment and feedback relate generally to a
variety of subjects, there have been calls for specific domain-focused research
(Bennett, 2011). Specifically, in mathematics education at the university level the
area of assessment and feedback seems to be under-researched. Of those studies
conducted in this area, an emphasis on summative assessment and closed-book
examinations has been noted (Iannone & Simpson, 2011; Iannone & Simpson, 2012;
Trenholm, Alcock & Robinson, 2015). Underpinning the need to conduct discipline
specific research in this area, Iannone and Simpson (2013) found that in contrast to
the general literature on assessment, mathematics students prefer traditional closed-
book examinations to more alternative assessment methods.
Engagement with feedback
There has been recognition that despite timely and informative feedback being
provided to students, students may not take action on it (Gibbs & Simpson, 2004;
Handley et al., 2011). Handley et al. (2011) emphasise the difference between the
student who skims and bins the feedback to one who takes “responsibility for
understanding, interpreting and applying assessment feedback” (p. 557). Price,
Handley and Millar (2011) discuss how a student may reject feedback “due to lack of
understanding, or based on identity or self-efficacy issues” (p. 892). They further
state that students may need more support in taking action on feedback. Of course,
feedback needs to be of an appropriate level to help the student. Similar to the
remediation process, Covic and Jones (2008) provided psychology students with the
opportunity to remediate corrected essay assignments. In this voluntary remediation,
48% of students opted to remediate for potentially higher marks. Students’ feedback
consisted of individual and group feedback on their corrected essays as well as an
initial grade. We have been unable to find an equivalent study in a mathematics
context.
MODULE CONTEXT AND DATA COLLECTION
For our analysis, we only considered students who were completing Maths for
Business for the first time and who sat the final examination in the module (n=470).
In Maths for Business, students have the choice of completing the module through
using online videos or by attending lectures or a combination of both (Howard,
Meehan & Parnell, 2017). The students are assigned to one of two lecture cohort
groups and have three lectures weekly. The lectures are designed to be partly
interactive with at least 15 minutes for in-class tasks. All students have access to 67
videos/screencasts which cover the entire module content and have an average length
of 7 minutes each. These videos were designed and developed by one of the two
module lecturers (and third author of the paper). There are no recommended
textbooks for this module. Students also have access to the Maths Support Centre,
and prior research has shown that students focus on using module resources with
very little if any use of external resources such as websites. A student’s final mark on
3

308 [Link]:indrum2018:174756
the module consists of 40% continuous assessment and 60% for the final
examination. To encourage consistent engagement with the module and mastery of
learning outcomes, Maths for Business has ten fifteen-minute weekly quizzes, with
each quiz usually consisting of two parts. Each quiz accounts for 5% of the final
mark, directly relates to the module’s content for the prior week, and is marked out
of five. However, only students’ best eight quizzes contribute towards their
continuous assessment mark of 40%.
One week after completion of a quiz, tutors returned the graded quiz to each student
with a mark for both parts of the quiz and the overall mark provided. Tutors also
provided oral feedback to their tutorial class on the most common errors made in the
quiz. There were approximately 50 students registered for each tutorial. On the
university’s Virtual Learning Environment (VLE) Blackboard, the lecturer provided
a video entitled “Most Common Errors” and a pdf copy of the quiz solutions that
also indicated the most relevant online videos from the module that a student may
wish to revise. In Semester 1 of 2016/17, students who did not receive the full five
marks on their quiz were given the opportunity to resubmit their remediated quizzes
for one additional mark. The following instructions were provided by the lecturer to
students:
• When your quiz is returned to you go over it and identify your errors. Write a
sentence beside each error on the quiz sheet so that when you are revising the
material again, you will have a note to yourself about where you went wrong.
• If it is the case that your errors were more than just a “slip”, then you should
write out the correct solution on the quiz sheet and write a sentence or two
beside the solution summarising the method.
• You should use a different colour pen so that the tutor can clearly distinguish
between what you wrote in the quiz and your remediation comments.
• Imagine you are correcting your friend’s quiz and you are explaining to your
friend where he/she went wrong.
The third author had difficulty in articulating these instructions, hence the second
research question. To remediate a quiz, students were encouraged to use any of the
resources available: “Most Common Errors” video; quiz solutions; relevant online
videos; Maths Support Centre; tutor feedback; and, friends. From the VLE, we were
able to record when a student accessed the first three resources listed, and from
Maths Support Centre records we had information on who attended the centre for
remediation purposes. We have no data for the number of students who sought help
from their friends, or those who made use of the tutor’s feedback comments.
Owing to the semester timetable, students could only remediate the first eight
quizzes. In total 1,746 remediation marks were awarded. We collected the

309 [Link]:indrum2018:174756
remediated quizzes, however, as some from the first quiz are missing, we only use
remediated quizzes two to eight inclusive in our qualitative analysis (n=1,511).

QUALITATIVE METHODOLOGY
Qualitative analysis of the remediated quizzes loosely followed the stages of
thematic analysis (Braun & Clarke, 2006). Initially, the first author examined the
remediated quizzes several times in order to familiarise himself with the data. The
remediated quizzes where students only provided a full, complete solution as
remediation were removed as limited information could be obtained from examining
them, especially since complete solutions were available as an online resource. This
left 687 remediated quizzes where students had done something other than provide
only a full solution. Guided by the instructions provided to students, the first author
analysed each of the remaining 687 quizzes in order to determine:
1. Has the student successfully identified each error?
2. Has the student provided a solution for each error?
3. Has the student explained their error, and if so, how?
Most quizzes consisted of two questions, therefore each question was analysed
separately. In order to ascertain whether a student had identified each of their errors,
the first author identified each error on a student’s quiz, and noted how many of
these the student identified. In terms of examining whether the student had provided
a solution for each error, two approaches taken by students were identified. Some
students provided a full solution to the complete question even if the error made only
related to part of it, whereas other students only wrote a solution for the specific
error made. Finally, the first author analysed if, and how, students explained the
errors made. Students seemed to vary in their approaches based on variables such as
the nature of the question asked, for example procedural or conceptual, and the quiz
mark received. We will elaborate on this further in the section below.

RESULTS
To allow for comparison between students with similar levels of achievement on the
quizzes, we divided students into four groups based on their average continuous
assessment mark (excluding remediation marks) received for the first eight quizzes
(4-5, 3-4, 2-3 and 0-2). These groups were of sizes 103, 188, 115 and 64
respectively. Of the 470 students, 47% were female. The following findings are
detailed in Howard et al. (under review): students who scored less than two on a quiz
were less likely to remediate; students in the two lower groups showed a limited
increase in final examination mark as a result of participating in the remediation
process in comparison to their peers; and, students who averaged 3-4 on their
continuous assessment, particularly benefited from participating in the remediation.

310 [Link]:indrum2018:174756
Students’ approaches to remediation
After coding was completed on all remediated quizzes, it was possible to compare
students’ remediation styles within a single quiz as well as the individual progression
through all submitted remediations. With regard to error identification, there was a
notable difference between students with high and low quiz marks. For students with
marks of 3 or 4 in a quiz, mistakes were usually simple calculation slips and thus
many students had only one or two errors to identify. This resulted in the majority of
students with these marks successfully identifying all their errors. However, for those
with lower marks, there were more conceptual misunderstandings to identify as well
as several calculation errors. It is not surprising that these students were less
successful at identifying every error. For students who received low grades on
average, this pattern was clearly evident, but in addition, higher grade students
exhibited this style on quizzes where they obtained a lower mark. There was a clear
contrast in the level of error identification between an individual students’ highest
and lowest scoring quiz. This supports our hypothesis that when students achieve
low marks, they are less successful at identifying their errors. In terms of provision
of solutions to errors, as noted above students either provided a complete solution to
the question even if the error only related to part of the solution, or they provided a
solution that related to a specific error.
In relation to how students explained their errors, three codes or types of
explanations were identified. Based on our knowledge of feedback, we define these
explanations as: diagnostic, instructional, and objective. Some students explained
errors in the context of incorrect notions or ideas that lead to them making a mistake.
We refer to this approach as being diagnostic for example, “I thought brackets
implied find the product but I should have used the chain rule”. Others focused on
providing advice or helpful tips to themselves to help prevent mistakes in any similar
questions they faced in future. We refer to this approach as instructional for example,
“Add powers together as they have the same base (multiply rule). Finally take away
the powers from each other (division/fraction rule)”). The final, and most common
approach, was an objective explanation of the error, simply describing the specific
error without referring to prior knowledge or providing instructions on how they
might answer future questions on the topic for example, “I compounded continuously
but the question asked for quarterly”.
Additionally, there was a less frequent code for whether a student provided an
incorrect statement, or identified their errors incorrectly. Having addressed the first
research question, we propose that in future the following instructions be given to
students:
1. With a different coloured pen to the one in which the quiz was completed, put
an “X” beside each error or, where relevant, indicate an omission in your
work.
6

311 [Link]:indrum2018:174756
 2. Provide an explanation for each of your errors, either describing the error or if
you can, elaborate on any incorrect notions you had which may have led to the
error.
3. Correct each of your errors by writing the correct version beside each one.
Avoid copying down the written solutions for each quiz without explicitly
consulting your areas of error.
4. If you are still unsure how to remediate your quiz or need help with the
questions, we suggest you visit the Maths Support Centre for help.
In addition to the above instructions, we would provide students with exemplars of
“good” remediations.
In-depth analysis of specific students’ remediations
We now turn to Research Question 3: What factors may influence the nature of a
student’s remediation? Owing to the large amount of data involved, we have chosen
to address this research question by examining a specific subset of the data. It is our
hope that this analysis will help us identify factors that may prove beneficial when
analysing the larger data set. We consider a subset of students (n=31) who
remediated at least six quizzes and achieved on average 3-4 quiz marks as these had
particularly benefitted from the remediation (Howard, et al., under review) and we
could investigate their style of remediating over several quizzes. These students
attended between 1-26 lectures and accessed between 11-232 videos with students
from the weaker mathematical backgrounds (based on the Irish State Examination
Mathematics results) accessing more resources than the others. All of these students
passed the end-of-semester examination achieving at least a B grade (60%) and 58%
were female. Twelve students accessed the Maths Support Centre and all but one of
these were female. Detailed records from centre show that at least five of these
students received help from the tutors on remediating the quizzes, including quizzes
where they received four out of five marks. Seeking help from the Maths Support
Centre for remediating the quizzes was uncommon among the larger Maths for
Business cohort. Overall, based on their module resource usage, these students seem
to work consistently throughout the semester.
To investigate how these 31 students remediated, their coded, remediated quizzes
were further analysed to examine any prevalent styles of remediation. Within this
group some students remediated consistently in the same manner each week, while
others exhibited various remediation styles. Due to the initial division made between
students writing only a full solution as remediation, and those who articulated some
form of further explanation, we decided to categorise a student’s overall remediation
style based on the amount of times a remediation consisting of just a solution was
submitted. Students could be split into roughly equal groups based on whether they
provided only full solution in every remediated quiz (n=10), more than half of their

312 [Link]:indrum2018:174756
remediated quizzes (n=8) or in half or less of their remediated quizzes (n=13).
Overall there was no obvious difference in the resources accessed (Maths Support
Centre, quiz solutions etc.) between the three groups of students. In terms of error
identification, these students were, in general, successful in identifying all their
errors. A prevalent remediation style consisted of students identifying their errors,
providing an objective explanation of the error, and writing a solution of the specific
error. With regard to the three main approaches to error explanation, the objective
explanation was the most common method. Interestingly, the diagnostic and
instructional approaches were rarely used consistently in remediations by individual
students.
The nature of the quiz question seemed to impact the remediation approach. The quiz
questions could be considered under the headings: conceptual (‘The following is a
graph of the first derivative f’(x) of a function f(x). Your friend is attempting this
problem...He asks you: “How can you tell that f has a minimum at x=2 just by
looking at the graph?”’), procedural (‘Find all first and second-order partial
derivatives of z = f(x,y) = ...) and economic context (‘Compute price elasticity of
demand, E... In one sentence, explain what your answer means’). Notably, for these
students, quiz questions that were more conceptual in nature, resulted in more
“solution only” remediations. As full solutions were provided as a resource to
students, it was possible to copy them and submit it as remediation. Thus, an increase
in the number of solution-only remediations for a given quiz question may allude to a
lack of student understanding of the module content. A number of quiz questions
also resulted in more diagnostic remediation responses. These questions were more
application based, and focused on application of techniques to economic contexts.
The two quizzes that contained these applied questions had the highest average
marks of all eight quizzes, with 21/31 quizzes obtaining full marks between the two.
We believe the increased use of diagnostic remediations may indicate students find it
easier to locate misconceptions in mathematical application than with more abstract
questions.

DISCUSSION AND CONCLUSION

Semester 1 2016/17 was the first time we implemented the remediation process.
Following from the definitions of Ramaprasad (1983) and Sadler (1989), students
were provided with their actual level of understanding (mark out of five), the desired
level of understanding (for example, pdf of quiz solutions), and were incentivised to
engage with feedback for the intention of closing the feedback gap. Gibbs and
Simpson (2004) propose that even if feedback is provided, students will not
necessarily use it. Despite assessment marks being provided as an incentive, on
average 70% of students engaged in the remediation process. Initially, we considered
providing remediation marks on a sliding scale based on a student’s initial quiz mark
received. This method would reflect the additional work of remediating for lower-
8

313 [Link]:indrum2018:174756
scoring students and offer them additional incentive over high-scoring students,
however, this system was not implemented as it required additional time and effort
from the tutors.
Qualitative analysis of the entire set of remediated quizzes allowed for the isolation
of certain properties that indicate whether a student is engaging with their assessment
feedback. We note that a substantial number of students provided the full solution
only for their remediation, however we advocate the three-step approach -
identifying, explaining and correcting errors encourages students to recognise their
own standard and utilise feedback to help close the gap between their own level and
the desired standard. It is for this reason that we suggest this style as a guideline for
future processes similar to remediation.
Owing to the brevity of the explanations provided in this paper on the different
remediation responses, it is pertinent to mention different avenues that will be
investigated as this research progresses. The in-depth analysis of the 31 students in
this paper suggests that task design and a student’s initial grade influences
remediation style. Different task types elicited different remediation responses, based
on whether the question was conceptually or application based. Perhaps, owing to a
lack of understanding of the content material or the inability to transfer knowledge
from one context to another, students tended to provide full solutions to conceptual
questions. In addition, quizzes where students received lower grades tended to result
in more solution only remediations. As these solutions are available as resources, this
may suggest that some students utilise solutions when they are unable to fully
identify and understand their errors on a quiz. From this, one can hypothesise that the
remediation process may have less benefit to weaker students as a level of baseline
knowledge may be required in order to engage with the process. In subsequent
research, task design and initial quiz mark will be examined further to discern any
influence on remediation response across the full set of remediated quizzes.
One limitation of our study is the lack of student perspective on the remediation
process. Also, while the remediation process was beneficial to the students of this
large non-specialist mathematics module, an investigation of the benefits/drawbacks
of the remediation process for mathematics modules for specialists would be a
constructive contrast with this research.
This study was completed with approval from the University College Dublin ethics
committee in accordance with ethics applications LS-E-17-20-Copeland-Meehan and
LS-16-48-Howard-Meehan.

REFERENCES
Bennett, R. E. (2011). Formative assessment: a critical review. Assessment in
Education: Principles, Policy & Practice, 18(1), 5-25.

314 [Link]:indrum2018:174756
Braun, V., & Clarke V. (2006). Using thematic analysis in psychology. Qualitative
Research in Psychology, 3(2), 77-101.
Covic, T., & Jones, M. K. (2008). Is the essay resubmission option a formative or a
summative assessment and does it matter as long as the grades improve?
Assessment & Evaluation in Higher Education, 33(1), 75-85.
Evans, C. (2013). Making sense of assessment feedback in higher education. Review
of Educational Research, 83(1), 70-120.
Gibbs, G., & Simpson, C. (2004). Conditions under which assessment supports
students’ learning. Learning and Teaching in Higher Education, 1, 3-31.
Handley, K., Price, M., & Millar, J. (2011). Beyond ‘doing time’: Investigating the
concept of student engagement with feedback. Oxford Review of Education, 37(4),
543-560.
Howard, E., Meehan, M., & Parnell, A. (2017). Live lectures or online videos:
students’ resource choices in a first-year university mathematics module.
International Journal of Mathematical Education in Science and Technology.
Howard, E., Meehan, M., & Parnell, A. (2018). Quantifying Participation in, and the
Effectiveness of, Remediating Assessment in a University Mathematics Module.
Under Review.
Iannone, P., and Simpson, A. (2011). The summative assessment diet: How we
assess in mathematics degrees. Teaching Mathematics and its Applications, 30(4),
186-196.
Iannone, P., and Simpson, A. (Eds.). (2012). Mapping University Mathematics
Assessment Practices. Norwich: HEA.
Iannone, P., & Simpson, A. (2013). Students' perceptions of assessment in
undergraduate mathematics. Research in Mathematics Education, 15(1), 17-33.
Price, M., Handley, K., & Millar, J. (2011). Feedback: Focusing attention on
engagement. Studies in Higher Education, 36(8), 879-896.
Ramaprasad, A. (1983). On the definition of feedback. Behavioral Science, 28, 4-13.
Sadler, R. (1989). Formative assessment and the design of instructional systems.
Instructional Science, 18, 119-144.
Trenholm, S., Alcock, L., & Robinson, C. (2015). An investigation of assessment
and feedback practices in fully asynchronous online undergraduate mathematics
courses. International Journal of Mathematical Education in Science and
Technology, 46(8), 1197-1221.

10

315 [Link]:indrum2018:174756
 1

Student Partners in Task Design in a computer medium to promote

Foundation students’ learning of mathematics
Barbara Jaworski, Stephanie Treffert-Thomas, Dave Hewitt,
Matt Feeney, Damma Shrish-Thapa, Dan Conniffe, Ayesha Dar,
Nikolaos Vlaseros, and Marinos Anastasakis
Loughborough University, Mathematics Education Centre, UK
A team consisting of three mathematics education teacher-researchers, four former
Foundation students (called Student Partners, SPs), and two analytic assistants
worked together to produce mathematical tasks in a computer medium for the
mathematical learning of current Foundation students (FSs). We have explored the
collaboration between the SPs and researchers, the processes and outcomes of task
design, and the contribution of the collaboration to tutorial teaching of FSs. We seek
insight into the learning of all concerned of mathematics, mathematics teaching, task
design and personal-professional development. The project is ongoing. Here we
introduce the project and present early findings – specifically related to task design
and the contribution of SPs.
Keywords: Teachers’ and students’ practices at university level; Novel approaches to
teaching; The role of digital and other resources in university mathematics education;
Students as partners in task design.

INTRODUCTION TO THE CATALYST PROJECT

We report from an exciting new development project (2016-18: The Catalyst
Projecti) in which mathematics teacher-educators collaborate with former Foundation
students as partners in designing computer-based mathematical tasks for current
Foundation students. Our research explores the collaboration of participants, the
design of tasks, teaching of students in tutorials, use of computer software for
teaching and learning mathematics and the learning of all concerned.
The mathematics learning of students in our university Foundation Studies
programme (henceforth Foundation Studies students – FSs) is the focus of this
developmental project. These are students who need a higher-level qualification than
they hold currently in order to be able to enter the first year of their desired
undergraduate programme (e.g., programmes in engineering or science). For such
programmes, mathematics is an essential component. All FSs are required to pass
their year-long module in mathematics. It has been observed that teaching and
learning in this module in the past has been rather procedurally based: students have
been introduced to and expected to learn the application of procedures to
mathematical problems and have been examined on their procedural competency. A

316 [Link]:indrum2018:172406
 2

current aim is to make the module more conceptually based, and to explore the use of
computer-based tasks for this purpose. Our previous research has shown that:
a) The involvement of peer learning-teaching in a mathematics module has resulted
in participating students gaining higher marks, with associated experiences reported
as positive to their understanding of concepts; (SYMBoL, e.g., Duah, Croft & Inglis,
2013; Solomon, Croft, Duah, & Lawson, 2014);
b) The cultural differences between teacher-researchers designing an innovation in
mathematics teaching-learning and students engaged in learning mathematics
through the innovation have contributed positively to outcomes/higher marks.
(ESUM, e.g., Jaworski, Robinson, Matthews, & Croft, 2012).ii
Beyond this activity, we have found very little other research involving students’
engagement through partnerships in mathematics teaching and learning. There is
relevant work in Higher Education more generally involving Partnership Learning
Communities (Healey, Flint and Harrington, 2014), but this does not include
mathematics specifically. A recent special issue of NoMAD, the Nordic Journal of
research in mathematics education, included several papers in which teachers were
involved in exploring their own practice. The reports address three themes, one being
‘innovative approaches to teaching and learning, with emphasis on student
participation in the educational process’ (Goodchild & Jaworski, 2017). One paper,
in particular, reported positively on students’ activity in oral presentations as a tool
for promoting metacognitive regulation in Real Analysis (Naalsund & Skogholt,
2014). Student engagement and understanding were seen to improve through their
participative activity. Searches to date have revealed no other relevant work.
Building on our experiences in SYMBoL and ESUM, we sought to design
mathematical tasks which would challenge the FSs in new ways, engage them
visually and actively and promote the beginnings of a new learning culture.
Recognising the value of student design of resources and peer support as
demonstrated in the SYMBoL Project we built both of these aspects into our Catalyst
project. One of the Catalyst aims was: To promote collaboration between staff and
students that results in higher degrees of confidence, motivation and learning in
mathematics and a new culture in the teaching-learning of mathematics (e.g., HEA,
2014). We hoped to learn from the various elements of this project in ways that
would have relevance beyond mathematics, particularly in the inclusion of former
FSs as Student Partners (SPs) in course design and teaching. Our own learning from
working with students in these ways was also a central aim, with the intention to
bring staff and student cultures into focus through our joint activity. The reflections
of the students on their activity, both SPs and FSs, were seen as an important
outcome of the project. Thus, our innovation in the Catalyst Project has two main
areas of inquiry:

317 [Link]:indrum2018:172406
 3

 The design of computer-based tasks in matrices and complex numbers for the
FSs using Autograph software.
 The involvement of former FSs (the SPs) in the design of teaching and in the
tutorial teaching of the current FSs.
According to its specification, the Foundation Studies programme provides “An
opportunity to make it onto a degree course at Loughborough University”. Within
the programme there is a wide range of student experience in mathematics from
GCSE grade C to A level grade A. We focus on a mathematics module called
‘Applicable Mathematics’ which prepares students to take up degree programmes in
Science or Engineering. The two semesters focus on the following topics:
Semester 1: Algebra, Logarithms, Inequalities, Functions, Trigonometry, Vectors,
Differentiation, Integration, Sequences; Semester 2: Polynomials, Partial Fractions,
Further Calculus, Conic Sections, Vectors, Matrices, Complex Numbers.
The project has focused on the teaching of Matrices and Complex Numbers in
Semester 2 in 2017. The three project leaders (PLs) have worked with four SPs to
design tasks using the computer software Autograph in the two topic areas. Tasks are
for use in tutorials with Foundation Students (FSs). SPs are former FSs: in the
previous year group they were successful in having achieved grades at the levels
required for transition to programmes in Mechanical Engineering, Chemical
Engineering, Physics and Chemistry. At the time of their recruitment, they were first
year students in their current programmes. In addition, two doctoral students in
Mathematics Education were recruited as “Analytical Assistants” to support data
collection and analysis. Thus, nine participants have been involved in the project,
with differing roles.

THEORETICAL BACKGROUND
We are concerned with learning at a number of levels.
 FSs learning of mathematics;
 SPs learning of mathematics, task design and participation with staff in
preparing for undergraduate learning;
 Mathematics teachers and researchers learning about the design of teaching in
partnership with students.
This learning is influenced by a wide range of factors which include the curriculum,
and institutional settings within the broader sociocultural setting. Some of these
factors we can seek to influence; others are less amenable to innovation. We take a
fundamentally Vygotskian (e.g., Vygotsky, 1978; Wertsch, 1991) perspective
recognising particularly mediation by people and tools that support learning; goal-
directed activity and action related to learning and teaching; scientific concepts that

318 [Link]:indrum2018:172406
 4

require pedagogic mediation; and the zone of proximal development in which

mediation fosters learning and development. We engage particularly with digital
tools, their design and use, and the ways in which they mediate the learning process
through both support and challenge for making sense of mathematical concepts.
An important theoretical concept is that of “partnership” between staff and SPs (e.g.,
Healey, Flint and Harrington, 2014). Relationships within the partnership have
resulted in the design of mathematical tasks and their use with the FSs. The nature of
this partnership is central to project outcomes, in terms of the designed tasks and
their use. We see ourselves as having formed a ‘Learning Community’ in which co-
learning is an important concept (Wagner, 1997), and which demonstrates tenets of a
Community of Practice, such as mutual engagement and joint enterprise (Wenger
1998) and a Community of Inquiry, such as critical alignment (Jaworski 2006).

METHODOLOGY
We take a developmental research approach, consistent with our Vygotskian theory,
which both studies project development and learning within the project and
contributes to development and learning (Goodchild, 2008; Jaworski, 2003).
Mediation and tool use, for example, can be seen in an interactive stance of
reflection and negotiation in which participants engage together in activity and
action with growth of mutual understanding and co-learning (Wagner, 1997).
Analysis begins in questioning of what is done and achieved and is formalised
through scientific inquiry addressing a range of data through recognised methods.
Research questions and data
Our Research Questions relevant to this paper are as follows:
 How have SPs engaged with task design and what has been the outcomes and
issues arising?
 How have FSs worked with the designed tasks?
 What have we learned about the FSs’ learning of mathematics with the
designed tasks? What issues arise?
Data, which are being analysed to address these questions include:
 The involvement of SPs and staff in the design process as shown through
recordings of project meetings, SP reflections/reports, documents (collected at
the design meetings and from the SPs’ own work).
 The tasks, and their use as seen through observation, screen capture and
discussion. The teacher’s narratives from her reviewing of tutorial data.

319 [Link]:indrum2018:172406
 5

Analytical approach to date

Reflection and negotiation have taken place through meetings, discussion, sharing
and review of designed tasks leading to increased awareness of issues in design and
communication. This collaborative co-learning has involved a bringing to collective
consciousness of key elements of the design process and issues to be resolved.
Analysis of collected data according to research questions has been qualitative,
focusing on data from the design meetings, and from tutorials with FSs in which
designed tasks have been used. A process of data reduction has summarised and
coded recorded data, allowing initial identification of key elements and issues
relating to research questions. The process is cyclic with developing depth of inquiry
and insight to significant issues expressed through analytical memos. The tasks
themselves have also been a focus of scrutiny which is ongoing. These analyses are
as yet in their very early stages, so what we report below is tentative and indicative.
Here, we discuss some emergent findings in task design and use of tasks with FSs.

EMERGENT FINDINGS
Task Design
The teacher/lecturer of the Foundation course provided course notes on the two
topics, Complex Numbers (CN) and Matrices. An expert in Autograph gave the
group an induction into its use and potential for mathematical representation and
exploration. SPs were asked to review the notes and think about possible tasks using
Autograph. Task design, in 2 SP pairs (one to each topic) took place over 6 weeks
and across 4 meetings – finding times for these meetings, from timetables of 9
partners, over a short time period was challenging. At the meetings, SPs’ presented
their ideas to the whole team, hesitantly in the beginning but with growing
confidence in response to expressed appreciation and suggestions from the team. The
pair working on tasks in complex numbers were quick to provide examples and to
modify them according to suggestions in meetings. The pair who worked on
matrices found it harder to get going. Tasks in complex numbers appeared to be more
readily achievable than in matrices where concepts seemed less amenable to digital
representation/questioning – it became necessary both to identify the barriers and to
find some resolution. Collaboration between the SP pair and the PLs, focusing on the
mathematics of matrices and the learner difficulties suggested by SPs, resulted in a
set of tasks on matrices. One of the PLs also designed a set of tasks in GeoGebra
focusing on matrix arithmetic. The emerging ‘raw’ tasks, consisting of an Autograph
(or GeoGebra) file with brief associated notes, were then ‘prepared’ by the FS
lecturer to make them ready for FS use. We see two examples of the prepared tasks,
one for each topic, in Figure 1. Certain characteristics, incorporated by the SPs into
these tasks can be seen in the examples; FSs have to:

320 [Link]:indrum2018:172406
 6

 use several display features of Autograph to “see” mathematical objects and

relationships;
 undertake some associated calculation by hand;
 relate movements on their screen to the handwork and theory involved;
 reflect on specific results to develop a more general awareness of concepts.
Some of these demands turned out to be a challenge for many FSs as we see below.
Question 2: Open the Autograph File Task 2
There are three complex numbers labelled z1, z2 and z. z1 is fixed while z2 and z can be moved.
Select z2 and move it until z reaches the position 3 + i.
(a) What complex number is z2?
(b) Right click and “Unhide All" to check your answer.
(c) What is the relationship between z1, z2 and z?
(d) Explore subtraction of complex numbers in Autograph.
(e) Now calculate by hand:
With z1 = -1 - 3i and z = 3 + i, find z2 such that z2 - z1 = z.
(f) Draw (by hand) all three complex numbers on an Argand diagram.
Give a geometric interpretation of the relationship between z and z1 and z2.

Question 7: Open the Autograph File Matrices 5

On this page you see two straight lines. Their equations are 4x - y = 14 and 7x + 4y = a
(a) By hand, using matrices, calculate the value of a so that the solution to the simultaneous
equations is 3 .
2
(b) When you have a solution, use the “constant controller” to vary the value of a until the point
(3, 2) is clearly displayed.
(c) Select both lines by holding down the “Shift" key. Both lines should have changed colour.
Go to “Object" in the menu bar and choose “Solve f(x) = g(x)".
A point is displayed. To see what the co-ordinates of this solution are, go to “View" in the menu
bar and choose “Results Box". Does your solution make Autograph show the intersection of the
lines to be x = 3 and y = 2?

Figure 1: Examples of prepared tasks in complex numbers and matrices

Use of tasks with Foundation students
The tasks were used in timetabled tutorials with the FSs, who were asked also to
comment on the tasks for the research. Those agreeing to participate were audio-
recorded in conversation with the teacher and one researcher. From analysis to date
we are gaining some insight into FS participation with the tasks. We have recorded
instances of FS requesting help from the teacher, asking questions, explaining their
solutions, revealing mathematical insight, surprise, or lack of understanding. In
interactions with the FS, the teacher responds to students, asks questions, explains,
and provides technical information. The dialogue in Figure 2 shows a teacher-student

321 [Link]:indrum2018:172406
 7

exchange, relating to Question 2 above, that reveals an issue in the student’s

engagement with the task that is typical of several such exchanges:
We see that the student had engaged with the task, moving z2 as instructed. However,
he did not know what is meant by “the relationship between z1, z2 and z” although
he had written that “z2-z1 gives you z”. It seems a case of not understanding the
meaning of the word “relationship”, although he had written the relationship. We are
learning here about language issues in working between the visual mode of
Autograph and the symbolic mode of expressing complex relationships.
T: What are you doing in Question 2? [She looks at what he has written] 3+i... Haa!
What did you find out? What relationship?
FS: I don't know what it means.
T: You don't know what that means? Well they are connecting aren't they? You must
be doing something with them. If they are moving together [z1, z2 and z are moving
together] we do something to them and then you get the third.
FS: I wrote that z2-z1 gives you z.
T: Figure
So you2:did write itofdown.
Example For between
dialogue [part] c, that is the relationship.
FS, working on Task 2,Yes,
andthat's what we
the teacher
meant.
Although
FS: OK the teacher had used Autograph, in demonstration mode, in her lectures,
students
T: So had not developed
the relationship a languageYou
is subtraction. to express what they
are subtracting could see
two complex visually.
numbers. So,
How
from what
did youtheknow
student said here,
that though? Did ityou
was
knowdifficult to gauge
that from the contribution
the picture? Or did you doof
Autograph to theelse?
something student’s understanding of complex subtraction. When pushed by
the
FS:teacher to explain
I worked it out. his process in getting an answer, he replied that he “worked
out”
T: theAh result. As actually
ok. So you the teacher interpreted,
did the [Link]=Teacher
might mean FS=that he calculated
Foundation Studentthe
result. Whether this is as well as discerning it from the movements in Autograph or
instead of this, is not clear. Thus, it might be that the student had used Autograph
effectively and made links with the symbolic forms. Or it may be that he had
sidestepped interpretation of the visual and instead had worked out the result
symbolically (the latter perhaps being a more familiar task).
The teacher’s reflective narrative relating to the recordings from the second tutorial
on CN reveals the following example [Teacher’s written words are italicised]:
[Student] found the additive relationship by adding separately the real parts and
then the imaginary parts. He says “is it bisecting the angle?” I reply “it’s
something to do with it”. [She asks about her related lecture presentation and
mentions a comparison with vectors.] The student correctly relates vector addition
to “adding head to tail” and that this forms a “triangle”. [She indicates that he
does not seem to understand what is meant by a “geometric interpretation”.]
The teacher’s reflection suggests a student engaging correctly with several concepts
including complex addition and vector representation. Yet, again, we see a problem
with language – the term “geometric interpretation” is unfamiliar to the student. One

322 [Link]:indrum2018:172406
 8

student, asked about his work on the Autograph tasks, replied that he found the
questions “hard to read” and could not “understand the way they are worded”.
The two examples quoted are typical of recorded exchanges. We see from these
a) Students not understanding the words or expression of ideas in the written
questions (e.g., “relationship”, “geometric interpretation”)
b) Students not articulating explicit conclusions from what they see on screen –
rather using the familiar forms of calculation to answer questions.
These observations lead us to question both the Autograph tasks and the wording of
the tasks. How might we have worded the tasks differently so that students would
engage with what they could see on screen and discern the mathematical
relationships that the task was designed to reveal? How might we wish to modify the
task itself so that students engage visually rather than depending on calculation? The
challenge for the team here is twofold: to design a teaching approach that introduces
the language we want students to use and enables students to become familiar with
its use; and to design tasks that are revealing of concepts in and of themselves, so
that students can see visually what they familiarly work out in calculation. These
seem to be important elements of the learning culture we are trying to foster.
We are aware that many FSs come to their university course from school or college
where their mathematical enculturation towards success in national final
examinations may have encouraged a procedural perspective on learning
mathematics (Minards, 2013) and that we see the results of this to some extent in
their response to the designed tasks. As well as looking critically therefore at the
design of the tasks, we have to consider the wider mathematical culture, the nature of
teaching that seeks to interact with this culture and the ways in which the tasks can
be incorporated within the teaching-learning interface.

CONCLUSION AND FURTHER RESEARCH

The developmental nature of the project can be seen through the development of
mathematical tasks for Foundation students by SPs and PLs in partnership, the
subsequent use of the tasks in tutorials with the FSs, and issues arising from task
design and use revealed both in practice and in analysis of data from the various
events. An aim of the project was to foster conceptual understanding of mathematics
by the FSs. We see above some issues arising from the nature of the tasks and the
ways in which they are written, and from the ways in which FSs’ mathematical
experience influences their engagement with the tasks.
Because analysis is in its very early stages, we are not yet in a position to report on
many of the aspects of learning in the project (such as aspects of the learning of the
SPs). However, already we can start to see indications of important learning and the

323 [Link]:indrum2018:172406
 9

feedback element of the developmental research. Co-learning has been demonstrated

between the FSs and the teacher-researchers – FSs’ learning of mathematics has
raised issues for the intention and preparation of tasks which offer challenges to the
researchers and for future work with the SPs. The tasks and their design have been
mediational tools not only for the mathematical learning of the students but also for
the awareness of the researchers about teaching-learning issues, not least the issue of
language in which tasks are expressed. The project is ongoing, both in terms of
teaching-learning development and of analysis of the data collected so far.
A major issue for the project has been the timescale as dictated by the funding body
and university organisation of teaching. We had barely half a semester to recruit
SPs, initiate the design process, hold 4 spaced meetings, prepare the tasks for FSs
and hold the tutorials. The project end coincides with the time for the next cohort to
reach the teaching of matrices and complex numbers, so we could not build this into
the project. We expect to use the same tasks again with the new cohort and collect
further data, informed by our experiences a year ago. Since our data is extensive, in
depth analysis is ongoing from which more in-depth reporting should be possible.
We expect to be reporting further on the many aspects of this project.

i
Supported by HEFCE (Higher Education Funding Council for England) Catalyst Fund: Innovations
in learning and teaching, and addressing barriers to student success A: Small-scale, ‘experimental’
innovation in learning and teaching.
ii
SYMBoL – Second Year Mathematics Beyond Lectures – was a project designed to support
teaching in two second year mathematics modules, Vector Calculus and Complex Analysis.
Students who had experienced these modules were employed over a summer period to design
resources in collaboration with mathematics staff. The resources were used in subsequent delivery
of the modules and a peer support system was initiated in which third year undergraduates
supported their second year counterparts.
ESUM – Engineering Students Understanding Mathematics – was a developmental research project
involving an innovation in mathematics teaching seeking to engage students more conceptually with
mathematics through inquiry-based activity, a computer-based learning environment, small group
tasks and an assessed small group project.

REFERENCES
Duah, F., Croft, T., & Inglis, M. (2013). Can peer assisted learning be effective in
undergraduate mathematics? International Journal of Mathematical Education in
Science and Technology, 45 (4), 552–565.
Goodchild, S. (2008). A quest for ‘good’ research. In B. Jaworski & T. Wood (Eds.),
International handbook of mathematics teacher education: Vol. 4. The

324 [Link]:indrum2018:172406
 10

Mathematics Teacher Educator as a Developing Professional (pp. 201-220).

Rotterdam, The Netherlands: Sense Publishers.
Goodchild, S. & Jaworski, B. (2017). Developing practice through research into
university mathematics education. NoMAD 22(4), 5-22.
Higher Education Academy (2014). Framework for Partnership in learning and
teaching in Higher Education. York, UK: HEA.
Jaworski, B. (2003). Research practice into/influencing mathematics teaching and
learning development: towards a theoretical framework based on co-learning
partnerships. Educational Studies in Mathematics 54(2-3), 249-282.
Jaworski B. (2006). Theory and practice in mathematics teaching development:
Critical inquiry as a mode of learning in teaching. Journal of Mathematics
Teacher Education, 9(2), 187-211.
Jaworski B, Robinson C, Matthews J, Croft A.C. (2012). An activity theory analysis
of teaching goals versus student epistemological positions. International Journal
of Technology in Mathematics Education 19(4), 147-152.
Minards, B. (2013). An exploration of high-achieving students’ experiences of A-
level mathematics. Unpublished PhD Thesis. Univ. of Birmingham: B’ham UK.
Naalsund, M. & Skogholt, J. (2017). Oral presentations as a tool for promoting
metacognitive regulation in real analysis. NoMAD 22(4), 103-119.
Solomon, Y., Croft, T., Duah, D. & Lawson, D., (2014). Reshaping understandings
of teaching-learning relationships in undergraduate mathematics: An activity
theory analysis of the role and impact of student internships, Learning, Culture
and Social Interaction, Vol. 3, 323-333.
Vygotsky, L. S. (1978). Mind in society. Cambridge MA: Harvard University Press.
Wagner, J. (1997). The unavoidable intervention of educational research: A
framework for reconsidering research-practitioner cooperation. Educational
Researcher, 26(7), 13-22.
Wertsch, J. V. (1991). Voices of the mind. Cambridge MA: Harvard Univ. Press.
Wenger, E. (1998). Communities of practice. Cambridge: Cambridge Univ. Press.

325 [Link]:indrum2018:172406
 Theoretical and empirical description of phases in the proving
processes of undergraduates
Katharina Kirsten1
1
University of Münster, Germany, [Link]@[Link]
In the presented study we adopt a process-oriented perspective on proving in order
to gain further insights into relevant actions and typical obstacles in
undergraduates’ approaches to proving. The primary aim is to theoretically and
empirically describe different phases, understood as bunches of intentionally closely
related actions of proving. Therefore, we suggest a theoretical model of the proving
process and confirm empirically that it can be used as an analytical tool for proving
approaches. Based on this model, several proving processes have been analysed. In
this paper we present first findings regarding the contribution of each phase to proof
construction as well as the general structure of the proving process.
Keywords: proof construction, proving process, phases, proving cycle.

THEORETICAL BACKGROUND
It is well known that undergraduates commonly have to deal with great difficulties in
constructing proofs (e.g. Weber, 2001; Moore, 1999). In order to gain additional
insights into the main obstacles in proving at university level, we focus on the
processes of proof construction. Preparing further projects, the presented study
serves as a pilot study, which initially aims at investigating the general structure of
proving processes on a macroscopic level. Therefore, we first give an overview about
the main characteristic of the proving process in contrast to its product. Based on
this, we deduce a theoretical model from literature in order to use this model as an
analytical tool for undergraduates’ approaches to proving.
The process of proving
Talking about proof construction, there is a consensus that a proof, especially at
university level, is a line of reasoning, which is strictly deductive and exclusively
based on theorems and axioms. In contrast to its result, the process of proving
contains not only deductive reasoning, but also includes inductive and explorative
processes (e.g. Hersh, 1993). Thus, developing a key idea and transferring it into a
formal proof is a complex and highly demanding process, which occurs in the tense
atmosphere of conviction and explanation as well as intuition and formality (Hemmi,
2008). In the present paper we will discuss on a theoretical and empirical basis, how
these different perspectives and processes interact in constructing proofs.
Mejia-Ramos and Inglis (2009) distinguish between three distinct types of
construction activities, which emerge from different external conditions. Although
all of these activities aim at constructing an appropriate proof, each is guided by a
specific goal: Exploration of a problem consists of working on an open-ended

326 [Link]:indrum2018:173782
question in order to discover and at least prove a new statement. However,
estimation of truth starts with an already prepared conjecture and justification begins
with a statement estimated to be true. In these cases, proof construction is rather
aimed at determining or verifying the truth-value of a statement instead of inferring
it. Meyer (2010) differentiates activities associated with mathematical proof
construction in a similar way based on the theory of abduction, induction and
deduction. According to his approach, each kind of inferring is related to a different
type of construction activity. Thus, external conditions and, especially, task designs
have a relevant influence on proving processes by stimulating different types of
inferring and construction activities. Combining both frameworks, we assume that
the activities described above are independent for the reason that they consist of
different types of processes and pursue different goals. In particular, justification is
not a part of exploration, which follows producing a conjecture, but requires specific
cognitive processes.
In order to describe those cognitive processes, various models have been developed,
which summarise the relevant actions and demands associated with proof
construction. These models mainly focus on processes related to the exploration of a
problem. As proving tasks at university level often consist of a statement estimated
to be true, these models only seem to be partially suitable for analysing
undergraduates’ proving processes. Hence, we suggest a model of proving, which is
mainly following existing models, but focusses on justification.
Models of proving processes
To analyse the cognitive processes of mathematical proof construction, there is a
need for abstraction. The complex structure of the process has to be reduced to the
relevant actions and, therefore, transferred onto a macroscopic level. Doing so, most
process-oriented models use the unit of episodes or phases in order to subsume
closely related actions in service of the same goal under a generic activity. Thus,
phases are “macroscopic chunks of consistent behaviour” (Schoenfeld, 1985, p. 292),
which summarise the relevant processes associated with proof construction.
Existing models of the proving process mainly differ in their amount of suggested
phases. The most cited models have been presented by Stein (1984) and Boero
(1999). Both models focus on the activity of problem exploration, which means that
the proving process is based on an open-ended problem area. In the first phase of
proving this problem area is explored regarding relevant conditions and regularities
in order to produce a conjecture. When the conjecture has been formulated as a
statement, the proving construction contains three phases: Identifying arguments for
the correctness of the statement, linking them into a deductive chain and formulating
an appropriate proof. As Stein’s (1984) model focusses on proving approaches of
students at secondary level and Boero (1999) describes the proof construction of
mathematicians, quality and formalisation of the proofs intended in the models differ
according to the mathematical standards shared in the particular context. Boero

327 [Link]:indrum2018:173782
(1999) even adds a further phase, in which mathematicians approach a formal proof.
Apart from this last step, which primarily seems to be relevant for experts, the phases
of proof construction described by Stein (1984) and Boero (1999) have been taken as
a basis for several research projects in mathematics education (e.g. Reiss & Renkl,
2002). According to this, we suppose the phases described in both models to be
relevant for undergraduates’ proving processes as well. However, proving tasks in
the initial phase of studies rather initiate justification instead of exploration. Due to
this, we suggest the following variation of Stein’s and Boero’s model:

Figure 1: Proving cycle as a model for proof construction

In contrast to the existing models, the proving cycle starts with a proving task
consisting of a statement estimated to be true. Although this statement can be similar
to the produced conjecture in other models, it plays another role in the proving
process. In case of problem exploration the statement is connected to insights from
conjecturing and exploration, but in case of justification no further information about
the statement is given. Hence, it is necessary to analyse the given statement, to
clarify the terms and conditions and to access previous conceptual or strategic
knowledge. Leaning on the approach of a situation model (Kintsch & Greeno, 1985;
Reusser, 1990), we suppose undergraduates to develop a mental representation,
which refers to their individual understanding of the given statement and guides
further processes. Based on this mental representation we suggest similar phases as
presented in the models of Stein (1984) and Boero (1999). The second phase aims at
exploring the statement area and discovering key ideas. By (re-)constructing
relations and objects or applying theorems reasons for the validity of the statement
can be identified. If some key ideas have been found, it is necessary to select
promising ideas, work out their details and structure single arguments in an
appropriate deductive order. This phase results in a proof outline, which contains the
main aspects of the proof, but can be fragmentary or difficult to understand for
someone else. In order to prepare a final proof, which meets the mathematical
standards of one’s community, one has to fill gaps and revise the linguistic and
formal arrangement of the proof. As a last phase the model includes validating
activities, which can be compared to a certain extent to Pólya’s stage of looking back

328 [Link]:indrum2018:173782
(Pólya, 1945). In this phase the final proof is reviewed regarding content, structure
and linguistics. Beyond that, one can consider further (shorter or more elegant)
proofs in this phase or reflect on the proving process, as, for example, thinking about
key ideas, difficulties and their solutions.
The single phases in the model are arranged as a cycle. In accordance with Stein’s
(1984) and Boero’s (1999) considerations, we assume that proof construction does
not necessarily proceed in a linear way. Instead, the proving process is shaped by
interruptions and revisions that cause transitions between different phases.

RESEARCH QUESTIONS
The primary aim of the presented study is to develop an analytic tool for proving
processes, which makes it possible to describe and compare students’ approaches to
proof construction on an individual, macroscopic level. Therefore, a proving cycle
has been derived from literature consisting of five phases estimated to occur in
undergraduates’ proving approaches. The unit of a phase seems to be well-suited to
describe proving activities in a clear and abstract way without neglecting specific
details and differences in the proving processes. Accordingly, the proving cycle may
be used to analyse proving approaches by describing the frequency, the duration and
the order of different phases occurring in the process. The presented study aims to
provide evidence of the utility of the proving cycle as an analytical tool and – if so –
to gain more detailed information about the process of proof construction at the level
of phases. In detail, the aims of the study described above lead to the following
research question:
1. Is the proving cycle an appropriate tool for analysing proving processes? That
means, is it possible to reconstruct the different phases stated in the proving
cycle empirically? Do further activities exist, which do not fit the theoretical
description?
2. How can the process of proof construction be described in general? That
means, which phases are taking a relevant share in the process? In which order
do undergraduates go through the different phases? Can the cyclic nature of
the proving process be confirmed?

METHOD
In accordance with the open-ended character of the research questions, an
explorative laboratory study has been designed. In this study proving processes of
undergraduates are initiated, observed and finally analysed. The concept of the study
and its conditions are described in more detail below.
Sampling and data collection
The study focuses on undergraduates and pre-service mathematics teachers (high
school) attending their first year of studies. Performing an informal unstructured

329 [Link]:indrum2018:173782
interview, participants are encouraged to work on proving tasks in the field of real
analysis. Doing this, they are told to prepare a joint solution that satisfies the
requirements of a proof in the initial phase of studies and that would be accepted by
a tutor or a lecturer. To encourage the participants to talk about their ideas and
approaches, the working processes are organised in pairs. In order to secure the same
conditions for all participants and, therefore, the comparability of the observed
processes, the interviewer offers no support. Instead, a commonly used textbook of
real analysis is provided. For investigation such proving tasks have been chosen that
require a one- or two-step proof and can be solved by applying a prominent theorem
of real analysis like the intermediate or mean value theorem. In order to explore the
proving cycle’s applicability, the proving tasks contain universal as well as
existential quantifications and can be proved directly or by contradiction.
While working on the proving tasks, the participants are videotaped. That means,
their proving approaches are recorded in sound and vision. Additionally, we
collected the final solutions as well as the notes of each pair. During the pilot phase
of the study seven pairs of undergraduates took part in the interviews and worked
each on one or two proving tasks. Processing time for a single task varies from 30 to
75 minutes. However, not all proving approaches were successful. While some
participants gave a proof, which was not completely correct, but contained useful
approaches, other students could not achieve any solution.
Preparation and analysis of data
To prepare the observed proving approaches for analysis, we transcribed each
videotape entirely. For precise investigations the transcripts of the dialog are
expanded by further information like non-verbal activities and notes. Combining the
transcription of natural conversation and written approaches, the protocols of the
proving processes are finally encoded according to Mayring’s (2014) structuring
content analysis. In accordance with the research questions a deductive category
formation with nominal categories has been chosen, which is closely related to the
proving cycle described above. The aim of the coding is to identify changes in the
participants’ behaviour in order to describe the structure of their proving process as a
sequence of transitions between different phases. Leaning on Schoenfeld’s (1985)
method of protocol analysis the coding consists of two steps: First, the proving
process is parsed, that is, making decisions regarding dividing lines of phases. Once
a proving process is partitioned into phases, each phase is characterised as one of the
theoretical stated phases in the proving cycle. The coding results in a macroscopic
description of the students’ proving processes that combines closely related actions
into phases and provides a summary of relevant activities.

RESULTS
In this section the results of analysing data from nine proving processes is presented.
For the moment, a case study is introduced to demonstrate the methodological

330 [Link]:indrum2018:173782
approach as well as possible results on an individual level. Based on this, findings
regarding the proving cycle being an appropriate analytic tool are discussed in
general. Beyond that, we present first assumptions concerning the frequency, the
duration and the order of different phases in undergraduates’ approaches to proving.
Case study of Michael and Leon
Michael and Leon are working on the following task, which can be solved by
applying the intermediate value theorem:

Partitioning their working process and characterising the identified phases, the
proving process of Michael and Leon can be described by the sequence of phases
presented in Figure 2.

Figure 2: Parsing of the proving process of Michael and Leon

After reading the task, Michael and Leon start their working process with a brief
brainstorming concerning useful theorems. Because the given function does not meet
the required preconditions, they experience difficulties applying these theorems,
which makes them reading the given statement a second time more carefully. Doing
so, they express confusion about the term fixed point and, hence, try to clarify its
meaning by looking it up and making a drawing. Based on an enhanced
understanding, Michael and Leon review their ideas and add new ones by leafing
through the book. Sooner or later, each of these ideas proves to be inadequate. While
Michael is reading in the book searching for applicable theorems, Leon makes
further drawings in order to visualise the preconditions and assertions given in the
task. Thereby, he remembers a strategic approach used in a proof before. Leaning on
this approach, he tries to construct an auxiliary function, which turns the original
problem of the existence of a fixed point into a similar one of a zero. Michael and
Leon work together on this approach until they recognise they have different
concepts regarding the connection between the domain and the range of the given
function. They resolve this disagreement by referring to the drawings and regarding
some exemplary points of the function. For more than half an hour Michael and Leon
have been working out the details of the proof now. At times they stop working on
new ideas and summarise previous insights in order to structure their arguments and
to identify gaps or inconsistencies in the proof outline. One of this structuring
activities sends Michael having doubts about the correctness of their formulation.

331 [Link]:indrum2018:173782
Leon is able to convince Michael of their approach, but has to admit that some
modifications are necessary like regarding an additional case. Later, another phase of
validating occurs as a result of identifying arguments. This time they recognise that
their auxiliary function is the identity function and that they can prepare their
arguments in a more elegant way. Michael and Leon are now able to construct a
satisfying proof by formulating and structuring alternately.
During the whole process of proof construction Michael and Leon are progressing
continuously until
they have achieved a
satisfying proof.
They only return to a
previous phase, when
their proving process
comes to a temporary
halt and they feel
there is a need for
reviewing some steps
done before in order
Figure 3: Proving process of Michael and Leon, validating
to specify or improve
activities are represented by stars
their mental
representation, their key idea or the proof outline. Michael and Leon go through a
phase of validating twice, because one of the student voices doubts about the
previous considerations being correct. They do not review their final proof to check
or improve details.
As displayed in figure 2, at times it has been inevitable to encode the same period of
time with two different categories. This kind of double-coding is necessary, if both
students work on their own in service of different goals or if some related activities
from different phases are done contemporaneously. The analysis of the proving
process presented in this section is quite typical for the sample. In the following
section similarities and differences between the individual proving processes are
discussed in general.
General observations
The analysis of data from nine interviews shows that the categories are applicable to
the proving protocols in a satisfactorily objective and reliable way. Interrater
reliability in coding is quite high (𝜅 = .73-.93). Therefore, the proving cycle seems to
be an appropriate tool for describing and analysing undergraduates’ approaches to
proof construction on a general macroscopic level. Regardless of the proving tasks
and a direct or non-direct approach to proving, each of the suggested phases could be
empirically confirmed in at least five of nine cases (Figure 4). In those cases, where
the phases identifying arguments and selecting and structuring arguments are
missing, participants have not been able to establish any serious approach due to a

332 [Link]:indrum2018:173782
Figure 4: Percentages of the different phases occurring in proving processes
poor mental representation or a lack of key ideas. However, quitting the working
process without any solution was only observed twice. More frequently it happens
that even in successful proving processes the phases of formulating and validating
are omitted. Instead of formulating a precise and clear proof, some participants
content themselves with quick but fragmentary solutions. Here, formulating the proof
according to mathematical standards does not seem to be as important as gaining
insight into the key ideas of a proof and verifying the statement for oneself. This
kind of view might be a general attitude towards mathematical formalism, but could
also be caused by the laboratory setting. Analysing the phase of validating, we
differentiate two kinds of action: Validating activities in service of reviewing and
enhancing a final proof could only be observed in one single case. In contrast,
activities like checking details and analysing suggestions occur continuously and are
connected to several other activities in the proving process. Although validating
activities are only listed in five of nine cases, it is reasonable that more validating
takes place without being encoded on a macroscopic level because of no sufficient
impact on the proving process. Comparing the percentages of different phases, it
emerges that in nearly all cases the phase of identifying arguments as a highly
creative and demanding action takes a large share in the proving process and,
therefore, seems to be one of the most significant parts of mathematical proof
construction. In contrast, the percentage of understanding and exploring the given
assertion varies greatly. While some participants spend a lot of time on drawings or
clarifications, others start with a brief glimpse on the task and continue with applying
theorems immediately.
In regard to the composition of proving processes, it has been assumed that the
underlying structure is a cycle. Due to that, there must be a high amount of
transitions and revisions in the proving approaches. In fact, only one third of the
encoded transitions is linear in that way that students move forward to the next phase
in the proving cycle. Though, the proving processes does not proceed as cyclic as
suggested. As illustrated in the case study of Michael and Leon a large shape of non-

333 [Link]:indrum2018:173782
linear transitions is made by transitions between consecutive phases, that is, moving
backwards to the phase before. Wider leaps from one phase to another are quite rare.
Most of these transitions between non-consecutive phases contain an interaction with
validating activities as students switch from any phase to validating and backwards.
This finding supports the assumption that validating is an activity, which is closely
related to other phases of the proving process. Only in a very few cases the key idea
is rejected at some point in the proving process and the participants restart
identifying arguments and exploring the given statement cyclically. Accordingly, we
suggest proof construction to be less cyclic than a linear process, which is interrupted
by several mini-cycles between consecutive phases.

DISCUSSION
In this paper a model for the process of mathematical justification at university level
has been derived from literature to develop a macroscopic analytical tool, which
describes a proving process as a sequence of phases. In a sample of nine processes
the proving cycle has proved suitable for describing undergraduates’ approaches to
mathematical proof construction. Each of the suggested phases could be empirically
confirmed. Analysing similarities and differences between individual proving
processes, there have been two key findings: Undergraduates’ proof construction
mainly proceeds on a straight line basis, which is interrupted at times by transitions
into immediately preceding phases in order to specify or improve considerations
done before. An exception to this are validating activities. Questioning, reviewing
and reflecting seem to be processes, which are rarely performed at the end of proof
construction, but are closely connected to other phases of the proving cycle. Hence,
initial results indicate that validating is not confined to the final proof, but relates to
the mental representation as well as the key ideas and the proof outline.
The presented study prepares a larger project by providing an analytical framework
for students’ approaches on proof construction. Based on the proving cycle, we
intend to analyse a larger sample including participants, who differ in progress and
performance. Comparing first-year and advanced students as well as successful and
non-successful proving processes might provide new ideas for fostering programs in
the introductory phase of studies. On the one hand investigations will remain on the
macroscopic level of phases in order to identify effective and less effective patterns
of proving processes. Therefore, the occurrence and the duration of a phase in a
proving process are compared with the quality of the corresponding proof. On the
other hand, further investigations are intended, which gain more inductive insights
on a microscopic level. Therefore, we aim at describing typical actions of each
relevant phase in detail. By doing so, frequent difficulties and potential obstacles of
an individual phase can be identified as well as effective and non-effective proving
strategies to overcome these obstacles. This information can help arranging effective,
process-oriented fostering programs.

334 [Link]:indrum2018:173782
REFERENCES
Boero, P. (1999). Argumentation and mathematical proof. A complex, productive,
unavoidable relationship in mathematics and mathematics education. International
Newsletter on the Teaching and Learning of Mathematical Proof (7/8). Retrieved
from [Link]
[Link]
Hemmi, K. (2008). Students‘ encounter with proof: the condition of transparency.
ZDM, 40, 413–426.
Hersh, R. (1993). Proving is convincing and explaining. Educational Studies in
Mathematics, 24 (2), 389–399.
Kintsch, W. & Greeno, J. G. (1985). Understanding and solving word arithmetic
problems. Psychological Review, 92 (1), 109–129.
Mayring, P. (2014). Qualitative Content Analysis. Theoretical Foundation, Basic
Procedures and Software Solution. Retrieved from [Link]
[Link]/urn:nbn:de:0168-ssoar-395173
Meyer, M. (2010). A logical view for investigating and initiating processes of
discovering mathematical coherences. Educational Studies in Mathematics, 74,
185–205.
Mejia-Ramos, J. P. & Inglis, M. (2009). Argumentative and proving activities in
mathematics education research. In F.-L. Lin, F.-J. Hsieh, G. Hanna, M. de Villers
(Ed.), Proceedings of the ICMI Study 19 Conference: Vol.2. Proof and Proving in
Mathematics Education (pp. 88–93). Taipei, Taiwan.
Moore, R. C. (1994). Making the transition to formal proof. Educational Studies in
Mathematics, 27 (3), 249–266.
Pólya, G. (1945). How to solve it. Princeton: Princeton University Press.
Reiss, K. & Renkl, A. (2002). Learning to prove. The idea of heuristic examples.
ZDM, 34 (1), 29–35.
Reusser, K. (1990). From text to situation to equation: Cognitive simulation of
understanding and solving mathematical word problems. In H. Mandl, E. De
Corte, N. Bennett & H. F. Friedrich (Ed.), Learning and Instruction: Vol 2.2.
Analysis of complex skills and complex knowledge domains (pp. 477–498).
Oxford: Pergamon Press.
Schoenfeld, A. (1985). Mathematical problem solving. Orlando: Academic Press.
Stein, M. (1984). Beweisen [Proving]. Bad Salzdetfurth: Franzbecker.
Weber, K. (2001). Student difficulty in constructing proofs. The need for strategic
knowledge. Educational Studies in Mathematics, 48, 101–119.

335 [Link]:indrum2018:173782
 Student use of resources in Calculus and Linear Algebra
Zeger-Jan Kock1 and Birgit Pepin²
1
Eindhoven University of Technology / Fontys University of Applied Sciences,
[Link]@[Link]; ²Eindhoven University of Technology, The Netherlands
In this study we have investigated the resources used by first year engineering students
in a technical university in the Netherlands, for their learning of Calculus and Linear
Algebra. Using a case study approach we have focused on how the resources and their
use (a) differed from upper secondary school as compared to university, and (b)
differed between the two university courses. The results indicate that, in terms of (a)
students built on secondary school experiences and emulated these into their university
courses, where some subsequently experienced difficulties. In terms of (b), we argue
that the course organization and the alignment of curriculum materials with the
learning goals had an impact on the students’ choice and use of resources. Human
resources played an important but varying role.
Keywords: Student use of resources, Case study, Transition from school to university,
Calculus, Linear Algebra.

INTRODUCTION
At university level a large diversity of resources is currently made available for
students learning mathematics. These include traditional curriculum resources (e.g.
readers, textbooks); digital (curriculum) resources (e.g. YouTube, websites, apps); and
also human resources (e.g. drop-in clinics run by tutors; setups for peer groups). The
ways in which university mathematics teachers interact with various resources has been
investigated by Gueudet (2017), for example, and several studies have been conducted
related to university students and their use of particular resources to learn mathematics
(Anastasakis, Robinson, & Lerman, 2017; Biza, Giraldo, Hochmuth, Khakbaz, &
Rasmussen, 2016; Inglis, Palipana, Trenholm, & Ward, 2011). However, relatively
little is known about how students of mathematics in their first year of university “cope
with” the plethora of available resources available to them, and how they organise and
coordinate them for their learning.
Using a case study approach, we have studied the resources in a Calculus (CS) and a
Linear Algebra (LA) course, and their use by students, in the context of a first year
Bachelor College programme at a technical university in the Netherlands. Moreover,
we have investigated, retrospectively, which resources were used by the students, and
how, in upper secondary school (as compared to university). Hence, we propose the
following research question:
Which kinds of resources are used by the students, and how, in first year university
Calculus and Linear Algebra courses, and how do these practices compare to students’
experiences at upper secondary school?

336 [Link]:indrum2018:174450
In this paper, first, we briefly outline selected insights from the relevant literature,
including our theoretical frame of “resources” (and their use) as a lens to develop a
better understanding of students’ mathematics learning. Second, we describe our
chosen methodology and data collection strategies, before we present our findings and
discuss our results in the third section. Fourth, we present our conclusions and outline
implications for the practice of university mathematics learning and teaching.

THEORETICAL FRAMEWORKS
Transition from secondary to tertiary education
In terms of mathematics learning, the transition from secondary school to university is
challenging for many students (Pepin, 2014), as discontinuities exist between
secondary and tertiary mathematics education. The literature reports numerous
differences between studying mathematics at school as compared to university. It is
said that in comparison to secondary education, at university: (a) the mathematical
content is introduced at a higher speed; (b) more mathematical autonomy is expected;
(c) the levels of generalization and abstraction are higher; (d) the approach is more
formal with an increased emphasis on proof; and (e) the institutional cultures at the two
institutions (secondary school, university) are different (Artigue, 2016; Gueudet,
2008). The ways the content is made available to students also differ between
secondary and tertiary education (e.g. Corriveau & Bednarz, 2017). University students
have to autonomously manage the various resources to learn mathematics, and it is
argued that secondary school does not prepare them well for this task (Williams, Black,
Davis, Pepin, & Wake, 2011). Thus, it can be expected that first year university
students have to find new ways of working with the resources they have access to, and
that are proposed to them in their courses.
Student use of resources
The use of resources by students has been the subject of relatively little research.
Selected studies (e.g. Anastasakis, et al., 2017) indicate that students, in their selection
of resources, have been predominantly motivated by the goal to be successful in
examinations (and to obtain high grades). The authors of this study made an inventory
of the resources used by students when studying for mathematics modules, and
explicitly related these to their learning goals. The most widely used resources were
those that the university provided for the students, and their own notes. The use of
particular resources, for example mathematics textbooks, was specifically linked to the
study of worked examples, which were said to help students to prepare for
examinations; albeit this often lead to emphasise the surface aspects of the examples
(Biza, et al., 2016). In their review study Biza, et al. (2016) identified several
limitations of tertiary mathematics textbooks, in particular the emphasis on formal
aspects of mathematics, at the cost of opportunities to develop intuitive meanings and
understandings. Relating the use of particular resources to examination grades, a study
by Inglis, et al. (2011) found that students who attended lectures or used the
university’s mathematics support centres had higher grades than students who often

337 [Link]:indrum2018:174450
watched online lectures. The authors suggest that students might need explicit guidance
on how to combine the use of various resources into an effective learning strategy.
Before this guidance can be given, or be reified in a blended learning environment,
more in-depth information on the actual use of resources by students is needed.
The lens of resources
In this study we use the notion of “re-source/s” that students have access to and interact
with in/for their learning. We assume that the ways university students learn
mathematics is influenced/shaped by their use of the various resources at their disposal.
By “use of resources” we denote, for example, which resources students choose
(amongst the many on offer) and for what purpose (e.g. revision); the ways they align
them (e.g. first lecture then checking the textbook, etc.); which ones seem central to
achieve particular learning goals (e.g. for weekly course work, examinations, for their
engineering topic area). However, we do not address the specific learning of CS and
LA, that is how students interact with particular (e.g. cognitive) resources to learn
particular topic areas in CS and/or in LA.
Gueudet and Pepin (in press) have defined student resources as anything likely to re-
source (“to source again or differently”) students’ mathematical practice, leaning on
Adler’s (2000) definition of mathematics “re-sources” (in Adler’s case used by
teachers). In this study we distinguish between (1) material resources, and (2) human
resources. (1) For material resources a further distinction has been made between (a)
curriculum resources (those resources proposed to students and aligned with the course
curriculum), and general resources (which students might find/access randomly on the
web). Curriculum resources are developed, proposed and used by teachers and students
for the learning (and teaching) of the course mathematics, inside and outside the
classroom (Pepin & Gueudet, 2014). They can include text resources, such as
textbooks, readers, websites and computer software, but also feedback on written work.
General resources are the non-curricular material resources mobilized by students, such
as general websites (e.g. Wikipedia, YouTube). (2) In terms of human resources we
refer to formal or casual human interactions, such as conversations with friends, peers
or tutors.

METHOD
Context
The study took place at a university of technology in the Netherlands, with a student
body of approximately 13000 engineering students. The university offers 15 bachelor
courses related to technology and engineering.
We selected two first year courses in the first term of the 2016-2017 academic year:
Calculus (CS); and Linear Algebra (LA). We purposefully chose these courses, as they
were different in size and target group: the CS course was obligatory for all first year
engineering students, approximately 2000 students, whereas the LA course was
targeted at “applied mathematics and physics” engineering students only,

338 [Link]:indrum2018:174450
approximately 130 students. The CS course was organized by the mathematics
department, differentiated at three levels (A, B, and C), according to perceived level of
difficulty and with varying level of emphasis on formal aspects of mathematics (e.g.
proof).
In CS, six hours of lectures were organised each week, and one hour of tutorials in
groups of eight students. In the course catalogue, and this was supported by lecturers,
the aim of CS was to give engineering students a “basis” to be able to “calculate
correctly”. It appeared that the aims of the CS course were to provide students with a
basic set of mathematical/computational tools they could subsequently use in their
engineering studies and in their future work as engineers.
In LA, four hours of lectures were organised each week, and four hours of tutorials, in
groups of approximately 30 students. As in CS, the LA learning aims were described
as the acquisition of mathematical skills. Moreover, aims of the course were to help
students develop the skills and realize the importance of correct mathematical
communication, including writing formal proofs. Completing a mathematical writing
assignment was part of the course requirements to reach this aim. It appeared that the
purpose of LA was to prepare students for higher mathematics (used in the mathematics
and physics courses).
Participants
In total, 24 students participated in the study: 18 CS students (involved in nine different
engineering programs and all taking the B level CS course); 1 CS student who dropped
out of university; 5 LA students (all studying for the ‘applied mathematics’ engineering
course). In terms of background, of the interviewed CS students 15 came from
secondary schools in the Netherlands, three came from other educational systems. For
the Dutch students the CS content was partly familiar, in particular for those who took
“strong mathematics” courses (Wiskunde D) at secondary school. Four of the five
interviewed LA students came from secondary schools in the Netherlands, one student
had attended secondary school in Belgium.
Data collection strategies
Data collection strategies included the following:
(1) Student interviews: The CS students were interviewed in four focus groups, and
one individual interview. During the interviews students were asked to make a drawing
of the resources they used for their mathematics course (Schematic Representation of
Resource System, SRRS - Pepin, Xu, Trouche, & Wang, 2017). These helped the
interviewer to understand the ways the resources were used, and for which purpose.
The LA students were interviewed in two groups of two, and one individual interview.
(2) Documents/curriculum resources: Relevant curriculum materials and documents
(digital and text materials) were collected and analyzed. These materials were provided
by the university for the students (e.g. examples of examinations, LA syllabus, LA

339 [Link]:indrum2018:174450
study guide, LA assignments, CS study guide, the CS textbook, course summaries in
the university’s course catalogue, video clips, videos of the lectures).
(3) Teacher interviews: Interviews with two CS lecturers and one CS tutor were
conducted, as well as one LA lecturer and one LA tutor.
For analysis, the interviews were transcribed and interview quotations were coded
using ATLAS-ti software. The codes were based on our knowledge from the literature
concerning the different curriculum resources and their use. In the next step of the
analysis the findings from CS and LA were compared, and subsequently these with
those from upper secondary school.

RESULTS
Resources at secondary school
In terms of curriculum materials/material resources the textbook was an important
resource for most secondary school students, and so were the graphical calculator (also
used in examinations) “to quickly plot graphs” (interview reference: CSS01), and past
examination papers for revision and practice to prepare for the national examinations.
The textbook was seen as the main source of exercises, which were done in class or at
home. Regarding homework one student remarked:
At school I didn't do my homework. There was homework but yeah, if you worked on
it during the class … you would get halfway and then at home I was like oh, I get it. I
don't have to do the remaining exercises (LAS01).
Online general resources (e.g. YouTube; Kahn academy) were hardly mentioned in
relation to secondary school.
In terms of human resources the teacher and classmates were mentioned as an
important support for secondary school students. Interestingly, teachers’ explanations
of the mathematical concepts were not important for all students: in some schools
students apparently worked largely independently (with the textbook) and the teacher
was only occasionally consulted.
In short, to learn the mathematics and pass the examinations, students reported that it
was sufficient to follow the teachers’ explanations, do (all) the exercises in the
textbook, and practice with the past examination papers. There were few resources, and
the ones provided could be straightforwardly accessed and used for solving the
problems posed.
Resources for CS
Figure 1 shows the typical resources used by a CS student. The lecture appeared to be
an important starting point for many students, albeit not for all. They provided an
orientation on the subject (“it’s easier for me to revise/practice when I have already
seen/heard about it”; Figure 1), and to some extent an enculturation into the world of
mathematical concepts and their usages. A lecturer said:

340 [Link]:indrum2018:174450
 I do the historical aspect too. Some things have been known for 3000 years, so you have
to know that too. So I also do applications and add historical things (..) I just want them
to be excited about the subject (CSLI03).
CS was supported by an almost 1100 page general CS textbook (authored by an
“external” author) chosen from commonly used university CS textbooks. It contained
the essential theory and part of the homework exercises (explained by someone
“outside” the students’ environment), but its content was not specifically aligned with
the lectures and the final examinations. Curriculum resources, such as the textbook,
were mainly used for exercises, whereas lecture notes (by the teacher) seemed
important for knowing about the content to be learnt. Students’ own notes were often
used for orientation (“I write down important stuff; also when the teacher says ‘this is
likely to be on your exam’, I write this down.”; Figure 1). Selected students used the
textbook for additional/different purposes: to read the theory in the textbook to prepare
in advance for the lecture.
At the same time digital resources, such as general online videos (e.g. YouTube; Kahn
academy) and video-recorded lectures (whole lectures or clips of particular moments),
were said to provide additional explanations. The course’s online tests and the weekly
coursework were used to check whether one had a good (basic) understanding of the
content. Moreover, students also mentioned human resources, such as the lecturer/tutor
and the roommate “to ask questions” (Figure 1). Friendship groups were important for
many students, and for some they were their first line of support (before the tutor or
peers in the tutorial/s), to work with on the coursework or to consult about difficult
theory or exercises.

Figure 1. Purpose and coordination of resources by a CS student

341 [Link]:indrum2018:174450
Resources for LA
In terms of material resources LA was supported by a 200 page course-specific reader,
authored by the lecturer and developed and improved over the years. It contained the
essential theory and the exercises required to prepare for the examinations - this was
“the backbone” of the course, according to the lecturer and the tutors. Other resources,
such as the lectures, the homework/coursework exercises, the lecture notes and the
videos with worked examples were all aligned with the reader, to become a
comprehensive and complete set of resources for the students. The importance of using
these resources to individually make sense of the mathematical content, and doing lots
of practice exercises “at home”, was emphasised by a student:
But now you have like a huge amount of homework and then you also have workgroups
where you can work on it, but then you don't get very far. (…) And if you don't do it at
home, you just won't get it and you won't make your tests really good. So you really
have to do a lot at home (LAS01).
As in the CS course, the lectures appeared a starting point for many students; they
provided an orientation on the subject, and an enculturation into the world of
mathematics (e.g. when mathematical proof was explained as an essential
mathematical thinking process). Moreover, in terms of human resources, students
relied on peer groups (e.g. they collaboratively solved problems during the weekly
tutorials), and on the tutors to provide help- this was an important support in the LA
course. Tutors were considered more approachable than the lecturers, although students
were generally positive about the possibilities to ask questions to lecturers.
Comparing resources and their use
School – university: Whilst selected resources (e.g. textbooks, past examination
papers) appeared to be part of the “staple diet” for every student at university or school
level, at university students tended to use more, and more varied resources than at upper
secondary school, including online lecture videos, video clips (of “difficult” notions),
online texts. In addition, selected resources, such as lecture notes, were not mentioned
in the secondary school context, where the theory would be taught by the teacher who
aligned his/her lessons with the book. Some of the additional resources, such as video
lectures, teachers’ lecture notes, readers on specific topics, and online tests, were part
of the curriculum resources made available by the university. Other resources, such as
online applets and videos, were identified by the students themselves.
In terms of human resources there were also differences: at secondary school for most
students the teacher provided practically all of the necessary guidance for learning the
mathematical topic. At university the lecturer provided the theory, an overview of what
was important for their learning of the topic area (and also for the examinations), and
selected worked examples. The practical guidance, i.e. how to solve particular
mathematical problems, was mainly provided in the tutorials and the
exercises/coursework accompanying the lectures. Hence, students had to find their own
learning/peer groups and supports for learning, as on their own it was not possible to

342 [Link]:indrum2018:174450
manage the amount of work and the pace it was taught. This situation was exacerbated
in the CS course, as only one hour of tutorial group work was offered, and students had
to work collaboratively outside this hour for completion of their tasks. Hence, many
CS students organized and coordinated their own support to work with their peers on
the coursework, or to consult about difficult theory or exercises.
The students reported that, compared to secondary school, at university: (a) the pace
was faster, (b) the content was more difficult to understand, and (c) the mathematical
content was offered in larger steps/sections. The interview data suggest that the role
and importance of resources changed as a result of this, as students needed more
support structures and feedback on their work. This was particularly pertinent with one
(autistic) student, who had dropped out of university. He claimed that he had done all
possible CS textbook exercises and interim tests – a practice he had succeeded with at
school, but he could not make sense of the questions when he sat the final examination.
He was lost in the immensity of resources on offer, which he could not possibly all trial
out and use for his learning. And he clearly missed the guidance and support given by
his schoolteacher, practices which had provided him with confidence for his learning,
and success.
CS- LA: Amongst the university curriculum resources, the student usages of the LA
reader and CS textbook differed. To come to understand the topic/s, most LA students
reported reading the reader, or the lecture notes, which were aligned with the reader.
CS students mentioned the textbook as one of their resources, mainly used for worked
examples and exercises. In CS, lecture notes and online resources were considered
practically as important as the textbook, as part of the provided resource system. This
can be understood in the light of the fact that the LA reader was very different, in
relation to the course, to the CS textbook: the LA reader was a “book” prepared by the
lecturer to align with his lecture, hence further lecture notes or online resources became
secondary/ complementary. The reader contained all information for students to pass
the examinations, and all other resources were related to/in line with the reader. In
contrast, the CS book was only a backup for the lecture notes (which provided the
essential notions to learn and study for the examinations), and students were only
expected to “dip into” it for clarification, explanation and/or further exercises. Hence,
the textbook did not provide a succinct support for CS students (e.g. to pass the CS
examinations).
An important difference between CS and LA was due to the different organization of
the tutor hours: in LA- 4 tutor hours/week, tutor groups of ca. 30 students/ tutor; in CS-
1 tutor hour/week, tutor groups of ca. 9 students/tutor. This meant that CS students had
to work on practically all of the problems by themselves, as there was less support from
the tutor. The fact that a wider variety of resources (used) was reported by the CS
students, can in part be understood in the light of the different course organisation: in
the LA case they adhere to the resources provided by the lecturer; in the CS case the
students had to identify and organize their own support (e.g. online resources,
friendship groups).

343 [Link]:indrum2018:174450
CONCLUSION
The results of this study have shown that the students built on secondary school
experiences and they took these as default positions into their courses. However,
learning mathematics at university was for most students different from learning
mathematics at secondary school. At secondary school the resources (text book, past
examinations, teacher) were well aligned and the teacher provided guidance and
support. At university more difficult content had to be understood in a shorter period
of time; and students had to identify and coordinate the relevant resources, and organise
their own support system (including human support such as friendship learning
groups), in particular in CS.
When comparing the two courses, the results indicate that (a) the course organization
and (b) the provisions and organisation of the curriculum materials (in line with the
learning goals) had an impact on which resources students used, and how they used
them. In the LA course, with aligned curriculum resources, four weekly tutorial hours
and group work, the use of resources largely corresponded to the intentions of use by
the university teachers. In the CS course resources were not clearly aligned (although
selected resources were recommended); it seemed that students were provided with a
“bag of tools” to choose from. Moreover, students had only one weekly tutorial hour
(plus six hours of lecture). This meant that students had (a) to identify which were the
relevant resources for their individual needs, and (b) to find and navigate their own
path through these resources, in order to work efficiently (with regards to
examinations) and effectively (with regards to the learning of the mathematics). In both
courses human resource, such as lecturers and tutors, peers and friends, played
important albeit changing roles for orientation and help seeking.
The results of this exploratory study indicate that in particular large courses, such as
CS (> 2000 students), could become better manageable for the students, if they were
supported and coached in their resource choice and organisation/management, so that
they can cater for their individual needs and preferences. This finding was less visible
in a smaller course, where less resources were on offer, and where resources and
resource use were more prescriptive and well aligned with the learning goals (e.g. the
LA course). However, when particular educational reforms are implemented (e.g.
towards more blended learning, with an abundance of digital learning tools on offer),
students need to be supported in their “use” of these resources, in particular at transition
from school to university. Course designers would also need to take this into
consideration.

REFERENCES
Adler, J. (2000). Conceptualising resources as a theme for teacher education. Journal
of Mathematics Teacher Education, 3(205 - 224).
Anastasakis, M., Robinson, C. L., & Lerman, S. (2017). Links between students’ goals
and their choice of educational resources in undergraduate mathematics.

344 [Link]:indrum2018:174450
 Teaching Mathematics and Its Applications, 36, 67-80. doi:
10.1093/teamat/hrx003
Artigue, M. (2016). Mathematics education research at university level: Achievements
and challenges. Paper presented at the First conference of International Network
for Didactic Research in University Mathematics, March 2016, Montpellier,
France.
Biza, I., Giraldo, V., Hochmuth, R., Khakbaz, A., & Rasmussen, C. (2016). Research
on teaching and learning mathematics at the tertiary level: State-of-the-art and
looking ahead ICME-13 Topical Surveys, doi:0.1007/978-3-319-41814-8_1
Corriveau, C., & Bednarz, N. (2017). The secondary-tertiary transition viewed as a
change in mathematical cultures: An exploration concerning symbolism and its
use. Educational Studies in Mathematics, 95, 1-19. doi: 10.1007/s10649-016-
9738-z
Gueudet, G. (2008). Investigating the secondary-tertiary transition. Educational
Studies in Mathematics, 67(3), 237-254.
Gueudet, G. (2017). University teachers’ resources systems and documents.
International Journal of Research in Undergraduate Mathematics Education, 3,
198-224.
Gueudet, G., & Pepin, B. (in press). Didactic contract at the beginning of university: A
focus on resources and their use. International Journal of Research in
Undergraduate Mathematics Education.
Inglis, M., Palipana, A., Trenholm, S., & Ward, J. (2011). Individual differences in
students’ use of optional learning resources. Journal of Computer Assisted
Learning, 27, 490-502.
Pepin, B. (2014). Re-sourcing curriculum materials: in search of appropriate
frameworks for researching the enacted mathematics curriculum. ZDM - The
International Journal on Mathematics Education, 46(5), 837-842. doi:
10.1007/s11858-014-0628-5
Pepin, B., & Gueudet, G. (2014). Curricular resources and textbooks. In S. Lerman
(Ed.), Encyclopedia of mathematics education. Berlin, Heidelberg: Springer.
Pepin, B., Xu, B., Trouche, L., & Wang, C. (2017). Developing a deeper understanding
of mathematics teaching expertise: an examination of three Chinese mathematics
teachers’ resource systems as windows into their work and expertise.
Educational Studies in Mathematics, 94(3), 257-274. doi: 10.1007/s10649-016-
9727-2
Williams, J., Black, L., Davis, P., Pepin, B., & Wake, G. (2011). Mathematics learning,
identity and educational practice: the transition into Higher Education
TransMaths Research Briefing: The University of Manchester.

10

345 [Link]:indrum2018:174450
Comparing male and female students’ self-efficacy and self-regulation
skills in two undergraduate mathematics course contexts

Juulia Lahdenperä
University of Helsinki, Department of Mathematics and Statistics
[Link]@[Link]

Students’ efficacy beliefs have a positive influence on students’ academic

achievement and retention, especially for female students. These beliefs are closely
linked to students’ ability to regulate their learning. In this quantitative study,
students’ self-efficacy and self-regulation of learning are compared in two university
mathematics courses that differ in content but also in their pedagogical setting: one
is a more traditional lecture-based course, and the other course is taught with the
Extreme Apprenticeship (XA) method. The analysis is based on the same cohort of
students in the two contexts (N=91). The results suggest that students have higher
self-efficacy levels in the course using the XA method. Also, the XA course seems to
diminish some gender differences present in the more traditional course setting.
Keywords: teachers’ and students’ practices at university level, novel approaches to
teaching, self-efficacy beliefs, self-regulation, instructional design

INTRODUCTION
According to Lave and Wenger’s (1991) view of situated learning, a set of skills as
well as a set of values and perspectives are needed for a holistic understanding of a
topic. Collins, Brown, and Holum (1991) suggest that the process of acquiring this
kind of understanding occurs most naturally within the community possessing this
knowledge. In this light, it is not irrelevant what kind of instructional practices are
implemented to teach university mathematics, and how do these practices offer
students opportunities to participate. As Greeno (1997, p. 9) states, “methods of
instruction are not only instruments for acquiring skills; they also are practices in
which students learn to participate”. Therefore, to enhance students’ learning,
instructional designs used to teach university mathematics should offer students both
an opportunity to acquire knowledge and an opportunity to become part of a
community.
Partly to answer to this need, lots of effort has been put into developing the
educational setting at the Department of Mathematics and Statistics at the University
of Helsinki (see eg. Oikkonen, 2009; Rämö, Oinonen, & Vikberg, 2015). The
department’s teaching has gone through major changes during the past few years as
many of the undergraduate level courses are now using the Extreme Apprenticeship
(XA) method as their pedagogical framework. In addition, some of the courses

346 [Link]:indrum2018:174796
working within the traditional lecture-based framework have been developed
towards a more interactive direction.
This paper approaches the development of university mathematics education from
the perspective of efficacy beliefs and self-regulation of learning. As a part of a
larger research project aiming at comparing university mathematics teaching
practices and transferring knowledge from research into practice, this paper
elaborates on students’ experiences of different instructional designs with the focus
on their efficacy beliefs and self-regulation of learning.

THEORETICAL FRAMEWORK
Self-efficacy is a person’s belief about how well they can perform a specific task in a
specific context; these beliefs determine how people feel, think, motivate themselves
and behave (Bandura, 1994). Self-efficacy beliefs play a crucial role in learning
mathematics as self-efficacy enhances academic achievement (Pajares, 1996; Peters,
2013), especially in female students (Raelin et al., 2014). The gender aspect of self-
efficacy is relevant as it affects female students’ career choices and increases their
retention in STEM fields (Pajares, 1996; Raelin et al., 2014). In terms of
instructional design, Peters (2013) shows that students’ self-efficacy is higher in
teacher-centred than in learner-centred classroom. However, Kogan and Laursen
(2014) argue that female students obtain affective gain from student-centred courses
as they are more confident in their mathematical abilities in these kinds of contexts.
The notion of self-regulation characterises how students regulate their cognition,
behaviour, motivation and emotions to enhance their personal learning processes
(Pintrich, 2004). Students are expected to learn self-regulation skills during their
university studies and therefore instructional designs should support the development
of these skills (Coertjens et al., 2013). The self-regulation process is cyclical in
nature as feedback from prior performance is used to adjust future learning
performances (Pintrich, 2004; Zimmerman, 2000). Consequently, the quality of self-
regulated learning is supported by motivational beliefs, such as self-efficacy beliefs
(Heikkilä, & Lonka, 2006; Pajares, 1996). As the social aspect of learning has a
significant role in learning self-regulation skills (Volet, Vauras, & Salonen, 2009),
instructional designs should also encourage student collaboration.
Aims and research questions
The aim of the paper is to compare students’ self-efficacy and self-regulation of
learning in two different course contexts. The further analysis focuses on gender as
previous research shows that especially female students benefit from more student-
centred course designs. The research questions are:
1. How do self-efficacy and self-regulation of learning differ between the two
course settings?

347 [Link]:indrum2018:174796
 2. How is the course setting related to male and female students’ self-efficacy
and self-regulation of learning?

METHODOLOGY
This study approaches the research questions with a quantitative analysis of students’
self-efficacy and self-regulation of learning in two course contexts. The following
subsections move on to describe the context, the data collection procedure and data
analysis in greater detail.
Context
The research was conducted in a research-intensive university in Finland. Data was
collected from two different courses that students usually take during the first
semester of their university mathematics studies. Both courses are proof-based, six-
week and five-credit (ECTS) courses with over 200 students. In addition to
mathematics majors, the courses are taken by many students studying mathematics as
their minor subject; these students are usually majoring in physics, computer science,
chemistry or education. The two courses, course A and course XA, are implemented
in accordance with different pedagogical frameworks. The main difference in the
course implementations are the role of lectures, design of the tasks, and the form of
support given to the students by the teaching assistants.
Course A is an analysis course. The main content of the course includes limit of a
function, continuity, derivative, and its applications. It is necessary to point out that
the course is an analysis course rather than a calculus course as exact definitions and
proof construction are emphasised. The course functions within the traditional
lecture-based setting. However, it has been developed over a decade towards a more
interactive direction to respond to students’ challenges in the beginning of their
university mathematics studies.
The course A consists of four hours of lectures and four hours of small group
sessions per week. The lectures are focusing on the main content of the course and
aim at creating deep understanding behind those concepts. Inspired by Tall’s three
worlds of mathematics (see e.g. Tall, 2014), the lectures are an active interplay
between the human and formal sides of mathematics. The small group sessions are
led by a teaching assistant, who is usually an older mathematics student. There are
two different kinds of small group sessions. The other one is allocated to the
problems students have solved prior to the class. The other small group session is
allocated to solving a new set of problems during the session together with other
students and with the help of the teaching assistant.
Course XA is a linear algebra and matrices course. The main content of the course
includes general vector spaces, subspaces, linear mappings, and scalar products. In
addition to mathematical content, the course emphasises skills such as reading
mathematical text, oral and written communication, and proof construction. The

348 [Link]:indrum2018:174796
course is taught with the XA method. The XA method is a student-centred
educational method developed in the Department of Mathematics and Statistics and
the Department of Computer Science at the University of Helsinki. The method
emphasises learning by doing, personalised scaffolding and continuous feedback,
and the core idea is to support students in becoming experts in their field by having
them participate in activities that resemble those carried out by professionals (see eg.
Rämö et al., 2015). The XA method is constructed upon the ancient process of
apprenticeship, where a skilled master supervises a novice apprentice, and its
theoretical background is in situated view on learning and Cognitive Apprenticeship
(Collins et al., 1991; Rämö et al., 2015).
In the XA method, students learn skills and gain knowledge by working on tasks that
have been divided into smaller and approachable goals, which are then merged
together as the students start to master a topic. The main method of teaching is
instructional scaffolding, and it is accompanied with continuous, bi-directional
feedback. Further, it supports students to establish relations within the communities
of practice which enhances the students’ integration into the community (Lave &
Wenger, 1991).
In practice, the teaching of the course consists of weekly problems, course material,
guidance and three hours of lectures per week. There is a flipped learning approach
as students start studying a new topic by solving a set of problems. These topics have
not yet been discussed during the lectures, so students need to read the course
material to complete the tasks. However, the tasks are designed to be approachable
and there are teaching assistants specifically to this course helping the students in
solving the problems. The teaching assistants guide the students in a learning space
in the middle of the department in drop-in basis approximately six hours a day.
Student collaboration is encouraged in the learning space. Students return written
solutions to the problems every week. Few problems are selected for inspection and
students get feedback for their solutions. The feedback focuses on solutions’ logical
structure, but also readability and language are evaluated, and students’ have the
possibility to improve and resubmit their solutions. Students are prepared when they
come to lectures as they have done pre-lecture tasks. Lectures focus on active
interaction as various small group activities are implemented and students’ active
participation encouraged. The aim is to form links between the topics and enhance
holistic understanding. After the lectures students get more challenging problems on
the topic.
Data collection
Quantitative data was collected on a five-point Likert scale (1=completely disagree,
5=completely agree) from students attending both courses. The questionnaire
included items measuring students’ approaches to learning, their experiences of the
teaching-learning environment, self-efficacy and self-regulation of learning. In this
paper, the analysis includes items measuring self-efficacy and self-regulation of

349 [Link]:indrum2018:174796
learning from students who answered the questionnaire for both course contexts
(N=91).
There are five items measuring self-efficacy. The items are slightly modified from
Pintrich (1991) and they are validated and highly used across disciplines in the
Finnish context (see e.g. Parpala & Lindblom-Ylänne, 2012). The 15 items
measuring self-regulation of learning are originally from the Inventory of Learning
Styles (ILS, Vermunt, 1994) and they have been modified to Finnish context
(Heikkilä, & Lonka, 2006). There self-regulation of learning is measured in four
scales: self-regulation of process, self-regulation of content, external regulation, and
lack of regulation. Self-regulation of process refers to a student’s ability to regulate
their own learning when facing challenges. Self-regulation of content measures
student’s seeking of additional literature beyond the course material. External
regulation measures to what extent the lecturer regulates student’s learning. Lack of
regulation refers to possible problems in regulation of learning, such as not knowing
how to proceed in the learning process or having challenges in finding ways to cover
the course content.
Data analysis
The data analysis is conducted by using IBM Statistics 24. The data in this paper is a
part of a larger data set. The results reported here are from the factor analyses
computed for the larger data set.
At first, exploratory factor analysis (EFA) was conducted with principal axis
factoring and a direct oblimin rotation. Based on the exploratory factor analysis,
there are four factors measuring self-regulation of learning. This is in accordance
with previous research (see eg. Heikkilä, & Lonka, 2006). Similarly, the factor
structure of the self-efficacy scale is like in previous studies (see e.g. Parpala, &
Lindblom-Ylänne, 2012) forming one factor. Boundaries used for Kaiser-Meyer-
Olkin measure of sampling adequacy (KMO) was 0.7, and for Bartlett's test of
sphericity p<0.001. One item measuring self-regulation of content was excluded
from the factor based on a low communality, a mixed factor loading and deviant
skewness and kurtosis. Every factor was then checked for internal consistency: the
Cronbach’s Alpha is 0.905 for the self-efficacy factor, 0.681 for the self-regulation
of process factor, 0.671 for the self-regulation of content factor, 0.708 for the
external regulation factor, and 0.661 for the lack of regulation factor. The
reliabilities are above the 0.65 level which can be considered acceptable.
As the current study follows a repeated measures design, the data was analysed by
using two-tailed paired samples t-test and Cohen’s effect size d. In addition, one-way
MANOVA with Wilk’s Lambda was used to analyse the interaction of independent
variables (gender) on the dependent variables (different course settings).

350 [Link]:indrum2018:174796
RESULTS
The data consists of 91 students (46 male, 45 female). These students attended both
course A and course XA. Students’ scores on self-efficacy and self-regulation scales
in both course contexts are presented in Table 1 with means, standard deviations,
mean differences, paired-samples t-test for statistical significance, and Cohen’s d for
effect size.
The biggest differences between the two course contexts lie in the self-efficacy and
lack of regulation factors. Students report statistically significantly higher self-
efficacy levels in course XA compared to course A (MD=0.58, t(90)=6.226,
p<0.001). This means that students are more confident in their abilities to succeed in
course XA compared to course A. The effect size (Cohen’s d=0.62) implies a
moderate role for the course context when measuring self-efficacy. A similar
phenomenon occurs in the lack of regulation factor; students report statistically
significantly less lack of regulation in course XA compared to course A (MD=0.48,
t(90)=6.987, p<0.001). In practice this means that on average, students report that
they lack regulation of learning more often in course A compared to course XA. In
other words, it was easier for students to find ways to handle large quantities of
content, self-evaluate their learning, and to meet the learning goals in course XA
compared to course A. The effect size (Cohen’s d=0.63) implies a moderate role for
the course context when measuring lack of regulation.
There are also smaller mean differences in the self-regulation of process and self-
regulation of content factors between the two course contexts (MD=0.14 and MD=-
0.19 respectively). These differences are statistically significant on a 0.05 level
(t(90)=2.189, p<0.05 and t(90)=-2.383, p<0.05 respectively). The results indicate
that on average, students in course XA seek more actively additional literature
beyond the course material and do more work than expected when compared to
course A (self-regulation of content). In addition, an average student in course XA
reports that they are more capable of regulating their learning processes when facing
challenges when compared to course A. However, one must notice that the effect
sizes are below 0.2 suggesting an insignificant role of the course contexts.
There is no statistically significant difference in the external regulation factor (MD=
0.06, t(90)=0.941, p=0.35). This means that on average, students report that the
lecturers’ instruction on how and in what order to proceed in learning the content
influences their learning similarly in both course contexts.
Course A Course XA
Variable Mean SD Mean SD Mean difference Effect
(XA-A) size
Self-efficacy 3.09 0.96 3.67 0.86 0.58*** 0.62
Self-regulation of 2.83 0.82 2.97 0.80 0.14* 0.17

351 [Link]:indrum2018:174796
 process
Self-regulation of 2.37 1.08 2.19 1.07 -0.19* 0.17
content
External regulation 3.59 0.78 3.65 0.72 0.06 0.08
Lack of regulation 3.28 0.78 2.80 0.75 -0.48*** 0.63

Table 1: Students’ scores, mean differences and effect sizes on the self-efficacy and self-
regulation factors in courses A and XA, as determined by two-tailed paired samples t-
test (* for p<0.05 and *** for p<0.001 significance levels) and Cohen’s d.
Let’s now move on to analyse both male and female students in one course context at
a time. In course A context, male and female students differ statistically significantly
in the self-efficacy factor (MD=-0.54, p<0.01, F(1,89)=6.602, partial η2=0.079). This
difference is not present in course XA context (MD=0.04, p=0.821, F(1,89)=0.038,
partial η2=0.001). These results indicate that female students have statistically
significantly lower self-efficacy in course A compared to male students; however,
this difference between genders is not present in course XA context.
There are statistically significant differences also in the external regulation factor. In
course A context, the mean difference between male and female students is not
statistically significant (MD=0.31, p=0.054, F(1,89)=3.827, partial η2=0.041).
However, the p-value is very close to the 0.05-significance level. In course XA
context, male and female students differ statistically significantly in the external
regulation factor (MD=0.51, p<0.001, F(1,89)=12.947, partial η 2=0.127). This means
that, in course XA contexts, an average female student applies more external
regulation compared to an average male student. In other words, female students
report that the lecturers’ instructions influence their learning more when compared to
male students.
There are no statistically significant differences between male and female students in
the two course contexts in self-regulation of process, self-regulation of content, and
lack of regulation factors.
Male Female
Factor Course Mean SD Mean SD Mean difference
Self-efficacy A 3.36 0.96 2.82 0.89 -0.54**
XA 3.65 0.96 3.69 0.72 0.04
External regulation A 3.43 0.80 3.74 0.74 0.31(*)
XA 3.40 0.72 3.91 0.61 0.51***

352 [Link]:indrum2018:174796
Table 2: Differences between courses A and XA based on students’ gender, as
determined by one-way MANOVA (* for p<0.05, ** for p<0.01, and *** for p<0.001
significance levels).

DISCUSSION
There are statistically significant differences between course A and course XA in
relation to the self-efficacy and self-regulation scales. The biggest differences are in
the self-efficacy and lack of regulation factors. The results show that an average
student has higher self-efficacy levels in course XA than in course A, and that an
average student lacks regulatory skills more often in course A than in course XA.
This is supported by the effect sizes implying a moderate role for the course contexts
when measuring self-efficacy and lack of regulation. The results bear significance as
self-efficacy has a strong positive and lack of regulation a strong negative relation to
academic performance (Pajares, 1996; Peters, 2013; Vermunt, 2005).
Gender has a statistically significant interaction with the factor measuring self-
efficacy. On average, female students report lower self-efficacy levels in course A
compared to male students. In contrast, the self-efficacy levels are very similar for
male and female students in course XA. In practice, an average female student is less
confident in her abilities to succeed in the course A context when compared to an
average male student. However, it seems that the change in course context
diminishes the difference as there is no statistically significant difference present
between genders in the self-efficacy factor in course XA context.
Female students report more external regulation compared to male students in both
course contexts. The results are statistically significant only in course XA context,
but the p-value is very close to the 0.05-significance level also in course A context.
This means that the lecturers instructions have more influence on female students’
learning processes than on male students’ learning processes. This is supported by
prior research (Vermunt, 2005), although the current study does give any
explanations to this phenomenon. However, in Vermunt’s (2005) study there was no
consistent interaction between students’ gender and their learning patterns, and
external regulation did not relate negatively to academic achievement. Further
research is needed to understand the motivations behind this phenomenon, as well as
its implications to instructional practices.
One of the major limitations of this study is that the two courses differ in content.
The limitation is caused by the choice of research design as it was not possible to
attain the same cohort of students in two different pedagogical settings with the same
course content. However, the different course contents do not fully provide an
explanation for the result that male and female students’ ability to regulate and
reflect on their learning processes is dependent on the course context. One can argue
that the gender differences are not caused by the characteristics of the mathematics
studied but by the characteristics of the learning environment used to study the

353 [Link]:indrum2018:174796
mathematics. The results of this study may also be affected by the fact that some
students are more capable to adopt themselves into new instructional designs. As
argued by Kogan and Laursen (2014), student-centred course settings often feature
collaborative work, problem-solving and communication, aspects known to be
effective for female students. Also, Vermunt (2005) states that female students like
cooperative learning more compared to male students. In addition, students who have
high confidence in collaboration seek more likely help from other students; these
help-seeking students then perform better in a flipped mathematics classroom
compared to students seeking less help (Sun, Xie, & Anderman, 2018).
Despite of the limitations, the findings of this study are supported by prior research.
To conclude, the results propose that student-centred course designs support
students’ self-efficacy and self-regulation, especially in female students. More
thorough analysis should be completed to understand the mechanisms and
motivations behind the differences in these two course contexts and to draw more
general conclusions regarding instructional designs used in teaching university
mathematics.

REFERENCES
Bandura, A. (1994). Self-efficacy. In V. S. Ramachaudran (Ed.), Encyclopedia of
human behavior (Vol. 4, pp. 71-81). New York: Academic Press.
Coertjens, L., Donche, V., De Maeyer, S., Vanthournout, G., & Van Petegem, P.
(2013). Modeling change in learning stratgies throughout higher education: A
multi-indicator latent growth perspective. PloS One, 8(7).
Collins, A., Brown, J., & Holum, A. (1991). Cognitive apprenticeship: Making
thinking visible. American Educator, 15(3), 6-46.
Greeno, J. G. (1997). On Claims that Answer the Wrong Questions. Educational
Researcher, 26(1), 5-17.
Heikkilä, A., & Lonka, K. (2006). Studying in higher education: students' approaches
to learning, self‐regulation, and cognitive strategies. Studies in Higher Education,
31(1), 99-117.
Kogan, M, & Laursen S. L. (2014). Assessing long-term effects if inquiry-based
learning: a case study from college mathematics. Innovative Higher Education,
39(3), 183-199.
Lave, J. & Wenger, E. (1991). Situated learning: Legitimate peripheral
participation. Cambridge: Cambridge University Press.
Oikkonen, J. (2009). Ideas and results in teaching beginning maths students.
International Journal of Mathematical Education in Science and Technology,
40(1), 127-138.

354 [Link]:indrum2018:174796
Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational
Research, 66(4), 543-578.
Parpala, A., & Lindblom-Ylämnne, S. (2012). Using a research instrument for
developing quality at the university. Quality in Higher Education, 18 (3), 313-328.
Peters, M. L. (2013). Examining the relationships among classroom climate, self-
efficacy, and achievement in undergraduate mathematics: a multilevel analysis.
International Journal of Science and Mathematics Education, 11(2), 459-480.
Pintrich, P. R. (1991). A manual for the use of the Motivated Strategies for Learning
Questionnaire (MSLQ).
Pintrich, P. R. (2004). A conceptual framework for assessing motivation and self-
regulated learning in college students. Educational Psychology Review, 16(4),
385–407.
Raelin, J. A., Bailey, M. B., Hamann, J., Pendleton, L. K., Reisberg, R., & Whitman,
D. L. (2014). The gendered effect of cooperative education, contextual support,
and self-efficacy on undergraduate retention. Journal of Engineering Education,
103, 599–624.
Rämö, J., Oinonen, L., & Vikberg, T. (2015). Extreme Apprenticeship –
Emphasising Conceptual Understanding in Undergraduate Mathematics. In K.
Krainer, & N. Vondrová (Eds.), Proceedings of the Ninth Congress of the
European Society for Research in Mathematics Education (CERME9, 4-8
February 2015) (pp. 2242–2248). Prague, Czech Republic: Charles University in
Prague, Faculty of Education and ERME.
Sun, Z., Xie, K., & Anderman, L. H. (2018). The role of self-regulated learning in
students' success in flipped undergraduate math courses. The Internet and Higher
Education, 36, 41-53.
Tall, D. (2014). How humans learn to think mathematically: exploring the three
worlds of mathematics. Cambridge: Cambridge University Press.
Vermunt, J. D. (2005). Relations between student learning patterns and personal and
contextual factors and academic performance. Higher Education, 49, 205-234.
Vermunt, J. D. (1994). Inventory of Learning Styles in higher education: Scoring
key. Tilburg University, Department of Educational Psychology.
Volet, S., Vauras, M., & Salonen, P. (2009). Self- and social regulation in learning
contexts: An integrative perspective. Educational Psychologist, 44(4), 215–226.
Zimmerman, B. J. (2000). Attaining self-regulation: A social cognitive perspective.
In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation
(pp. 13–39). San Diego, CA: Academic Press.

355 [Link]:indrum2018:174796
 Situational interest in university mathematics courses:
similar for real-world problems, calculations, and proofs?
Stefanie Rach1, Stefan Ufer², and Timo Kosiol²
1
University of Paderborn, Institute of Mathematics, Germany, rach@[Link]-
[Link]; ²LMU Munich, Department of Mathematics, Germany
In the educational context where this study was conducted, the transition from school
to university is associated with changes of the learning domain mathematics. In high
school, solving real-world problems and performing calculations are dominating
practices, whereas in university, the main focus is on proving. Successful learning
processes are associated with appropriate prerequisites, e. g. interest. The question is
which component of interest concerning which practice is important for successful
processes. We developed an instrument to differentiate these facets of students’ situa-
tional interest. A study with 339 first-semester students in mathematics partially con-
firms the expected factorial structure of this instrument. Precise information concern-
ing learners’ interest may help us to support students at this challenging transition.
Key words: Students’ interest; Feeling- resp. value-related component; Mathematical
practices; Transition to and across university mathematics; Teachers’ and students’
practices at university level.

INTRODUCTION
The transition from school into a university mathematics programme is a challenging
phase for many students. Reports mentioning high dropout rates in academic study
programmes with a focus on mathematics (OECD, 2010) illustrate this fact. Reasons
for students’ problems in the first year of university study are primarily attributed to
the changes of the learning domain: while the school subject mathematics strongly
relies on performing calculations and solving real-world problems, mathematics in
university is presented as a scientific discipline with a focus on proving (Rach,
Heinze, & Ufer, 2014). It is yet not clear which of the individual prerequisites, that
students bring from school, support successful learning processes in university
courses. Many researchers assume that subject-specific interest, in the sense of a
person-object or a person-situation relationship, is an important learning prerequisite
in general (Wigfield & Cambria, 2010). Common instruments to survey interest in
mathematics are questionnaires. Mostly, these use the word “mathematics” to
describe the objects of interest. At the transition from school to university, however it
is not clear if students refer to school or university mathematics when reporting their
interest in this questionnaires. To obtain a more differentiated insight into students’
interest with a stronger relation to concrete learning situations, we develop a task-
based instrument to measure students’ interest concerning different mathematical
practices. In this contribution, we present the conceptualisations of this instrument in
detail and report results of an empirical study with 339 first-semester students that
investigates the factorial structure of the instrument.

356 [Link]:indrum2018:173528
THEORETICAL BACKGROUND
In the first section, we give an overview about the motivational construct “interest”
by addressing common definitions and results concerning its role in learning
processes. In the second section, we summarize ideas to the objects of interest,
mathematics.
Interest in mathematics learning processes
Researchers agree that interest is an important motivational variable. Its role as a
learning prerequisite, a learning processes measure, and as a learning outcome has
been put forward repeatedly (Rotgans & Schmidt, 2017; Sonnert & Sadler, 2015;
Wigfield & Cambria, 2010). Interest is defined as a specific relationship between a
person (here: a student) and an object or a situation (here: mathematics resp.
mathematical practices) (Krapp, 2002; Schiefele, 2009). The objects of interests may
be concrete objects, topics, or school subjects. Whereas Schiefele, Krapp, and
Winteler (1992) conceptualize interest mainly as a relatively stable trait: “Individual
interest is conceptualized as a relatively stable affective-evaluative orientation toward
certain subject areas or objects” (Schiefele, 2009, pp. 198), Linnenbrink-Garcia et al.
(2010, 2013), in contrast, see interest as specific to a situation: “Situational interest is
a temporary state aroused by specific features of a situation, task, or object (e. g.,
vividness of a text passage)” (Schiefele, 2009, pp. 197–198). Situational interest is
limited at a point of time, e. g., in a certain learning situation and may or may not
develop into individual interest (Hidi & Renninger, 2006). Schiefele (2009)
distinguishes between two components of interest: (1) a feeling-related component,
related to fun or other positively experienced emotions, and (2) a value-related
component, related to a high importance of the objects of interest for oneself (adapted
from Linnenbrink-Garcia et al., 2010, 2013). Linnenbrink-Garcia and colleagues
(2010, 2013) could confirm in empirical studies the two-factor structure of
(maintained) situational interest and separate the two components from each other.
The separation of interest in a feeling-related and a value-related component is also in
line with expectancy-value models that distinguish between an intrinsic and a utility
value (Eccles & Wigfield, 2002).
Several researchers assume that interest can trigger successful learning processes:
interest can lead to more engagement and to an enhanced use of deeper learning
strategies which then result in better learning achievement (Hidi & Renninger, 2006).
However, empirical studies could not confirm this assumption for learning processes
in undergraduate mathematics courses (Rach & Heinze, 2017). The conceptualisation
and operationalisation of interest in mathematics may be one reason for these
conflicting findings: firstly, the interest object “mathematics” changes its character at
the transition from school to university. So the objects of reported interest may differ
from the learning object in the first semester (see next section). Secondly, many of
the used instruments measure the individual interest rather than the situational
interest, although the situational interest is a more proximal variable to the specific
learning situation.

357 [Link]:indrum2018:173528
Mathematical practices at the transition from school to university
In the literature (Gueudet, 2008; Engelbrecht, 2010), changes in two relevant aspects
of the learning environment at this transition are described: a shift in the character of
the learning domain, mathematics, and a change in the learning opportunities and
their use. As interest is a person-object relationship resp. a person-situation
relationship, the change of the learning domain is important for the conceptualisation
of interest. The specific differences of mathematics at school and at university might
vary between countries due to traditions concerning learning goals etc. The
subsequent presentation refers to the situation in Germany which was analysed in
empirical studies (Rach et al., 2014) and which is relevant for the empirical study we
present below. Nevertheless, several of the described features might hold for other
countries as well. One central goal of teaching and learning at school is to apply
mathematics for solving real-world problems (CCSSI, 2010; OECD, 2016). Thus,
describing realistic situations mathematically, performing computations, and applying
mathematical procedures are central. On the contrary, university mathematics is
usually taught as a scientific discipline based on formal definitions of concepts and
formal-deductive proofs (Gueudet, 2008; Nardi, 1996). Mathematics at university is
often presented in a DTP (Definition-Theorem-Proof) structure to emphasise its
logical rigidity (Engelbrecht, 2010; Weber, 2015). These conclusions are mainly
based on observations of lectures (Weber, 2004) or of tutorials (Nardi, 1996),
interviews with involved parties (Nardi, 2008), or analyses of tasks (Gueudet, 2008).
In general, mathematical tasks are often used to describe and examine learning
situations and their cognitive demands (e. g., Stein & Lane, 1996).
Theoretical analyses and anecdotal evidence from students and teaching staff have
described a change of the learning domain (e.g., Engelbrecht, 2010; Gueudet, 2008;
Thomas & Klymchuk, 2012). However, empirical studies supporting the role of this
shift and its effects on student learning are scarce. In particular, reliable evidence on
the role of students’ motivational variables, especially of students’ interest, is scarce.
The SISMa project
The goal of the SISMa project (“Self-concept and Interest when Studying
Mathematics”) is to contribute evidence concerning students’ interest and self-
concept with regard to certain mathematical practices (Ufer, Rach & Kosiol, 2017).
As a first step of the project, we developed interest and self-concept measures that
focus central mathematical practices. Using these measures, the development of these
variables during students’ learning processes and their effect on student learning in
undergraduate mathematics programmes are investigated.
In this contribution, we focus on the measurement of situational interest. We
developed a questionnaire of situational interest that is based on the conceptualisation
of interest as a person-object (situation) relationship, and on the assumption that
interest concerning a specific task is an indicator of situational interest. Here, we
follow the idea of Schukajlow and colleagues (2012) who operationalize the two

358 [Link]:indrum2018:173528
components of situational interest with the constructs enjoyment and value (see also
Schukajlow & Krug, 2014). In this questionnaire, mathematics tasks are presented to
students and they are asked to imagine solving this task (not to solve the task), and
then to state their anticipated enjoyment and value appraisals when working on the
task. For our questionnaire, we designed mathematical tasks concerning the topic
derivatives that each prototypically represent one of the three practices “solving real-
world problems” (resp. applying mathematics, sample item: “Using metal, you should
produce a cylindrical can with a prescribed volume. For which radius is the material
consumption minimal?”), “performing complex calculations” (sample item: “Let f be
. Calculate the extrema of the function f.”), “and “proving”
(sample item: “Let be a differentiable function. Show that f is
continuous.”). After a pilot study (see Rach et al., 2014), we used 12 tasks. While the
first and second practice are considered central in mathematical lessons at school, the
last practice predominates in university courses. After reading each task, students rate
their agreement to one statement concerning the feeling-related component (item: “It
would be fun to me to work on this task.”) and one statement concerning the value-
related component of interest (item: “Even if the task is not part of an exam, it is
important to be able to solve the task.”). So in sum, situational interest is divided into
six subscales – one for each component and each mathematical practice (see table 1).
Solving real-world Performing complex Proving (4 tasks)
problems, applying calculations
mathematics (4 tasks) (4 tasks)
Feeling related Feel Apply Feel Calc Feel Proof
Value related Value Apply Value Calc Value Proof

Table 1: Two components of situational interest and three practices.

RESEARCH QUESTIONS
The aim of this contribution is to investigate the structure of students’ situational
interest as measured with our instrument. To achieve this, we applied the instrument
to students from a first semester mathematics course. The questions focused in this
contribution address students’ reported levels of interest and the empirical structure
of the questionnaire:
1. What level of feeling- and value-related situational interest do students from a
first semester mathematics course report concerning the three practices?
Since prior research has shown similar trends, we expected that students
would report lower interest levels for each component and practice after six
weeks, compared to the first day of their studies. One reason might be that high
demands in the first semester lead to a decrease in interest.
2. Is the theoretical structure of subscales, that guided their development,
reflected in the factorial structure of the newly developed instruments?

359 [Link]:indrum2018:173528
 We expected that the two components of situational interest, the feeling- and
the value-related component, are reflected in the empirical data. Moreover, we
expected that subscales concerning the three different practices can be
separated from each other.

METHOD
Design and sample
We present data from two measurements, at the beginning of the first semester (T1)
and in the middle of the first semester (T2). Our sample consists of 339 mathematics
students (162 female) of one German university of the course “Analysis I” which is
compulsory for first-semester mathematics students. In this course, mathematics is
presented as a scientific discipline with a strong focus on formal concept definitions
and deductive proofs. The students were enrolled in the bachelor’s programmes
“mathematics” (n = 90), “business mathematics” (n = 91), or a mathematics teacher
education programme for the highest attaining secondary school track in Germany
(n = 104) – for the remaining students, we have no information about their study
programme. The participation in the study was voluntary.
Instruments
The mathematical tasks were arranged in a fixed, random order. The questionnaire
was submitted to the students with the following instructions: “Imagine how you
solve these tasks. Do not solve the tasks, but report your agreement to the following
statements”. Students rated each statement on a four-point likert scale from
0 (disagree) to 3 (agree). The individual mean value of a single student on a scale was
computed if this student had answered at least half of the items of the scale.

RESULTS
Table 2 shows the means, standard deviations, and Cronbach’s Alpha of the six scales
at the first and second measurement point.
Beginning of the first semester (T1) Middle of the first semester (T2)
N M (SD) α N M (SD) α
Feel Apply 323 2.22 (0.61) .73 226 2.02 (0.67) .76
Feel Calc 323 2.16 (0.71) .79 230 2.01 (0.74) .80
Feel Proof 331 2.18 (0.61) .77 237 2.00 (0.61) .82
Value Apply 323 2.25 (0.64) .82 224 2.04 (0.71) .82
Value Calc 325 2.26 (0.70) .86 232 2.00 (0.76) .86
Value Proof 331 2.31 (0.61) .88 237 2.06 (0.67) .86
Table 2: Means, standard deviations, and Cronbach’s Alpha of the interest scales.
Likert scale from 0 (disagree) to 3 (agree).

360 [Link]:indrum2018:173528
Concerning all 24 items (3 practices, 4 tasks, 2 components), there are no floor or
ceiling effects. Reliability analyses underpin the internal consistency of the six
subscales. For T1 (N = 323–331), Cronbach’s alpha of each scale ranges from .73 to
.88, for T2 (N = 224–237), from .76 to .86. Indeed, students reported higher value and
feeling ratings concerning all practices in the first as compared to the second
measurement (cf. table 2; t(194–197) = 5.42–6.83; p < .001; d = 0.39–0.49; measured
with students who were present at both measurements). Levels of value- and feeling-
related interest were relatively similar between the three practices.
Correlation analyses (see table 3) show that, as expected, the feeling-related and
value-related scales concerning each practice strongly relate to each other (Solving
real-world problems: T1: r = .55, T2: r = .65, Performing complex calculations: T1:
r = .49, T2: r = .53, Proving: T1: r = .53, T2: r = .60). Moreover, the three feeling-
related scales (r = .43–.60) resp. value-related scales (r = .59–.79) correlate strongly.
The correlations of feeling-related scales with value-related scales for different
practices are moderate (T1: r = .28–.44, T2: r = .31–.39).
Feel Feel Feel Value Value Value
Apply Calc Proof Apply Calc Proof
Feel Apply .47 .43 .55 .31 .31
Feel Calc .47 .60 .28 .49 .32
Feel Proof .44 .54 .38 .44 .53
Value Apply .65 .31 .31 .66 .68
Value Calc .30 .53 .39 .59 .79
Value Proof .37 .37 .60 .66 .79
Table 3: Correlations between the interest scales. Over the diagonal T1 (N = 334–339),
under the diagonal T2 (N = 239–243). All correlations significant with p < .01.
As some of the situational interest scales correlate strongly, we investigated whether
our expected scale structure would be replicated by exploratory factor analyses. The
results of Principal Component Analysis with Varimax rotation for every
measurement point partially support our expected structure of our scales. For T1,
results indicate four factors that explain nearly 63% of the variance. Table 4 shows
the factor loadings of the items on the four identified factors.
As expected, the feeling-related items concerning each of the three practices load
strongly on one of the factors two to four each. Contrary to the theoretical
construction of our questionnaire, the first factor includes the value items for all three
practices. Some value-related items have cross-loadings on the feeling-related scales
for the same practice. For T2, the factor analysis shows similar results, with the four
factors explaining 65% of the variance. Thus, it also seems to be possible to combine
all value items into one value scale. This value-related scale with twelve items has an
excellent reliability of α = .93 (N = 317, T1) resp. α = .92 (N = 223, T2).

361 [Link]:indrum2018:173528
 Factor 1 Factor 2 Factor 3 Factor 4
Feel Apply 1 .35 .36 .42
Feel Apply 2 .31 .72
Feel Apply 3 .44 .64
Feel Apply 4 .81
Feel Calc 1 .50 .41
Feel Calc 2 .38 .67
Feel Calc 3 .78
Feel Calc 4 .35 .74
Feel Proof 1 .71
Feel Proof 2 .62
Feel Proof 3 .65
Feel Proof 4 .70
Value Apply 1 .70 .44
Value Apply 2 .61 .50
Value Apply 3 .60 .56
Value Apply 4 .55
Value Calc 1 .75
Value Calc 2 .75 .32
Value Calc 3 .69 .46
Value Calc 4 .76 .35
Value Proof 1 .73 .35
Value Proof 2 .82
Value Proof 3 .71 .32
Value Proof 4 .76 .35

Table 4: Factor loading on the four extracted factors (T1, N = 313). Loadings under .30
not shown. The expected loadings on a factor are bolded.

DISCUSSION
Students’ interest is an important variable in successful learning processes, including
the transition to university mathematics. Based on prior research, we argue that to
measure students’ interest validly, one requires situated, e.g. task-related, measures,
to cover feeling- and value-related components of interest and to take into account the

362 [Link]:indrum2018:173528
changing role of different mathematical practices during this transition (Rach &
Heinze, 2017). Contrary to the findings of Schukajlow et al. (2012), the results of our
empirical study show that the relations between the feeling-related components for
different practices are relatively low as compared to the respective correlations
between the value-related components. This partially supports the expected six-factor
model (c. f. Schiefele, 2009), and underpins the different roles the three practices play
during the transition in the educational context in Germany. Beyond this, factor
analyses indicate that the value-appraisals seem to be fairly consistent over all three
practices and the exploratory factor analyses also allow a model assuming only one
value factor over all practices. This observation applies to data from the first day of
university study, but also to data collected after six weeks into the semester. After six
weeks, students report across all scales lower approvals. Interestingly, even though
students seem to differentiate their interest reports by practices, we find almost no
differences in their mean levels of interest between the different interest scales.
Of course, the results of our study rely on students’ self-reports about their
anticipated situational interest in a set of specific tasks. Even though the differences
between school and university mathematics have been described internationally, it
might be interesting to replicate the studies in other educational systems. In future
studies, the relation between the anticipated situational interest and the actual
engagement in learning situations should be considered. However, Schukajlow and
Krug (2014) found only slight differences in prospective and retrospective ratings of
interest in working with mathematical tasks. In sum, the newly developed instrument
may provide more differentiated insights into students’ interest concerning different
mathematical practices. Further research should investigate which facets of
situational interest indeed go along with learning gain in the study entry phase. In
particular, it is an open question if interest in practices that are typical for university
mathematics, such as proving, are more important for learning gain than other interest
facets. In the future, this instrument may help to evaluate support courses in the study
entrance phase, to analyse the development of students’ interest (c. f. Hidi &
Renninger, 2006), and to investigate the impact of different interest facets on study
success.

REFERENCES
CCSSI (2010). Common core state standards for mathematics. Retrieved from
[Link]
Eccles, J. S., & Wigfield, A. (2002). Motivational beliefs, values, and goals. Annual
Review of Psychology, 53, 109–132.
Engelbrecht, J. (2010). Adding structure to the transition process to advanced
mathematical activity. International Journal of Mathematical Education in Science
and Technology, 41, 143–154.

363 [Link]:indrum2018:173528
Gueudet, G. (2008). Investigating the secondary-tertiary transition. Educational
Studies in Mathematics, 67, 237–254.
Hidi, S., & Renninger, K. A. (2006). The Four-Phase Model of Interest Development.
Educational Psychologist, 41, 111–127.
Krapp, A. (2002). Structural and dynamic aspects of interest development: theoretical
considerations from an ontogenetic perspective. Learning and Instruction, 12,
383–409.
Linnenbrink-Garcia, L., Durik, A. M., Conley, A. M., Barron, K. E., Tauer, J. M.,
Karabenick, S. A., & Harackiewicz, J. M. (2010). Measuring Situational Interest in
Academic Domains. Educational and Psychological Measurement, 70, 647–671.
Linnenbrink-Garcia, L., Patall, E. A., & Messersmith, E. E. (2013). Antecedents and
consequences of situational interest. British Journal of Educational Psychology,
83, 591–614.
Nardi (2008). Amongst Mathematicians: Teaching and Learning Mathematics at
University Level. New York: Springer.
Nardi, E. (1996). The novice mathematician’s encounter with mathematical
abstraction: Tensions in concept-image construction and formalisation. (Doctoral
dissertation). Retrieved from [Link]
4ed4-ab98-3be18da31e16/datastreams/THESIS04.
OECD (2016). PISA 2015 Mathematics Framework. In PISA 2015 Assessment and
analytical framework: Science, reading, mathematics and financial literacy
(pp. 63–78). Paris: OECD Publishing.
OECD (2010). How many students drop out of tertiary education? In OECD (Ed.),
Highlights from Education at a Glance 2010 (pp. 22–23). Paris: OECD Publishing.
Rach, S., & Heinze, A. (2017). The Transition from School to University in
Mathematics: Which Influence Do School-Related Variables Have? International
Journal of Science and Mathematics Education, 15, 1343–1363.
Rach, S., Heinze, A., & Ufer, S. (2014). Welche mathematischen Anforderungen
erwarten Studierende im ersten Semester des Mathematikstudiums? [Which
Mathematical Demands do Students Expect in the First Semester of Their Study?]
Journal für Mathematik-Didaktik, 35, 205–228.
Rotgans, J. I., & Schmidt, H. G. (2017). The Role of Interest in Learning: Knowledge
Acquisition at the Intersection of Situational and Individual Interest. In P. A.
O'Keefe & J. M. Harackiewicz (Eds.), The Science of Interest (pp. 69–96). Cham:
Springer International Publishing.
Schiefele, U. (2009). Situational and individual interest. In K. Wentzel & A. Wigfield
(Eds.), Handbook of motivation at school (pp. 197–222). New York: Routledge.

364 [Link]:indrum2018:173528
Schiefele, U., Krapp, A., & Winteler, A. (1992). Interest as a predictor of academic
achievement: A meta-analysis of research. In K. A. Renninger, S. Hidi, &
A. Krapp (Eds.), The role of interest in learning and development (pp. 183–212).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Schukajlow, S., & Krug, A. (2014). Are Interest and Enjoyment important for
students' performance? In C. Nicol, S. Oesterle, P. Liljedahl, & D. Allan (Eds.),
Proceedings of the Joint Meeting of PME 38 and PME-NA 36, (Vol. 5, pp. 129–
136). Vancouver, Canada: PME.
Schukajlow, S., Leiss, D., Pekrun, R., Blum, W., Müller, M., & Messner, R. (2012).
Teaching methods for modelling problems and students’ task-specific enjoyment,
value, interest and self-efficacy expectations. Educational Studies in Mathematics,
79, 215–237.
Sonnert, G., & Sadler, P. (2015). The Impact of Instructor and Institutional Factors on
Students’ Attitudes. In D. Bressoud, V. Mesa, & C. L. Rasmussen (Eds.), Insights
and Recommendations from the MAA National Study of College Calculus (pp. 17–
29). MAA Press.
Stein, M. K., & Lane, S. (1996). Instructional Tasks and the Development of Student
Capacity to Think and Reason: An Analysis of the Relationship between Teaching
and Learning in a Reform Mathematics Project. Educational Research and
Evaluation, 2, 50–80.
Thomas, M. O. J., & Klymchuk, S. (2012). The school-tertiary interface in
mathematics: teaching style and assessment practice. Mathematics Education
Research Journal, 24, 283–300.
Ufer, S., Rach, S., Kosiol, T. (2017). Interest in mathematics = Interest in
mathematics? What general measures of interest reflect when the object of interest
changes. ZDM Mathematics Education, 49, 397–409.
Weber, K. (2015). Effective Proof Reading Strategies for Comprehending
Mathematical Proofs. International Journal of Research in Undergraduate
Mathematics Education, 1, 289–314.
Weber, K. (2004). Traditional instruction in advanced mathematics courses: a case
study of one professor’s lectures and proofs in an introductory real analysis course.
Journal of Mathematical Behavior, 23, 115–133.
Wigfield, A., & Cambria, J. (2010). Students’ achievement values, goal orientations,
and interest: Definitions, development, and relations to achievement outcomes.
Developmental Review, 30, 1–35.

365 [Link]:indrum2018:173528
Determining your own grade – Student perceptions on self-assessment
in a large mathematics course
Jokke Häsä, Juuso Henrik Nieminen, Johanna Rämö, Henri Pesonen, and Matti Pauna
University of Helsinki, [Link]@[Link]
In our poster, we report on students’ perceptions on self-assessment in large undergraduate
course context taught with the DISA model (Digital Self-Assessment) in which the final exam
is replaced by self-assessment. The students found that practising self-assessment skills was
valuable and encouraged them to study for themselves. At the same time, self-assessment
was considered unfamiliar and therefore difficult. We argue that the DISA model
encouraged students to take more responsibility on their own learning.
Keywords: Assessment practices in university mathematics education, Novel approaches to
teaching, Self-assessment, Student perceptions, Ownership of learning.

INTRODUCTION AND THEORETICAL BACKGROUND

Self-assessment skills are often mentioned as an important ingredient in becoming an
effective life-long learner (Boud, 2000). Good self-assessment skills are vital also
during studies, as they form a key part of self-regulation (Zimmerman, 2002). In
university mathematics, self-assessment practices have been shown to promote
learner autonomy and mathematical communication skills (Stallings & Tascione,
1996). In a study for engineering students, self-assessment was linked to better time
management and more effective learning (Friess and Davis, 2016). In earlier studies,
the use of self-assessment is usually limited to low-stakes training exercises or
controlled self-grading of the final exam. We, however, do not want merely to
introduce a reflection tool, but instead question thoroughly the traditional concept of
external assessment in mathematics. The aim of our study is to develop the theory of
self-assessment in the specific field of mathematics.
In the DISA model, the traditional end exam is removed, and final course grades are
awarded by students' self-assessment based on a learning objectives matrix. We use
Yan and Brown’s (2017) model of cyclical self-assessment process as our theoretical
framework. In the DISA model, the development of self-assessment skills is
supported with various feedback methods throughout the course. The students’
reflective skills are formatively practised with several self- and peer-assessment
exercises. In an earlier study (Nieminen, Rämö, Häsä, & Tuohilampi, 2017), we
found that mathematics students connected the self-assessment training and the lack
of exam with deep learning approach. The purpose of this poster is to present
students’ perceptions on self-assessment in the DISA model.

DATA COLLECTION AND ANALYSIS

The participants of the study were students of a linear algebra course taught with the
DISA model in a Finnish university in autumn 2017. Data was collected after the

366 [Link]:indrum2018:174829
course with a questionnaire with open ended questions (n=113). A qualitative
content analysis was conducted to study the student perceptions on self-assessment.

RESULTS
Data analysis of qualitative survey data resulted in two main categories. One
category was “Ownership of learning”, containing a theme of positive perceptions on
the self-assessment method and another theme concerning developing one’s own
reflection skills. The other category was “New kind of learning culture”. Students’
comments in this category were either “micro level“ comments concerning the course
arrangements or “macro level” comments concerning self-assessment as part of a
new kind of learning culture. Our results indicate that the theory of cyclical self-
assessment can be applied to the field of mathematics. Furthermore, we argue that
digital learning environments can be used in large mathematics courses to practice
self-assessment skills.

REFERENCES
Boud, D. (2000). Sustainable assessment: rethinking assessment for the learning society. Studies in
continuing education, 22(2), 151-167.
Falchikov, N. (2013). Improving assessment through student involvement: Practical solutions for
aiding learning in higher and further education. Routledge.
Friess, W. A., & Davis, M. P. (2016). Formative Homework Assessment Strategies to Promote
Student Self-Reflection and Improve Time Management: A Pilot Study. In Proceedings of the
ASEE NE 2016 Conference, Rhode Island.
Nieminen, J. H., Häsä, J., Rämö, J., & Tuohilampi, L. (2017). Replacing Exam with Self-
Assessment: Reflection-Centred Learning Environment as a Tool to Promote Deep Learning.
20th Meeting of the MAA Special Interest Group on Research in Undergraduate Mathematics
Education. San Diego: RUME.
Stallings, V., & Tascione, C. (1996). Student self-assessment and self-evaluation. The Mathematics
Teacher, 89(7), 548-554.
Yan, Z., & Brown, G. T. (2017). A cyclical self-assessment process: towards a model of how
students engage in self-assessment. Assessment & Evaluation in Higher Education, 42(8), 1247-
1262.
Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory into practice,
41(2), 64-70.

367 [Link]:indrum2018:174829
 Seminar groups as part of first-semester mathematics teaching:
What did the students learn?
Kerstin Pettersson1 and Niclas Larson2
1
Stockholm University, Sweden, [Link]@[Link];
2
University of Agder, Kristiansand, Norway.

In this study, the learning encouraged by teaching activities in a small-group setting

was investigated through the analysis of students’ responses to survey and interview
questions. The results indicate that the students perceived an increased ability to
communicate mathematics in written form, but to a lesser extent developed their ability
to discuss mathematics and build conceptual understanding.
Keywords: Teachers’ and students’ practices at university level, teaching and learning
of specific topics in university mathematics, communication, small-group teaching.
Empirical research on mathematics teaching at tertiary level is an area of sustained
growth and interest (Biza, Giraldo, Hochmuth, Khakbaz, & Rasmussen, 2016). Besides
studies on the discourse of mathematics teaching in lectures (e.g. Viirman, 2015), there
are studies focusing on small-group tutorial sessions (e.g. Mali, Biza, & Jaworski,
2014). Inquiry-based small-group activities were found to have reasonable effect in a
project aiming to improve the conceptual understanding of engineering students
(Jaworski, Robinson, Matthews, & Croft, 2012). However, there is still a need for
research on the use of small-group teaching as a complement to lectures across
different national and institutional practices. In the study presented here, we analysed
students’ experiences of learning from teaching activities in a small-group setting at a
large Swedish university. Our research question is: How do the students perceive their
mathematical abilities to have developed through the small-group seminars?
In the first-semester mathematics course, besides whole-class lectures, students once a
week took part in so-called seminars with 10–15 students in each group. The stated aim
of the seminars was “to learn to discuss mathematics, and to present mathematics orally
and in written form”, but also to “support the learning of algebra and calculus”. Each
seminar group was led by a lecturer. Before each seminar, students handed in a written
solution to one or two tasks. During the seminars, they presented the solutions orally
and received comments from the seminar leader on both their oral and their written
presentations. Furthermore, the students discussed conceptual tasks given by the
seminar leader or discussed questions posed by the students themselves.
A survey was conducted in a lecture after 10 of the 13 seminars had been held. Of the
49 students present, 42 completed the questionnaire. Students were asked to indicate
on a Likert scale the extent to which the seminars had facilitated their learning of
presenting and discussing mathematics, solving problems, and understanding concepts.
Seven students volunteered for individual interviews undertaken three weeks after the
survey. The focus was on students’ experiences from the seminars and questions were

368 [Link]:indrum2018:174463
posed about their learning. In the analysis, we searched for utterances where students’
perceived learning was pronounced. These utterances were then grouped together in
categories due to the abilities mentioned. This was a deductive analysis drawing on the
notion of competencies (e.g. Niss & Jensen, 2002). For each category we then
interpreted how the students perceived that their mathematical abilities had developed.
The analysis of data is ongoing. However, the preliminary analyses indicate that
students perceived that they developed their ability to communicate mathematics in
written form, but to a lesser extent developed their ability to solve problems and discuss
mathematics. In the survey, students marked high on the Likert scale for learning to
make written presentations (mean 3.5, scale 1–4 with 4 as ‘a lot’), a little lower on
solving problems (3.1), understanding concepts (3.0), oral presentations (2.9) and
discussions (2.9).
The categories revealed by the interviews were: understanding of concepts, communi-
cation, solving problems, procedural knowledge, and reasoning. Analyses of the inter-
views confirmed the results from the survey; the perceived learning from the seminars
largely concerned the written presentations. While some students talked about their
aim to build conceptual understanding, they mentioned also how the focus on details
and technical aspects in the written communication distracted them from this aim, a
finding in line with the distracting role of aspects such as examination forms found by
Jaworski et al. (2012). In the interviews, we also met descriptions of different ways in
which seminar leaders had chosen to use representations and problem solving (cf. Mali
et al., 2014). Further data collection is planned to answer questions on the role of these
characteristics and the seminar leaders’ mathematical discourse (cf. Viirman, 2015).

REFERENCES
Biza, I., Giraldo, V., Hochmuth, R., Khakbaz, A., & Rasmussen, C. (2016). Research
on teaching and learning mathematics at the tertiary level: State-of-the-art and
looking ahead. Springer Open.
Jaworski, B., Robinson, C., Matthews, J., & Croft, A. C. (2012). An activity theory
analysis of teaching goals versus student epistemological positions.
International Journal for Technology in Mathematics Education, 19, 147–152.
Mali, A., Biza, I., & Jaworski, B. (2014). Characteristics of university mathematics
teaching: Use of generic examples in tutoring. In P. Liljedahl, S. Oesterle, C.
Nicol, & D. Allan (Eds.), Proceedings of the joint meeting of PME 38 and PME-
NA 36 (Vol. 4, pp. 161–168). Vancouver, Canada: PME.
Niss, M., & Jensen, T. H. (Eds.). (2002). Kompetencer og matematiklæring: Ideer og
inspiration til udvikling af matematikundervisning i Danmark (Vol. 18).
Copenhagen: Undervisningsministeriet.
Viirman, O. (2015). Explanation, motivation and question posing routines in university
mathematics teachers’ pedagogical discourse: A commognitive analysis.
International Journal of Mathematical Education in Science and Technology,
46, 1165–1181.

369 [Link]:indrum2018:174463
 Students' usage patterns of video recorded lectures in an
undergraduate mathematics course
M. Wold
Department of Engineering Sciences, University of Agder, Norway,
[Link]@[Link]
It is customary for some universities to offer recordings of live lectures to
students. Whether this improves learning and academic performance is debated
in the literature. As with most technology, correct usage can lead to increased
academic performance, but there are also usage patterns that can be considered
counter productive, especially for learning mathematics. In order to investigate
patterns in students' usage of such online recorded live lectures, we analyze log
files from a server holding the recordings of an undergraduate mathematics
course. This poster presents results from the statistical analysis and discusses
some of the usage patterns found in light of Moore's theory of transactional
distance.
Keywords: teachers' and students' practices at university level, the role of
digital and other resources in university mathematics education, video recorded
lectures, streaming media.
INTRODUCTION AND THEORETICAL FRAMEWORK
It is becoming customary for some universities to stream lectures and make
recordings of lectures available online, in particular for large enrolment courses.
Students appreciate the flexibility that access to recorded live lectures give
(Yoon & Sneddon, 2011), but it is unclear whether it improves learning and
academic achievement, particularly in subjects such as mathematics where
conceptual understanding is more important than rote learning. Several studies
indicate a weak association between frequent use of online recorded lectures and
poorer performance (Brooks, Erickson, Greer, Gutwin, 2014; Howard, Mehan,
& Parnell, 2017; Inglis, Palipana, Trenholm, & Ward, 2011, Trenholm, Alcock,
& Robinson, 2012 and references therein). Yoon & Sneddon (2011) suggests
that easy accessibility of recorded lectures may give some students a false sense
of security, resulting in procrastination and missing the lecture completely in the
end. One of the three variables in Moore's theory of transactional distance
(Moore, 1993) is autonomy, the ability a student has to manage their own
learning (the other two variables are structure and dialogue). Within this theory,
students with a high degree of autonomy would be able to utilize recorded
lectures better than students on the other end of the autonomy-spectrum.
RESEARCH TOPIC AND RESEARCH QUESTIONS
The research topic of this poster is students' usage patterns of online recorded
lectures in an undergraduate mathematics course. Data are collected from log

370 [Link]:indrum2018:174838
files of viewings of recorded lectures in a first year (second semester) course in
linear algebra and differential equations at the Department of Engineering
Sciences at the University of Agder held in spring 2016, with a total of 381
enroled students. We analyze the data to answer the following research
questions: How many students watch the recordings? Do students watch the
recordings regularly or in a more random fashion? How much of each recorded
lecture is watched? When do students watch the recorded lectures? What is the
delay between when the lecture was given and when the recording is watched?
RESULTS
We present the results of the analysis in this poster in terms of histograms and
bar plots with accompanying texts. In short, our results are as follows: Out of
381 enroled students, approximately 80 students (21%) use the recordings on a
regular basis, with a marked increase in viewings around sports- and Easter
holiday. Closer analysis shows that typically all lecture content in the videos are
watched, suggesting that most students watch the whole recording instead of just
smaller parts of it. The maximum number of viewings on a day happens on the
same day as the live lecture, and the viewings then drop during the following
two-three weeks after a lecture. There are relatively few viewings before the
exam suggesting that students do not use the recordings for exam preparations.
REFERENCES
Brooks, C., Erickson, G., Greer, J., Gutwin, C. (2014). Modeling and
quantifying the behaviours of students in lecture capture environments.
Computers and Education, 75, 282–292. doi:10.1016/[Link].2014.03.002
Howard, E., Meehan, M., Parnell, A. (2017). Live lectures or online videos:
students' resource choices in a first year university mathematics module.
International Journal of Mathematical Education in Science and Technology,
1–24. doi:10.1080/0020739X.2017.1387943
Inglis, M., Palipana, A., Trenholm, S., & Ward, J. (2011). Individual differences
in students' use of optional learning resources. Journal of Computer Assisted
Learning, 27, 490–502. doi:10.1111/j.1365-2729.2011.00417.x
Moore, M.G. (1993). Theory of transactional distance. In D. Keegan (Ed.),
Principles of Distance Education (pp 22–38). Routhledge, New York.
Trenholm, S., Alcock, L., Robinson, C.L. (2012). Mathematics lecturing in the
digital age. International Journal of Mathematical Education in Science and
Technology, 43(6), 703–716. doi:10.1080/0020739X.2011.646325
Yoon, C., Sneddon, J. (2011). Student perceptions of effective use of tablet PC
recorded lectures in undergraduate mathematics courses. International
Journal of Mathematical Education in Science and Technology, 42(4), 425–
445. doi:10.1080/0020739X.2010.543165

371 [Link]:indrum2018:174838
TWG 5: Teachers’ practices
 Helping lecturers address and formulate teaching challenges:
an exploratory study
Ignasi Florensa1, Marianna Bosch², Jordi Cuadros² and Josep Gascón3
1
Escola Salesiana Universitaria de Sarrià in Barcelona, Spain, iflorensa@[Link];
²IQS School of Management, Univ. Ramon Llull, Barcelona, Spain; 3Dep.
Matemàtiques, Univ. Autònoma de Barcelona, Spain
In this paper we present an exploratory study on the kind of questions or difficulties
lecturers point out at the beginning of an educational course – to be addressed in it.
These questions happen to be very general and are poorly connected to the
knowledge to be taught. We suggest a twofold interpretation of this phenomenon. On
the one hand, and in the line of the didactic transposition theory, teachers do not
allow themselves to raise questions about the knowledge that is supposed to be their
main field of expertise. On the other hand, the prevailing institutional pedagogy does
not provide teachers with a fruitful enough conceptual frame to formulate this kind of
questions. From the experience of several lecturer education courses, we postulate
that didactics can help university teachers better interpret their practice and question
it in a more productive way.
Key-words: university teacher education, pedagogy, knowledge to be taught, scale of
levels of codeterminacy, teacher problems.
INTRODUCTION: THE PROBLEMS TEACHERS FACE
Since 2009, our research group has designed and implemented two different courses
to provide university lecturers and research assistants with educational tools enabling
them to better design, implement and analyse teaching and learning processes. The
first course took place at IQS – Universitat Ramon Llull in Barcelona, an institution
offering degrees and master programmes in engineering and management. This first
course was addressed to PhD students teaching at the institution or planning to teach
soon. The PhD students’ research domains and subjects taught were diverse and
included econometry, finance, mathematics and engineering, among others. The
course was structured into thirteen 2-hour sessions and lasted 4 academic years. In
the first session, the participants were asked to raise teaching questions they would
like to address in the course. The collected questions were then classified according
to the level of co-determinacy they affected (Chevallard, 2002, we will come back to
this notion later). The subsequent sessions were each devoted to addressing the
questions that belonged to one of the levels, starting from the general ones
(Civilisation, Society) and finishing with the content-specific ones (Domain, Sector,
Theme, Question). At the end of the course students were asked to design a teaching
project for a subject of their specialty, including a syllabus, the planning of learning
goals and a detailed description of three teaching activities: a lecture, a student-
centred task and an autonomous out-of-class activity.

373 [Link]:indrum2018:174767
The second course was held at EUSS-UAB in Barcelona, an engineering school
offering Mechanical, Electronic, Electrical and Management engineering degrees.
The participants were in-service teachers with different educational backgrounds and
research fields. The course was organised in six 2-hour sessions. It was based on a
study and research path for teacher education, an inquiry-based teaching format
focused on the study of a professional teaching question (Florensa, Bosch, & Gascón,
2017). The question addressed was “Could modelling be the main motivation of my
subject?” It was approached through different phases where participants experienced
an inquiry study process in the position of the students, then analysed the process
experienced and finished by designing an inquiry study process for their subjects.
Both courses started by asking the participants to provide a list of questions or
difficulties they would like to address with the help of the educators. In all the cases,
we were surprised to find there was only a small number of questions that dealt
specifically with the knowledge to be taught. Teachers mainly mentioned general
issues related to assessment, class management, coordination or student
characteristics (diversity, lack of motivation, the role of mathematics in their subject,
etc.). They rarely included their subjects in the questions and, when they did, the
problems formulated were very general.
We compared this result with an investigation research carried out by Cirade (2006)
in pre-service teacher education in France within the anthropological theory of the
didactic (ATD). In this research, during 3 editions of a 25-week course in 3 academic
years, the participants who were doing an internship in secondary schools were asked
to formulate a question every week. These “questions of the week” constituted the
basis of the course, despite the fact that only a small sample of them could be
addressed – all in all, more than 7,000 questions were collected. Cirade provides a
systematic gathering and analysis of the teacher-students’ spontaneous questions and
uses them to identify the mathematical difficulties teachers encountered and their
trouble in making them explicit. The kind of questions raised at the beginning of the
course – which coincided with the beginning of the academic year – were initially
very general, and were related to how to behave in class, how to manage the students’
behaviour, what to do in a meeting with parents, etc. Then, as the teacher education
course progressed and certain tools coming from the field of didactics of mathematics
were introduced, teacher-students became more and more able to state questions
related to the knowledge to be taught. In a sense, we can say that they stopped taking
the knowledge to be taught as a given and dared to state questions about their own
field of expertise. For instance, they ended up asking questions such as “How to
justify the need of sketching functions given their analytical expression?”, or “Why
do we need to measure angles in radians in addition to degrees?”, etc.
Following Cirade (Chevallard & Cirade, 2010; Cirade, 2006), we postulate that
educational courses for university teachers cannot ignore the way teachers
problematize their professional practice and teachers should take their concerns and

374 [Link]:indrum2018:174767
difficulties as the starting point of educational processes. Besides, as researchers in
mathematics education, we also agree with the importance of approaching these
questions from a discipline-based level. As stated by Berthiaume (2009, p. 215):
For some time now, educational researchers have investigated the idea that, in order to be
effective, higher education teaching may have to be ‘discipline-specific’. In other words,
teaching in higher education has to take into account the specific characteristics of the
discipline being taught. This means that developing an understanding of teaching and
learning is not sufficient to become an effective teacher in higher education. Rather, one
must also develop an understanding of the teaching and learning requirements of one’s
own discipline. This has been termed ‘discipline-specific pedagogical knowledge’.
We consider essential for university teachers to be able to formulate their difficulties,
not only as general issues concerning students and class management, but also
including the knowledge to be taught as a key element. Even if teaching problems are
initially perceived as general in their manifestation, the way to address them will
necessarily involve knowledge-based activities. From the perspective of the ATD,
taking the knowledge to be taught into account means more than simply including it
as a variable or parameter of the problem formulation. It also means considering it as
an institutional construction, questioning its current shape and searching for possible
new reorganisations, taking into account – without assuming – the epistemologies
and pedagogies prevailing at the university (Barquero, Bosch, & Gascón, 2013).
The aim of our study, which is still at an exploratory stage, is to analyse the kind of
questions university teachers are able to state at the beginning of an educational
course – as the ones we implemented – and locate their questions at different levels of
specificity/generality regarding the knowledge to be taught. We postulate that
knowledge in didactics is important to provide university teachers with conceptual
and methodological tools not only to improve their professional practice, but also to
describe, interpret, conceive and question it in a more productive way. The first step
to make progress in this direction is to start understanding how lecturers
spontaneously formulate the challenges faced during their daily practice.
WHAT PROBLEMS DO LECTURERS SET FORTH?
We collected a total of 143 questions from the 4 courses, 35-40 per course, each of
which was attended by 10-15 participants. In all of the cases, teachers attending
the course were asked the following: “Write down two or three problems,
difficulties or doubts that you find, or you think you may find during your
teaching practice.” There was a lot of redundancy in the questions, so we
eliminated repetitions even if the phrasing was different. We are presenting this
selection according to the questions’ generality, using the scale of levels of
didactic co-determinacy. This tool was introduced by Chevallard (2002) in the
didactic analysis to include aspects of the institutional organisation of teaching
and learning processes that are usually taken for granted (Artigue & Winsløw

375 [Link]:indrum2018:174767
 2010; Chevallard & Sensevy, 2014). It helps distinguish the conditions and
constraints affecting teaching and learning processes that are originated within
the discipline, and the generic levels common to the teaching of any discipline:

Upper levels Lower levels

Humanity 
 Discipline
Civilisation 
 Domain
Society 
 Sector
School 
 Theme
Pedagogy 
 Question
Figure 1. Scale of levels of didactic co-determinacy
Civilisation and Society
The upper levels of the scale refer to the conditions that are set up by our society or,
when these are common to several societies, by the civilisation they belong to. We
identified the following questions at this level:
- What to do in a culture in which effort and reward are no longer related?
- How does a social situation influence the effectiveness of a course or a teaching
format? For example, the students’ attitude seems different in times of crisis…
- To what extent should study plans be aligned with the labour market?
- What is “academic freedom” and what are its limits?
- Where do competencies come from? How are they established?
- Clashing of (generational or social?) values: sometimes it is difficult to act as a
teacher, a guide or a referent when our own values seem to be obsolete (or to strongly
contrast with those of our students). For example: the value of effort, the gratification
of work well done, the fact that money cannot buy everything or that not everything
is on the web, the importance of culture, of thinking, that there are things that are
“right” and others that are “wrong” (e.g. cheating in an exam is “wrong”), etc.
- What to do with students who act as “clients”?
- The application of Bologna is an adaptation of the learning process and an evolution
or change: more participative students, more teacher-student interaction, etc.
Adaptability is therefore considered a consequence of an evolution, but if we analyse
it, we are giving the same classes, in the same environment, with the same student
profile. Can we do anything to make the context change?
- How is the knowledge of the different subjects selected and what criteria are used?
- How far should we, as teachers, arrive in our role in and out of class? What are
teachers educated for?

376 [Link]:indrum2018:174767
As we can see, all these questions refer to dimensions or difficulties related to
university teaching that do not depend on the specific institution considered – many
of them can be extended to any kind of teaching and to other educational levels. What
is questioned is the way our societies – more or less explicitly – conceive, organise,
and manage the dissemination of knowledge, and the general roles assigned to
teachers as guides, leaders or knowledge disseminators.
School (here University)
The School level includes the conditions and constraints that depend on the specific
teaching institution considered, in our case, the University with its own specificities:
- Is the number of students per class important in terms of effectiveness of the
teaching? Is there an optimal number? Are there exceptions?
- How are decisions regarding time-schedule, session duration, etc. taken?
- To what extent are university facilities important? Are there optimal premises? How
to adapt to the ones available?
- How to respond to the pressure of introducing ICT in the classroom? Is it used
because of real educational reasons or is it cost saving? Is it just a trend?
- How to ensure a good coordination between teachers of the same subject? What
happens when they have a different conception of the subject to teach?
- How to establish more synergies between colleagues, sharing methodologies?
- How to ensure a coherent programme? What relationships exist between subjects?
- How to integrate the different subjects to obtain a more global education?
These questions also reveal the aspects teachers think can be changed and the ones
they take for granted, not even considering them questionable. For instance, in the
fifth question, only the coordination with teachers of the same subject is considered,
according to the traditional compartmentalization of knowledge in higher education.
Together with the sixth question, they reveal the lack of a professional culture that
might include coordination between teachers. However, the sixth question seems to
consider that this coordination only affects “methodologies”, which again appears to
be a vague and general dimension of teaching. The last two questions are content-
related, but only with respect to the relationships between subjects, as these are
considered to be previously determined – and, therefore, untouchable.
Pedagogy
The level of pedagogy is common to the different subjects or disciplines that can be
taught. It includes all the resources, formats, and strategies teachers and students
activate – many times without even noticing it – for teaching and learning processes
to take place. We gathered numerous questions that can be located at this level of the
scale and organised them into two main blocks: students and lecturers.
Students

377 [Link]:indrum2018:174767
 - How to manage long projects, where students slack off and decrease their work
intensity? How to reach a balance between establishing milestones and letting
students work independently?
- What to do with the students that chat, are unfocused, use their mobile phone, do
other things than what is requested of them?
- How to deal with student diversity?
- How to arouse the students’ interest in subjects that are not at the core of the degree?
- How to motivate students beyond the minimum required “pass” grade?
- How to encourage students to participate in a large group?
- Should students be monitored closely or should they work more independently?
- How to motivate students to behave in class?
- How to encourage students to be more competitive through the intrinsic values of the
subject that is taught?
- Students are not previously taught how “to learn”. How will this affect our job?
- How to teach students to listen and maintain their attention?
Lecturers
- How to improve oral and body expression?
- How to organise assessment in a fair and impartial way?
- How to assess core (non-disciplinary) competencies?
- Is it better to use final examinations or continuous assessment? How to measure
long-term student learning outcomes?
- How to design contents, planning and methodologies of the subject that take into
account the student diversity?
- How to reach all the students and not only those who have more knowledge,
excluding the ones that got lost? How to find the balance between maximising
student learning and the amount of information provided?
- How to ensure an individualised methodology considering the time limitations?
- How to become the best teacher for each student?
- How to deal with the so-called “decline in student knowledge”?
- Does the decline in student knowledge correspond to something real or is it just what
each generation says about the previous one?
- Does it make sense to give lectures nowadays?
- How to improve teaching resources and methodologies using ICT?
- What to do after the class? How to analyse what happened and what the teacher did?
How to assess teacher performance?
As can be seen from the questions above, most of them focus on specific teaching
practices, but they do not refer to the difficulties of the corresponding subjects. The
questions are mainly related to what the teachers can or might do, and they are very
general. Only two of them refer to specific teaching formats: projects and lectures.
There is no mention of the activities organised (labs, tutorials, problem solving or
case study sessions, outdoor activities, etc.) and the way to better implement them.

378 [Link]:indrum2018:174767
The need to implement new kinds of activities is not mentioned. The questions
mainly have to do with the teacher and the teacher’s actions. For instance: assessment
is always considered as a lecturer’s task; “motivation” is assumed to be generated
(only) by the teacher. The questions thus reveal many features of the traditional
pedagogical contract, which seems to be fully assumed by the lecturers.
Discipline
As said before, we were astonished to find so few questions at this level, which
corresponds to the conditions and constraints directly linked to the content taught and
learned. They can be related to the general discipline the content belongs to
(Mathematics, Engineering, Economics, Management, etc.) or to the different
components of the discipline, according to the way it is structured or delimitated in
the considered institution. The general terms used to specify these levels are:
domains, sectors, themes, tasks or questions. The divisions and boundaries
established in a discipline or field of knowledge are institutional constructions. They
vary from one institution to another and from one historical moment to another. The
collected questions at this level remain very general; none of them specifies the
difficulties related to the teaching or learning of a given piece of content. The first
one, for instance, is very similar to those located at the School level: it depends on
whether we interpret the question as affecting the design of an entire programme or
the possible actions in one discipline:
- How to better connect the different subjects of the programme?
- How to highlight the multi-disciplinary nature of the subjects?
- How to select the learning goals of the subject? What content should be included?
- What should the level of the learning goals be?
- How to relate the subject with the real world?
- How to balance learning goals between specialisation and generalization?
WHY THESE QUESTIONS? AN INTERPRETATION FROM THE ATD
The assumed educational contract between lecturers and educators
The first reason that came to our mind when trying to understand why lecturers did
not ask any content-related question is the kind of implicit didactic contract that was
assumed by them at the beginning of the course. Given the fact that the course was
about university teaching, they might have expected to learn certain generic tools to
help them in their teaching practice; not something related to their specific subject.
The educators were seen as specialists in teaching processes and the questions were
stated at this general level. Either way, this shows a first important phenomenon:
lecturers expect to receive help with general teaching practices that are only a part of
their daily practice. A lot of their teaching work (elaboration of the syllabus; choice
of textbooks, reference books and other kind of resources or materials; selection,
design and organisation of activities, cases or problems; decisions about the kind of

379 [Link]:indrum2018:174767
in-class and out-of-class activities students should carry out; renewal of the subject
matter; etc.) does not seem to have been included in the objectives of the course.
A problem of legitimacy
The second reason we put forward is related to what we call a problem of legitimacy.
University teachers are often also researchers or at least experts in the subject they
teach. Therefore, they may be reluctant to accept the idea that their teaching
difficulties might come from problems with the subject matter they are supposed to
master. Their lack of expertise can only be attributable to what is external to the
discipline they teach. This reinforces the previous reason about their expectations
from the educational course.
The divide between pedagogy and didactics
There is another important and more general factor that may explain the lack of
content-related questions. It corresponds to the dominant interpretation of teaching
and learning phenomena that has been called “pedagogical generalism” (Gascón &
Bosch, 2007) or the “didactic divide” between pedagogical and subject-matter
knowledge (Bergsten & Grevholm, 2004). It tends to introduce a strict separation
between instructional processes and the “content” of these processes, that is, using
the scale of didactic co-determinacy, between the level of Discipline and the level of
Pedagogy. The main point in crossing the boundary between the two levels is the way
knowledge is conceived in the considered teaching process or, in other words, which
aspects of the subject-matter are questioned and which ones are assumed as a given.
When a teacher – or a lecturer – is asked to teach a given piece of knowledge k, the
first question she will first ask herself is “what should I do to teach k?”, not “what is
this k I should teach?” What the theory of the didactic transposition (Chevallard,
1985; Chevallard & Bosch, 2014) states is that instructional processes rely on the
fiction that there is only one way to define k and that this is the k that is taught and
learned. Questioning the knowledge to be taught, asking about its origin, selecting
and applying a given instructional process rarely occur. This is why it is normal the
participants of the course did not set forth questions of that kind. In the questions
stated, knowledge always appears as a given, not as a variable.
The “pedagogical generalism” that is found in the teachers’ questions is not an
isolated fact. If we look at the teaching support some universities offer their (new)
faculty, we see that only the Pedagogical level is addressed, and possibly some
aspects of the School level. For instance, in the Teacher Training in Higher
Education (FDES)1 programme proposed by the Autonomous University of
Barcelona, the structure of the programme is presented as follows:

1
[Link]
docent/programa-fdes/[Link]

380 [Link]:indrum2018:174767
 Activity 1. Teaching in the new context of learning and teaching
Activity 2. Practicing oral discourse
Activity 3. How to assess university students’ learning?
Activity 4. Experiences in educational innovation
Activity 5. Observation in the classroom
Activity 6. Teaching planning: from study programmes to syllabus
Activity 7. Teacher’s portfolio
Similar programmes can be found at other universities. For example, some years ago,
the Teaching Engagement Program of the University of Oregon posted a list of
frequently asked questions (FAQs) organised according to the following headings:
“Getting ready to teach; Presenting and facilitating information; Motivating students;
Questions of respect; Assessment; Managing the classroom climate”. None of the
questions was content-related. It seemed as if, once certain answers were provided to
the pedagogical issues, their specification to the subject-matter was considered
evident or, at least, non-problematic.
CONCLUSION: A LACK OF TOOLS AND NOTIONS
One of the consequences of “pedagogical generalism”, that can partially be seen in
the questions stated by university teachers, is the lack of terms and concepts to go
below the level of Pedagogy and start questioning the levels of Discipline. University
teachers develop their professional activity at institutions where little is said about the
way knowledge should be selected, arranged, updated, organised, “elementarised”,
put-into-practice, problematized, etc. in order to teach it or to help students to learn it.
This is a crucial aspect in which the results obtained from research in Didactics of
Mathematics, both practical and theoretical, can assume an important function.
From the experience of the courses here presented, we have seen how introducing
certain elements of the ATD (the notions of praxeology, didactic contract, didactic
moments, Herbartian schema, media-milieu dialectics, didactic ecology, etc.)
provides lecturers with a productive enough framework to talk about and start
questioning a larger part of their teaching activities (Florensa, Bosch, Gascón, &
Ruiz-Munzon, 2017). The more is said about didactic processes, the more dimensions
of these processes can be questioned and tentatively changed. Our hope is that the
education of lecturers, as is the case with primary and secondary school teacher
education, will have the power to make this state of things evolve. Our experience
with the courses presented lets us be moderately optimistic in this respect.
ACKNOWLEDGMENTS
Funded by project EDU2015-69865-C3-1-R (MINECO/FEDER) from the Spanish
Ministry of Economy, Industry and Competitiveness and the European Regional
Development Fund.

381 [Link]:indrum2018:174767
REFERENCES
Artigue, M., & Winsløw, C. (2010). International Comparative Studies on
Mathematics Education : a Viewpoint From the Anthropological Theory of
Didactics. Recherches en Didactique des Mathématiques, 31(1), 47-82.
Barquero, B., Bosch, M., & Gascón, J. (2013). The ecological dimension in teaching
of mathematical modelling at university. Recherches en Didactique des
Mathématiques, 33, 307-338.
Bergsten, C., & Grevholm, B. (2004). The didactic divide and the education of
teachers of mathematics in Sweden. NOMAD, 9(2), 123-144.
Berthiaume, D. (2009). Teaching in the disciplines. In H. Fry, S. Ketteridge, & S.
Marshall (Eds.), A Handbook for Teaching and Learning in Higher Education
(pp. 215-225). Oxon: Taylor & Francis.
Chevallard, Y. (1985). La transposition didactique. Du savoir savant au savoir
enseigné. Grenoble, France: La Pensée sauvage.
Chevallard, Y. (2002). Organiser l’étude. Écologie & regulation. 11ème École d’Été
de Didactique des Mathématiques (pp. 41-56). Grenoble: La Pensée sauvage.
Chevallard, Y., & Bosch, M. (2014). Didactic transposition in Mathematics
Education. In S. Lerman (Ed.), ACM Encyclopedia of Mathematics Education
(pp. 170–174). Dordrecht, The Netherlands: Springer.
Chevallard, Y., & Cirade, G. (2010). Les ressources manquantes comme problème
professionnel. In G. Gueudet & L. Trouche (Eds.), Ressources vives. Le travail
documentaire des professeurs en mathématiques (pp. 41-55). Rennes, France:
PUR and Paris: INRP.
Chevallard, Y., & Sensevy, G. (2014). Anthropological Approaches in Mathematics
Education, French perspectives. In S. Lerman (Ed.), ACM Encyclopedia of
Mathematics Education (pp. 38-43). Dordrecht, The Netherlands: Springer.
Cirade, G. (2006). Devenir professeur de mathématiques: entre problèmes de la
profession et formation en IUFM . Les mathématiques comme problème
professionnel. (Doctoral Dissertation). Université d’Aix-Marseille.
Florensa, I., Bosch, M., & Gascón, J. (2017). Formación didáctica del profesorado
universitario: análisis de un curso. In J. M. Muñoz-Escolano, A. Arnal-Bailera, P.
Beltrán-Pellicer, M. L. Callejo, & J. Carrillo (Eds.), XXI Simposio de la Sociedad
Española en Investigación en Educación Matemática (pp. 237-247). Zaragoza.
Florensa, I., Bosch, M., Gascón, J., & Ruiz-Munzon, N. (2017). Teaching didactics to
lecturers : a challenging field. In 10th Conference of European Research in
Mathematics Education - CERME10. Dublin.
Gascón, J., & Bosch, M. (2007). La miseria del “generalismo pedagógico” ante el
problema de la formación del profesorado. In L. Ruiz-Higueras, F. Estepa, & F.
J. Garcia (Eds.), Sociedad, Escuela y Matemáticas. Aportaciones de la teoría
Antropológica de lo Didáctico (pp. 201-240). Jaén, Spain: Servicio de
Publicaciones de la Universidad de Jaén.

382 [Link]:indrum2018:174767
An Analysis of University Mathematics Teaching using the Knowledge
Quartet
Sinéad Breen1, Maria Meehan², Ann O’Shea3, and Tim Rowland4
1
Dublin City University, Ireland; ²University College Dublin, Ireland; 3 Maynooth
University, Ireland; 4 Norges teknisk-naturvitenskapelige universitet, Norway.
([Link]@[Link]) 3
We analyse accounts written by three mathematics lecturers on their practice using
the Knowledge Quartet framework. This framework has been used to study how a
teacher’s knowledge of mathematics and mathematics pedagogy influences his/her
actions in the classroom at both the primary and secondary level. We consider how
the framework could be used to study university level teaching, and we report on the
dimensions of teacher knowledge that were made visible by this framework.
Keywords: Knowledge Quartet, teacher knowledge, university mathematics teaching.
INTRODUCTION
The first three authors of this paper are mathematics lecturers at three universities in
Ireland, who also engage in mathematics education research. Over the course of two
years, they wrote accounts of incidents which occurred during their teaching as part
of a professional development project using the Discipline of Noticing (Mason,
2002). In this paper, we report on our more recent use of a different theoretical
framework, the Knowledge Quartet framework (Rowland, Huckstep & Thwaites,
2005), to analyse these accounts. The Knowledge Quartet categorizes situations from
classrooms where mathematical knowledge surfaces in teaching. There has been one
previous attempt to use the framework to analyse university mathematics teaching
(Rowland, 2009). The focus of that paper was the knowledge-grounded foundation
beliefs of the university lecturer, about mathematics and about teaching and learning.
The purpose of our current study is twofold. Firstly, we are interested in whether the
Knowledge Quartet framework could be applied to study teaching at university level.
Secondly, we would like to know what features of university teaching are highlighted
when our set of accounts are analysed using the Knowledge Quartet. Previously, the
first three authors had analysed their accounts to study the many decision points that
arose while teaching and in O’Shea, Breen and Meehan (2017) these decision points
and their triggers were categorised. We were interested to see if using the Knowledge
Quartet framework would draw attention to other aspects of teaching in the accounts.
In this article, we will first of all consider the literature on teacher knowledge,
especially at university level. We will then expand on the Knowledge Quartet
framework, and give some results from our analysis using this lens. Finally, we will
discuss our findings and suggest some future avenues for research.

383 [Link]:indrum2018:174552
LITERATURE REVIEW
The Knowledge Quartet is a theoretical tool for observing, analysing and reflecting
on actual mathematics teaching. Ball, Thames and Phelps (2008) also studied
mathematics classrooms to develop a theory of mathematical knowledge for teaching
that built on the work of Shulman (1987). This resulted in the identification of an
important subdomain of content knowledge - ‘specialized content knowledge’. This is
distinct from ‘common content knowledge’ and is unique to the work of teaching.
Independently, Ainley and Luntley (2007) suggested that experienced teachers draw
on ‘attention-dependent knowledge’ in addition to subject knowledge and
pedagogical knowledge (both general and subject-specific). Few research studies
have been concerned with the knowledge employed in university mathematics
teaching. McAlpine and Weston (2000) conducted a research study with six
professors considered exemplary in their teaching and found that all the professors
drew on pedagogical knowledge, pedagogical content knowledge, content knowledge
and knowledge of learners (following Shulman (1987)) while monitoring their own
actions and making decisions during lectures. This was despite the fact that three of
the professors were mathematicians who had no pedagogical training (while the
remaining three were mathematics educators or trained teachers). McAlpine and
Weston (2000) hypothesised that the mathematicians constructed this knowledge
largely through experience and reflection, and that their lack of training led them to
depend more on their experience than the mathematics educators did.
On the other hand, Wagner, Speer and Rossa (2007) examined the knowledge, other
than content knowledge, required by a mathematician teaching an undergraduate
course. They reported that he was unable to anticipate how students would respond to
particular activities and how the content or sequence of individual classes contributed
to the instructional goals of the entire course. The authors claim these findings lend
support to the assertion that there is knowledge particular to teaching that is distinct
from, and not easily constructed from, knowledge of content.
Speer and Wagner (2009) focussed on whole-class discussions and examined the
nature of the knowledge that a mathematician could employ to make effective use of
undergraduates' mathematical contributions in a way that furthered the goals for the
class. Their analysis focussed on the role of (a lack of) pedagogical content
knowledge and specialized content knowledge in the difficulties experienced by the
instructor in scaffolding student learning while orchestrating such discussions.
THEORETICAL FRAMEWORK
The Knowledge Quartet
The Knowledge Quartet is a ‘theory’ in the sense that it proposes a way of thinking
about mathematics teaching in the usual institutional settings (lessons/classes), with a
focus on the disciplinary content (mathematics) of the lesson.

384 [Link]:indrum2018:174552
The Knowledge Quartet (KQ) was the outcome of empirical research at the
University of Cambridge, UK (Rowland et al., 2005), in which 24 mathematics
lessons were videotaped and scrutinised. The research team identified aspects of the
teachers’ actions in the classroom that could be construed as being informed by their
mathematics subject matter knowledge or pedagogical content knowledge. This
inductive process initially generated a set of 18 codes (later expanded to 21),
subsequently grouped into four broad, super-ordinate categories or dimensions.
The first dimension of the KQ, foundation, consists of teachers’ mathematics-related
knowledge, beliefs and understanding. The second dimension, transformation,
concerns knowledge-in-action as demonstrated both in planning to teach and in the
act of teaching itself. The third dimension, connection, concerns the ways by which
the teacher achieves coherence within and between sessions. The final dimension,
contingency, is witnessed in classroom events that were not envisaged in the teachers’
planning. Essentially, it is the ability to “think on one’s feet”.
Conceptualising the Knowledge Quartet
The concise conceptualisation of the KQ which now follows is a synthesis of the
characteristics of its four dimensions.
Foundation
The first member of the KQ is rooted in the foundation of the teacher’s theoretical
background and beliefs. It concerns their knowledge, understanding and ready
recourse to what was learned in preparation (intentionally or otherwise) for their role
in the classroom. The key components of this theoretical background are: knowledge
and understanding of mathematics per se; knowledge of significant tracts of the
literature and thinking which has resulted from systematic enquiry into the teaching
and learning of mathematics; and espoused beliefs about mathematics, including
beliefs about why and how it is learnt. The remaining three categories focus on
knowledge-in-action as demonstrated both in planning to teach and in the act of
teaching itself.
Transformation
At the heart of the second member of the KQ is Shulman’s observation that the
knowledge base for teaching is distinguished by “ … the capacity of a teacher to
transform the content knowledge he or she possesses into forms that are
pedagogically powerful” (Shulman, 1987, p. 15, emphasis added). This dimension
picks out behaviour that is directed towards a student (or a group of students), and
which follows from deliberation and judgement informed by foundation knowledge.
The choice and use of examples has emerged as a rich vein for reflection and critique,
and one of the most prevalent codes observed in practice (Rowland, 2008).
Connection
The next dimension concerns the coherence of the planning or teaching displayed
across an episode, lesson or series of lessons. Our conception of connection includes

385 [Link]:indrum2018:174552
the sequencing of topics of instruction within and between lessons, including the
ordering of tasks and exercises. To a significant extent, these reflect deliberations and
choices entailing not only knowledge of structural connections within mathematics,
but also awareness of the relative cognitive demands of different topics and tasks, and
the implementation of strategies to remove (or lessen) obstacles to learning.
Contingency
Our final dimension concerns the teacher’s response to classroom events that were
not anticipated in the planning. This dimension of the KQ is about the ability to
‘think on one’s feet’: it is about contingent action. Whilst the teacher’s intended
actions can be planned, the students’ responses cannot. The teachers’ response to
students’ unexpected contributions is one of the most low-inference codes of the KQ.
Many moments or episodes within a session can be understood in terms of two or
more of the four units; for example, a contingent response to a student’s suggestion
might helpfully connect with ideas considered earlier. Furthermore, the application
of content knowledge in the classroom always rests on foundational knowledge.
The KQ is a lens through which the observer ‘sees’ classroom mathematics
instruction. It offers a four-dimensional framework against which mathematics
lessons can be discussed, with a focus on their subject-matter content, and the
teacher’s related knowledge and beliefs.
This framework has been used in different contexts and for different purposes. For
instance, Rowland (2012) used the KQ to examine situations in which mathematical
knowledge surfaces in primary and secondary mathematics. He concludes that
elementary mathematics teaching poses challenges which are qualitatively different
from those confronting secondary mathematics teachers. However, the mathematics
knowledge primary mathematics teachers must possess is neither less profound nor
easier to acquire than that of secondary teachers. Turner and Rowland (2011)
describe a project in which the framework was used to guide pre-service teachers in a
process of personal reflection on their teaching. The participants found that the KQ
helped them to focus more effectively on the mathematical content of their lessons
and its enhancement. The authors reported that this enhanced focus on mathematical
content knowledge had a positive influence on its further development. There was
also evidence that the KQ helped the participants to develop a more learner-centred
view of teaching and one in which conceptual understanding rather than procedural
fluency was emphasised. Other recent studies using the KQ have focussed on
contingent moments in the classroom (e.g. Rowland & Zazkis, 2013).
METHODOLOGY
The accounts which form the data for this study were written using the Discipline of
Noticing (Mason, 2002). This advocates that practitioners write ‘brief-but-vivid’
accounts of incidents that they have noticed in their practice. Mason (2002) defines a
brief-but-vivid account as

386 [Link]:indrum2018:174552
 one which readers readily find relates to their experience. Brevity is obtained by omitting
details which divert attention away from the main issue. The aim is to locate a
phenomenon, so the less particular the description, the easier this is, without becoming so
general as to be of no value….Thus description is as factual as possible. (p.57)
He advises that these accounts should also avoid justification of incidents or actions,
and should therefore be ‘accounts of’ rather than ‘accounting for’ a particular
situation. The first three authors of this paper had written brief-but-vivid accounts of
their teaching over a two-year period. These focused on notable incidents that
occurred while they were teaching, but are not reflections or descriptions of a whole
lecture. For more details, see O’Shea, Breen and Meehan (2017).
For this paper each of the three lecturers chose one of their modules; only the
accounts relating to that module which contained references to mathematical
knowledge were analysed (20 accounts for Lecturer 1, 29 for Lecturer 2, and 38 for
Lecturer 3). Lecturer 1 chose a one-semester Introduction to Analysis module for 27
second-year students (this module was delivered separately to 7 Pure Maths students
and 20 Science students), Lecturer 2 also chose a one-semester Introduction to
Analysis module for a group of 75 second-year students, while Lecturer 3 chose a
year-long Differential Calculus module for a group of 49 first-year students. All three
lecturers aimed to foster dialogue in their classrooms, perhaps because of their
interest in educational research and the relatively small class sizes in these modules.
When coding the data we compared our accounts with the descriptions of each of the
21 codes associated to the KQ framework, with reference to the examples available at
[Link]. We began the coding process by first coding a small set
of accounts together. Then each lecturer coded her own set of accounts and passed on
her analysis to the other two lecturers in turn. They coded the accounts independently
before comparing their analysis with that of the original instructor. All three
discussed any discrepancies and agreed on the final coding.
During the coding process, we felt that the names of a few of the codes did not fully
reflect the terminology used in teaching mathematics at the university level. We
interpreted the code Teacher Demonstration (to explain a procedure) to also
encompass teacher demonstration to explain a proof. We chose to use the code
Choice of Example (CE) to include particular instances of an abstract concept or a
general procedure; and, as the rehearsal of a procedure or ‘exercise’ (Rowland, 2008),
and also for non-routine tasks. We also applied the code Responding to Students’
Ideas (RSI) from the Contingency Dimension to encompass instances where the
lecturer had to respond to a lack of students’ ideas.
RESULTS
A summary of the number and percentage of codes found in each of the four
categories of the KQ for each author is given in Table 1 below. While a number of
codes could be applied to some events, the one which we judged to be predominant
was what was counted in this table.

387 [Link]:indrum2018:174552
 KQ Dimension Lecturer 1 Lecturer 2 Lecturer 3
Foundation 10 (20.83%) 4 (11.11%) 41 (34.74%)
Transformation 14 (29.17%) 17 (47.22%) 32 (27.12%)
Connection 7 (14.58%) 8 (22.22%) 24 (20.34%)
Contingency 17 (35.42%) 7 (19.45%) 21 (17.8%)
Total 48 (100%) 36 (100%) 118 (100%)
Table 1. Number and percentage of codes in each KQ dimension for each lecturer
On coding the accounts it became apparent that all three lecturers frequently wrote
accounts about giving a task to the class or instigating a whole class discussion
around a task, recording some students’ responses (or lack of responses) in relation to
the task, and noting what the lecturer thought or learned about student thinking.
Therefore it is not surprising that the code Choice of Example (CE) in the
Transformation Dimension was the most frequently occurring code for all three
lecturers, while Responding to Students’ Ideas in the Contingency Dimension was in
each of their top three most frequently occurring codes. In many accounts, the
lecturer contrasted student learning on a task with learning on the same task the
previous year or with students in a different class, often noting what students found
easy or difficult. The tasks were usually designed and planned by the lecturer with
specific aims for student learning in mind, thus the codes Anticipation of Complexity
(AC) in the Connection Dimension and Awareness of Purpose (AP) in the
Foundation Dimension frequently appear for all three lecturers. Lecturer 3, who was
simultaneously conducting a research project on mathematical tasks, was often
explicit about the pedagogical rationale behind a given task. Consequently, another
significant code for her accounts was Theoretical Underpinning of Pedagogy (TUP)
in the Foundation Dimension.
By way of example, the following is an account from Lecturer 1, coded as RSI. She
struggles to understand what the student is asking but still feels she has to respond:
A student asked a question in the middle of a complicated proof. I didn't understand the
question and asked him to ask it again. He tried but I still couldn't understand. So I
explained the proof again as best I could paying attention to what I thought he had had
problems with. However I realised I had made a choice. I could have continued probing
until I figured out what he was asking. I decided not to do that so as not to embarrass him,
but maybe I didn't really answer his question in the end.
While the majority of accounts were on incidents during lectures, some relate to
preparation of tasks and lessons, or conversations with students outside of class. The
following is an example of an account by Lecturer 2 coded as CE, which describes a
task given to students to work on during the second lecture of the semester.
I handed out the first Inclass Exercise of the module. It contained the following statement:
There exists a university in the world, where every Analysis student achieves a final mark of

388 [Link]:indrum2018:174552
 at least 90% in the module. The instructions were as follows: Write down what you would
need to do to prove that Statement B is false. At the end of the class, a student came up to
me and said that suppose there were infinitely many universities in the world, then you
couldn’t actually disprove the statement because you wouldn’t be able to get around to all of
them to check the Analysis grades. I was impressed with how he extended the statement.
Given that CE was the most frequently occurring code for all three lecturers, the
following account provides another example of a task given, this time by Lecturer 1,
to help students propose conjectures about the relationship between bounded and
convergent sequences.
I was talking about bounded sequences with the class today. I got them to come up with some
bounded and some unbounded sequences. I tried to get the class to make conjectures by asking them
to guess what the next theorem would be, or what it definitely wouldn't be. They immediately
realised that there would be no theorem that said that every bounded sequence converges and then
conjectured that every convergent sequence is bounded. They seemed to enjoy the process.
Next we present an account from Lecturer 3, coded as AP. She is explicit in her
intentions to engage students in mathematical sense-making and on challenging
students’ views of mathematics as a set of rules to be learned and applied.
Today I continued with sketching graphs of functions and asked the students to draw the
graph of f(x)=1/x on its natural domain, among others. I circulated the room as they were
doing this and noticed that a number of what I had considered to be the more able students
were drawing the graph incorrectly (possibly confusing f(x)=1/x and g(x)=1/x^2). I have
been trying to put across the idea of Calculus as a ‘science’ from the point of view that
‘experiments/trials’ can be undertaken to check ‘hypotheses’, results can be ‘replicated’ and
so on, but it appears some students are disregarding this and still regard it as a collection of
facts to be learnt and remembered.
Finally, we present an account from Lecturer 3. Her pedagogy is underpinned by
having students take a guided-discovery approach as a classroom community (TUP).
I tried to use a ‘guided-discovery’ approach to facilitate students’ realization that the graph
of a function and its inverse are mirror images of each other in the line y=x. However, each
step of this took a lot longer than I envisaged. Moreover, I wasn’t convinced at the end that
the students would retain this particular piece of information longer or understand it better
for having discovered it themselves as a class community.
DISCUSSION
In this paper we have used the KQ to analyse a set of accounts written as part of a
professional development project that involved engaging with the Discipline of
Noticing (Mason 2002). This is not the usual type of data that has been used in
previous KQ studies. Typically, the researchers in those studies had access to
classrooms (of either pre-service or experienced teachers), and have been able to
record and analyse entire lessons. Our data is different in two key ways. Firstly we do
not have recordings of entire lessons but the brief-but-vivid accounts of the instructor

389 [Link]:indrum2018:174552
herself on some aspect of the class, which was memorable to her. This is a limitation
because we may have chosen not to include some relevant aspects of our classes, or
of our students’ experience and reactions, but the accounts do shed some light on the
‘attention-dependent knowledge’ of the instructor (Ainley & Luntley, 2007). We did
not write our accounts in order to give a representative view of our teaching, rather
we concentrated on aspects which were troublesome to us. However, we do have
accounts from almost every lecture in the modules considered whereas previous
studies have data only from a very small number of classes with a given teacher.
In his KQ analysis of university mathematics teaching, Rowland (2009) refers to only
one lecture. The analysis homes in on the foundation dimension and in particular on
the beliefs of the lecturer (about mathematics and pedagogy), but does not explore the
other three dimensions. Our analysis has shown that all four dimensions were present
in our data. It should be noted that all three lecturers pursued an interactive approach
in their classes, and perhaps the same spread of codes would not be present in an
analysis of a more stereotypical university lecture.
On the other hand, the prevalence of the use of the responding to student ideas code
for the accounts discussed here suggests that the traditional image of a lecture (in
which a lecturer delivers from a pre-prepared script, rarely deviating from it, and
interacts minimally with students) is not always accurate and highlighted this element
of our practice for us.
In addition, given our previous focus on decision points in these accounts (O’Shea,
Breen & Meehan, 2017), we may have expected the contingency dimension to be
dominant but this was not the case. The KQ highlighted the importance of the other
three categories in our accounts, especially the transformation dimension in the
choice of examples. We found the framework provided a lens through which the
knowledge brought to bear in the preparation and teaching of lessons could be viewed
in a coherent and comprehensive manner.
Each of the first three authors is a mathematician and while none has any formal
pedagogical training, all three conduct research in mathematics education. Many of
the accounts suggested an awareness of purpose on the lecturers’ behalf or a
theoretical underpinning to the pedagogy used when teaching. Perhaps this is a
consequence of their familiarity with the research literature. However, the fact that
the instructors often contrasted student learning in the lectures for which accounts
were written with that of other cohorts lends some support to the hypothesis of
McAlpine and Weston (2000) that a teaching mathematician can construct knowledge
of learners and pedagogy through experience and reflection.
In several accounts the three lecturers highlighted what they noticed about student
thinking on a given task and reflected on this after the lecture. These reflections could
be said to inform their knowledge about mathematics pedagogy, particularly their
knowledge of content and students and knowledge of content and teaching (Ball et
al., 2008), which is a component of the Foundation Dimension. Although we chose

390 [Link]:indrum2018:174552
not to code these reflections since they occurred after lectures, they point to a growth
of knowledge as a consequence of reflection on teaching. It seems that use of the KQ
as a reflection tool could afford mathematicians (with no formal pedagogical training)
an opportunity to develop pedagogical knowledge. It is interesting to compare this
with Turner and Rowland’s (2011) finding that the KQ afforded preservice primary
teachers (typically non-specialists in mathematics) an opportunity to develop
mathematical content knowledge, illustrating the usefulness of KQ to mathematics
teachers of a variety of backgrounds.
Some KQ codes did not appear in our analysis. For example there were no accounts
coded as displaying behaviour such as adherence to a textbook or concentration on
procedures. This may be because of the nature of the mathematics taught in the given
modules. We also found very few references to use of mathematical terminology and
overt display of subject knowledge, which is not to say that the lecturers did not use
terminology or show their subject knowledge during classes but that they did not talk
about it in their accounts (perhaps because it was normal and not problematic). We
used the code identifying student errors sparingly, even though many accounts
contained instances of a lecturer noticing a problem with student understanding. In
our accounts the lecturers seemed to focus more on how to respond to a student rather
than being able to tell when a piece of mathematics was wrong, and so we coded
these episodes using the responding to student ideas code. We also used this code
when the lecturer was faced with a lack of student ideas, for instance when she asked
a question but received no replies. It may be that this is a situation that occurs more
frequently in university than in school, where the size of classes can result in
unwillingness to take part in discussions. We found that the type of specialist
knowledge required to teach abstract mathematics at university was accounted for in
the KQ with many of the codes mentioned earlier as well as others such as choice of
representation, recognition of conceptual appropriateness and making connections
between concepts or representations.
Even though there are some differences in the prevalence of codes at school and
university level, we believe that the KQ offers a useful lens with which to study
undergraduate teaching. It has drawn our attention to the importance of different
facets of lecturers’ mathematical knowledge which we may otherwise have
overlooked. It would be interesting to explore the relationships between the four
dimensions of the Quartet, for example how the underpinning dimension of
foundation knowledge influences the lecturers’ choices made in the other three
dimensions, and how it is in turn influenced by knowledge generated by the lecturer
in a contingent moment. We used the KQ to code reflective accounts written by
mathematics lecturers as they reflected on their teaching. However, we suggest it
could also be used to guide the reflective process and the writing of the accounts. It
would be interesting to explore whether such an approach would lead to a change in
the lecturers' perspectives on teaching similar to those described by Turner and
Rowland (2011).

391 [Link]:indrum2018:174552
REFERENCES
Ainley, J. and Luntley, M. (2007) The role of attention in expert classroom practice.
Journal of Mathematics Teacher Education 10(1) 3-22.
Ball, D. L., Thames, M. H. and Phelps, G. (2008) Content knowledge for teaching:
What makes is special? Journal of Teacher Education 59(5) 389-407.
Mason, J. (2002). Researching Your Own Practice: The Discipline of Noticing.
London: RoutledgeFarmer.
McAlpine, L. and Weston, C. (2000) Reflection: Issues related to improve professors’
teaching and students’ learning. Instructional Science 28(5) 363-385.
O’Shea, A., Breen, S., and Meehan, M. (2017). Decision points in mathematics
lectures. Innovations in Education and Teaching International. DOI:
10.1080/14703927.2017.1321495.
Rowland, T. (2008) The purpose, design and use of examples in the teaching of
elementary mathematics. Educational Studies in Mathematics 69(2) 149-163
Rowland, T. (2009) Beliefs and actions in university mathematics teaching. In M.
Tzekaki, M. Kaldrimidou and C. Sakonidis (Eds.) Proceedings of the 33rd
Conference of the International Group for the Psychology of Mathematics
Education, Volume 5, 17-24. Thessaloniki, Greece: PME.
Rowland, T. (2012) Contrasting knowledge for elementary and secondary
mathematics teaching. For the Learning of Mathematics 32(1) 16-21.
Rowland, T., Huckstep, P. and Thwaites, A. (2005) Elementary teachers’
mathematics subject knowledge: the Knowledge Quartet and the case of Naomi.
Journal of Mathematics Teacher Education 8(3) 255-281.
Rowland, T. and Zazkis, R. (2013) Contingency in the mathematics classroom:
Opportunities taken and opportunities missed. Canadian Journal of Science,
Mathematics and Technology Education 13(2), 137–153.
Shulman, L. S. (1987). Knowledge and teaching: Foundations of the new reform.
Harvard Educational Review, 57, 1-22.
Speer, N. M. and Wagner, J. F. (2009) Knowledge needed by a teacher to provide
analytic scaffolding during undergraduate Mathematics classroom discussions.
Journal for Research in Mathematics Education 40(5) 530-562.
Turner F., Rowland T. (2011) The Knowledge Quartet as an Organising Framework
for Developing and Deepening Teachers’ Mathematics Knowledge. In: Rowland
T., Ruthven K. (eds) Mathematical Knowledge in Teaching. Mathematics
Education Library, vol 50. Springer, Dordrecht.
Wagner, J. F., Speer, N. M. and Rossa, B. (2007) Beyond mathematical content
knowledge: A mathematician's knowledge needed for teaching an inquiry-oriented
differential equations course. Journal of Mathematical Behaviour 26(3) 247-266.

392 [Link]:indrum2018:174552
 Knowledge of the Practice in Mathematics in University Teachers
Rosa Delgado1 and Diana Zakaryan1
1
Pontificia Universidad Católica de Valparaíso, Chile
rosamdelgadorebolledo@[Link]
This research is part of the study of university teachers’ knowledge that has emerged as
a new line of research, which aims to understand the components of this knowledge, its
development, and how it is reflected in university teachers’ teaching practice. This study,
using the Mathematics Teacher’s Specialized Knowledge model, seeks to characterize a
university teacher’s knowledge of practice in mathematics in the content area of
mathematical analysis. Based on an instrumental case study, through classroom
observation, we provide indicators of the teacher’s knowledge of ways of reasoning,
validating, and proceeding in mathematics, contributing to the understanding of the
nature of this teacher knowledge.
Keywords: teachers’ and students’ practices at university level, teaching and learning of
analysis and calculus, university teachers’ knowledge, knowledge of the practices in
mathematics, Mathematics Teacher’s Specialized Knowledge model.
INTRODUCTION
Research on mathematics teachers’ knowledge began to be carried out in the nineties (e.g.
Fennema & Franke, 1992; Broome, 1994) and currently continues to be developed with
great force. Studies in this line of research have focused on teachers’ knowledge of various
concepts, such as fractions (Llinares & Sanchez, 1991) and functions (Even & Markovits,
1991), and in different mathematical domains, such as algebra (McCrory, Floden, Ferrini-
Mundy, Reckase, & Senk, 2012) and geometry (Herbst & Kosko, 2012). We also find
proposals of models of teacher knowledge such as the Knowledge Quartet (Rowland,
Huckstep, & Thwaites, 2005), Mathematical Knowledge for Teaching (Ball, Thames, &
Phelps, 2008) and, more recently, Mathematics Teacher’s Specialized Knowledge
(Carrillo, Climent, Contreras, & Muñoz-Catalan, 2013). These models have been used for
studying mathematics teachers’ knowledge, principally in primary and secondary
education, with scarce accounts of studies of university teachers, as noted by Speer, King,
& Howell (2014).
Currently, it can be said that research in higher education has gone from being centered
on students to having a more balanced interest in both, students and teachers (Artigue,
2016), to the point that research on university teachers’ knowledge has emerged as a new
line of research from which it is asked what is understood as knowledge, how this
knowledge is developed, and how it is reflected in the teaching practice of university
teachers (Biza, Giraldo, Hochmuth, Khakbaz, & Rasmussen, 2016).

393 [Link]:indrum2018:171617
In agreement with the above and taking into account that the Mathematics Teacher’s
Specialized Knowledge model (MTSK) has been shown to be useful for studying
university teachers’ knowledge (e.g. Vasco, 2015), we carry out our research based on the
MTSK with the aim of characterizing the knowledge of a university mathematics teacher
who teaches content in the area of mathematical analysis. In this text, some results are
presented in relation to knowledge of the practice in mathematics, one of the sub-domains
of knowledge considered in the model.
MATHEMATICS TEACHER’S SPECIALIZED KNOWLEDGE
Mathematics Teacher’s Specialized Knowledge (MTSK) is an analytical model for
understanding mathematics teachers’ knowledge and at the same time a methodological
tool that allows analyzing teachers’ teaching practices (Carrillo et al., 2013). The model
was developed based on a theoretical, empirical, and reflective work proposed by
Shulman (1986) regarding foundational knowledge for teaching and the refining of
Mathematical Knowledge for Teaching (Ball et al., 2008).
In MTSK, two domains of teacher knowledge are distinguished, mathematical knowledge
(MK) and pedagogical content knowledge (PCK), considering that all of this knowledge
is specialized, that is, it derives from the teaching profession and is conditioned by
mathematics itself. Consequently, MTSK does not include, for example, general psycho-
pedagogical knowledge included in Shulman’s works. Also, the model takes into account
that teacher’s beliefs and conceptions about mathematics, its teaching, and its learning
permeate the organization and the use of knowledge (Carrillo et al., 2013).

Figure 1: Mathematics Teacher’s Specialized Knowledge (Carrillo et al., 2013).

394 [Link]:indrum2018:171617
Mathematical Knowledge comprises three sub-domains, knowledge of topics (KoT),
knowledge of the structure of mathematics (KSM), and knowledge of the practice in
mathematics (KPM).
KoT includes knowledge of concepts, properties, procedures, classifications, formulas,
and algorithms, with their respective meanings and foundations. For example, knowing
the topological property, density of the rational numbers in R, lies in this sub-domain.
KSM includes knowledge of the interconceptual connections that can be established
among mathematical concepts. So, knowledge of relationships between infinity and the
Archimedean property of the real numbers belongs to this sub-domain. KPM includes
knowledge of how to proceed, reason, and establish validity in mathematics. Knowledge
of how proofs are made using different methods is part of teachers’ KPM.
Pedagogical Content Knowledge contains three sub-domains: knowledge of mathematics
teaching (KMT), knowledge of features of learning mathematics (KFLM), and knowledge
of mathematics learning standards (KMLS).
KMT includes knowledge about didactic resources, strategies, tasks, and examples for
making mathematical contents understandable. A teacher using an example to illustrate
the meaning of a necessary condition forms part of his or her KMT. KFLM addresses the
teacher’s knowledge about mathematical contents as an object of learning, for example,
teacher’s knowledge about students’ difficulties in understanding proofs belongs to this
sub-domain. KMLS describes what students should achieve in a given course, conceptual
and procedural capacities and mathematical reasoning that are promoted in given
educational moments. The sequencing of the topics completeness theorem,
characterization of the greatest element, and the Archimedean property of the real
numbers is an example of KMLS.
KNOWLEDGE OF THE PRACTICE IN MATHEMATICS
The idea of knowledge of the practice in mathematics (KPM) comes from the works of
Schwab (1978), Ball (1990), and Ball & Bass (2009) regarding syntactic knowledge of
mathematics, which implies that the teacher should know how to reason mathematically,
know different kinds of reasoning, and know in which mathematical contexts a particular
kind of reasoning is more adequate than others. In this regard, within KPM, the importance
of the teacher not only knowing established mathematical results, but also how to proceed
and think in mathematics to arrive at these results is highlighted.
Knowledge that makes up part of this sub-domain is propositional logic, mathematical
language and its precision, how definitions are made and used in the construction of
mathematical knowledge, knowledge about different kinds of proof and their internal
logics, the role that examples and counterexamples play in proofs, different kinds of
heuristic reasoning, how knowledge is created in mathematics, how it is validated,
reasoned, and generalized, and the role of mathematical conventions and symbols.

395 [Link]:indrum2018:171617
Although this knowledge is important, research carried out in MTSK with teachers at
different educational levels reports scarce evidence of KPM in their mathematics classes
(e.g. Montes, 2014; Vasco, 2015). In this regard, this investigation contributes to the
understanding of the nature of this knowledge in university teachers.
METHODOLOGY
In this investigation, based on an interpretive paradigm and a qualitative methodology
(e.g. Denzin & Lincoln, 2000), we carry out an exploratory study such that our results
correspond to a first approximation for characterizing a university teacher’s knowledge of
the practice in mathematics (KPM). Our work is supported by an instrumental case study
(Stake, 1995) for which we chose as an informant a teacher who taught a real analysis
course for prospective teachers of mathematics during one semester in a Chilean
university.
The teacher, who we will call Diego, is a mathematics researcher with a Ph.D., with more
than 20 years of teaching experience, and this is the sixth time in recent years that he has
taught the real analysis course. These academic characteristics of Diego make it likely that
he possesses plentiful knowledge in elements of KPM.
The data was obtained through video recording while Diego taught the system of real
numbers. The video recordings were transcribed and later subjected to content analysis
(Bardin, 1997), identifying the units of analysis associated with the KPM sub-domain and
considering the differentiation between evidence and indication of knowledge (Moriel-
Junior & Carrillo, 2014). An evidence is an element that supports the presence of teacher
knowledge, while an indication provides suspicion of the existence of knowledge but
requires additional information in order to be confirmed as evidence.
RESULTS
In this section, we present episodes from one of Diego’s classes on the properties of real
numbers that allow us to observe his knowledge of the practice in mathematics (KPM).
Diego begins the class enunciating the property of the density of Q and R\Q in R. For the
proof of this property, Diego takes an interval [𝑎, 𝑏] with 𝑎 and 𝑏 in R and considers the
case 𝑎 = 0, then he generalizes it for any positive a. In fact, he takes 𝑎 and 𝑏 positive
numbers and indicates that with this supposition there is no loss of generality.
Teacher: We are going to assume that 𝑎 and 𝑏 are two positive numbers, which
is not a great assumption. If they were negative, for example, I work
with −𝑎 and −𝑏.
If I have 𝑎 and 𝑏 here [indicating on the number line], If I take −𝑎 and
−𝑏 then they are on the other side of the zero.

396 [Link]:indrum2018:171617
 And if I find a rational number here [indicating on the number line
between 𝑎 and 𝑏], then its opposite, will also be rational.
And if I find an irrational number, its opposite will be irrational.
And if one number is positive and the other is negative, then zero is
rational, and it is between them, and the irrational will come from a
proposition.
In the previous episode, Diego’s knowledge of the different cases that should be taken
into account to make a proof whose hypothesis possesses an implicit or explicit
disjunction is observed, as is his knowledge that a particular case can be sufficient for
showing the behavior of other cases in the proposition. Following this, the teacher’s
knowledge about the consideration of cases to particularize or generalize, is a way of
proceeding in mathematics and can be considered as an indicator of KPM. The teacher’s
knowledge that the proof goes beyond the example of a concrete case and addresses all
the possible cases the statement can include (e.g. Brodie, 2010; Montes, 2014) is part of
his KPM.
Additionally, Diego emphasizes that in mathematics for a fact to be considered valid it
must be proven:
Teacher: Now, a question, do I know of any irrational number that is less than 1
and greater than 0?
Student: The square root of square root of 2.
Teacher: Ah! Why is it less than 1? We have to prove that it exists, and, wait, we
have to prove that the root exists and that it is a number less than 1. And
prove something else, because they didn’t ask us for a number between
0 and 1, but rather for an irrational number between 0 and 1.
So, you can take a real number, prove that its root exits, but you still
don’t know if it is irrational or not.
In the prior episode, Diego’s knowledge of proof as a way of validating in mathematics is
shown. As Brodie (2010) maintains, in addition to knowing that a kind of proof exists that
confirms the truth of a statement, it is necessary to know how this type of proof works.
Furthermore, Diego points out the importance of this role of proofs (de Villiers, 1990), as
he is in front of a course for prospective mathematics teachers and after proving the
density property he says:
Teacher: So, now you can say to your students with certainty that between any
two rational numbers there is always an irrational number.
The comment above shows as indicator of KPM, the teacher´s knowledge of the necessity
and importance of proofs as a way of validating in mathematics (e.g. Balacheff, 2000).
Continuing with the density proof, Diego discusses with the students about taking the
maximum or minimum of a set of natural numbers. In the following episode, we observe
Diego’s knowledge about the convenience of selecting a certain element to develop an
argument in a proof.

397 [Link]:indrum2018:171617
 Teacher: The issue is that if I take either of the two [the maximum or
minimum] I can argue that there is a rational number in between
them.
What I have to guarantee before choosing one of the two is that in
fact there is something to choose, so, which is easier to guarantee?
The smallest or the largest?
Student: The smallest.
Teacher: Why?
Student: Because of the well-ordered principle.
Teacher: Of course, the minimum exists by the well-order principle.
Nevertheless, on occasions, rather than choose, is necessary construct an element that
allows developing an argument. Diego also gives evidence of this knowledge when
discusses with the students about irrationality of a number that belongs to the interval they
are working on. He uses √2 to construct √2⁄𝑚 as shown below.
Teacher: I know that √2 does not belong to the rational numbers. We proved it.
It is real, it exists, and it is not rational.
If I now divide this number by a whole number 𝑚, will it continue to be
not rational? will it become a rational number?
That is, this is not rational, but is this [pointing to √2⁄𝑚 ] not rational
either?
According to above episodes, an indicator of KPM is the teacher´s knowledge of the
construction or selection of elements for developing an argument in a proof as a way of
proceeding in mathematics.
Regarding to √2⁄𝑚, Diego prove its irrationality and establishes this affirmation as a
lemma that he uses in different moments in the proof of the density property:
Teacher: So, if I take the Archimedean property for 𝛿 = 𝜀⁄√2, then exists an 𝑚 ∈
𝑁 such that 0 < 1⁄𝑚 < 𝛿, and this implies that, √2⁄𝑚 is smaller than
𝜀 and does not belong to the rational numbers.
So, between 0 and 𝜀, no matter how small 𝜀 is, there is always a rational
number, and there is always an irrational number.
Ok, so this affirmation that I just wrote, we’re going to write it as a
lemma.
This episode gives us evidence that the teacher knows that establishing preliminary results
is a way of proceeding in mathematics that facilitates the development and the
communication of a long and/or complex proof.
On the other hand, regarding the existential quantifier present in the Archimedean
property and the well-ordered principle, Diego expresses that the existential quantifier
only gives information about the characteristics or properties of an element, but does not
say what the element is like or which element it is, only that it exists

398 [Link]:indrum2018:171617
 Teacher: When the for all quantifier is given, I can always choose an element that
works best for me, but when it is existence, it’s “whatever you’ve gotten”, no
more. Then as exists gives you “whatever you’ve gotten”, you have to do
some work so that what you get, is what you need.
An indicator of KPM in this episode is the teacher’s knowledge of the role and the
meaning of the quantifiers when they are found in the hypothesis or in the conclusion of
a proposition.
Summarizing, indicators of the teacher’s knowledge of the practice in mathematics related
to ways of reasoning, validating, and proceeding in mathematics has been evidenced. In
the exposed episodes, the teacher teaches mathematical reasoning in order to give meaning
to the mathematical activity (e.g. Brodie, 2010), not only for students to understand and
acquire sensibility regarding how to establish truth in mathematics, but also, they find
meaning in the need to do so (Montes, 2014).
CONCLUSIONS
A mathematics teacher is a professional whose knowledge of the discipline he or she
teaches has a level of deepening, organization, and structure that is greater than what the
students are going to receive (Ma, 1999). In this regard, KPM is necessary knowledge for
the teacher, as it provides logical thinking structures that help to understand the function
of diverse aspects of mathematics (Flores-Medrano, Escudero-Avila, Montes, Aguilar, &
Carrillo, 2014). As observed in the case studied, the teacher’s knowledge of processes of
particularization-generalization, of the necessity and importance of proof for validation,
and the knowledge of different ways of proceeding in mathematics are closely linked to
the transition to advanced knowledge (e.g. Pino-Fan, Godino, Castro, & Font, 2012) and
are related to the particular way of mathematical work.
In line with the ideas above, we consider that KPM allows the teacher to promote in
students the construction of mathematical knowledge and the acquisition of abilities for
reasoning, proof, and problem solving that are considered important for learning
mathematics at all levels of education (e.g. Flores-Medrano et al., 2014).
With reference to the indicators of KPM obtained in this investigation regarding ways of
reasoning, validating, and proceeding in mathematics, we agree with Sosa, Flores-
Medrano, & Carrillo (2015) that indicators shown directly in empirical data can contribute
to identifying, understanding, and analyzing teachers’ knowledge in their discipline.
Given that teacher knowledge is a complex and multidimensional construct, more studies
are necessary to deepen the understanding of its different components, in order to advance
in interrogations proposed regarding how this knowledge is developed and how it is
reflected in the teaching practice of university teachers (e.g. Biza, Giraldo, Hochmuth,
Khakbaz, & Rasmussen, 2016).

399 [Link]:indrum2018:171617
ACKNOWLEDGEMENTS Beca de Doctorado año 2017 CONICYT, folio 21170442.
REFERENCES
Artigue, M. (2016). Mathematics Education Research at University Level: Achievements
and Challenges. In E. Nardi, C. Winslow, & T. Hausberger (Eds.), Proceedings of the
First Conference of International Network for Didactic Research in University
Mathematics (pp. 11-27). Montpellier, France.
Balacheff, N. (2000). Procesos de prueba en los alumnos de matemáticas [Proof process
in mathematics students]. Bogotá: Una empresa docente. Universidad de Los Andes.
Ball, D. (1990). The mathematical understandings that prospective teachers bring to
teacher education. The Elementary School Journal, 90(4), 449-466.
Ball, D., & Bass, H. (2009). With an eye on the mathematical horizon: Knowing
mathematics for teaching to learners’ mathematical futures. In The 2009 Curtis Center
Mathematics and Teaching Conference. University of California, LA.
Ball, D., Thames, M. H., & Phelps, G. (2008). Content knowledge for teaching: What
makes it special? Journal of Teacher Education, 59(5), 389-407.
Bardín, L. (1997). Content analysis. São Paulo: Livraria Martins Fontes.
Biza, I., Giraldo, V., Hochmuth, R., Khakbaz, A., & Rasmussen, C. (2016). Research on
Teaching and Learning Mathematics at the Tertiary Level: State-of-the-art and
Looking Ahead. Switzerland: Springer.
Brodie (2010). Teaching mathematical reasoning in secondary school classrooms. New
York: Springer.
Bromme, R. (1994). Beyond subject matter: A psychological topology of teachers’
professional knowledge. In R. Biehler, R. Sholz, R. Strässer, & B. Winkelman (Eds.),
Didactics of Mathematics as a Scientific Discipline (pp. 73-88). Dordrecht: Kluwer
Academic Publisher.
Carrillo, J., Climent, N., Contreras, L. C., & Muñoz-Catalán, M. C. (2013). Determining
Specialised Knowledge for Mathematics Teaching. In B. Ubuz, Ç. Haser, & M. A.
Mariotti (Eds.), Proceedings of the CERME 8 (pp. 2985-2994). Antalya, Turkey.
de Villiers, M. (1990). The role and function of proof in mathematics. Pythagoras, 24, 17-
24.
Denzin, N., &. Lincoln, Y. (Eds.), The handbook of qualitative research. London: Sage.
Even, R., & Markovits, Z. (1991). Teachers’ pedagogical knowledge: the case of
functions. In F. Furinghetti (Ed.), Proceedings of the 15th PME international
conference, (Vol.2, pp. 40-47). Assisi, Italy.

400 [Link]:indrum2018:171617
Fennema, E., & Franke, M. (1992). Teachers’ knowledge and its impact. In D. A. Grows
(Ed.), Handbook of research on mathematics teaching and learning: A project of the
NCTM, (pp.147–164). New York: MacMillan Publishing.
Flores-Medrano, E., Escudero-Avila, D., Montes, M., Aguilar, A., & Carrillo, J. (2014).
Nuestra Modelación del Conocimiento Especializado del Profesor de Matemáticas, el
MTSK. [Our modelling of Mathematics Teacher’s Specialized Knowledge, the MSKT]
In J. Carrillo, N. Climent, L.C. Contreras, M. Montes, D. Escudero-Ávila, & E. Flores-
Medrano (Eds.), Un Marco Teórico para el Conocimiento Especializado del Profesor
de Matemáticas [A Theoretical Framework for the Mathematics Teacher’s Specialized
Knowledge] (pp. 71-93). Huelva: Universidad de Huelva Publicaciones.
Herbst, P., & Kosko, K. (2012). Mathematical knowledge for teaching high school
geometry. In L. R. Van Zoest, J. J. Lo, & J. L. Kratky (Eds.), Proceedings of the 34th
PME-NA (pp. 438-444). Kalamazoo, MI: Western Michigan University.
Llinares, S., & Sanchez, M. (1991). The knowledge about unity in fractions tasks of
prospective elementary teachers. In F. Furinghetti (Ed.), Proceedings of the 15th PME
international conference, (Vol 2, pp. 334-341). Assisi, Italy.
Ma, L. (1999). Knowing and teaching elementary mathematics: Teachers’ understanding
of fundamental mathematics in China and the United States. Mahwah, NJ: Erlbaum.
McCrory, R., Floden, R., Ferrini-Mundy, J., Reckase, M., & Senk, S. (2012). Knowledge
of algebra for teaching: A framework of knowledge and practices. Journal for Research
in Mathematics Education, 43(5), 584-615.
Montes, M. (2014). Conocimiento especializado del profesor de matemáticas acerca del
infinito. Un estudio de caso [Mathematics Teacher’s Specialized Knowledge about
infinity. A case study] (Doctoral Dissertation). University of Huelva, Spain.
Moriel-Junior, J. G., & Carrillo, J. (2014). Explorando indícios de conhecimento
especializado para ensinar matemática com o modelo MTSK. [Exploring indications of
specialized knowledge for mathematics teaching through MTSK model] In M. T.
González, M. Codes, D. Arnau, & T. Ortega (Eds.), Investigación en Educación
Matemática XVIII (pp. 465-474). Salamanca: SEIEM.
Pino-Fan, L., Godino, J., Castro, W., & Font, V. (2012). Conocimiento didáctico-
matemático de profesores en formación: Explorando el conocimiento especializado
sobre la derivada. [Prospective teachers’ didactic-mathematics knowledge: Exploring
the specialized knowledge about the derivative] In A. Estepa, Á. Contreras, J. Deulofeu,
M. C. Penalva, F. J. García, & L. Ordóñez (Eds.), Investigación en Educación
Matemática XVI, (pp. 427 - 434). Jaén: SEIEM.

401 [Link]:indrum2018:171617
Rowland, T., Huckstep, P., & Thwaites, A. (2005). Elementary teachers’ mathematics
subject knowledge: The knowledge quartet and the case of Naomi. Journal of
Mathematics Teacher Education, 8(3), 255-281.
Schwab, J. (1978). Education and the structure of the disciplines. In I. Westbury & N. J.
Wilkof (Eds.), Science, curriculum, and liberal education (pp. 229-272). Chicago, IL:
University of Chicago Press.
Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching.
Educational Researcher, 15(2), 4-14.
Sosa, L., Flores-Medrano, E., & Carrillo, J. (2015). Conocimiento del profesor acerca de
las características de aprendizaje del álgebra en bachillerato. [Teacher knowledge about
the features of learning algebra in high school]. Enseñanza de las Ciencias: revista de
investigación y experiencias didácticas, 33(2), 173-189.
Speer, N., King, K., & Howell, H. (2014). Definitions of mathematical knowledge for
teaching: using these constructs in research on secondary and college mathematics
teachers. Journal of Mathematics Teacher Education, 18(2), 105-122.
Stake, R. (1995). The art of case study research. London: Sage.
Vasco, D. (2015). Conocimiento especializado del profesor de álgebra lineal: un estudio
de casos en el nivel universitario [Teacher specialized knowledge of linear algebra: A
case study in university level] (Doctoral Dissertation). University of Huelva, Spain.

402 [Link]:indrum2018:171617
 Exploration of new post-secondary mathematics teachers’
experiences: preliminary results of a narrative inquiry
Sarah Mathieu-Soucy1, Claudia Corriveau², and Nadia Hardy3
1
Concordia University, Montreal, Canada, [Link]@[Link]; ²Laval
University, Quebec, Canada; 3Concordia University, Montreal, Canada
This paper reports on a pilot study that has taken place during the winter semester of
2017, in the context of a larger project whose goal is to contribute to studying the
transitions from being a university mathematics student to becoming a post-
secondary mathematics teacher. With very scarce literature on these transitions at
this specific level, this pilot study acts as an exploration into new teachers’
significant experiences that may be involved in shaping their relationships with
mathematics, and its teaching and learning. We conducted a narrative inquiry with
three new post-secondary mathematics teachers who were interviewed on a regular
basis during a semester. Those interviews provided an insight into the new teachers’
experiences by pointing out themes that are relevant to them. We conclude with a
discussion on what remains to be achieved to conduct this research.
Keywords: Teachers’ and students’ practices at university level, Preparation and
training of university mathematics teachers, Narrative Inquiry, Becoming a teacher.
INTRODUCTION
This paper presents a pilot study that took place during the winter semester of 2017.
It addresses the transition from being a university mathematics student to
becoming/being a post-secondary mathematics teacher. More precisely, the context is
set in cegep institutions (general and vocational colleges), the first step in post-
secondary education in the province of Quebec, Canada. The focus of this study is on
new cegep mathematics teachers who, with an education mainly – sometimes
exclusively – in mathematics, negotiate the transition from being a mathematics
student (undergraduate or graduate) to teaching at post-secondary level. We start with
a brief description of cegep institutions, followed by a discussion on the literature on
becoming a teacher at post-secondary level. The theoretical framework and the object
of study are then described followed by the methodology and the main results. We
conclude with the next steps in the study and our expectations for the results.
A Few Words on the Context of Study
Cegeps are general and vocational colleges. Two-year pre-university programs are
offered in sciences, arts or social sciences; three-year technical programs, such as
nursing, computer science, building engineering technology, etc., are also offered. All
programs include compulsory general courses, such as philosophy, French, English
and physical education and multiple programs offer mathematics courses. Some
technical programs include mathematics courses that are specific to the field of study.
Those courses are offered mostly by mathematics instructors, even if they are not

403 [Link]:indrum2018:174832
very familiar with the profession their students are aspiring to do. Pre-university
programs in sciences and social sciences offer two calculus courses (differential and
integral calculus) and one linear algebra course, slightly different from one program
to the other, depending on the institution. Social science students are required to take
a quantitative methods course. Some institutions also offer optional mathematics
courses for science students, such as multivariable calculus, probability and statistics
or discrete mathematics. Therefore, from advanced calculus for science students to
quantitative methods for social science students, as well as mathematics applied to
computer science, cegep mathematics teachers are required to teach a wide array of
courses, to a wide range of programs and students.
To teach in a cegep institution, individuals are (officially) required a 3-year
undergraduate degree in mathematics or connected field (Conseil Supérieur de
l’Éducation, CSÉ, 2000). However, each institution can add other requirements, such
as a master or a doctorate in mathematics. A handful of universities in the province
offer a one-year graduate certificate in cegep teaching, unique for all disciplines
taught in cegep. Institutions will sometimes see it as an asset when hiring although
cegep instructors who have completed one of these certificates qualify them as far
from the reality of cegep teaching (CSÉ, 2000).
This context is quite different than that of elementary and secondary institutions. To
teach at these institutions individuals are officially required a 4-year degree in
education, which sometimes include minimal formal mathematics training. Also,
cegep institutions are often a place of transition from secondary to university
education; students are introduced to more formal mathematics and some level of
autonomy is expected from them (schedule less structured, no mandatory attendance
to classes).

Teaching at Post-Secondary Level

Little is known about becoming a mathematics teacher at post-secondary level (e.g.
Speer & Hald, 2008). More and more research about post-secondary mathematics
education is being conducted, arguing this level is key to the training not only of
future teachers but also scientists, engineers and mathematicians (Hodgson, 2001).
However, many recognize the need for change in the way the mathematics are taught
at this level. Most research emphasizes the mistakes and negative traits of university
teaching and claim the inadequacy of the existing training programs on creating
change and improving post-secondary teaching, while only few seem to offer
concrete solutions (e.g. Beisiegel, 2009; Belnap, 2005; Speer, 2001; Speer, Gutmann
& Murphy, 2005). About this, Beisiegel’s work (2009) focused on how graduate
students “developed a sense of themselves […] as post-secondary teachers of
mathematics” (p. 2) during a teaching assistantship. The author looks closely into the
graduate students’ lives and their journey into becoming teachers. She concludes that
existing training is inadequate and puts forward the need to investigate closely the
experiences of these students becoming new teachers in order to effectively study

404 [Link]:indrum2018:174832
post-secondary teaching and training (2009, p. 25). This recommendation is the
starting point for our work with new cegep teachers who were not so long-ago
university mathematics students. Our goal is to better understand their experiences in
their life as new mathematics teachers and in the transition from being a mathematics
student to becoming a post-secondary teacher.
The challenge – and partially the trigger for our pilot study – is that the literature on
post-secondary teaching in general, and in cegep in particular, is very scarce. Looking
at research on elementary and high school teachers, we see that it is heavily based on
the education they received, namely a degree that aims at training them as teachers
(e.g. Ambrose, 2004; Ensor, 2001; Franke & al., 1998). However, post-secondary
mathematics teachers are mainly, if not solely, trained to become mathematicians.
Therefore, we cannot transpose or extend the results encountered in the literature
about teachers’ training for elementary and secondary levels to the group we are
interested in. With this lack of background on the transition under study, we chose to
conduct a pilot study to inform our main research project.

THEORETICAL FRAMEWORK AND OBJECT OF STUDY

The focus on experience of this study led us to consider Dewey’s philosophy (1938)
to frame it. For Dewey, one learns through experience and experiences shape how
one goes about the world and about new experiences (1938, p. 35). Dewey also sees
being faced with challenging experiences as a key aspect of life, growth and change:
“growth depends upon the presence of difficulty to be overcome by the exercise of
intelligence” (1938, p. 79). Indeed, faced with a familiar context, one can have an
idea of how to act, and the consequences of those actions, based on knowledge
acquired through past experiences. However, more reflection is needed in an
unfamiliar context, where one could have to connect many different, apparently
unrelated, experiences in order to know how to act and the related consequences
(1938, p. 68). This results in new knowledge that could be applied to one’s action per
respect to future experiences.
Because of the importance of new and challenging experiences in life, growth and
change, we chose them to be the focus of our study. In the context of our research,
becoming a mathematics teacher at post-secondary level is seen as a trigger for new
and possibly challenging experiences. And because a situation may be new and
challenging for some but not for others depending on one’s past experiences, it is
important to put the focus on the experiences lived by the new teachers that are
challenging and educational for them, and not all experiences that are lived or that
might be assumed challenging and educational by others. We call those significant
experiences.
However, this approach opens the door to a broad and varied spectrum of experiences
that may be significant in the transition under study. In our project, we focus on
significant experiences that may shape new teachers’ relationships with mathematics
and its teaching and learning. We will use the abbreviation RWMTL for the

405 [Link]:indrum2018:174832
remaining of this paper to refer to the relationships under study. These RWMTL
entail the visions, opinions, beliefs, attitudes and feelings about mathematics, and its
teaching and learning.
In this sense, we hypothesize that new teachers possess a RWMTL, developed
through years of doing, learning, being taught, and sometimes teaching mathematics,
and that these RWMTL play an important role in their becoming post-secondary
mathematics teachers. This hypothesis is supported by the work of Speer (2001) and
Beisiegel (2009). Speer’s work emphasizes that graduate mathematics students have
beliefs about mathematics and mathematics education, and that their RWMTL play a
role on how new teachers will act in their new profession, how they will deal with,
interpret and react to what happens to them. In particular, she claims that teaching
assistants’ beliefs, especially about mathematics and undergraduate students, have a
significant impact on how they teach and interact with students. These RWMTL are
influenced by the education they received in university and by the implicit teacher
training they received during their time as students (Beisiegel, 2009, p. 42). It creates
a relationship, conscious or not, with education through their experiences as students.
Indeed, Beisiegel (2009) claims that graduate mathematics students’ life experiences
in a mathematics department, shape their “views of the discipline and teaching”
(p. 43), as well as how they view what and how they should become. She argues that
the experiences those students have can play an important role into how they define
their role as mathematics teachers: “it appears that in the lives of mathematics
graduate students there exists a complex and intricate interplay between the structures
that they encounter, their feelings about mathematics and themselves and their ideas
of their future role as mathematics instructors or professors” (p. 43). This last quote
reinforces our hypothesis that the RWMTL play an important role in individuals
becoming postsecondary mathematics teachers. In general terms, the goal of our main
study is to investigate these RWMTL; in particular, our goal is to identify the nature
of the significant experiences that shaped new teachers’ RWMTL. Incidentally, the
goals of the pilot study this paper is reporting on were to come up with a list of
themes relevant to, or in relation with, the significant experiences of the new
mathematics cegep teachers, in order to support and guide the main study later on.

METHODOLOGY
Our methodology is based on ideas brought forward by narrative inquiry (“NI”,
Clandinin & Connelly, 2000). NI argues that the essence of human experience
happens narratively and choosing this method means acknowledging that people
make sense and give meaning to their lives narratively – they lead storied lives
(Clandinin, 2013, p. 13). Furthermore, NI takes into account the wholeness of
someone’s life while allowing the researcher and the subject to collaboratively
investigate and distinguish what makes it unique and specific.

NI was developed specifically to understand and inquire into experiences, in relation

to the people who have them and the physical and social context where they are

406 [Link]:indrum2018:174832
(Clandinin, 2013). This is to emphasize how past experiences and social and
individual matters influence how someone lives an experience. In our case and for the
research project, we use NI to determine what significant experiences shape new
teachers’ RWMTL. In the context of the pilot study we are reporting on, we use NI to
shed a first light into these significant experiences; collaboratively working with the
subjects in thematising them, their nature and their shaping role.

With this in mind, we built open and semi-structured interviews with broad questions,
so teachers would account for what is actually relevant or important in their journey
of becoming teachers. This gives a chance for the unexpected to arise.

The Process of Investigation

Weekly meetings were planned with three cegep teachers during a whole semester
(January to May 2017), but we ended up meeting with each of them 15, 10 and 4
times respectively over the whole semester. The meetings, which lasted between 30
minutes and 90 minutes, were audio-recorded. The first teacher was starting his third
year as a teacher, had completed a master in mathematics and a one-year certificate in
pedagogy. The second participant was starting his second year of teaching, had
completed a master in mathematics and had also completed a one-year certificate in
pedagogy. The third participant was teaching his second course while pursuing their
second year of Ph.D. in mathematics. They were asked to share stories of events they
lived in the week prior to the meeting, which they identified as significant for their
RWMTL, and of reflections they made about these events. After each meeting, an
account was written in the form of a story of the meeting, using as much as possible
the words of the teacher, with only slight changes for clarity purposes. Those
accounts constitute the data that was analyzed.

RESULTS
Emerging themes
The pilot study served as a way for us to learn about significant experiences of new
teachers in relation with mathematics, teaching and learning. Our goal was to
circumscribe areas, contexts or topics, we call them themes to be concise, which
seemed to play an important part into their lives, thoughts and reflections. We wanted
to find themes that seemed to foster experiences that were significant for them. This
list is not meant as “a list of understandings” but rather as a list of words to help us
think to understand the stories (Clandinin, 2013, p. 39). Of course, those ideas will
not be final as we do want to stay open to what will come up as we meet new people
next year, for the main study, and learn about their experiences.
First, as we considered the teachers’ relationship with teaching, two categories arose:
“being a teacher” (1) and “teaching to students” (2). The former (1) addresses the
very personal aspects of the journey into becoming a teacher. It includes themes such
as the new teachers’ expectations of the profession of teaching mathematics in a

407 [Link]:indrum2018:174832
cegep institution. This unfolds in reflections in relation to whether or not they
attended cegep as students, the lifestyle they expect to have while holding such a
position and the expectations arising from their memories of their own teachers.
Another theme that came up is about being part of an institution and being part of a
community of teaching, who respectively involves reflection on the role of cegeps in
the society, and integrating in a team of established teachers. Finally, the ability to
teach, which includes their reflections in relation to their own and others’ abilities,
and the initial training they should have or want to have, were topics that played a
role in their everyday life and reflections.
The latter subdivision (2) of the themes in relation with teaching that we found
addresses themselves as teachers in relation with their students. This includes the role
they (should, can) have in their classroom, whether it is to pass on knowledge, make
it interesting or getting them prepared for a job or university. They also reflected on
the assessment of the level of difficulty of concepts or problems as very central in
their daily life, whether it comes from their own judgment or from a formal source
such as ministerial specifications. In the same line of thought, reflections on the
assessments were playing a huge role in their experiences with questions such as how
many, when, weight of each, and also how to prepare the students for them and how
to assess exactly what needs to be assessed. Finally, their teaching method, using
technology for example, and the ability to adjust to the students, to their ways of
thinking and being, were present in the reflections expressed.
As we considered the teachers’ relationship with learning, the following themes
emerged. First, the expectations of the level of the students, as far as what they should
know from high school or from previous cegep courses, and the expectations in
relation to the students fulfilling their role in the classroom, such as participating in
class and doing their homework, were central topics for the teachers. On another
note, the teachers found challenging to manage the students’ expectations in regard
of their teacher, as far as the act of teaching and the level of the material they were
expecting. Finally, the teachers mentioned choosing the attitude towards the students
as being a challenge, especially when it comes to differences between countries and
provinces, since many teachers are not from Quebec, where the relationship between
a teacher and its students might be different according to different cultural norms.
As far as the teacher’s relationship with mathematics, we did not come up with a list
of themes as we found that our data did not allow us to do so. Indeed, as we met the
participants over and over, we realized that it was not spontaneous for the new
teachers to talk about mathematics when asked about their daily lives. They favoured
talking about teaching and their students. We conjecture that they have a number of
things to think about other than the mathematics, such as the themes mentioned
above, which seem to take much more work to be mastered or managed than to
master the mathematics they are teaching. In other words, it made us realize that for
(these) new teachers, it is difficult to talk about mathematics spontaneously when
questioned about their experiences as new teachers; considerations in relation to

408 [Link]:indrum2018:174832
teaching and learning seem to take all their time. Among the three participants, only
one would address mathematics when discussing their everyday activities; the
individual still in the midst of completing their Ph.D. in mathematics. This made us
think that the new teachers that were full-time, not actively conducting research in
mathematics nor taking mathematics courses, seemed to have in the front of their
mind matters that were farther from mathematics and closer to teaching and (their
students’) learning. The two other participants, when asked directly, struggled to talk
about mathematics for more than a few moments.
Finally, we found multiple themes what were present across mathematics, teaching
and learning. Indeed, it seemed that some institutional aspects specific to cegep were
having an impact on the teachers experience at multiple levels. First, as mentioned in
the introduction of this paper, cegep institutions offer science, social science and
technical programs. The teachers expressed multiple times that there were challenges
on multiple levels to teaching to different audiences. For example, in a differential
calculus class for science, a teacher has to teach to future engineers and future
doctors, two groups who do not have the same interests and goals. Also, this same
teacher may teach in the same semester differential calculus for social science
students (the course covers almost the same material), who have again very different
interests, goals and aspiration, to study psychology or economics for example, in
relation with mathematics and, for some of our participants, very different ways of
thinking. Finally, that same teacher may have to teach mathematics to technical
programs, where the goal is for the students to learn the mathematics they would need
to apply in their field. Therefore, the teachers we met expressed how difficult it was
to navigate teaching to those different audiences: how to teach them according to
their expectations and aspirations, how they learn best, and to determine what kind of
mathematics, and what in the mathematics, they really needed.
Another key aspect across mathematics, teaching and learning is the fact that cegeps
offer courses during the day and courses at night as part of continuing education.
Indeed, teaching in continuous education meant that the teacher was not assigned an
office space, and was not active in the department (they were not necessarily invited
to department meetings). This also meant that they were often left on their own to
discover the ways of the institution, would rarely come across other instructors and
seemed to lack opportunities for asking advice. It is also important to mention that
new cegep teachers often have to teach at multiple institutions during the first few
years, sometimes even during a semester, before they can be guaranteed work at one
place. This isolation and constant change seemed to be heavy for teachers who enjoy
working with others, reflect through discussions on teaching, get advice from more
experienced colleagues or just be active in their workplace. The students are also
different, as most day-students are young adults who go to school full-time and,
night-students are older and work full-time during the day. And again, to try to adapt
to all those aspects turned out to foster a number of significant experiences for our
participants.

409 [Link]:indrum2018:174832
CONCLUSION: WHAT’S NEXT
This pilot study helped us understand and circumscribe themes that were at the heart
of becoming a teacher in cegep institutions.
However, through the 29 interviews we conducted, very few experiences were
complete enough to be talked about in terms of how they played a role in the new
teachers’ lives, mainly because the reflection in relation to an experience was often
missing or incomplete. Indeed, from our data, we found that the reflective experience
can take a long time, longer than the academic term in which the experience
happened. This made us realize that our goal of understanding changes in the
RWMTL could not be tracked over this short period of time. Indeed, our pilot left us
believing that reflection on recent experiences is often not mature enough for teachers
to be comfortable verbalizing it and sharing it with us. Also, we were unable to know
if the recent experiences would play a significant role in the transition under study on
a longer term. Therefore, to be able to hear about significant experiences, we need to
go another route. Our conjecture is that by introducing a theme first and asking
participants to share past experiences in relation to this theme, we could be able to
grasp some of these significant experiences and the role they play in shaping new
teachers’ RWMTL.
Indeed, by asking the participants to talk about experiences that have shaped their
RWMTL in relation to a specific area, context or topic, we hope to hear about the
experience and the reflection associated to it. As Dewey (1938) would say, the
quality of an experience comes mostly from what it opens to, i.e. the resulting tools
and how they help understanding and acting towards new experiences (p. 27). This
takes time, since we think about our past experiences differently with time, reflection
and probably other experiences. We therefore want to understand how the new
teachers evolved in this transition, how their way of living and acting towards new
experiences evolved with time and new experiences. As Clandinin (2013) said, we,
the participants and the researchers, “are always interpreting [our] pasts from [our]
present vantage points” (p. 46).
The pilot interviews also revealed that the participants had a hard time focusing on
experiences specifically related to mathematics and often tended to discuss
experiences related to class management, institutional constraints, etc. This lead us to
design new protocols with the hope of instilling the norm that the goal is to talk about
experiences with mathematics, and its teaching and learning.
To conclude, the pilot study helped us in creating protocols for the main study that
would guide our endeavour to understanding some of the experiences that shape new
cegep teachers’ RWMTL, and more precisely, what kinds of experiences are
significant as they transition from being mathematics students to cegep teachers, all
while staying open to the new lives we will meet as the main data collection unfolds.

410 [Link]:indrum2018:174832
REFERENCES
Ambrose, R. (2004). Integrating change in prospective elementary school teachers’
orientations to mathematics teaching by building on beliefs. Journal of
Mathematics Teacher Education, 7, 91-119.
Beisiegel, M. (2009). Being (Almost) a Mathematician: Teacher Identity Formation
in Post-Secondary Mathematics. Doctoral thesis, University of Alberta.
Belnap, J. (2005). Putting TAs into context: Understanding the graduate mathematics
teaching assistant. Unpublished doctoral dissertation, University of Arizona.
Clandinin, D. J., & Connelly, F. M. (2000). Narrative Inquiry: Experience and story
in qualitative research. The Josey-Bass education series, a Wiley Imprint, San
Francisco.
Clandinin, D. J. (2013). Engaging in Narrative Inquiry. California, Left Coast Press.
Conseil Supérieur de l’Éducation [CSÉ] (2000). La formation du personnel
enseignant du collégial : un projet collectif enraciné dans le milieu. Québec:
Conseil Supérieur de l’Éducation.
Dewey, J. (1938). Experience and Education. New York: Collier Books.
Ensor, P. (2001). From preservice mathematics teacher education to beginning
teaching: A study in recontextualizing. Journal for Research in Mathematics
Education, 32(3), 296-320.
Franke, M., Fennema, E., Carpenter, T., Ansell, E. & Behrend, J. (1998).
Understanding teachers' self-sustaining, change in the context of professional
development. Teaching and Teacher Education, 14, 67-80.
Hodgson, B. (2001). The mathematical education of schoolteachers: Role and
responsibilities of university mathematicians. In D.A. Holton (Ed.) The Teaching
and Learning of Mathematics at the University Level: An ICMI Study (pp. 501-
518). Boston, MA: Kluwer Academic Publishers.
Speer, N. (2001). Connecting teaching beliefs and teaching practices: A study of
teaching assistants in reform-oriented calculus courses. Unpublished doctoral
dissertation, University of California, Berkeley.
Speer, N, M., Gutmann, T. & Murphy, T. J. (2005). Mathematics teaching assistant
preparation and development. College Teaching, 53(2), 75-80.
Speer, N. M. & Hald, O. (2008). How do mathematicians learn to teach? Implications
from research on teachers and teaching for graduate student professional
development. In M. C. & C. Rasmussen (Eds.), Making the Connection: Research
to Practice in Undergraduate Mathematics Education (pp. 305-318). Washington:
Mathematical Association of America.

411 [Link]:indrum2018:174832
 Inquiry-based learning and pre-service teachers education
Josep Gascón1 and Pedro Nicolás²
1Universitat Autònoma de Barcelona, España; ²Universidad de Murcia, España pe-
dronz@[Link]
We use Hintikka’s analysis of the notion of inquiry to point out some obstacles that
the implementation of inquiry-based learning might find, and to suggest a way to
overcome these difficulties via the use of the Reference Epistemological Models deve-
loped in the framework of the Anthropological Theory of the Didactic.
Keywords: inquiry-based learning; teacher’s and students’ practices; novel approa-
ches to teaching
INQUIRY-BASED LEARNING
Inquiry-based learning (IBL), roughly conceived as a teaching method in which stu-
dents are invited to learn in a way similar to that of a scientist or a mathematician,
has been suggested, with different levels of precision, by several researchers, institu-
tions and pedagogical and didactic approaches. Artigue and Blomhøj (2013) include
as instances of IBL the teaching proposals made from the following approaches in
Mathematics Education: problem-solving, theory of didactical situations, realistic
mathematics education, modelling perspectives, anthropological theory of the didac-
tic, and dialogical and critical approaches. For its part, European Union, through re-
ports prepared by experts (see, for instance, (Rocard, Csermely, Jorde, Lenzen, Wal-
berg-Henriksson, & Hemmo, 2007)) and projects (see, for instance, PRIMAS, http://
[Link]), has supported also the implementation of IBL in educative
european institutions.
It would be incorrect to believe that these IBL proposals have been suggested as a
means to reach the very same educational end. Let us mention some proposal aiming
different ends. The Theory of Didactical Situations proposes the IBL, embodied in
the notion of situation, as a means to achieve a real knowledge of mathematics
(Brousseau, 1997, p. 22). For its part, the Anthropological Theory of the Didactic
proposes the IBL, via the notion of Study and Research Path, as a means to transform
the cognitive ethos of our society, that is to say, to provide new attitudes and habits
with respect to the acquisition of knowledge (Chevallard, 2015). European Union, in
reports and projects like the ones mentioned above, stands up for IBL as a means to
remedy the declining interest of youth in science, confirmed by some OECD reports,
and the subsequent lack of technological innovation in Europe. According to this
point of view, IBL would contribute to change science and mathematics learning into
a motivating activity. On the other hand, Europen Union also promotes IBL, for ins-
tance in the PRIMAS project, as a means to prepare students for a future in which “it
is no longer sufficient (…) to learn facts” but “to be able to solve non-routine pro-
blems, to analyse data, to discuss with colleagues, to communicate their result and to
work autonomously” (Maaß & Reitz-Koncebovski, 2013, p. 10). Therefore, the ends

412 [Link]:indrum2018:174392
to which different proposers aim to arrive through IBL are as varied as the very ends
of regulated education: the acquisition of knowledge, habits, attitudes, values, etc.
KINDS OF KNOWLEDGE
To analyse possible uses of IBL in the acquisition of knowledge we have to wonder:
are there different kinds of knowledge? In Epistemology it is customary to distin-
guish between three kinds of knowledge (Ichikawa, Steup, 2017; Fantl, 2017): know-
ledge by acquaintance, knowledge how and propositional knowledge. Knowledge by
acquaintance is the kind of knowledge you have when you can identify something or
someone (for instance, the label criterion of divisibility by 3) and the corresponding
name, description, formulation, etc. (in this case: a natural number is divisible by 3 if
and only if the result of adding all its digits is divisible by 3). Knowledge-how is the
kind of knowledge you have when you carry out a series of intentional actions (for
example, to apply the criterion of divisibility by 3) towards the attainment of an end
(to know whether a given number is divisible by 3, in this case). Finally, propositio-
nal knowledge is the kind of knowledge you have when you know why a certain pro-
position is true (for example, when you know why it is true the criterion of divisibilty
by 3). Notice that in the knowledge by acquaintance you also seem to know the truth
of a proposition (for instance, the proposition asserted by the statement “The criterion
of divisibility by 3 is the statement A natural number is divisible by 3 if and only if
the result of adding all its digits is divisible by 3”). But in this case this is just a con-
tingent truth, existing by convention.
INQUIRY AND PROPOSITIONAL KNOWLEDGE
In order to analyse possible obstacles to the implementation of IBL, we shall use a
theoretical model of the notion of inquiry, namely, the one developed by the logician
and philosopher Jakko Hintikka and collaborators in several works. Of course, one
can also find other theoretical models of the notion of inquiry in some approaches to
Mathematics Education. Anyway, here we choose the one provided by Hintikka due
to its enlightening use of Logic, which seems to be an essential ingredient for a true
comprehension of an inquiry process.
Interrogative Model of Inquiry
To get a better understanding of the concept inquiry, as claimed by Hintikka (1982),
is not enough to study individual acts instantiating (…) whatever the concept in question
may be. We also have to study the more complicated rule-governed behavioral complexes
in which their “logical home” is.
For this, Hintikka has developed a game-theoretical model, the so-called Interrogati-
ve Model of Inquiry (IMI), which presents a formal approach to inquiry. A good refe-
rence for this model, which has been explained here and there, is (Hintikka, Halo-
nen, & Mutanenet, 2002). In few words, inquiry is regarded as a game with two pla-
yers: Inquirer and Nature. The game starts with a pair, (T, Q), where T is a given
theoretical premise, and Q is a question. The game finishes when the Inquirer finds

413 [Link]:indrum2018:174392
an answer to Q. Along the game the Inquirer is allowed to make two kinds of moves:
questions to Nature, and deductions. The only moves of Nature are answers to Inqui-
rer’s questions. We will not explain here all the details of Hintikka’s IMI, but at least
we will point out two of its main features:
1) The principal goal of an inquiry is the acquisition of propositional knowledge.
2) Inquiry is a process in which a complex dialectics between deductions and ques-
tions takes place.
We emphasize here these two aspects because, as we hope to show below, they seem
to be crucial for practical didactic considerations about IBL.
The goal of inquiry is to acquire propositional knowledge
The inquiry finishes if we can deduce (based on the premise and on Nature’s answers
to our questions) an answer A to the initial question Q (Hintikka et al., 2002, Theo-
rem 1). Of course, the deduction procedure, which is carried out through a series of
familiar rules (Hintikka et al., 2002, § 2), has to do with semantics. Indeed, these ru-
les are such that if the premise T and Nature’s answers are true in a given model M,
then the answer A is also true in this model M. This is why we can read in (Hintikka,
1996, p. 38) that what the winning of an “interrogative game” shows is “knowledge
of truths”.
Later, we will speak of the implications of the presence of the model M in the use of
IBL for teaching. For the moment, I would like to stress the fact that, according to
Hintikka’s IMI, the result of an inquiry is a proposition. Inquiry is a quest for propo-
sitional knowledge, which is the kind of knowledge linked to non-contingent truth
and falsehood. This does not prevent, in the course of an inquiry, the acquisition of
knowledge-how and knowledge by acquaintance. But we would not consider an in-
quiry finished if, along this inquiry, we would have found a successful technique
(knowledge-how), whose success remains unexplained. In other words, in the course
of an inquiry we can get knowledge-how, but we then would pursue the correspon-
ding propositional knowledge able to explain the success (and even its limitations,
portability to other contexts, etc.) of this knowledge-how.
Dialectics between deductions and questions
In a game of inquiry not all the questions can be asked at any moment. On the con-
trary, one is forced to pay attention to the presuppositions of that questions. As ex-
plained by Cross and Roelofsen (2016), there are different logical kinds of questions:
whether-questions, which-questions, why-questions, etc. Each kind of questions de-
termines its own kind of presuppositions. For example, the presuppositions of which-
questions (e.g., “Which is the smallest prime number bigger than 7?”) are existential
statements (e.g., “There exists a smallest prime number bigger than 7”).
According to Genot and Gulz (2015), based in turn on (Hintikka & Hintikka, 1989),
the Inquirer’s range of attention, at a given moment of inquiry, is the set of questions
such that:

414 [Link]:indrum2018:174392
- The corresponding presuppositions are already available or can be obtained by a
deduction move.
- Inquirer is aware of those presuppositions.
The main problem of interrogative inquiry is the mismatch between Inquirer’s range
of attention and available Nature’s answers, since sometimes Inquirer might ask a
question which has no answer, or fails to ask a relevant question whose answer is
available.
In the literature on IBL it is usual to distinguish between several kinds of inquiries
depending on two parameters: how open is the initial question Q, and how guided is
the inquiry (PRIMAS, 2011, pp. 11-12).
Concerning the first parameter, a serious study of types of questions and degrees of
openness in relation to the development of the subsequent inquiry is still to be done.
Our point is that IMI provides a suitable theoretical framework to carry out such a
study.
Concerning the second parameter, we refer to (Hintikka, 1982) in order to learn about
all the possible moves the teacher can do in a game broader than an inquiry-game, in
which the Student-Inquirer and Nature are but two players among many others (the
Teacher being one of them). In particular, a kind of move the Teacher can do in order
to guide the inquiry is to manipulate Inquirer’s range of attention so that to avoid the
mentioned mismatch. Among the possible kinds of manipulation one finds: to ask
whether we can deduce a certain proposition from the available information (in order
to use it further as a presupposition of a question), to attract attention to a certain (al-
ready proved) proposition and to incite to consider it as a presupposition of a ques-
tion, etc. One might wonder to what extent the Teacher can avoid this manipulation in
a IBL process channeled to the acquisition of some propositional knowledge. Con-
cerning this, Genot and Gulz (2015) proved that “a trade-off between success [of the
inquiry] and autonomy is unavoidable” (p. 1). Indeed, in one hand, with no manipula-
tion of Inquirer’s range of attention success might not be achieved. On the other hand
The IMI [Interrogative Model of Inquiry] does however warrant the following conclu-
sion: a guaranty that an inquiry learner will be able to solve interrogatively a problem can
always be obtained by manipulating the learner’s range of attention. (Genot & Gulz,
2015, p. 18)
OBSTACLES TO IBL
Admittidly, to plan for and support IBL is difficult due to the presence of obstacles of
different nature: political, cultural, concerning teacher’s view of her own profession,
concerning teacher’s training, epistemological, etc. Let us use the IMI as a microsco-
pe to inspect here two of them.
Propositional knowledge and models
According to (PRIMAS, 2011, p. 22):

415 [Link]:indrum2018:174392
 (…) many teachers find that IBL comes into conflict with the way they learnt science and
mathematics in school and at the university, and even with the way they have been tea-
ching science and mathematics for many years. That is, with their beliefs about the nature
of mathematics/science and/or their beliefs about teaching of mathematics/science. Pro-
bably this will be one of major obstacle you will find.
This is strongly related to the fact that content taught in formal education is, typically,
knowledge by acquaintance and knowledge-how, with a bleak lack of genuine propo-
sitional knowledge. Therefore, in agreement with the analysis of inquiry previously
exposed, this content, not being propositional knowledge, can not be learnt through a
proper inquiry. This is the case of the knowledge by acquaintance consisting in kno-
wing the statement of theorems (Pythagorean theorem, Thales theorem, etc.) without
proof. It is also the case of the following examples of knowledge-how, which lives
just in the realm of syntax, of manipulation of symbols, without being supported by
meanings: all kind of algorithms of addition, subtraction, multiplication and division
of several kind of numbers, algorithms to calculate the greatest common divisor and
least common multiple, divisibility criteria for natural numbers, etc. Summarizing, for
a piece of mathematics to be learnt in an inquiry-based fashion, it must be presented
as propositional knowledge and, insofar as this kind of knowledge is concerned with
true propositions, the presence of models fixing meanings and guiding the inquiry is
unavoidable. In other words, without models there are no meanings, without mea-
nings there are no truths, without truths there is no propositional knowledge, and so
there is no inquiry.
One might argue that, after all, these mathematical objects do have a meaning. For
instance, the fraction 2/3, applied to an object, refers to any portion of this object
equivalent (with respect to a certain, previously fixed, magnitude: volume, mass,
area, etc.) to the one obtained after performing the following steps:
i) Split this object into three parts so that they were equivalent with respect to the
fixed magnitude.
ii) Take any two of these three parts.
We would eventually agree that some objects, like natural numbers, positive frac-
tions, etc., are attached to a meaning in regular teaching. But I claim that this mea-
ning is not used later to support techniques. This makes the difference: whereas an
unexplained technique (for example, to multiply fractions) is just mechanical know-
ledge-how, a justified one becomes propositional knowledge. Let us illustrate my
claim with the case of the product of fractions.
Typically, we say that the product of two fractions, a/b and c/d, is the fraction (a·c)/
(b·d), whose numerator (respectively, denominator) is the product of the two numera-
tors (respectively, denominators). Notice that this mirrors the common definition in
formal mathematics, where a fraction is defined as an ordered pair (a, b) of integers,
with b different from zero, and the product of two fractions, (a, b) and (c, d), is (a·c,
b·d). This formal definition, taking part in the praiseworthy human enterprise of

416 [Link]:indrum2018:174392
founding Mathematics in Set Theory, does not need any further proof. Actually, being
a definition, it can not be proved. But, for the existence of a proof, there is a more
important obstacle other than the fact that it is a definition: there is no model-inter-
pretation neither of the term fraction nor of the term product referring to fractions.
Concerning this we would like to underly two points:
- Once you have a model-interpretation of these terms you can (at least) try ‘to prove
your definition’, namely, you can try to ask the question “Is it true that, according
to the fixed meaning of the terms fraction and product, the product of two fractions
is calculated by following the former procedure?”
- The way you answer the question strongly relies on the model-interpretation of the
terms.
Imagine, for instance, that your interpretation of the term fraction is the one above:
the denominator indicates the number of equivalent (according to a fixed magnitude)
parts into which a given object has been split, and the numerator indicates the number
of these parts you are considering. You still have to give an interpretation of the term
product. This is a hard task. To beging with, we can say that a product is the result of
a multiplication. Now, what is a multiplication? In natural numbers, a multiplication
is what we do to calculate an amount of magnitude which has been expressed as a
whole amount of a whole amount, for instance, to calculate the cardinal of a set which
results from the union of 27 sets, each of which has 63 elements. Similarly, we can
say that the multiplication of fractions is what we do to calculate an amount of mag-
nitude which has been expressed as a fraction amount of a fraction amount. For ins-
tance, to know which is the total fraction we are considering when we calculate 2/3 of
4/5 of some amount of a given magnitude? I would know how to answer it if I knew
how to answer in the case of 1/3 of 4/5. Similarly, this will not be a problem if I knew
how to calculate 1/3 of 1/5. But it is not difficult to calculate that if each of the 5
fifths were divided into 3 parts, then the initial amount would be divided into 15
parts. Thus, 1/3 of 1/5 is 1/15. Now, since 1/3 of 4/5 is 1/3 of 4 times 1/5, we get 4
times 1/15, which is 4/15. And 2/3 of 4/5 is 2 times 1/3 of 4/5, that is to say, 2 times
4/15, which is 8/15. One can see, in this and other examples, that the numerator (res-
pectively, the denominator) of the final fraction can be directly obtained from the first
two fractions just by multiplying their numerators (respectively, denominators).
These considerations show that only after having an interpretation of the term product
(again, it is the result of a multiplication, and a multiplication is what you do to calcu-
late an amount of magnitude which has been expressed as a fraction amount of a
fraction amount) you can prove the truth of the following proposition: “the product of
the fractions a/b and c/d is (a·c)/(b·d)”. The good news is that models allow to go
from knowledge-how to propositional knowledge. The bad news is that the acquisi-
tion of this propositional knowledge strongly relies on the chosen model. This is not a
minor issue. What to do if we have many possible interpretations of our theoretical
terms? Which of them should be considered? Is there any didactic criterion (for ins-

417 [Link]:indrum2018:174392
tance, to overcome some learning obstacle) to choose among the different possible
interpretations? For instance, the interpretation given above to the term fraction is the
most usual, but it is not the only one. Should we also consider the others interpreta-
tions of the term fraction? If so, how to prove the multiplication formula for these in-
terpretations? We face the problem to choose criteria for attaching meanings such
that: 1) they help us to fight against undesirable didactic phenomena specific to the
usual teaching of fractions; 2) they are compatible with the meanings attached to ot-
her numerical fields, like negative or real numbers.
How to plan and support inquiries for students?
Even if all the questions raised in the previous paragraph about the implementation of
semantics in mathematics syntax are answered, we still have to deal with further dif-
ficult problems.
A very first one is: what could be the initial question Q of the inquiry? This question
seems to be very difficult to find as it is intended to initiate an inquiry through which
many kinds of mathematic entities would appear. Among them, notably, concepts. In
this direction, we find the following teacher’s claim which figures in (Maaß & Reitz-
Koncebovski, 2013, p. 12):
What about conceptual knowledge -surely students cannot be expected to reinvent mat-
hematical or scientific concepts for themselves?
In terms of Hintikka’s IMI, to search for Q amounts to looking for a question such
that, together with the premises T (student’s previous knowledge) and Nature’s ans-
wers (derived from mathematical examples possibly examined by students), allows to
deduce (as a final product but also as something obtained in the course of inquiry) the
aimed mathematical propositions.
But still, even if Q is already clear, as it is said in Anderson’s study (as cited in PRI-
MAS, 2011, p. 20), teachers have difficulties with managing “the challenges of new
teacher roles and new student roles”. According to Walker’s work (cited in PRIMAS,
2011, pp. 20 - 22),
Teacher loses control: although it depends on the degree of freedom teacher gives to stu-
dents, it is clear that in IBL students should take control of the lesson.
Also, although the next quotation refers to IBL of science, it is perfectly translatable
to mathematics:
Inquiry based lessons might not “work”: there is the risk that experiments do not work,
that students collect wrong data and that they will get a wrong idea. In the classical use of
experiments, these are carefully planned so that they always work and offer the right
exemplification of the phenomena that is at stake.
Therefore, teachers have questions concerning their role in and control of students’
inquiries. In terms of Hintikka’s IMI, the question is: which would be the dialectics
between deductions and questions in the expected inquiry? To have this dialectics re-
latively planned contributes to prepare the teacher in her duty of manipulating stu-

418 [Link]:indrum2018:174392
dent’s range of attention, which, in turn, grants the success of the inquiry in order to
provide the acquisition of the aimed propositional knowledge.
PROPOSAL FOR TEACHING IBL TO PRE-SERVICE TEACHERS
In order to solve the problem of finding, given a piece of mathematics to be taught, a
good model for the theoretical terms of this content, a good initial question Q and a
way to handle the corresponding inquiry, PRIMAS has the so-called professional de-
velopment modules [1]. There is, for instance, a module with examples of questions
to initiate an inquiry. There is a module with strategies to promote students’ questio-
ning. There is, also, a module with examples of ways of acquiring concepts. But still
there is a need of complete examples, including all together an initial question, the
relevant moves students should do in the corresponding inquiry game, and moments
in which the teacher could manipulate well enough student’s range of attention in this
very game.
As we said at the beginning of this work, the Anthropological Theory of the Didactic
(ATD) is one of the proposers of IBL. ATD suggests the implementation of the so-ca-
lled paradigm of questioning the world (Chevallard, 2015). According to it, formal
education would be carried out by means of study and research paths (SRP). Succin-
ctly, a SRP is the process you follow to find the answer A to a question Q. ATD emp-
hasizes that, along this process, you are involved in different kinds of activities: stud-
ying possible (perhaps partial) answers to Q, formulating new auxiliar questions, etc.
Although less philosophically informed, ATD’s analysis of the notion of inquiry, via
the notion of SRP, runs almost parallel with that of Hintikka, via the IMI.
There is a continuous spectrum of types of SRP, the extremes of which are what we
could call open and closed SRP. The open ones are those in which the teacher is not
specially interested in leading the students towards a particular piece of knowledge
O. In contrast, the closed ones are those in which the question has been selected with
the intention of leading to the natural emergence, in the course of the SRP, of a cer-
tain piece of knowledge O previously selected.
In formal mathematics this knowledge O is not expressed as the output of an inquiry,
but as a series of axioms, definitions, theorems, examples and standard techniques.
Therefore, closed SRP demand, at least, to reorganise the piece of knowledge O to be
found along the inquiry. The corresponding reorganisations are what ATD calls Refe-
rence Epistemological Models (REM) (see, for instance, Sierra, 2006).
Typically a REM is expressed in terms of praxeologies (Chevallard, 2006), that is to
say, in terms of: types of tasks, techniques devoted to face these types of tasks, a te-
chno-logical considerations about each technique (a detailed description, a justifica-
tion, a study of its scope and reliability, possible enhacements) and, possibly, also
some theoretical considerations about the situations under study (our metaphysical
description of them: basic entities, basic properties, etc.). It is a key feature of a REM
that essentially everything in it appears motivated by the study of the types of tasks.
It is worth mentioning that normally the construction of a REM on a piece of know-

419 [Link]:indrum2018:174392
ledge O is not only guided by the aim of using it as a basis of a SRP, but also by the
intention of counteracting some undesirable didactic phenomena specific to the usual
teaching of O (Gascón, Nicolás, in press).
It is still an open question, but our point is that each REM [2] implicitly provides a
complete example of a possible inquiry, and so a solution to the obstacles to the im-
plemenation of the IBL previously mentioned, namely, the need of interpretations-
meanings, the need of a good initial question and the need of knowledge about when
and how to guide the students in their inquiry. More precisely, we think each REM
implicitly proposes: a model for the theoretical terms and syntax appearing in the
mathematical content O to be studied, a question Q to initiate an inquiry, a proof (ba-
sed on Logic and Game Theory) of the fact that the corresponding inquiry would fu-
lly cover O, and a set of moments of the inquiry at which teacher should evaluate, and
possibly manipulate, Inquirer’s range of attention. In future works we will try to pro-
vide evidences for this point via a logical analysis (IMI-like, in terms of deductions,
questions and answers) of some of the published REM.
NOTES
1. Available at [Link]
[Link]
2. For the moment, there are, among others, published REM on natural numbers, integer numbers,
decimal numbers, proportionality, algebra and differential calculus. See [Link]
grupo-tad/
REFERENCES
Artigue, M., & Blomhøj, M. (2013). Conceptualizing inquiry-based education in
mathematics. ZDM Mathematics Education, 45, 797—810.
Brousseau, G. (1997). Theory of didactical situations in mathematics. Dordrecht:
Kluwer.
Chevallard, Y. (2006). Steps towards a new epistemology in mathematics education.
In M. Bosch (Ed.) Proceedings of the 4th Conference of the European Society for
Research in Mathematics Education (CERME 4) (pp. 21-30). Barcelona: FUN-
DEMI-IQS.
Chevallard Y. (2015). Teaching Mathematics in Tomorrow’s Society: A Case for an
Oncoming Counter Paradigm. In: S. Cho (Ed.) The Proceedings of the 12th Inter-
national Congress on Mathematical Education (pp. 173-187). Springer, Cham
Cross, C., & Roelofsen, F. (2016). Questions, The Stanford Encyclopedia of Philo-
sophy (Spring 2016 Edition), Edward N. Zalta (Ed.), URL = <[Link]
[Link]/archives/spr2016/entries/questions/>.
Fantl, J. (2017). Knowledge How, The Stanford Encyclopedia of Philosophy (Fall
2017 Edition), Edward N. Zalta (Ed.), URL = <[Link]
fall2017/entries/knowledge-how/>.

420 [Link]:indrum2018:174392
Gascón, J., & Nicolás, P. (in press). Economía, ecología y normatividad en la teoría
antropológica de lo didáctico [Special issue]. Educaçao Matemática Pesquissa.
Genot, E., & Gulz, A. (2015). The interrogative model of inquiry and inquiry lear-
ning. In B. Can (Ed.), Perspectives on Interrogative Models of Inquiry: Develop-
ments in Inquiry and Questions Springer.
Hintikka, J. (1982). A dialogical model of teaching. Synthese, 51, 39 - 59.
Hintikka J., Hintikka M. B. (1989) Reasoning about Knowledge in Philosophy: The
Paradigm of Epistemic Logic. In: The Logic of Epistemology and the Epistemology
of Logic. Synthese Library (Studies in Epistemology, Logic, Methodology, and
Philosophy of Science), vol 200. Springer, Dordrecht
Hintikka, J. (1996). The Principles of Mathematics Revisited. Cambridge: Cambridge
University Press.
Hintikka J., Halonen, I., & Mutanen, A. (2002) Interrogative Logic as a General
Theory of Reasoning. In D. M. Gabbay, R. H. Johnson, H. J. Ohlbach, and J.
Woods (Eds.), Handbook of the Logic of Argument and Inference, vol 1 (295 -
337). Elsevier
Ichikawa, J. J., & Steup, M. (2017). The Analysis of Knowledge, The Stanford Ency-
clopedia of Philosophy (Fall 2017 Edition), Edward N. Zalta (Ed.), URL =
<[Link]
Maaß, K., & Reitz-Koncebovski, K. (Eds.) (2013) Primas (Promoting Inquiry in
Mathematics and Science Education Across Europe). Inquiry-based learning in
maths and science. Final report prepared by PRIMAS, an international project wit-
hin the Seventh Framework Program of the European Union. ([Link]
[Link])
PRIMAS (2011). PRIMAS guide for professional development providers. Retrieved
from [Link]
[Link]
Rocard, M., Csermely, P., Jorde, D., Lenzen, D., Walberg-Henriksson, H., & Hemmo,
V. (2007). L’enseignement scientifique aujourd’hui: une pédagogie renouvelée
pour l’avenir de l’Europe. Commission Européenne, Direction général de la re-
cherche, Science, économie et société.
Sierra, T. A. (2006). Lo matemático en el diseño y análisis de organizaciones didácti-
cas. Tesis Doctoral, Universidad Complutense de Madrid.

421 [Link]:indrum2018:174392
 A model for instructional design in mathematics implemented in
teacher education
Heidi Strømskag
Norwegian University of Science and Technology,
Department of Mathematical Sciences, Trondheim, [Link]@[Link]
This paper presents a model for instructional design in mathematics and reports from
student teachers’ experimentation with it in secondary school. The model is framed
within the theory of didactical situations in mathematics. It contributes to the
research field through the notion of an epistemological model, which provides an
enhanced understanding of the stage of ‘conception and a priori analysis’ of
didactical engineering. Findings show how the model is instrumental in creating
student teachers’ awareness of the impact of the milieu on the nature of the
knowledge developed by the pupils.
Keywords: teachers’ and students’ practices at university level, novel teaching
approaches, epistemological model, adidactical situation, milieu.
INTRODUCTION
The mathematical performances that represent the ambitions of the mathematics
classroom and curriculum are constituted through teacher and student participation in
activities stimulated by mathematical tasks designed (or selected) by the teacher for
the realisation of an instructional purpose (Clarke, Strømskag, Johnson, Bikner-
Ahsbahs, & Gardner, 2014). The centrality of tasks as instruments in mathematics
classrooms is reported in the TIMSS 1999 Video Study: in the eighth-grade
classrooms that were investigated (in seven countries), at least 80% of lesson time, on
average, was spent on solving mathematical tasks (Hiebert et al. 2003). In this paper,
I present a model for instructional design in mathematics, where tasks are embedded
in situations that preserve meaning for the target knowledge. Further, I report from
student teachers’ utilising the model to teach an optional piece of mathematical
knowledge in secondary school. The research question addressed in the paper is:
What do student teachers learn from using the model for instructional design?
The model for instructional design is applicable when there is an intention of teaching
someone some particular mathematical knowledge. It is rooted in the theory of
didactical situations in mathematics, TDS (Brousseau, 1997), the main concepts of
which I present briefly in the next section.
A BRIEF INTRODUCTION TO TDS
TDS provides a systemic framework for investigating teaching and learning
processes and for supporting didactical design in mathematics, where the particularity
of the knowledge taught plays a significant role. In the following I explain concepts
of TDS (based on Brousseau, 1997) that are central to the model for instructional
design.

422 [Link]:indrum2018:174897
The milieu represents the elements of the material and intellectual reality with which
the students interact when solving a problem. An adidactical situation is a situation
in which the students as a group try to solve the problem given to them on the basis
of features of the milieu, without significant help from the teacher. The milieu of an
adidactical situation is called an adidactical milieu. An appropriate adidactical milieu
provides feedback to the students, whether their responses are adequate with respect
to the knowledge at stake. The teacher has three main roles in the broader didactical
situation: first is the devolution of an adidactical situation to the students, which
means to present the adidactical situation and the problem to be solved and make the
students accept this transfer of ownership; a second role is the regulation, which
means to handle the evolution of the adidactical situation and its milieu; and, a third
role is the institutionalisation of the knowledge developed in the adidactical situation,
which means to transform the contextualised responses produced by the students into
scholarly knowledge aimed at by the institution.
The didactical contract refers to the phenomenon that the interaction between the
teacher and students in a didactical situation is regulated by rules related to the
knowledge at stake. These rules form a set of implicit reciprocal obligations and
mutual expectations. In devolution, the teacher (implicitly) negotiates a contract that
involves a temporary transfer of responsibility for the knowledge at stake, from the
teacher to the students.
After devolution, four situations follow where the role of the teacher and the status of
knowledge change: Situations of action, formulation, and validation are
(intentionally) adidactical, whereas the situation of institutionalisation is didactical.
The adidactical situations are designed with milieus that are supposed to give
feedback to the students, as mentioned above. The situation of action is where the
students engage in the situation on the basis of its milieu without the teacher’s
involvement—that is, they construct an implicit solution (a strategy) that guides them
in their decisions. The situation of formulation is where the students’ formulations are
useful in order to indirectly solve the problem—that is, formulation of an explicit
solution that enables somebody else to operate on the material milieu. Here, the
teacher’s role is to make different formulations “visible” in the classroom. The
situation of validation is where the students try to explain a phenomenon or verify a
conjecture. Here, the teacher’s role is to lead a whole-class discussion and trying to
make the students use precise mathematical notions. The situation of
institutionalisation is where the teacher’s role is to connect the contextualised
knowledge built by the students with scholarly forms of knowledge.
A MODEL FOR INSTRUCTIONAL DESIGN IN MATHEMATICS
The model for instructional design introduced here is based on the theoretical
analysis presented in (Strømskag, 2017). It contains four phases: epistemological
analysis; development of an epistemological model; implementation; and,
institutionalisation. They correspond to design and implementation phases of
didactical engineering (Artigue, 2015): epistemological analysis corresponds to

423 [Link]:indrum2018:174897
‘preliminary analyses’; development of an epistemological model corresponds to
‘conception and a priori analysis’; and implementation followed by
institutionalisation corresponds to ‘realization’. The model contributes to the research
field through the notion of an epistemological model, which provides an enhanced
understanding of the phase of ‘conception and a priori analysis’ of didactical
engineering. Its elements are illustrated in Figure 1 (the dotted curve signifies that the
epistemological analysis informs the institutionalisation).

Figure 1. A model for instructional design (reproduced from Strømskag, 2017, p. 912)
Figure 1 shows how TDS concepts are constituents in design and implementation of a
didactical situation that aims at some particular mathematical knowledge. For an
account of its methodology, see (Strømskag, 2017). The phases of the model are
described briefly in the following sections.
Epistemological analysis
Epistemological analysis of the knowledge at stake involves two components: an
analysis of the knowledge itself, and a didactical analysis. The aim of analysing the
knowledge is to identify possible epistemological obstacles, to find out what this
knowledge is for, what questions that motivated its genesis, and how its validity can
be justified. A didactical analysis aims at surveying research on teaching and learning

424 [Link]:indrum2018:174897
of the knowledge at stake. The purpose of the epistemological analysis is to figure out
what conditions must be fulfilled for a situation to implement the target knowledge
(i.e., what problem would have the target knowledge as its solution), and by this
inform the next phase, development of an epistemological model.
An epistemological model
An epistemological model of the knowledge at stake is a construct that consists of
three components: first, a model of the target knowledge—possibly an iconic
representation; second, a situation that preserves meaning—involving a problem that
can be solved in an optimal manner using the target knowledge; and, third, milieus of
situations of action, formulation and validation—designed so as to make students’
knowledge progress towards gradually more explicit and mathematical forms, based
on an image of students’ adaptations to the milieus. The second and third component
together can be considered a model of the generic and epistemic student’s intended
learning—it is developed according to the conditions that must be fulfilled for a
situation to implement the knowledge it defines. The point is to design milieus so that
the responses produced by the students will become gradually more explicit and
mathematical, and ultimately can be institutionalised to become the scholarly
knowledge aimed at by the teacher.
Implementation
The third phase involves implementation in the classroom, where the epistemological
model is the basis for the teacher’s devolution and regulation of an adidactical
situation (including a problem) aiming at students’ interaction with the milieus of
situations of action, formulation and validation. Students’ learning during
implementation is understood as independent adaptation.
Institutionalisation
The fourth phase involves institutionalisation of the solution to the problem into
scholarly and decontextualised forms of knowledge. Students’ learning in this phase
is understood as acculturation, which is intended to enable them to know the place,
importance, and future of the mathematical knowledge reached.
The two processes of learning—adaptation and acculturation—are governed by the
didactical contract, and the relationship between them is understood as devolution
and institutionalisation. The model presented in Figure 1 displays the teacher’s roles
in the design and realization of a didactical situation: the process of didactical
transposition (Chevallard, 1989) transforms the knowledge at stake into an
epistemological model; the process of devolution transforms the epistemological
model into a problem embedded in a situation; and, the process of institutionalisation
transforms the situated knowledge used to solve the problem into scholarly forms of
knowledge aimed at by the institution.

425 [Link]:indrum2018:174897
AN EPISTEMOLOGICAL MODEL EXEMPLIFIED
I give here a brief presentation of an epistemological model, where the target
knowledge is the general numerical statement expressing that the sum of the first n
odd numbers is equivalent to the square of n, possibly represented by
. I have implemented the epistemological model exemplified here
in a teacher education programme, a detailed account of which is given in
(Strømskag, 2017).
A model of the theorem
A model of the target knowledge is created using a dissection of a square into L-
forms consisting of consecutive odd numbers of unit squares (where 1 is represented
by one unit square, hence a degenerated L). A generic example is given in Figure 2,
illustrating that . It is made of a dissection of the fourth square
into the first four odd numbers. The model of the target knowledge is not to be shown
to the students. It is a tool for developing a model of the students’ intended learning,
parts of which I describe below.

Figure 2. A model of the target knowledge

A situation that preserves meaning of the target knowledge
I invented a situation based on an imaginary company called TILEL selling a special
kind of tile formations that can be used to cover squares. The tile formations have
shapes as L-forms, and consist of an odd number of unit squares. The idea of this
situation is derived from the model of the target knowledge, shown in Figure 2.
The task has two main parts:
- Finding a method for building a square of side length a natural number, using L-
forms from TILEL.
- Explaining why the method will work for any natural number.
The milieu
The milieu of action is the material milieu on which the students are supposed to
operate. The material milieu consists of ten paper cut-outs with unit
squares that represent the first ten odd numbers, as illustrated in Figure 3. The task in
the situation of action is to find a method of building a square of side length a natural
number up to ten, using L-forms of different sizes. The features of the milieu
(common for action, formulation, and validation) are the following:
- The material milieu does provide feedback: it is visible for the students whether or
not they succeed in building a square of the intended size, using the L-forms.

426 [Link]:indrum2018:174897
- There is an obligation to use L-forms of different sizes. This is part of the
didactical contract, and is to ensure that the students engage with odd numbers.
- It is only the size of the resulting square that matters. This is part of the didactical
contract, and means that different configurations of L-forms should not be
distinguished (i.e., it is not a combinatorial task).

Figure 3. The material milieu for the intended theorem (paper cut-outs)
For an account of the milieus of formulation and validation, and the intended
knowledge progress based on the TILEL-situation, see (Strømskag, 2017).
STUDENT TEACHERS’ EXPERIMENTS IN SECONDARY SCHOOL
Methodical approach
In 2017, within a master’s course in mathematics education, I gave a group of 13
student teachers (hereafter ‘students’) the task of using the presented model to design
and implement (with pupils in secondary school) a situation that preserved meaning
for a chosen piece of mathematical knowledge. The students were in the 4th or 5th
year in a 5-year teacher education programme in mathematics and natural sciences.
The preparation at campus for the experiment (spanning 4 lessons, á 90 minutes)
entailed first, my introduction of necessary concepts from TDS and presentation of
the model itself. Then the students worked on a task on percentages that I had
designed (using the design principles of the model) to let them experience and solve a
problem embedded in a TDS situation. The next step was my comparison of the a
priori and a posteriori analyses of the TILEL-situation, implemented with 20
students in a teacher education programme (this experiment is reported in Strømskag,
2017). This was followed by a brief presentation of two other models of
mathematical knowledge: models of convergent geometric series, shown in Figures 4
and 5. The students discussed what series they were models of, and what the limit of
each of them was.
The last lesson comprised students’ work in four groups on an assignment that
involved using the model for instructional design to design and implement a teaching
situation aiming at some particular mathematical knowledge (suitable for the class in
which they were to do the experiment). I was present for guidance during the initial
design part, where they created an adidactical situation with a milieu and a problem.

427 [Link]:indrum2018:174897
 0.5

0.5

Figure 4. Figure 5.

Data used to explore the utility of the proposed model consist of students’ reports
from the experiments, which I received their consent to use (texts from the reports are
translated into English by H. S.). Even if the students worked in groups to plan and
conduct the experiment, they wrote individual reports. A demand was that the report
contained a comparison between a priori and a posteriori analyses of the
implemented situation—that is, how the situation was suitable to develop the target
knowledge.
Results
A general finding from the 13 reports is that the implemented situations did not have
a satisfactory adidactical potential. The students described how they—contrary to the
ideal roles of the teacher in TDS situations—had to intervene during the pupils’
work. This was explained by too little attention or effort to create an appropriate
milieu for the intended knowledge. In the following, I present how three students’—
through their comparison of a priori and a posteriori analyses—in each case,
reflected on limitations of the milieu of the implemented situation, and suggested
improvements that might possibly strengthen its adidactical functioning. In the last
case, the strength of the designed situation is also accounted for.
The target knowledge for Uma’s pupils (upper secondary first year) was a formula for
the number of components of the n-th element of a shape pattern,
. This knowledge was the solution to a problem embedded in a
situation with a barbeque area of variable size with fixed shape, covered by tiles as
illustrated by the shape pattern in Figure 6.

Figure 6. The material milieu for Uma’s barbeque area problem (iconic representations)
Uma described how emphasising the structure of the elements of the pattern by, for
example, using different colours, might have made it easier for the pupils to
implement the knowledge aimed, without the teacher’s guidance and help. This idea
of Uma can be seen as stimulating algebraic thinking. She also commented on the
importance of the devolution for the adidactical functioning of the milieu.

428 [Link]:indrum2018:174897
The target knowledge for Tracy’s pupils (upper secondary second year) was the
theorem expressing that the sum of two consecutive triangular numbers is equivalent
to the square number of rank the same as the largest of the triangular numbers. The
theorem was symbolised as , where
Tracey presented a model of the theorem, a generic example illustrating that
, as shown in Figure 7. The target knowledge was the solution to a
problem embedded in a situation with production of quadratic birthday cards, made
by two staircase figures (representing triangular numbers) as shown in Figure 8.

Figure 7. Tracey’s model of the target Figure 8. The material milieu for the
knowledge birthday card problem (paper cut-outs)
Tracey explained that the colours of the material milieu were an obstacle to pupils’
engagement with the knowledge at stake; the pupils referred to staircase figures by
their colours instead of by their structure (sum of natural numbers) and rank. Further,
generalisation of the birthday card problem was supposed to be brought into effect by
the task: “Formulate a method to make a birthday card of an arbitrary size, and argue
why the method works”. This part was problematic because the pupils understood the
concept arbitrary (in Norwegian: “vilkårlig”) in the meaning of any birthday card
made by the figures available in the milieu, instead of any birthday card made by
imaginary staircase figures of any size (as intended).
The target knowledge for Tina’s pupils (grade 8) was an explicit formula for the
number of components of the n-th figure of a shape pattern (shown in Figure 9). This
knowledge was the solution to a problem at a four-way crossing, where a car in the
middle blocked the crossing and caused one car at each of the four ways to be stuck
every second. The situation after two seconds is shown in Figure 10 (this was the
material milieu along with a pile of paper cut-outs of cars).

Figure 9. Tina’s model of the target Figure 10. Material milieu for the crossing
knowledge situation (paper cut-outs)
For small numbers of seconds, Tina expected the pupils to use a recursive approach.
Hence, she tried to motivate for an explicit formula by asking for the number of cars

429 [Link]:indrum2018:174897
after 1 minute. Some pupils, however, used a strategy where they treated the situation
as a direct proportionality (representing the four cars added each second) and added 1
in the end for the car in the middle. One pupil reasoned that since there were 21 cars
after 5 seconds, he multiplied 20 by 2, then multiplied 40 by 6, and added 1 in the
end. This gave the right answer, but it was not what Tina was after. She was after a
functional relationship between the seconds that had passed and the total number of
cars. Tina explained that in planning it was not anticipated that anyone would use the
mentioned method, so the milieu did not have adequate feedback to handle it.
In conclusion, Tina wrote: “The results show that the work in advance of
implementation is very important regarding the knowledge actually developed by the
pupils […] The results show, however, that feedback from peers has been
instrumental in helping pupils to correct misunderstandings in adidactical situations.
This I see as a strength of the designed milieu.” Another strength reported by Tina is
that in institutionalisation, the pupils were exposed to a new shape pattern, the
relationships of which they were able to represent through algebraic symbols. Tina
claimed that this was evidence that the knowledge developed through the experiment
had in fact been generalised beyond the designed situation.
DISCUSSION
There were institutional constraints on the study reported here: the time available for
instruction at campus was only 4 lessons (a modifiable condition); and the students
had no knowledge of the pupils before the experiment (an unmodifiable condition).
The fact that the students—contrary to their intention—had to intervene during the
pupils’ work is not unique to the model for instructional design. What the
methodology of the model offers, however, is validation of an implemented teaching
situation based on comparison between a priori and a posteriori analyses; this is
similar to the methodology of didactical engineering (Artigue, 2015). In the model
for instructional design, the a priori analysis is constituted by an epistemological
analysis and development of an epistemological model, whereas the a posteriori
analysis is an analysis of the realisation of the designed situation.
The answer to the research question posed—about what the students learned from
using the model for instructional design—is that the students were able to identify
relationships between the adidactical milieu and the knowledge progress in each case.
The points addressed in the students’ reports were about features of the milieu that
constituted limitations on generalisation processes: In Uma’s case it was about the
devolution, and a material milieu that failed to focus on structure and algebraic
thinking. In Tracey’s case it was about features of the material milieu that suppressed
focus on structure, and pupils’ not knowing the meaning of the term “arbitrary”. In
Tina’s case it was about a milieu without feedback for a linear generalisation method
used by many of the pupils. However, Tina also emphasised a quality of the milieu:
how pupils’ progress in adidactical situations were enabled by feedback from peers.

430 [Link]:indrum2018:174897
Such comparisons between a priori and a posteriori analyses will identify
relationships between the nature of the milieu and the knowledge progress in
adidactical situations, and thereby create awareness of what enables and what
prevents pupils from reaching the target knowledge. In this way, the model has
potential to inform future design (aiming at the same knowledge): it might provide
deeper insight into the nature of the knowledge at stake, and a better understanding of
necessary conditions on the milieu concerning its adidactical functioning.
In conclusion, the impact of the study on my practice as a teacher educator is that the
students’ experiments are used as objects of study (with the same or other students),
where the task is to revise the experimented situations and their milieu to better
match them with their original didactic intention, and subsequently, to implement
them with pupils to validate them. In this way, the experimented situations are
considered material milieus for instructional design. This represents a meta-level
perspective on instructional design in mathematics teacher education.
REFERENCES
Artigue, M. (2015). Perspectives on design research: The case of didactical
engineering. In A. Bikner-Ahsbahs, C. Knipping, & N. Presmeg (Eds.),
Approaches to qualitative research in mathematics education (pp. 467-496).
Dordrecht, The Netherlands: Springer.
Brousseau, G. (1997). The theory of didactical situations in mathematics: Didactique
des mathématiques, 1970-1990. Dordrecht, The Netherlands: Kluwer.
Chevallard, Y. (1989). On didactic transposition theory: Some introductory notes. In
H. G. Steiner & M. Hejny (Eds.), Proceedings of the International Symposium on
Selected Domains of Research and Development in Mathematics Education (pp.
51-62). University of Bielefeld, Germany, and University of Bratislava, Slovakia.
Clarke, D., Strømskag, H., Johnson, H. L., Bikner-Ahsbahs, A., & Gardner, K (2014).
Mathematical tasks and the student. In P. Liljedahl, C. Nicol, S. Oesterle, & D.
Allan (Eds.), Proceedings of the 38th Conference of the International Group for
the Psychology of Mathematics Education (Vol. 1, pp. 117-143). Vancouver,
Canada: PME.
Hiebert, J., Gallimore, R., Garnier, H., Givvin, K. B., Hollingsworth, H., Jacobs, J.,
… Stigler, J. W. (2003). Teaching mathematics in seven countries: Results from
the TIMSS 1999 video study. Washington, DC: National Center for Education
Statistics.
Strømskag, H. (2017). A methodology for instructional design in mathematics—with
the generic and epistemic student at the centre. ZDM Mathematics Education,
49, 909-921. [Link]

431 [Link]:indrum2018:174897
 University Teachers-Researchers’ Practices: the Case of Teaching
Discrete Mathematics
Theresia Tabchi
University of Reims Champagne-Ardenne, CEREP EA 4692, Reims, France,
[Link]@[Link]

Our work is part of a thesis which deals with the teaching practices of university
mathematics professors. Practices of university professors concur specificities: an
example is the articulation between teaching and research. We seek to characterize
the research impact in the teaching practices at university. We choose discrete
mathematics as an object of study. It is an area whose didactic transposition is not
fully achieved while the links with other mathematical fields remain undefined. These
elements make the choice of resources to be mobilized and conceived quite complex.
Our exploratory study draws on interviews with university professors aiming to
characterize the interaction with resources; which can help us clarify the research
activities impact on teaching practices in the case of discrete mathematics.
Keywords: Teachers’ practices at university level, discrete mathematics,
undergraduate education, impact of research, resources.

INTRODUCTION
Our work is part of an ongoing thesis which deals with the practices of university
mathematics professors. The study belongs to the growing body of research on
university professors’ practices (Biza, Giraldo, Hochmuth, Khakbaz, & Rasmussen,
2016). At this level of our study, we designate by “University professors” the
university mathematics teachers who occupy teaching/research positions.
While we are interested in teaching practices at university, we try to characterize the
different factors that impact them (institutional, didactic and epistemological), in
particular, the impact of the research activity on the teaching practices. Throughout
the text, we will be relying on the following definitions:
- “Teaching practices” and “university teachers’ practices” to refer to the “teaching”
aspect of the university professors’ work;
- “Research activity” to refer to the research aspect of the profession.
Our study focuses on the place of discrete mathematics in undergraduate level.
Discrete mathematics holds an epistemological importance, and it is a branch whose
topics are not stabilized in mathematics curricula (Heinze, Anderson, & Reiss, 2004;
Hart & Sandefur, in press). Hence, it is particularly interesting to study university
mathematics professors’ practices in the field of discrete mathematics, a choice that
we will justify later in the document.
In a study of ICME [1], Biza et al. (2016) conducted a survey of existing research on
mathematics education at university. Drawing on journal publications and conference
proceedings on university mathematics education published since 2014, Biza et al.
1

432 [Link]:indrum2018:174774
identified the current trends and the most recent advances in the field. They classified
the existing work according to the topic of research. Among these, the influence of
teachers’ research activity on their teaching approaches in particular contexts: the use
of examples in mathematics tutorials and the use of graphic representations. The
study highlights the emerging interest in taking into consideration resources in
teaching at university, in particular, interactions with resources for mathematics
education and their impact on teachers’ professional development. The study points
as well the need for further research on practices of university mathematics teachers
and the possible impact of their research activity on their teaching practices.
The work that we present in this text is a contribution to the growing body of research
in mathematics education on the practices of mathematics teachers at university level.
We will present first the context of the study and the research questions. Then, we
will describe the first steps in the construction of our theoretical framework and the
collection of data. Finally, we will present the methodology we used in our study as
well as some extracts, along with some preliminary results and perspectives.

CONTEXT AND RESEARCH QUESTIONS

The study of the practices of university professors is part of a field of research
currently in development, but still largely unexplored. There are similarities in many
aspects of the practices of secondary and university teachers: the preparation of the
courses and tutorials, the design of instruction, the conception of evaluations, the
classroom management, and the interactions with students. However, the practices of
university professors conquer specificities, from an institutional point of view and
given their training and their academic background.
University professors have some freedom in the design of their courses and the
choice of contents to be taught, and in the selection and development of resources.
Moreover, their research activity can impact (or not) different aspects of their
teaching practices (Biza et al, 2016; Mali, 2015) such as the instructional approach,
the form of interaction with students, and the choice of new knowledge to be taught.
University professors, in their teaching practices, maintain a constant interaction with
various resources (Gueudet, 2017) derived from their teaching practices as well as
their research activity (old resources, software, online resources, videos and podcasts,
manuals, research articles, etc.). The interaction with the resources can take place at
different moments of their teaching practices: in the design of the sessions, the choice
of contents to be taught, the teaching in the classroom, and the evaluation of learning.
In the working environment of a university professor, the same resources can be used
in his teaching practices and his research activity (numerical computation software
for example). Other interactions between teaching and research can take place, but in
a less tangible way. The study of the university professors’ interactions with the
resources could be a first step in clarifying the nature of the relationship between
their teaching practices and their activity of research: in the choice of contents to be

433 [Link]:indrum2018:174774
taught, in the teaching mode of different mathematical contents, and in the learning
they wish to develop with students (reasoning, application of properties, etc.).
In mathematics education, teaching practices are not considered independently of the
contents taught. The relationship between university teachers’ practices and their
research activity may depend on several parameters, including the mathematical
branches addressed in teaching and research; discrete mathematics, in particular, offer
an interesting experience. Discrete mathematics have a real epistemological
importance as they unfold in different mathematical domains; the objects are easy to
access and can contribute to the understanding of other branches of mathematics
(Grenier & Payan, 1998, Maurer, 1997). They provide an introduction to modeling,
optimization, operational research, and experimental mathematics (Grenier & Payan,
1998, Maurer, 1997). They promote the learning of proof and the development of
heuristic processes in students (DeBellis & Rosenstein, 2004, Goldin, 2004).
Discrete mathematics is a field in expansion, with significant achievements in society.
Attempts to integrate discrete objects in the curricula appear in several countries at
the international level. In France, the place of discrete mathematics in the curriculum
is still not sufficiently stabilized and takes various forms depending on the
educational context (Ouvrier-Buffet, 2014); which makes this branch very interesting
to study the relationship between research activity and teaching practices of
university professors. Hence the research questions can be formulated as follows:
- How do university professors interact with the resources in / for the teaching of
discrete mathematics?
- How can we characterize the relationship between the teaching practices of discrete
mathematics at the tertiary level and research activities in the same field?
Trying to answer these questions requires specific theoretical and methodological
developments.

TOWARDS THE CONSTRUCTION OF A THEORETICAL FRAMEWORK

We conducted a first theoretical choice: the documentational approach (Gueudet &
Trouche, 2009). We will justify the choice in what follows. We note that the choice
of the documentational approach constitutes a tool for a first exploration of the field.
Gueudet (2017) based her work on the documentational approach (Gueudet &
Trouche, 2009) to analyze the interactions of university professors with resources
derived from their teaching practices. In the documentational approach, there is a
distinction between resources and documents. University professors select, mobilize
and use various resources (contents to be taught in class, written evaluations, old
resources, etc.). This interaction with the resources generates a document, which is
the association of resources and a scheme of use of these resources. A scheme is used
here as it was defined by Vergnaud (2009) as the invariant organization of conduct

434 [Link]:indrum2018:174774
for a set of situations having the same aim. According to Vergnaud (1998), a scheme
is a dynamic that has four interacting components:
- An aim that can be easily identified and that indicates intentionality in the
organization of the activity;
- Rules of actions which are the ways of acting generated by the scheme in order to
achieve a specific aim;
- Operational invariants that influence the rules of action. They can be theorems-in-
action (propositions considered as true by the subject, but may be true or false) or
concepts-in-action (considered as relevant in a given situation);
- Possibilities of inferences, that are the adaptations that the subject can bring to his
activity in order to respond to the specificities of a situation corresponding to an aim.
A scheme developed by a subject is associated with a class of situations (Vergnaud,
2009). A class of situations includes all situations having the same aim. In her work,
Gueudet (2017) considered classes of situations for specific aims (for example,
preparing an assessment in linear algebra) and thus corresponding to a single
document, and larger classes of situations that are independent of the mathematical
content (for example, preparing an assessment). She made these choices in order to
observe the organization of the resources of university mathematics professors
globally, and to analyze the operational invariants and the rules of actions associated
to mathematical contents in more restricted classes of situations. Although Gueudet
(2017) recognizes the impact that university professors’ research activity can have on
their teaching practices, she did not focus on this aspect in her study.
We rely on the methodology developed by Gueudet (2017). We seek to analyze the
choices, of contents and resources, of university professors in their teaching of
discrete mathematics. Furthermore, we are interested in the impact of the research
activity on the teaching practices. Therefore, we will adapt and develop Gueudet’s
methodology (2017) in a way to be able to consider the particular field of discrete
mathematics and to take into consideration the research activity’s impact on teaching.

ELABORATION OF INTERVIEW GUIDELINES AND DATA COLLECTION

The methodology of our exploratory study is based on interviews. We consider

university professors whose field of research is discrete mathematics in Lebanon and
in France. The interviews, of approximately 90 minutes, took place in workplaces of
the Lebanese professors in order to have access to their resources. For French
professors, the interviews were conducted via Skype.
We have developed the interview guidelines in order to characterize the interactions
of university professors with various resources, resulting from their teaching practice
and activity of research. We asked questions such as “What are the main resources
you use for your teaching? How do you select the contents of your courses?” these

435 [Link]:indrum2018:174774
questions intended to give a panoramic view of the resources selected, mobilized and
used in the different classes of situations (preparation of lectures and tutorials,
evaluation); We believe that the type and nature of the resources (resources resulting
from research or teaching) can inform us about the relations between the research
activity and the teaching practices of university professors; furthermore, the choice of
contents (lectures, tutorials and evaluations) can give us insight about the different
rules of action and the operational invariants corresponding to the different classes of
situations. Another question was “What are the conditions and constraints that guide
your different choices?” Learning about the institutional conditions and constraints
(related to the educational institution or to the research institution) can contribute in
characterizing the research activity’s impact on teaching practices. Other questions
were about the links with other mathematical fields, the collective work, the
experimentation in classrooms, etc.

We conducted pilot interviews with two discrete mathematics university professors,

one in Lebanon and the other in France, to test the interview guidelines. We recorded
and transcribed the interviews. To keep the anonymity of the professors interviewed,
we will call them Michel (Lebanon) and Bertrand (France) throughout the document.

THE METHODOLOGY OF ANALYSIS

Our methodology of analysis consists of two steps which we will be presenting next.
The first step is analyzing the practices of university professors in terms of their
interactions with resources. For that, we detected the aims, mentioned by the
university professors in the interviews, related to their teaching practices. In other
words, we identified possible classes of situations based on statements, such as
“preparing a lecture” or “preparing an assessment”. For each aim, we identified the
associated resources (manuals, books, scientific articles, computers, online resources
etc.) that were explicitly mentioned in the interview. For the rules of action, we relied
on the university professors’ declarations; they described the ways they behaved in
order to achieve an aim, and the adaptations they brought to their actions in
accordance with the peculiarities of each situation. These are the conditions and
actions that can be expressed in statements of the form “if (condition) ... then (rules of
action) ...” or “for (aim) … I (action) ...”. The regular ways of acting in specific
situations reveal the presence of operational invariants that university professors do
not always express in their discourse. To detect them, we identified in the interviews,
the statements justifying the rules of actions, the propositions held to be true by the
university professors (theorems-in-action) and the logical reasoning underlying the
choice of actions to be conducted according to the specificities of each situation.
We constructed a table for the documents of each university professor; horizontally,
in each table, we come across a document associated with a given class of situations

436 [Link]:indrum2018:174774
(the resources used and the scheme of use). Vertically, we find the sets of the
resources used, the rules of action and the operational invariants.
The second step consists in studying the impact of the research activity on the
teaching practices in classes of situations related to teaching. For each class of
situations, first we choose to detect, in the set of resources used, the presence /
absence of an impact of the research activity (for example, the choice of contents to
be taught in class might be influenced by the research activity of the university
professors or not). Then, we will define a typology of operational invariants; it would
allow us to classify the reasons behind some rules of actions of the university
professors and behind the “transfer” (use, selection and modification) of some
resources from their research activity to their teaching: beliefs developed by the
activity of research, gestures acquired with professional experience or institutional
constraints and conditions. We will eventually search for additional tools for a deeper
study of the impact of research activity on teaching practices, following the results of
the pilot interviews.

ANALYSIS OF THE FIRST PILOT INTERVIEW

Our first interview was with Michel (Lebanon), whose research domain is graph
theory. Michel teaches graph theory for students in masters’ degree and discrete
mathematics for students majoring in mathematics in year 2. We detected, in Michel's
declarations, the resources produced and / or used in his teaching practice.
Aims Resources Rules of actions Operational invariants
Prepare a well-known He tries to convey The contents must be aligned
course of books in the the basic ideas in with the official instructions of
discrete world discrete the major.
mathematics mathematics. Definitions in discrete
in second He uses many mathematics are simple, there is
year Scientific
publications examples. a need for applications to

Typical He chooses contents initiate the work.

texts that allow students Understanding the ideas is very
to discover ideas. important in discrete
mathematics.
Implementing well-known He engages students It is important that students
a discrete books in the in writing. practice writing.
mathematics world He encourages It takes examples and
course in students to read applications to familiarize
class typical texts. students with discrete
Typical
texts Before writing the mathematics.
proof of a theorem, It is very important to illustrate
6

437 [Link]:indrum2018:174774
 he illustrates the the ideas in discrete
ideas. mathematics.
Table 1- Extract of Michel’s documents table from his teaching practices (aims, resources,
rules of action and operational invariants)
The operational invariants that justify the actions of Michel can be classified as
follows. From a didactical point of view, the importance of understanding and writing
ideas, as well as illustration; and the importance of examples and applications in
familiarizing students with the field of discrete mathematics. From an epistemological
point of view, the characteristics of proofs in discrete mathematics such as the
accessibility of discrete objects and the simplicity of definitions. From an institutional
point of view, the alignment of contents to ensure a smooth progressivity.
In developing the contents of his courses, Michel uses examples and “technical
proofs”. He believes that students need time to get used to reasoning in discrete
mathematics. He tries to convey the basic ideas in the theory of graphs such as the
definitions and characteristics of objects; a condition that guides his choice of
contents is the compliance with the instructions of the major.
In teaching, Michel stresses the importance of proof and illustration, and engages his
students in reading and writing proofs. He teaches discrete mathematics in a
theoretical way without real life applications; there is a lack of experimentation in his
courses. He recognizes the importance of computers for research (for expanding the
empirical space), but he does not use them for teaching in his classes.
The interview with Michel was set up to test the guidelines. We have identified areas
for improvement and adapted the guideline before the second interview.

ANALYSIS OF THE SECOND PILOT INTERVIEW

The second interview was with Bertrand. Bertrand (France) is a researcher in graph
theory and a research director at the National Center of Scientific Research. He
teaches by choice: an optional course of combinatory games and mathematical
reasoning to students in first and second year and a course of graph theory for
master’s degree students.
We are particularly interested in his teaching practices at university. Therefore, we
will identify the different classes of situations related to his teaching practices. We
present in Table 2, an extract from the table of documents of Bertrand.
Aims Resources Rules of actions Operational invariants
Develop a Research He selects problems It is important that students
course of problems arising from research experience research problems.
combinatory such as situations. Problems in discrete
games and hunting the mathematics are easily
7

438 [Link]:indrum2018:174774
mathematical beast and He makes choices accessible.
reasoning for problems that can put Discrete mathematics makes it
students in issued from emphasis in his possible to teach mathematical
first and situations teaching on proofs. knowledge and knowledge
second year proposed in related to the proof.
of university workshops.
Implement a Research He focuses more on Doing mathematics means
discrete problems proofs than on working on proofs.
mathematics such as mathematical
course in hunting the objects. It is important that students
class beast and He puts the students experience research problems
problems in the position of in mathematics.
issued from researchers. In discrete mathematics, it is
situations
proposed in He uses material important to experiment with
workshops. objects for material objects.
experimentation. A computer can be interesting
He does not use for research purposes, but does
computers in not contribute in the teaching
teaching. of discrete mathematics.

Table 2 - Extract of Bertrand’s documents table from his teaching practices (aims,
resources, rules of action and operational invariants)
The operational invariants that justify Bertrand’s actions can be classified as follows.
From a didactical point of view, the role of discrete mathematics in developing
reasoning skills, the importance of experiencing research moments for the students.
From an epistemological point of view, the importance of proofs, the characteristics
of problems in discrete mathematics (fun and accessible). From an institutional point
of view, the freedom of a researcher in the choice of contents of his courses.
To develop a course in discrete mathematics, Bertrand selects research problems and
problems arising from situations experienced in workshops. The course of discrete
mathematics to students in first and second year of university is optional and students
are not all in scientific majors. Bertrand has some freedom in the choice of contents.
He is convinced that the course should develop students' reasoning and knowledge
about proofs, as well as knowledge related to discrete mathematics; these convictions
guide his choice of contents. He focuses in his teaching on the activity of proof.
Bertrand’s classroom management is based on putting students in situations of
research. He does not provide the answers, he believes that they must experience
research situations like real scientists. He engages students in experimentations with
games such as hunting the beast. He thinks that computers are not useful in class.

439 [Link]:indrum2018:174774
RESULTS AND DISCUSSION
First results at the interview grid and difficulties encountered in the analysis
In the analyses, we realized the impact of the level of education (Bachelor or Master)
on the interactions of university professors with the resources in their teaching
practices and on the relationship between teaching practice and research activity.
We relied on the methodology developed by Gueudet (2017) to study the interactions
of university professors with resources derived from their teaching practices. It
seemed to us that the institutional context has a large impact on teaching practices in
university (distinction between Bachelor/Master, progressivity in certain fields,
training of non-scientists, etc.). Taking into account the institutional context implies
new theoretical and methodological choices.
First results at the research questions
Our first research question is “How do university professors interact with the
resources in / for the teaching of discrete mathematics?” There is a big difference
between Michel and Bertrand. Michel’s course of discrete mathematics in the second
year of university aims to prepare the students for a Master’s degree in graph theory;
a constraint on the choice of contents is the compliance with the instructions of the
major. On the other hand, Bertrand's course is optional, which allows him a larger
degree of freedom in the selection of contents. Bertrand uses, in teaching, resources
derived from his research activity, which may be due to his involvement in research.
Both professors insist on the importance of reasoning and proofs, therefore they
select contents that contribute to develop students’ ability to write proofs. Bertrand
uses games and material objects for experimentation while Michel’s approach focuses
more on theories. Both professors state that the use of software cannot contribute to
reach the teaching objectives in class, although it can be useful for research purposes.
The second research question is “How can we characterize the relationship between
the teaching practices of discrete mathematics at the tertiary level and research
activities in the same field?” Bertrand holds a position of research. He adopts a
researcher’s attitude in class; he engages students in research and adapts the plan of
his course according to their results. On the other hand, Michel follows a stiffer plan
in his teaching. Quoting Bertrand “A teacher has to have the knowledge and pass it
on, a researcher does not know everything, seeking answers is part of the research”.
The analysis of the pilot interviews shows that the choice to consider the interaction
with the resources is revealing for the study of the relations between teaching
practices and research activity. The operational invariants identified are based on
experience in teaching as well as experience in research. It seems interesting to
define, according to the interviews, a typology of operational invariants (for classes
of teaching situations). We plan to complete the data collection with a questionnaire
that will be disseminated in universities in Lebanon and France for an analysis of the
9

440 [Link]:indrum2018:174774
institutional context of discrete mathematics education. The results of this
questionnaire will give us a map that will allow us to better study the case.
[1] International Congress on Mathematical education

REFERENCES
Biza, I., Giraldo, V., Hochmuth, R., Khakbaz, A., & Rasmussen, C. (2016). Research
on Teaching and Learning Mathematics. ICME-13 Tropical surveys. Hamburg:
springer.
DeBellis, V. A., & Rosenstein, J. G. (2004). Discrete Mathematics in Primary and
Secondary Schools in the United States. ZDM, 36(2), 46-55.
Goldin, G. A. (2004). Problem Solving Heuristics, Affect, and Discrete Mathematics.
ZDM, 36(2), 56-60.
Grenier, C., & Payan, D. (1998). Spécificités de la preuve et de la modélisation en
mathématiques discrètes. Recherche en Didactique des Mathématiques, 18/1, 59-
100.
Gueudet, G. (2017). University Teachers’ Resources Systems and Documents.
International Journal of Research in Undergraduate Mathematics Education,
40753, 1 – 27.
Gueudet, G., & Trouche, L. (2009). Towards new documentation systems for
mathematics teachers? Educational Studies in Mathematics, 71(3), 199-218.
Hart, E. W., & Sandefur, J. (in press, Eds). Teaching and Learning Discrete
Mathematics in the School Curriculum Worldwide. an ICME-13 Monograph .
Springer.
Heinze, A., Anderson, I., & Reiss, K. (2004). Discrete mathematics and Proof in the
High School . Zentralblatt für Didaktik der Mathematik, 36(2), 44-84.
Mali, A. (2015). Characterising university mathematics teaching. In K. Krainer & N.
Vondrová (Eds.), Proceedings of CERME9 (pp. 2187–2193). Prague, Czech
Republic: Charles University in Prague, Faculty of Education and ERME.
Maurer, S. B. (1997). What is Discrete Mathematics? The many answers. In Discrete
Mathematics in the Schools (pp. 121-132). American Mathematical Society.
Ouvrier-Buffet, C. (2014). Discrete mathematics teaching and learning. In S. Lerman
(Ed.), Encyclopedia of Mathematics Education (pp. 181-186). Dordrecht: Springer.
Vergnaud, G. (1998 ). Towards a cognitive theory of practice . In K. Jeremy, & S.
Anna (Eds.), Mathematics education as a research domain: A Search for identity
(pp. 227-241). Dordrecht: Kluwer academic publisher.
Vergnaud, G. (2009). The theory of conceptual fields. Human development, 52, 83-
94.
10

441 [Link]:indrum2018:174774
Characteristics of participation - A mathematician and a mathematics
educator collaborating on a developmental research project
Olov Viirman
University of Gävle, Sweden, [Link]@[Link]
MatRIC, University of Agder, Norway
In this paper, a developmental research project involving offering mathematical
modelling (MM) activities to university biology students is presented, and a
particular aspect is studied, namely the project as a collaboration between
mathematicians and mathematics educators. The aim of the paper is to investigate
what characterizes their participation in the project, and how the characteristics of
the project and its development might influence this participation. Interview data as
well as observation data from the MM sessions are analysed, and findings show that
the mathematics educator served as a broker influencing the practice of the
mathematician. It is hoped that the findings of the study can be of use when planning
future collaboration between mathematicians and mathematics educators.
Keywords: teaching and learning of mathematics in other fields, teachers’ and
students practices at university level, cross-disciplinary collaboration, mathematical
modelling, communities of practice.
Recent developments in science as well as in higher education have led to a greater
need for cross-disciplinary collaboration. For instance, the changing needs of a
biological science more and more dependent on mathematical methods has led
several authors to suggest a closer integration of mathematics and biology in the
education of future biologists (e.g. Brewer & Smith, 2011). This was part of the
motivation for a recent collaborative developmental research project between two
Norwegian centres of excellence in higher education (The Centre for Research,
Innovation and Coordination of Mathematics Teaching, MatRIC; and the Centre for
Excellence in Biology Teaching, bioCEED) in which biology-related mathematical
modelling (MM) activities were introduced to undergraduate biology students as a
means to increase their appreciation for, and competence in, mathematics. In this
paper, however, I will use the project as a case for studying a different kind of cross-
disciplinary collaboration, namely between the mathematicians and mathematics
educators developing and conducting the project.

MATHEMATICIANS AND MATHEMATICS EDUCATORS

There is a close relationship between mathematics and mathematics education – as
remarked by Kilpatrick (1998, p. 36), “neither can exist without the other”. However,
collaboration between mathematicians and mathematics educators is still relatively
rare, although it has received attention in the literature (e.g. Fried & Dreyfus, 2014),
and its importance and relevance to both communities have been emphasized (e.g.

442 [Link]:indrum2018:174854
Nardi, 2008). Dörfler (2003) has pointed out several obstacles to collaboration. For
instance, the communities of mathematics and mathematics education are often quite
separated: mathematicians and mathematics educators work in different departments,
teach different subjects and have different educational backgrounds. Furthermore,
there are few arenas where they can meet professionally: they mostly publish in
different journals and attend different conferences. Dörfler also highlights prejudices
about the other field as a possible obstacle to collaboration, something also discussed
by other researchers (e.g. Ralston, 2004). On a related note, the mathematician in
Nardi’s (2008) book states the need for mathematics educators to “be able to talk to
mathematicians about mathematics”, for there to be a basis for collaboration (p. 270).
At the same time, “there is no such thing as the mathematician, the mathematics
educator, or the mathematics teacher” (Thompson, 2014, p. 319). What is described
above is only the general picture – there are numerous examples of functioning
collaborative relationships. Most of these, however, take place on the individual
rather than the organizational level. Indeed, the importance of individual
relationships and trust is pointed out repeatedly in the literature. As Thompson
(2014) puts it: “successful collaboration requires mutual trust and respect among
collaborators in the context of a shared commitment to solving a problem” (p. 331).

UNIVERSITY MATHEMATICS TEACHING AS SOCIAL PRACTICE

The MM activities forming the basis of the project were developed and conducted by
a MatRIC team consisting of a mathematician and three mathematics education
researchers, and in order to study what characterized this collaboration I will adopt a
Communities of Practice perspective (Wenger, 1998). A community is “a group of
individuals identifiable by who they are in terms of how they relate to each other,
their common activities and ways of thinking, and their beliefs and values” (Biza,
Jaworski & Hemmi, 2014, p. 162), and a community of practice (CoP) is a
community characterized by mutual engagement, joint enterprise and shared
repertoire (Wenger, 1998, p. 73). Mutual engagement concerns, for instance, norms
and social relationships within the community; joint enterprise refers to common
understandings of the aims and ideals of the practice; and shared repertoire concerns
what and how various resources are used in the practice (Biza et al., 2014, p. 163).
Participation in a CoP can vary from the central participation of an experienced “old
timer” to the more peripheral participation of a newcomer (ibid, p. 162). An
individual’s sense of belonging to a CoP involves engagement – active involvement
in mutual negotiation of meaning; imagination – extrapolating from your own
experiences to form an image of your own place within the CoP; and alignment –
coordinating your activities to fit within the structures of the CoP and contribute to
the enterprise (Wenger, 1998, p. 173-174).
The CoP perspective has been used successfully in research on university
mathematics teaching (Biza et al., 2014). For example, Jaworski and Matthews
(2011), in a study of university teachers’ lecturing, found little indication of a joint

443 [Link]:indrum2018:174854
enterprise of teaching. Also originating in research using elements of social practice
theory to investigate university mathematics teaching is the Spectrum of Pedagogical
Awareness (SPA) (Nardi, Jaworski, & Hegedus, 2005). The SPA provides a means
of describing and reflecting upon university mathematics teachers’ pedagogical
thinking and practices, and consists of four levels ranging from Naïve and
Dismissive, through Intuitive and Questioning and Reflective and Analytic to
Confident and Articulate (ibid, p. 293). There are also several case studies using
socio-cultural perspectives to investigate university mathematics teachers’ teaching
practices (e.g. Treffert-Thomas, 2015). However, there is little research on the type
of situation investigated in the present paper. Here, a team of mathematicians and
mathematics educators collaborate on developing and conducting teaching activities
at the university level, not primarily aimed at introducing new mathematics but rather
at getting students to apply the mathematics they already know in new contexts. Of
relevance for its focus on the collaborative aspect is the study by Cooper and
Zaslavsky (2017), investigating a mathematician/mathematics educator co-teacher
partnership in a course on mathematical proof.
The present study seeks to answer the following questions: What characterizes the
participation of the mathematician and mathematics educator in the project? How
might the characteristics of the project and its development influence this
participation?

THE PROJECT – METHODS, DATA AND PARTICIPANTS

As mentioned above, the project involves MM activities aimed at improving biology
students’ motivation for, interest in, and perceived relevance of mathematics in
biological studies. The planning and teaching was carried out by a MatRIC team
consisting of one mathematician and three mathematics educators. The
mathematician, Yuriy Rogovchenko, is a professor of mathematics with extensive
experience of MM both in teaching and in research. He is also coordinator of the
MatRIC network on MM. Yuriy had the main responsibility for planning the
mathematical content of the sessions and selecting tasks. The mathematics educators
were: Simon Goodchild, professor of mathematics education and leader of MatRIC,
with extensive experience of teaching mathematics in secondary school, but whose
teaching experience at the university level mostly consists of courses in mathematics
education; the author, a post-doctoral researcher in mathematics education with a
background in mathematics, whose role in the project was mostly data collection and
research support to Yuriy; and a doctoral student in mathematics education who was
mostly acting as an observer and research assistant in this iteration, in preparation for
the second iteration which would form a major part of his doctoral project. Since
writing about their collaboration at the level of detail necessary for this paper would
make anonymization practically impossible, professors Goodchild and Rogovchenko
kindly agreed not only to participate in the research, but also to make their identities
known. For simplicity and brevity, however, in what follows they will be referred to

444 [Link]:indrum2018:174854
as ME and M, respectively. The project was conducted at a well-regarded Norwegian
university where biology students take one compulsory mathematics course, taught
in the first semester, designed not specifically for the biology undergraduate program
but for students from about twenty different natural science programs. Hence, the
course provides little opportunity for focusing on issues specific to biology.
So far, the project has been through two cycles of development; however, in this
paper only the first iteration is considered. It consisted of a pilot, where the team met
with 10 volunteer students for one three-hour session, a regular sequence of four
three-hour sessions with a different group of 11 volunteer students concurrently with
their compulsory mathematics course, and a follow-up meeting with the second
group of students the following semester. All sessions were taught in English, but
student group-work was conducted in Norwegian. The basic structure of the sessions
was similar throughout, with an introductory lecture conducted by M introducing
some ideas and tools of MM, followed by small-group work on various MM tasks set
in a biological context, with whole-group follow-up, led by M but with some input
from ME. Examples of tasks included estimating the density of a rabbit population
based on the number of roadkill rabbits, and investigating the growth of a yeast
culture in a petri dish. All sessions were video recorded. Data analysis is still
ongoing, but initial results have been presented elsewhere (Viirman & Nardi, 2017).
The analysis presented in the present paper draws on the video-recorded data from
the four regular sessions in the first cycle of the project. To provide further insight
into the collaborative process involved in developing the project, after the conclusion
of the second iteration of the project the author of the present paper conducted an
audio-recorded, semi-structured group interview with the two principal members of
the team, M and ME. It was decided not to include the fourth member of the team
(the doctoral student) in the interview, since his involvement in the first iteration of
the project was peripheral. A thematic content analysis was then conducted, of the
interview data and the video data from the whole-class sessions, using the CoP
framework as a tool for structuring themes, focusing on signs of mutual engagement,
shared enterprise and joint repertoire, central and peripheral participation, and
possible obstacles to a CoP developing. For instance, concerning shared enterprise, I
looked at what M and ME said about their aims when developing the project, and
then examined the video-recordings of the sessions looking at how these aims
translated into what was emphasised in their teaching practice, comparing and
contrasting the practices of M and ME. Some consideration of my own role in the
project is appropriate. Although I was involved in the project from an early stage,
and am well acquainted with its aims, my role has been that of the researcher. I did
not participate in the planning of the content of the sessions, and my involvement in
the sessions was for the most part restricted to data collection.

445 [Link]:indrum2018:174854
RESULTS
Looking at the video recordings from the sessions, certain patterns in the behavior of
M and ME could be discerned, where M often had a very clear focus on the
mathematics and the tasks, while ME assumed responsibility for the students’
wellbeing and the nature of their learning. For instance, in the second session work
on the first task had taken longer than expected. Still, M, enthusiastic about the
mathematics and eager to get through all that he had planned for the session, begins
introducing the next task:
M: You will be solving an important medical problem. You are
ME: Yuriy?
M: Yes?
ME: Are you aware of the time?
(…)
ME: I’m talking to our student friends. Are you ready to go on and have a look at
this task, or are you saying “Hey, hang on a minute, it’s five o’clock?”
Similar situations occurred on several occasions during the sessions, with M getting
carried away by the mathematics, and ME intervening on the students’ behalf. At the
same time, this pattern was not entirely consistent. There were occasions where M
showed a clear concern about the students’ wellbeing and enjoyment, telling jokes
and striving to make the students feel at ease. Similarly, there were occasions where
ME got to present his solutions to some of the tasks, displaying an obvious
enthusiasm for and a deep engagement with the mathematics. Still, the overall
tendency was relatively clear.
Another more distinctive difference between M and ME concerned their engagement
with student contributions to the solution of the MM tasks. Again, M had a strong
focus on the mathematics, to the extent that he displayed signs of what Ralston
(2004) calls the One Right Answer Syndrome, clearly having one particular solution
in mind. This affected the way he lead discussions. For instance, when students came
up with solutions or suggestions that fit with this expected solution, they were
received as being the “correct” ones, as in these examples from sessions 1, 2 and 3,
respectively:
M: Any other assumptions you were using in your work?
S: Yeah, that 97 dead rabbits were per 24 hours
M: OK, that’s a good one
S: Because it said that they were easily recognizable
M: Correct, you read exactly what was meant there
(…)

446 [Link]:indrum2018:174854
 M: I will show you how the reasoning goes
---
M: When we look at the answers, the second group gave the correct one
---
M: This is the perfect one (…) the only thing we need to check is, how far is
this from my computation
Particularly the last quote displays how M used his own solution as the yardstick
against which student solutions were measured. On the other hand, when student
contributions did not resemble what he had in mind, M often did not comment on
them at all, or used a different terminology: “Well, that was a very constructive way
to solve the problem.” This contrasts with how ME talked about the students’ work
on the tasks, as in this example from session 3:
ME: We really, really, REALLY would like you to share with us solutions that
you have worked on within groups on those tasks. Now, it doesn’t matter at
all whether they are good or wonderful solutions, perfectly correct
solutions, whatever perfectly correct means, whether they look like his
[points at M] solutions or look like my solutions, which will be quite
different – it really doesn’t matter. What matters is the engagement, and the
thinking processes, and sharing with each other, and with us, the thinking
processes.
In line with this, on the few occasions during the sessions where ME conducted
discussions, he instead let the students lead. He interrupted students’ presentations
less frequently, and was not so quick to evaluate their contributions. When
presenting solutions to tasks he emphasized the process of doing mathematics over
the result, presenting his mistakes and dead-end attempts, seeing these as
opportunities for students’ learning.
Still, there were signs of M gaining an increased understanding of students’ needs
and potential for contribution as the project progressed. Preparing the last session, he
decided to involve the other team members more in the session, providing support to
the groups in their work, stating: “We should have done this in the first session”. He
also expressed this willingness to reflect on his own practice in the interview:
M: Feedback was essential, so it was like my teaching was shaped during those
four sessions towards some goals – I was experimenting, I was changing –
partly I was listening to what you [ME and the author] were saying, partly I
was trying to do what I felt was appropriate (…) I had very useful advice
which might have affected my way of teaching.
Interviewer: In what way?
M: I probably more often look at what I’m doing from the other side. From the
students’ perspective.

447 [Link]:indrum2018:174854
At this point, ME commented that this was “very much the didactician’s view”, and
that it showed “a readiness to reflect” upon his own teaching. This readiness to
reflect and change was further emphasized by how M later in the interview talked
about his plans for a new course he would be teaching, where he intended to
incorporate more student-centred teaching methods.
In the interview, when discussing the collaboration M emphasized the particular
qualities of himself and ME: “We are both weird people”. When asked what this
“weirdness” consisted in, he alluded to the stereotypical views mentioned by Dörfler
(2003), saying that he has a (for mathematicians) uncommon interest in teaching,
partly due to the influence of ME, while ME is (for a mathematics educator)
uncommonly knowledgeable in mathematics. Furthermore, both M and ME strongly
emphasized the value they place on their professional relationship – that two people
from adjacent, but still different fields can work so productively together. M added
that in no department where he had worked (and he has worked in numerous places,
in many different countries) had educational matters been on the agenda. The
collaboration was always on the individual level, never on the departmental, and as
ME added, this holds true even at their current university, where mathematicians and
mathematics educators work in the same department. At the same time, M expressed
the view that the differences between mathematicians and mathematics educators had
helped improve teamwork within the current project, but emphasized how this was
dependent upon all participants being able to engage with the mathematics.
Concerning other aspects of successful collaboration, ME highlighted the need for
mutual motivation: “Maybe this interdepartmental work has worked so well because
bioCEED wants something, and we [MatRIC] want something, and those two wants
coincided.” When asked what the mutual motivation was that enabled M and ME to
work together, ME highlighted the will to see MatRIC succeed, and how this success
was in part dependent on the collaboration between ME and M. M struggled with
formulating an answer, but emphasized interest in and curiosity about mathematics.
The main obstacle to collaboration mentioned in the interview was lack of time for
preparation:
ME: In terms of preparation, one of the things that I (…) found a bit frustrating,
was that the preparation was very often done between three and four o’clock
in the morning on the day of the session, and therefore there was very little
opportunity for discussion between the three of us [M, ME and the author]
about what was going to happen in that session.
ME attributed this to differences in individual styles of working, but still pointed out
that because of this and other time constraints, discussions about the mathematical
and didactical aspects of the sessions rarely managed to go below the surface level.
In response to this, M stated that his initial plan was to have the whole sequence of
activities pre-planned, but that the feedback he got while conducting the sessions led
to continuous refinement and change. Still, he agreed that providing the rest of the

448 [Link]:indrum2018:174854
team with the tasks further in advance would have been preferable. Generally, both
M and ME stated that the main difficulty with the project was the lack of time –
everyone involved were so busy with other things that they found it difficult to
engage with the project as deeply as should have been needed.

CONCLUSIONS AND DISCUSSION

In the findings from the actual teaching sessions, M was seen to a large extent
emphasising the mathematics and ME showing more concern with the students, their
wellbeing and learning experiences. These characteristics of the participation of M
and ME in the sessions align well with observations made by Cooper and Zaslavsky
(2017), where a similar pattern could be seen. At the same time, these characteristics
were not uniform throughout the sessions, and the overall picture of their
engagement with the content and the students is not quite as simplistic as that. Still,
the differences are discernible and should not be downplayed.
Furthermore, one might ask whether a CoP developed around the project. From what
M and ME expressed in the interview there are definite signs of mutual engagement
– they both emphasize the close professional relationship, and there is a dedicated
support between the members of the project in making it work. There is however,
less evidence when it comes to the joint enterprise and shared repertoire. In
particular, practical circumstances regarding the project – the late preparation, the
lack of time for discussion and reflection – made the development of a shared
repertoire difficult. The responsibility for developing activities and tasks fell mainly
on M, and the ways of interacting with students differed quite substantially, as the
findings above show. It is also difficult to find evidence of joint enterprise. On a
surface level, this could of course be said to be the project itself, but looking at what
M and ME say about the aims and motivations behind the project, the existence of a
joint enterprise is less clear. For ME, as leader of MatRIC, it is about the success of
the centre, whereas for M it has more to do with engagement in mathematics at a
personal level. Hence, it is difficult to claim that a CoP has developed.
Rather, what happened was that ME, occupying a more central position within the
mathematics education community, acted as a broker (Wenger, 1998, p. 105),
introducing elements of mathematics education practice and thereby contributing to
changing M’s practice. In this way, M displays the beginnings of a trajectory from a
peripheral position towards a more central. In the interview, M clearly states how
participation in the project has changed the way he views teaching – he talks about
“seeing things through students’ eyes”, described by ME as “a didactician’s view”.
These changes are perhaps less clearly visible in the video recordings from the
sessions, but as mentioned above there are signs of an increased awareness of
students’ needs as the project progressed. This could also be understood in terms of
the Spectrum of Pedagogical Awareness. When M talked about reflecting about his
own practice, and “seeing through students’ eyes”, this fits well with the Reflective

449 [Link]:indrum2018:174854
and Analytic level of the spectrum, which is characterized in part by “awareness in
starting to articulate pedagogical approaches and of reflection enabling making
strategies explicit” (Nardi, Jaworski & Hegedus, 2005, p. 293). At the same time,
much of what M did in the sessions would rather fit with the Intuitive and
Questioning level of the spectrum, with a less explicitly articulated pedagogical
thinking and a more intuitive recognition of students’ difficulties and needs (ibid, p.
293). Hence, a movement along the spectrum could be discerned as the project
progressed, and there is some support for claiming that M’s pedagogical awareness
has increased as a consequence of the project.
Regarding what factors contribute to functioning collaboration between
mathematicians and mathematics educators, what has been reported here is only a
single case, and one should be wary of drawing strong conclusions. Still, it is worthy
of note how well findings from literature, about mathematicians demanding that
mathematics educators know the mathematics (Nardi, 2008) and the need for
individual relationships and trust (Thompson, 2014) resonate with what M and ME
say in the interview. Both point out these factors as crucial to their collaboration,
while lack of time for joint preparation and reflection is pointed out as the main
obstacle. An awareness of the importance of these factors will be highly useful when
planning and conducting further collaborative efforts involving mathematicians and
mathematics educators. Indeed, providing institutional means and resources for
preparing and developing projects of the kind discussed in this paper might be one
way of moving parts of the responsibility for establishing collaboration from the
individual to the institutional level, something which is necessary if such
collaboration is to be sustainable.

ACKNOWLEDGMENT
My most sincere thanks are due to my collaborators in the project, both within
MatRIC and bioCEED, and to the students who kindly agreed to participate. Thanks
also to the participants of the NORMA 17 symposium “Researching university
mathematics education”, in particular Professor Elena Nardi, for insightful critique
of an earlier version of this paper.

REFERENCES
Biza, I., Jaworski, B., & Hemmi, K. (2014). Communities in university mathematics
education. Research in Mathematics Education, 16(2), 161–176.
Brewer, C. & Smith, D. (Eds.). (2011). Vision and change in undergraduate biology
education: A call to action. Washington DC: American Association for the
Advancement of Science.
Cooper, J. & Zaslavsky, O. (2017). A Mathematics Educator and a Mathematician
Co-Teaching Mathematics: Affordances for Teacher Education. In T. Dooley & G.
Gueudet (Eds.), Proceedings of the 10th Conference of European Research in

450 [Link]:indrum2018:174854
 Mathematics Education (CERME10). (pp. 2025-2032). Dublin, Ireland: DCU
Institute of Education and ERME.
Dörfler, W. (2003). Mathematics and mathematics education: Content and people,
relation and difference. Educational Studies in Mathematics, 54, 147–170.
Fried, M. & Dreyfus, T. (Eds.) (2014). Mathematics and mathematics education:
searching for common ground. Dordrecht: Springer.
Jaworski, B. & Matthews, J. (2011). How we teach mathematics: Discourses on/in
university teaching. In M. Pytlak & T. Rowland (Eds.), Proceedings of the 7th
Conference of European Research in Mathematics Education (CERME7) (pp.
2022-2032). Rzeszów, Poland: ERME.
Kilpatrick, J. (2008). The development of mathematics education as an academic
field. In M. Menghini, F. Furinghetti, L. Giacardi, & F. Arzarello (Eds.), The first
century of the international commission on mathematical instruction (1908–
2008): reflecting and shaping the world of mathematics education (pp. 25–39).
Rome: Istituto della Enciclopedia Italiana.
Nardi, E. (2008). Amongst mathematicians: Teaching and learning mathematics at
university level. Dordrecht: Springer.
Nardi, E., Jaworski, B., & Hegedus, S. (2005). A spectrum of pedagogical awareness
for undergraduate mathematics: From “tricks“ to “techniques“. Journal for
Research in Mathematics Education, 36(4), 284-316.
Ralston, A. (2004). Research mathematicians and mathematics education: A critique.
Notices of the American Mathematical Society, 51, 403–411.
Thompson, P. W. (2014). Reflections on collaboration between mathematics and
mathematics education. In M. N. Fried & T. Dreyfus (Eds.), Mathematics and
mathematics education: searching for common ground (pp. 313–333). Dordrecht:
Springer.
Treffert-Thomas, S. (2015). Conceptualising a university teaching practice in an
activity theory perspective. Nordic Studies in Mathematics Education, 20(2), 53-
77.
Viirman, O. & Nardi, E. (2017). From ritual to exploration: The evolution of Biology
students’ mathematical discourse through Mathematical Modelling activities. In T.
Dooley & G. Gueudet (Eds.), Proceedings of the 10th Conference of European
Research in Mathematics Education (CERME10). (pp. 2074-2081). Dublin,
Ireland: DCU Institute of Education and ERME.
Wenger, E. (1998). Communities of practice: Learning, meaning and identity.
Cambridge: Cambridge University Press.

451 [Link]:indrum2018:174854
 A case study of a university teacher of Calculus 1
Ibrahim Alzubaidi1 and Keith Jones²
1
Ibrahim Alzubaidi, Umm Alqura University, Saudi Arabia, iaaa1g15@[Link];
²Keith Jones, University of Southampton, UK
This case study of a university teacher of calculus1, based on data from a
questionnaire, semi-structured interview, and observation, illustrates how the
teacher used their knowledge of calculus teaching to sequence the building blocks of
mathematical theories (BBMT) of the concepts of calculus.
Keywords: teachers’ practices, university level, teaching calculus.

INTRODUCTION
Calculus is important at the university level because understanding calculus is an
essential step in understanding how the world works. It is a foundation on which
other skills can be built and is often a requirement or prerequisite for STEM
programmes. Nevertheless, it is a challenging area regardless of which educational
institution is offering the instruction (Petropoulou et al, 2016). As such, while there
is much research on calculus learning and teaching, Rasmussen et al (2014) identify
teacher knowledge as an under-developed topic. In this study our research questions
are; how does a calculus teacher demonstrate their pedagogical content knowledge
(PCK), to achieve their aims, to develop instructional strategies (ISs), and to assess
students’ understanding? Here we present elements of a single case study.

RESEARCH ON PCK OF UNIVERSITY TEACHERS

Numerous studies, from Shulman (1987) to Khakbaz (2015), have indicated that
teachers require different types of knowledge in the classroom. PCK was proposed
by Shulman (1987) as an essential component of teacher knowledge, defined as a
“special amalgam of content and pedagogy that is … understanding” (p.8). Later
work by, for example, Khakbaz (2015), confirms that it is necessary to go beyond the
subject (e.g. Calculus) and examine how the teacher interprets the subject matter
(e.g. the interpretation of calculus) and how this is linked to their role in facilitating
learning in the classroom. Research on PCK of university teachers remains limited.
THEORICAL FRAMEWORK AND METHODOLOGY
For the purpose of this study, we adopted Lessing’s (2016) PCK framework. Based
on this, we take PCK as a combination of knowledge of purposes of teaching
calculus, of the building blocks of mathematical theories (BBMT) and of ISs, while
taking into account knowledge of learners’ conceptions and difficulties.
A mixed methods multiple case study design was adopted for this research.
Qualitative data was gathered by semi-structured interview and observation of eight
taught sessions of the module from 15 April to 15 July 2017 to explore the ways in

452 [Link]:indrum2018:171659
which calculus teachers implemented their PCK in the classroom setting.
Quantitative data was gathered using a questionnaire (multiple choice) on their
knowledge of ISs, learners’ conceptions in teaching and learning calculus, and
learning difficulties. The qualitative data was analysed by coding and categorising
according to the theme in order to identify the ways that the participating teachers
demonstrated their PCK during their calculus teaching.
THE CASE OF TEACHER JOHN
John is a mid-career mathematician with four years of experience of teaching
calculus1 at a university. The calculus teaching strategies employed by John were a
pattern of topic-specific ISs based on using the BBMT in teaching calculus. From
our data, it was clear that John approached each calculus topic through the BBMT
using axioms, definitions, theorems and proof. Therefore, the lecture structure
chosen by John often aligned with the mathematical concept used in the topic. Here,
John often started the lecture with a definition, then theorem, and then sometimes the
proof and then examples illustrated with graphs. On one occasion, John asked the
students to read a proof. In an interview, John explained that, in the lectures, students
are sometimes asked to read a proof because John wants to understand the students’
conceptions, and misconceptions, of proofs. On one occasion, in a part of the topic
‘continuous functions on an interval’, John gave some examples and asked students
to give a definition. At other times, John used ISs such as diagnostic techniques,
(through class discussion, etc.), reviewing previous lessons as a way to introduce
subsequent lessons, and using various mathematical representations.
CONCLUSION
The analysis of John illustrates how knowledge of calculus teaching (i.e. calculus
PCK) was used in sequencing the BBMT of the concepts of calculus. The next stage
of the research is documenting other cases as part of the multiple case study.
REFERENCES
Khakbaz, A. (2016). Mathematics university teachers’ perception of pedagogical
content knowledge (PCK). IJMEST, 47(2), 185-196.
Lesseig, K. (2016). Investigating mathematical knowledge for teaching proof in
professional development. International Journal of Research in Education and
Science, 2(2), 253-270.
Petropoulou, G., et al. (2016). Addressing large cohorts of first year mathematics
students in lectures. In Proceedings of the 1st INDRUM, (p390-399). Montpellier.
[Link]
Rasmussen, C., Marrongelle, K., & Borba, M. C. (2014). Research on calculus: what
do we know and where do we need to go? ZDM, 46(4), 507-515.
Shulman, L. (1987). Knowledge and teaching: Foundations of the new
reform. Harvard Educational Review, 57(1), 1-23.

453 [Link]:indrum2018:171659
Who Becomes U.S. Mathematics Teachers: A Longitudinal Analysis of
First-time Praxis® II Mathematics Content Knowledge Examinees
David Glassmeyer1, Kimberly Cortes2, Jie Hao3, Herman Ray4, and Gregory
Rushton5
1
Kennesaw State University, Department of Secondary & Middle Grades Education,
dglassme@[Link]
2
Kennesaw State University, Department of Chemistry and Biochemistry,
klinenbe@[Link]
3
Kennesaw State University, Department of Statistics and Analytical Science,
jhao2@[Link]
4
Kennesaw State University, Department of Statistics and Analytical Science,
hray8@[Link]
5
Stony Brook University, Institute for STEM Education,
[Link]@[Link]
The U.S. uses the Praxis® tests to measure academic skills and the subject-specific
content knowledge needed to be a teacher. This study examined Praxis®
Mathematics Content Knowledge scores from 2006 to 2016 to describe patterns in
who are becoming certified to teach mathematics in the country (n=89,693).
Longitudinal analyses were used to discern patterns in the demographics of
examinees and trends in exam performance across several demographic
characteristics. The results reveal substantial differences in performance and pass
rates between examinees of different genders, races, undergraduate majors,
undergraduate GPAs, and census regions. From our analyses, we suggest several
measures for the improvement of mathematics teacher preparation.
Keywords: assessment practices, longitudinal analysis, mathematical content
knowledge, teacher preparation.

DESCRIPTION OF POSTER CONTENT

This poster will provide information on the U.S. Praxis® (ETS, 2017) tests to an
international audience. A small section will describe the theoretical perspective (Hill,
Blunk, Charalambous, Lewis, Phelps, Sleep, & Ball, 2008; Shulman, 1986) and how
it was used in this study. Our research questions will be included on the poster, as
well as how data were collected and analyzed to answer these questions. The middle
portion of the poster will include several graphics depicting the model resulting from
the analysis, specifically the statistically significant factors in whether test takers
pass the exam and the demographic characteristics of those who pass. These figures
will illustrate the substantial differences in performance and pass rates between
examinees of different genders, races, undergraduate majors, undergraduate GPAs,
and census regions.

454 [Link]:indrum2018:175003
We will offer several measures for the improvement of the mathematics teaching
workforce and establish potential leaks in the teacher pipeline that may impact the
quality and diversity of U.S. mathematics teachers (Gitomer, Brown, & Bonett,
2011). Specifically mathematics education faculty would do well to recognize that
students who have sat in their classrooms and lecture halls are represented in this
sample and that these scores are at least partially reflective of what they have
retained from their courses (Kleickmann, 2013). The adoption of instructional
strategies that promote students’ long-term conceptual understanding over those that
validate rote memorization may be a critical choice on the part of mathematics
faculty in improving the preparedness of future mathematics educators.

NOTES
In full disclosure, the U.S. Educational Testing Service reviewed this paper, though the analyses
1.
were conducted by our research team.
REFERENCES
Gitomer, D. H., Brown, T. L., & Bonett, J. (2011). Useful signal or unnecessary
obstacle? The role of basic skills tests in teacher preparation. Journal of Teacher
Education, 62(5), 431-445.
Hill, H. C., Blunk, M. L., Charalambous, C. Y., Lewis, J. M., Phelps, G. C., Sleep,
L., & Ball, D. L. (2008). Mathematical knowledge for teaching and the
mathematical quality of instruction: An exploratory study. Cognition and
instruction, 26(4), 430-511.
Kleickmann, T., Richter, D., Kunter, M., Elsner, J., Besser, M., Krauss, S., &
Baumert, J. (2013). Teachers’ content knowledge and pedagogical content
knowledge: The role of structural differences in teacher education. Journal of
Teacher Education, 64(1), 90-106.
Educational Testing Service (2017). Praxis® Subject Assessments Overview.
Retrieved May 5, 2017, from [Link]
Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching.
Educational researcher, 15(2), 4-14.
U.S. Department of Education. (2017, January 18). Teacher certification. Retrieved
October 16, 2017, from
[Link]

455 [Link]:indrum2018:175003
 Cross-linking maths – using keynotes to structure a curriculum for
future teachers
Tanja Hamann1 and Barbara Schmidt-Thieme²
12
Universität Hildesheim, Institut für Mathematik und Angewandte Informatik,
Germany, 1 hamann@[Link]
The paper gives a frame for a concept for university maths teacher education, that is
based on three keynotes: central scientific notions, history and language. Amongst
other benefits, the keynotes serve as cross-links between the different courses the
students go through in their studies.
Curricular and institutional issues concerning the teaching of mathematics at
university level, Teaching and learning of specific topics in university mathematics,
teacher education, history of mathematics, language education.

INITIAL CONDITIONS AND REQUIREMENTS

Subject-specific content knowledge is an essential component of professional teacher
knowledge (for the concept of professional knowledge see for example Schwarz
(2013)). It forms a basis for the ability to judge a specific topic’s significance for an
entire subject and thus forms an important requisite for didactical reflections.
Considering this, mathematics teachers’ content knowledge must comprise not only
advanced insight into singular topics but also a netlike overview of mathematics as
one, consisting of notions and relations between them as well as methods and basic
principles. To achieve this, a university curriculum for future teachers demands
cross-links between the single courses. Basic notions (such as sets, functions,
algorithms…) that are found in diverse mathematical subdomains seem suitable for
serving as cross-links between scientific lectures, both, in a horizontal and a vertical
way. This would transfer Bruner’s (1969) concept of fundamental ideas and a spiral
curriculum onto teacher education.
Basic ideas and fundamental notions can be found when one looks into the history of
a subject as they should appear throughout time (Schweiger, 1982), though maybe in
different disguises. Apart from that aspect, including the history of mathematics into
teacher education has been suggested long-since, presumably bearing a whole lot of
further benefits (Schubring, 2000; Katz, 2000; Jankvist et. al., 2016), for example
experiencing mathematics as a process conducted by human beings rather than as a
mere product.
Generation, clarification and precision of mathematical concepts and notions require
the use of language (e.g. Morgan et al., 2014). Without language, definitions and
propositions could not be worded and not be taught either, so language forms another
essential part of content knowledge, concerning scientific as well as educational
knowledge. Due to increasing heterogeneity in pupils’ language abilities, it is

456 [Link]:indrum2018:172106
necessary that future teachers learn to illustrate subject matters on differing language
levels and that they become able to vary by those levels.

THE HILDESHEIM CONCEPT OF LEARNING ALONG KEYNOTES

We believe that integrating the three keynotes – basic notions, history and language
of mathematics – into a teachers’ curriculum and thus cross-linking the courses
supports the construction of a professional teacher content knowledge as required.
Therefore, they should become a constitutional part of all scientific lectures
throughout studies. Our poster presents a yet draft concept of how the three keynotes
might constitute a scaffolding for a spiral curriculum, where historical and language
elements serve as enrichment as well as embedding of basic contentual and
methodical concepts. The concept seems convenient for linking subdomains as well
as different levels of challenge throughout a curriculum that is adapted to the specific
needs of future teachers. A research frame for evaluating efficacy of the concept has
not yet been developed.
Examples of implementation by emphasis on basic ideas in the lectures, integration
of historical tasks in practices as well as inclusion of training of language skills (for
more details on this see the poster presented by Schmidt-Thieme at INDRUM on the
project SuM_MaSt) will be presented on the poster.

REFERENCES
Bruner, J. S. (1969). The process of education. Cambridge, MA: Harvard Univ.
Press.
Jankvist, U. T., Mosvold, R., & Clark, K. (2016). Mathematical Knowledge for
Teaching Teachers: The case of history in mathematics education. Paper
presented at History and Pedagogy of Mathematics 2016, Montpellier, France.
Katz, V. J. (2000). Using history to teach mathematics: An international perspective.
Washington, D.C.: Mathematical Association of America.
Morgan, C., Craig, T., Schuette, M., & Wagner, D. (2014). Language and
communication in mathematics education: an overview of research in the field.
ZDM, 46(6).
Schubring, G. (2000). History of mathematics for trainee teachers. In Fauvel, J., &
van Maanen, J. (Eds.), History in Mathematics Education: The ICMI Study (91-
142). Dordrecht, The Netherlands: Springer Science+Business Media.
Schwarz, B. (2013). Professionelle Kompetenz von Mathematiklehramts-
studierenden: Eine Analyse der strukturellen Zusammenhänge. Wiesbaden,
Germany: Springer Spektrum.
Schweiger, F. (1982). “Fundamentale Ideen“ der Analysis und handlungsorientierter
Unterricht. BzMU 1982, 103-111.

457 [Link]:indrum2018:172106
 BiMathTutor – The program for further training of new tutors in
University Mathematics
Judith Huget
Bielefeld University, Faculty of Mathematics, IDM, jhuget@[Link]
In the Faculty of Mathematics at Bielefeld University, new tutors have to attend the
further training BiMathTutor, which took place for the first time in the winter term
2017. At this program, they learn about the essentials of teaching. This program is
being researched to the effects on tutors and their students. The poster presents the
theoretical framework, the concept of BiMathTutor, the research design and first
results.
Keywords:1. Teacher’s and student’s practices at university level, 7. Preparation
and training of university teachers, 4. Novel approaches to teaching.

INTRODUCTION
In this paper, we focus on the development of a further training for tutors. The main
aim of a further training for tutors is to improve teaching skills and students’
learning. In most cases, the tutors are experienced students who are interested in
passing down their knowledge and, of course, earning some money to support
themselves. These teaching skills should be based on a qualification program to
establish and maintain standards of teaching and learning.
In the Faculty of Mathematics at Bielefeld University, starting in October 2017, new
tutors have to attend the program for further training called “BiMathTutor”
(Bielefeld Mathematics Tutor Program). This program has been reformed over the
past year and evaluation is happening under different aspects, namely, the change of
attitude towards teaching, the expectations of the tutors of what is important for the
tutorials and the effects of “BiMathTutor” on the exam results of the students
attending the tutorials (see figure 1).

Figure 1: Time table of BiMathTutor

458 [Link]:indrum2018:174822
RESEARCH QUESTIONS
The object of research are the effects of is the tutor program. Hence, the research
questions for the survey are the following:
1. How do new tutors assess the further training?
2. What expectations do students visiting the tutoring class have?
3. What is the grade of effectiveness of the further training on the students’ exam
performance?
The first hypothesis is that new tutors will know the typical process and the
organisational aspects of a tutoring class, will be able to write and apply a horizon of
expectation, will execute different methods to engage students in a learning process,
will establish characteristics of a good tutoring class, will be aware of students’
difficulties dealing with university mathematics and will investigate their own
strengths and weaknesses. On the second research question, one will expect that
students have the goal to pass the exam. They will not be aware of the learning
process involved and the possible simplification by a tutor, who attended
BiMathTutor. The third hypothesis is that the further training has an effect on the
students’ exam result.

THEORETICAL FRAMEWORK
The theoretical framework is still in process of elaboration. The design of
BiMathTutor was made by Dee Fink’s theory “Designing courses for significant
learning” (2003). The key components “Learning goals”, “Teaching and Learning
activities” “Feedback and Assessment” have to be in unison. Additionally, the
situational factors must be considered.

DATA COLLECTION
The flowsheet (see figure 1) shows the different elements of the intended surveys.
Module 1-3 and the observation session plus feedback conversation is stated on the
top. The semester week shows a typical semester at Bielefeld University. There are
two general cohorts, one being the tutors and the other being students visiting the
tutorials. We have several times of measurement indicated by black bordered boxes.
The study uses a combination of quantitative and qualitative methods. The sample
size of the tutor survey is N1=24. The sample size of the student survey is N2= 597.
The test instruments are in form of questionnaires and exam results.

REFERENCES
Fink, L. D. (2003). A self-directed guide to designing courses for significant
learning. University of Oklahoma, 27.

459 [Link]:indrum2018:174822
Getting language awareness: A curriculum for language and language
teaching for pre-service studies for teachers of mathematics
Barbara Schmidt-Thieme1
1
University of Hildesheim, Germany, bst@[Link]
As language is an important medium as well as an aim of mathematics lessons at
school, teachers of mathematics have to achieve the required competencies in
language and language teaching. Therefore, we designed a curriculum, which offers
students opportunities to get these competencies throughout their academic studies.
Some examples and first results will be shown.
Keywords: Teaching and learning of specific topics, Curricular issues, Language
awareness, Special language; Explaining.
LANGUAGE AND LANGUAGE TEACHING COMPETENCIES AS PART OF
TEACHER PROFESSIONAL KNOWLEDGE
Teaching and learning mathematics is based on the use of language: for introducing
and defining new mathematical objects, discussing different ways of calculating,
documenting the results of a proof or explaining how to handle with teaching
materials, different representations are used, but almost always accompanied by
language. Language as a medium and an aim of mathematics lessons at school is a
well-known object of didactic research and laid down in several publications (see
Morgan, Craig, Schuette, & Wagner, 2014). But in the last years one can observe a
focus being laid on differentiations and transitions between different languages or
registers of one language used in mathematics (Duval 2006). On the other side, the
professional knowledge of teachers, and questions on how this can be developed in
pre-service and in-service lessons gained considerable interest (Gniffka & Roelcke,
2016; Kunter et al., 2013). This leads to following questions: (1) Which language-
related competencies a teacher has to have? How do they interact with other parts of
his professional knowledge? (2) When, where and how can he achieve these
competencies? For this in the last years several universities in Germany designed and
developed exemplary trails in pre-service teacher-studies. In most of the case these
are extra, but obligatory courses on “German as second language”. Another approach
is followed for example by a project called “Umbrüche gestalten”
([Link]
A “LANGUAGE CURRICULUM” (SUM_MAST)
Following the approach of this project we developed a pre-service “language curri-
culum” for teachers of mathematics (based on experiences out of the implementation
of language studies in teacher studies above mentioned) following four requirements
for the learning opportunities (used to guide the curriculum): they should
(a) be spread from the beginning to the end of the academic studies, like a vertical
spiral curriculum; (b) be integrated – besides explicit language related courses - in

460 [Link]:indrum2018:171923
mathematical lessons, for the technical language of mathematics is best learned by
doing mathematics; (c) evoke an active and reflective handling with language in
learning situations and be applied and tested in authentic situations; (d) include
individual feedback and allow some comparative measurements. (And in addition: It
should be transferable to other designs of academic studies for teachers at other
universities.) Our suggestions is, that a curriculum following these leads to a
connection of linguistic and mathematical knowledge as a theoretical base for
teaching practise.
After some pre-studies the design based research project SuM_MaSt started fully
2016 at the University of Hildesheim. Besides schemas of linguistic requirements of
mathematics lessons (in multilingual classrooms) and the language-related
competencies of teachers of mathematics (as a part of PCK), the poster will give an
overview over the types of “linguistic tasks” (theory/practice, real/simulated, ...) and
the curriculum implemented at the University of Hildesheim. Four examples will
clarify the tasks, the type of language competence, which can be achieved, and show
some first results of the evaluation of the above mentioned suggestion :
(a) linguistic tasks in mathematical lectures (examples: geometry, arithmetic, number
theory; evaluation (qual.): correctness of the mathematical content and adequacy of
language),
(b) input and tasks on “Explaining” in the lecture “arithmetic” (evaluation (qual.):
describing the development of linguistic competencies),
(c) contents and tasks of an explicit course “Language and Mathematics” (evaluation
(qual.): analysis of the products during the lessons),
(d) questionnaire for reflecting and self-assessment of the competence (evaluation
(quant.): questionnaire).
REFERENCES
Duval, R. (2006). A cognitive analysis of problems of comprehension in a learning of
mathematics. Educational Studies in Mathematics, 61, 103-131.
Gniffka, G, Roelcke, T. (2016). Fachsprachenvermitltung im Unterricht. Paderborn:
Schöningh.
Kunter, M., Baumert, J., Blum, W., Klusmann, U., Krauss, S., & Neubrand, M.
(Eds.). (2013). Cognitive activation in the mathematics classroom and professional
competence of teachers. New York: Springer.
Morgan, C., Craig, T., Schuette, M., & Wagner, D. (2014). Language and
communication in mathematics education: an overview of research in the field.
ZDM, 46(6).

461 [Link]:indrum2018:171923
 A university lecturer’s role and interactions in a flipped classroom
Erlend Sten Syrdalen
University of Agder, Norway, erless15@[Link]
This poster explores the role of the lecturer in a university level mathematics class
utilizing the flipped classroom approach to teaching. The role of the lecturer is
investigated using the Systematic Classroom Analysis Notation (SCAN) tool and the
theoretical framework of instrumental orchestration.
Keywords: Novel approaches to teaching, lecturers’ practices at university level,
flipped classrooms, instrumental orchestration, analysis of lecturers’ actions

NEW RULES OF INTERACTION IN FLIPPED CLASSROOMS

University lecturers traditionally presents new mathematical topics in class, while
students work on related tasks at home or in colloquiums arranged by the university.
The Flipped Classroom approach (FC) flips this around, having the students watch
pre-recorded videos at home, while working in a more collaborative and student-
centred manner while in class. This flipping of the classroom changes the rules of the
classroom drastically, as the lecturer now has to prepare videos for the students’ out-
of-class to “prime” them for an in-class active phase of learning (Fredriksen,
Hadjerrouit, Monaghan & Rensaa, 2017). To explore the new role of the lecturer in a
FC, I will look at a lecturer’s role using the framework of instrumental orchestration,
and analysing the in-class interactions of the lecturer using a modified version of the
Systematic Classroom Analysis Notation (SCAN) (Beeby, Burkhardt & Fraser,
1979).

THEORETICAL FRAMEWORK
To conceptualize the lecturers’ new role as a result of flipping the classroom I use
the framework of instrumental orchestration (Trouche, 2004). The term is defined as
the lecturer’s intentional and systematic organization and use of the various available
artifacts in a learning situation – in this case both out-of-class and in-class – in order
to facilitate students’ instrumental genesis – the process during which an artifact or
object is transformed into the psychological construct of an instrument, the
combination of artifacts and schemes to be used for different specific types of task.
Within instrumental orchestration it is possible to distinguish between three
elements: a didactic configuration, an exploitation mode and a didactical
performance (Drijvers, 2010). The didactic configuration is the arrangement of the
teaching setting and the artifacts involved in it. The exploitation mode is the manner
in which the lecturer uses a given didactic configuration for the benefit of his/her
didactic intentions. The didactic performance takes into consideration the decisions
the lecturer must make on the fly while performing in the given didactic

462 [Link]:indrum2018:173647
configuration with its didactic intentions. Using these three elements of
orchestration, it should possible to explore the lecturers’ new role as a result of FC.

RESEARCH QUESTIONS
The research questions address both the in-class and out-of-class aspects of the
lecturer’s new role in a FC.
1. How does the lecturer in a FC interact with the students in the lesson?
2. How does the lecturer perceive his new role, and how does this compare to
his observed interactions while in-class?

METHODOLOGY
To look into these questions, I have observed and interviewed a lecturer utilizing the
FC approach in a mathematics course for engineers. During the course of a week,
four 90 minute in-class sessions were videotaped. These sessions were part of a
module on mathematical series. Short videos, out-of-class preparation for problem-
solving activities in-class, facilitated the learning of Taylor series and its applications
and tests of series convergence. The study employs an interpretative research
paradigm with qualitative research methods. The videotapes of sessions will be
analyzed utilizing SCAN (Beeby et al., 1979) to code the teacher’s interactions with
students. SCAN is based on “time-slicing”, and works on three different time-scales
– “event”, “episode” and “activity”. A prominent feature is a set of qualifiers for each
activity, which evaluates the depth of demand and level of guidance in lecturer
interactions. The interview will be analyzed using open coding.

REFERENCES
Beeby, T., Burkhardt, H. & Fraser, R. (1979) Systematic Classroom Analysis
Notation for Mathematics Lessons. Nottingham: Shell Centre for Mathematical
Education.
Drijvers, P., Doorman, M., Boon, P., Reed, H., & Gravemeijer, K. (2010). The
teacher and the tool: Instrumental orchestrations in the technology-rich
mathematics classroom. Educational Studies in Mathematics, 75(2), 213-234.
Fredriksen, H., Hadjerrouit, S., Monaghan, J., & Rensaa, R. J. (2017) Exploring
Tensions in a Mathematical Course for Engineers utilizing a Flipped Classroom
Approach. In Dooley, T. & Gueudet, G.. (Eds.) Proceedings of the Tenth Congress
of the European Society for Research in Mathematics Education (CERME10,
February 1 – 5, 2017). Dublin, Ireland: DCU Institute of Education and ERME.
Trouche, L. (2004). Managing the complexity of human/machine interactions in
computerized learning environments: Guiding students’ command process through
instrumental orchestrations. International Journal of Computers for Mathematical
Learning, 9(3), 281-307.

463 [Link]:indrum2018:173647
 Programming in mathematics teacher education – a collaborative
teaching approach
Olov Viirman, Irina Pettersson, Johan Björklund, and Jonas Boustedt
University of Gävle, Sweden, [Link]@[Link]
Keywords: novel approaches to teaching, teaching and learning of mathematics in
other fields, team teaching, algorithmic thinking, programming.
The digitalization of modern society is having a major impact on schools and
schooling, and in recent years, programming and algorithmic thinking is increasingly
seen as topics to be included in curricula at all levels. In Sweden, for instance, the
National Agency for Education have instigated a revision of the national curricula to
emphasize what they denote as “digital competence” (Skolverket, 2017a, b). In
particular, the mathematics curriculum now includes the use of algorithmic thinking
and programming as tools for problem solving. Since these topics have not been a
mandatory part of the education of mathematics teachers in Sweden, this creates a
need for such courses aimed both at practicing and prospective teachers. Given the
direction taken in the revised curriculum, these courses will need to take an
integrated approach, focusing on how programming and algorithmic thinking can be
used to learn mathematics and to solve mathematical problems. This places particular
demands on the instructors teaching these courses. Not only do they need to have
knowledge of programming, but also of mathematics and mathematics education.
Most probably, courses in programming aimed at practicing mathematics teachers
will attract teachers from all levels of the educational system, placing additional
demands on the mathematical competence of the instructors. In the light of these
challenges, we are considering the idea of collaborative, or team, teaching. Having
mathematicians and/or mathematics educators collaborating with computer scientists
can provide different viewpoints on the topics considered and problems posed in the
course, thus serving to more closely connect the programming to the mathematics.
Moreover, having two teachers in the classroom naturally leads to discussions,
showing by example the type of questions experts pose when engaging with a topic,
and hopefully fostering a classroom climate where students also engage in
discussion. Before outlining the planned project, we sketch what team teaching is
about, and exemplify how it has been used in university mathematics education.
Various forms of collaborative teaching are already used in schools, where teachers
collaborate on course preparation, implementation, and assessment. There are several
models for collaborative teaching (Friend, Cook, Hurley-Chamberlain &
Shamberger, 2010, p. 12), ranging from “one teacher, one assistant” models, via
parallel and alternative teaching through to team teaching proper, with two teachers
in the classroom together, taking shared responsibility for content. The research
literature on team teaching of mathematics at the university level mostly consists of
case studies in the context of teacher education. For instance, Clarke and Kinuthia

464 [Link]:indrum2018:174857
(2009) describe a project where two lecturers collaborated on the planning of courses
on mathematical methods and instructional technology, emphasising the cross-
disciplinary character of the courses. However, one teacher for each course did the
actual teaching. What we aim to do is slightly different. A team of instructors with
backgrounds in mathematics/mathematics education and computer science will co-
teach a course on programming in a mathematical context, sharing responsibility for
content, planning and assessment. For part of the course, two teachers will lecture
jointly, presenting and discussing content together, providing different perspectives
on programming in school mathematics. For instance, in the context of probability
theory, the computer scientist can show how to program simulations of probabilistic
experiments, with the mathematician pointing out how the underlying mathematical
ideas influence the design of the simulation. The mathematician can then go deeper
into the mathematical theory, with the computer scientist highlighting the
algorithmic aspects. In this way, we aim to provide students with opportunities to
develop a deeper understanding of how to integrate programming and computational
methods in their mathematics classes. The programming course is still in its planning
stages, but we are currently piloting the team teaching approach in the context of a
first-year calculus course within the engineering programs.
From a research perspective, taking a discursive view on learning (Sfard, 2008) we
wish to investigate to what extent the team teaching might support students’ active
participation in mathematical and computational discourse. We are also interested in
how the discourse of the students develops through participation in such a course,
particularly concerning the interplay of mathematical and programming discourse.
Furthermore, we view team teaching as having great potential for teacher
development, with participating teachers being able to learn from one another as well
as jointly developing innovative teaching practices.

REFERENCES
Clarke, P. A. J., & Kinuthia, W. (2009). A collaborative teaching approach: Views of
a cohort of preservice teachers in mathematics and technology courses.
International Journal of Teaching and Learning in Higher Education, 21(1), 1-12.
Friend, M., Cook, L., Hurley-Chamberlain, D., & Shamberger, C. (2010). Co-
teaching: An illustration of the complexity of collaboration in special education.
Journal of Educational and Psychological Consultation, 20(1), 9-27.
Sfard, A. (2008). Thinking as communicating. Human development, the growth of
discourse, and mathematizing. New York, NY: Cambridge University Press.
Skolverket (2017a). Få syn på digitaliseringen på grundskolenivå. Stockholm,
Sweden: Skolverket.
Skolverket (2017b). Få syn på digitaliseringen på gymnasial nivå. Stockholm,
Sweden: Skolverket.

465 [Link]:indrum2018:174857
TWG 6: Transition to
and across university
 Different goals for pre-university mathematical bridging courses –
Comparative evaluations, instruments and selected results
1
Rolf Biehler, 1Elisa Lankeit, 1Silke Neuhaus, 2Reinhard Hochmuth, 2Christiane
Kuklinski, 2Elena Leis, 2Michael Liebendörfer, 1Niclas Schaper, and
1
Mirko Schürmann
1
University of Paderborn, Germany; biehler@[Link], elankeit@[Link],
nsilke@[Link], ²University of Hannover, Germany.
To ease difficulties in the transition from school to university, bridging courses are
implemented at many German universities. In this paper, we present instruments we
have developed for evaluating those bridging courses. We also show selected results
from six bridging courses at five German universities, comparing their different
goals and achievements.
Keywords: mathematical bridging courses, instruments for evaluation, transition to
and across university mathematics, teachers’ and students’ practices at university
level

INTRODUCTION
The transition from school to university is a big challenge for many students, espe-
cially in mathematics (Biehler, Hochmuth, Fischer, & Wassong, 2011; Gueudet,
2008). Several supportive measures such as pre-university bridging courses or
mathematical support centres are implemented at German universities to ease stu-
dents’ difficulties in the transition phase (Hoppenbrock, Biehler, Hochmuth, & Rück,
2016). But we often do not know how effective these supportive measures are as de-
tailed studies on the effects and success conditions are missing.

THE WIGEMATH-PROJECT
At this point, the ongoing WiGeMath project (Effects and success conditions of
mathematics learning support in the introductory study phase), a joint project of the
universities of Hannover and Paderborn (Liebendörfer et al., 2017) in collaboration
with 14 universities, comes in. We distinguish four types of support: pre-university
bridging courses, mathematics support centres, newly designed bridging lectures in
the first semester, and support systems accompanying traditional lectures such as e-
learning material or extra tutorials. The WiGeMath project’s goals are developing a
theoretical framework in order to be able to describe, analyse and compare support
measures, investigating effects and success conditions and elaborating recommenda-
tions for effective designs for mathematical support measures in the introductory
study phase. The theoretical framework for the examinations is the 3P model of
Thumser-Dauth (2007). It describes a programme evaluation for higher education
measures based on Chen’s theory-driven evaluation approach (Chen, 1990). Based
on transition literature from mathematics education, we refined this framework to

467 [Link]:indrum2018:174093
make it content-specific. Interviews with our collaborating universities and docu-
ment analysis were used to locate the specific measures in the framework
(Liebendörfer et al., 2017). The reconstructed programme theories contain goals,
procedures, circumstances and expected effects of the measures. Based on the theo-
retical framework, instruments were developed for evaluating the success of the
measures.
Bridging courses in Germany
Most German universities provide bridging courses in mathematics for various kinds
of beginning students shortly before the first semester. They differ in length, struc-
ture, amount of e-learning, content, audience, and goals. Some courses focus on the
repetition of school mathematics while others aim at introducing students to univer-
sity mathematics content and working methods (Bausch et al., 2014; Biehler &
Hochmuth, 2017).
One main aim is to evaluate the success of bridging courses by assessing short term
and medium term effects on attitudes and mathematical knowledge of the students.
Therefore, we questioned the students at the beginning of the course, immediately
after the course and after two months in the first semester.
Sample: Selected Bridging courses in the WiGeMath study
The following six bridging courses at five German universities are included in the
analysis.

University Online (O) or Duration Aimed at…

Attendance (A) in weeks
A O 5 Math., Comp. Sci., Engineer-
ing, teacher ed.
B A 2 Engineering
C A 2 Math., Physics, teacher ed.
DA A 4 Math., Comp. Sci., teacher ed.
DO O 4 All math programs (except
econ. and physics)
E A 2 Math., teacher ed.
Table 1: Overview over the investigated bridging courses

RESEARCH QUESTIONS
In this paper, we will focus on the post-test directly after the course. Apart from
evaluating different instruments used in the post-test questionnaire the research ques-
tions are:

468 [Link]:indrum2018:174093
 1. Which goals do lecturers of bridging courses set for their courses? How can
the profiles of the courses be compared and located in the WiGEMath frame-
work?
2. To which extend do students think they achieved explicit or implicit goals of
their bridging course?
3. How much do the results of two different instruments measuring to which ex-
tend the students think they achieved different goals in the bridging course dif-
fer?
4. How do the (theoretical) profiles set up by the lecturer differ from the empiri-
cal profiles of the course?

METHODS AND INSTRUMENTS FOR EVALUATING BRIDGING

COURSES
Instruments based on the WiGeMath Framework

The 13-pages questionnaire for the post-test contains about 205 Items – usually 6-
level Likert-scale from “strongly agree” to “strongly disagree”. The following tables
illustrate the scales with exemplary items. Most of our scales had a reliability above
0.6 in the majority of cases at all locations.
Category of goals Scale name Example item
School math. knowledge Identifying and overcoming “I got to know my indi-
and competencies deficiencies in school vidual deficiencies in
mathematics. school mathematics.”
Recapitulating and elaborat- “School mathematical
ing school mathematics topics were repeated.”
University math. knowledge University mathematics “I learned new mathe-
and competencies knowledge and competen- matical topics.”
cies
Mathematical terminology Mathematical terminology “I have learned new
mathematical symbols”
Table 2.1 Knowledge goals

Category of goals Scale name Example item (of 2 to 4 per scale)

Mathematical Process-related com- “I have learned how to read mathemati-
modes of working petences concerning cal texts.”
math. texts
Metaknowl. for math. “I know how to recapitulate a mathe-
modes of working

469 [Link]:indrum2018:174093
 matical lecture.”
Working autono- “I can work on mathematical tasks and
mously on math. topics on my own for some hours. “
tasks
University modes Organizing university “I learned how to organise my daily
of working (*) routine routine at university on my own.”
Learning strategies New ways for learn- “I learned about new ways to study
(*) ing mathematics mathematics.”
Learning and Study groups (*) “I learned to work in study groups.”
working behaviour
Knowledge about “I know the digital learning platforms
digital tools and how used at my university.”
to use them
Table 2.2 Behavioural (action-oriented) goals. (*) only one item

Category of Scale name Example item (of 3 to 7 per scale)

goals
Beliefs Metaknowl. and beliefs “In the course, I recognised the role
concerning higher maths. of proofs in higher mathematics.”
Relevance of Estimating how relevant “In the course, I became aware that
school maths. for school mathematics are school mathematics provides a basis
future studies for future studies and for my further studies.”
later profession
Table 2.3 Attitudinal goals

Category of Scale name Example item (of 2 to 6 per

goals scale)
Social contacts Social contacts between students “I met fellow students.“
Perceived social integration “I think the other students of the
(Rakoczy, Buff, & Lipowsky, course would help me, if neces-
2005) sary.”
Studying together with fellow “If I have an idea for a solution,
students (Liebendörfer et al., I will discuss it with other stu-
2014) dents.”
Making uni- Gaining insight into university “I gained insight into higher

470 [Link]:indrum2018:174093
versity study learning/teaching methods re- mathematics learning and teach-
demands garding mathematics ing methods at university.”
transparent Getting to know possible diffi- “I heard about possible difficul-
culties at the beginning of uni- ties at the beginning of my stud-
versity and how to solve them ies.”
Table 2.4 System-related goals
Additionally, we asked the participants about some affective characteristics (such as
mathematical fear and self-regulation), these items are not used in the analysis for
this paper.
Instruments adapted to the explicit goals of the course
The WiGeMath instruments are based on a comprehensive framework of potential
goals of a bridging course. As a supplement we used a learning outcome oriented
evaluation system, called BiLOE, proposed by Frank and Kaduk (2015). For the
BiLOE, each lecturer is asked to specify her/his three to six major learning goals in
his/her own words. Additionally, the lecturers had to specify up to seven study ac-
tivities that should help the students to achieve these goals. Students are asked to
evaluate these goals and activities. An important further element of the BiLOE is that
the students have to state their personal goals for the course and are asked to which
extend they think they achieved them. Those students who did not believe they
achieved a learning goal were asked to give reasons for this at the end of the ques-
tionnaire. The BiLOE also requests the students to evaluate the relevance of the lec-
turer’s goals and how much certain activities helped them to achieve those goals.

SELECTED RESULTS
RQ 1: The different profiles of the investigated bridging courses
We categorized the major goals provided by the lecturers in the BiLOE from the per-
spective of the WiGeMath framework. The results can be found in table 3. Quite dif-
ferent profiles become visible.
Category A B C DA DO E
School math. knowledge and competencies 1 1 0 0 1 0
University math. knowledge and competen- 1 1 1 1 1 3
cies
Mathematical terminology 1 0 0 0 0 1
Mathematical modes of working 0 0 0 0 0 2
University modes of working 0 0 1 0 1 0
Learning strategies 0 0 0 0 0 0
Learning and working behaviour 0 2 1 0 2 0
Social contacts 1 0 0 1 0 0

471 [Link]:indrum2018:174093
Making university demands transparent 0 0 0 2 0 0
Table 3: Number of learning objectives in the respective category mentioned by the
lecturers
It is striking that no attitudinal goals were mentioned among the major goals, neither
beliefs, nor affective features, nor mathematical enculturation. Likewise, none of the
lecturers mentioned teaching learning strategies as a goal of their bridging course.
RQ 2: Goal achievements
The results of the WiGeMath and the BiLOE instruments provide valuable informa-
tion for every single lecturer. The broader spectrum of the WiGeMath results will
moreover provide information on the effects of the course from the perspective of its
participants that the course lecturer may not have explicitly thought of in the selected
major goals. This analysis provides empirical profiles and assesses the success of the
various bridging courses.
In all cases, we calculated the percentage of students who rather agree up to fully
agree (meaning greater than 3 in Likert scales with 4 steps or greater than 4 in Likert
scales with 6 steps, respectively).
The following tables show the percentage of participants agreeing to the WiGeMath
scales concerning the respective categories.
Category of goals / Scale A B C DA DO E
School math. knowledge and competencies
Identifying and overcoming deficiencies in school math. 59 81 58 33 71 34
Recapitulating and elaborating school math. 86 89 69 34 87 35
University math. knowledge and competencies
University mathematics knowledge and competencies 60 89 100 95 73 95
Mathematical terminology
Math. terminology 63 86 99 97 74 96

Table 4: Results knowledge goals: Rounded percentage of participants agreeing to the

WiGeMath scales (n=651)
Category of goals / Scale A B C DA DO E
Mathematical modes of working
Process-related competences regarding math. texts 36 42 74 46 31 56
Meta knowledge for mathematical modes of working 34 72 70 66 46 65
Working autonomously on mathematical tasks 70 73 76 65 64 71

472 [Link]:indrum2018:174093
University modes of working
Organizing university routine 45 58 42 41 46 49
Learning strategies
New ways for learning mathematics 55 58 74 51 44 69
Learning and working behaviour
Study groups 25 49 88 47 17 82
Knowledge about digital tools and how to use them 58 80 54 67 77 45

Table 5: Results action-related goals: Rounded percentage of participants agreeing to

the WiGeMath scales (n=651)
Category of goals / Scale A B C DA DO E
Beliefs
Meta knowledge and beliefs towards higher mathematics 50 70 94 90 59 90
Relevance for eventual profession and for subsequent studies
Estimating how relevant school mathematic is for university and pro- 63 74 63 54 71 44
fession

Table 6: Results attitudinal goals: Rounded percentage of participants agreeing to the

WiGeMath scales (n=651)
Category of goals / Scale A B C DA DO E
Social contacts
Social contacts between students 34 84 98 76 64 86
Perceived social integration 46 82 93 88 74 89
Studying together with fellow students 36 71 81 66 57 88
Making university demands transparent
Gaining insight in university learning/ teaching methods regarding 34 82 95 85 44 91
mathematics
Getting to know possible difficulties at the beginning of university and 31 67 81 42 62 63
how to solve them

Table 7: Results system-related goals: Rounded percentage of participants agreeing to

the WiGeMath scales (n=651)
RQ 3: Differences between the two evaluation tools
To compare BiLOE and WiGeMath data, we first matched the learning goals given
by the lecturers with the framework categories. The BiLOE results are mostly similar
to the WiGeMath results. There are only six cases with differences of more than 15

473 [Link]:indrum2018:174093
percentage points. We reported back the interesting differences to the respective lec-
turers but these are relevant only for the individual and provide the general insight
that the WiGeMath framework is sufficient for the evaluation.
RQ 4: Comparison of theoretical and empirical profiles
With these empirical results, a re-evaluation of the profiles based on the formulated
learning goals of the respective bridging courses is possible. We will evaluate in
which categories the percentage of agreeing participants are high or low and com-
pare these results to the theoretical profiles. Here, “high” means an agreement to the
WiGeMath scales of more than 80% of the participants and “low” is an agreement of
less than 40%.
Course A. Based on the formulated learning goals, bridging course A has many
goals. School mathematics, university mathematics, mathematical modes of opera-
tion, and social contacts are aimed at equally strongly. The empirical results differ:
The only category with high agreement is school mathematics. There are some cate-
gories with low agreement, including social contacts, which was originally formu-
lated as a goal by the lecturer.
Course B. Goals in various categories were stated as well. The empirical profile is
similar but even broader: high agreement is reached in school mathematics, univer-
sity mathematics, mathematical terminology, social contacts, and gaining insight in
university learning and teaching methods. No goals concerning the last three catego-
ries were stated by the lecturer.
Course C formulated various goals. The empirical results show that there is high
agreement in the categories university mathematics and study groups. There is also
high agreement in the categories mathematical terminology, meta knowledge and be-
liefs towards high mathematics, social contacts and making university demands
transparent.
Course DA’s empirical results also fit the theoretical classification very well. Addi-
tionally, high agreement is reached for mathematical terminology and meta knowl-
edge and beliefs towards high mathematics. The only category with a low percentage
is school mathematics, which was not an explicit learning goal, however.
Course DO. The empirical profile of this bridging course differs significantly from
the profile based on learning goals. The only category with a high percentage of
agreement is school mathematics. Therefore, the focus of the course seems to be
more on school mathematics than on university mathematics. Based on the learning
goals, both could have been seen as equally strong.
Course E was the only one with a clear profile based on the formulated learning
goals which was on university mathematics (including mathematical terminology).
This is reflected in the empirical results, which however show a broader spectrum.
Additionally, there is also a high percentage in the categories studying in study

474 [Link]:indrum2018:174093
groups, meta knowledge and beliefs towards high mathematics, social contacts, and
gaining insight in university learning and teaching methods. Low agreement was
found concerning school mathematics, which was not a formulated learning goal.
SUMMARY AND DISCUSSION
The presented results are an intermediate step in communicating back to those who
were responsible for the respective bridging courses with two goals. The immediate
goal is to give feedback in order to improve and change the profile of the course – if
desired. The second goal is to redesign our instruments so that the future instrument
combines scales from the WiGeMath framework and more specific goals of the lec-
turers. The lecturers’ goals given are quite diverse but all goals could be classified
into the theoretical framework of the WiGeMath project. Some WiGeMath catego-
ries remained empty, however, e. g. learning strategies. We asked for the most im-
portant 5 goals, so it may be the case that our lecturers regarded them as minor ones.
Additionally, some lecturers stated more specific goals, while other stated general
ones. This may be due to the lack of experience with formulating learning goals as
most of the lecturers do not work in the field of didactics. It seems necessary and
valuable to extend the phase of specifying BiLOE goals by informing the lecturer in
more depth about the WiGeMath framework as a supportive frame for specifying
their own goals.
The students in the different courses differ when referring to their achievement of
certain goals and categories. This is no surprise. For example, an online bridging
course will not provide as much social contact to other students as an attendance
based bridging course. It is important to mention that all answers are based on the
students’ self-assessment. The instruments developed (termed WiGeMath scales) and
the BiLOE mostly yield similar results, sometimes the results differ. That can be ex-
plained by the BiLOE items being more specific or some of the learning goals only
having a corresponding category but no perfectly fitting scale was found, as the
questionnaire was already up to 13 pages long. The BiLOE is limited by the number
of goals a lecturer can state, while the developed instruments of the WiGeMath pro-
jects allow providing a general survey of the bridging course. Additionally, the
(theoretical) profiles set up by the lecturer differ from the empirical profiles of the
course. For example, no lecturer mentioned an attitudinal goal. Nevertheless, the
WiGeMath scales show that there is very high agreement in this category in relation
to courses for students in mathematics and mathematics teacher education. In these
courses the focus explicitly lies on university mathematics and not on the recapitula-
tion of school mathematics.

ACKNOWLEDGEMENTS
The WiGeMath project is supported by the German Federal Minister for Science and
Research (BMBF) (FKZ 01PB14015B and 01PB14015A).

475 [Link]:indrum2018:174093
REFERENCES
Bausch, I., Biehler, R., Bruder, R., Fischer, P. R., Hochmuth, R., Koepf, W.,
Schreiber, S., Wassong, T. (2014). Mathematische Vor- und Brückenkurse.
Konzepte, Probleme und Perspektiven. Wiesbaden: Springer Spektrum.
Biehler, R., & Hochmuth, R. (2017). Relating different mathematical praxeologies as
a challenge for designing mathematical content for bridging courses. In R.
Göller, R. Biehler, R. Hochmuth, & H.-G. Rück (Eds.), Didactics of
Mathematics in Higher Education as a Scientific Discipline – Conference
Proceedings (pp. 14-20). Kassel: Universitätsbibliothek Kassel.
Biehler, R., Hochmuth, R., Fischer, P. R., & Wassong, T. (2011). Transition von
Schule zu Hochschule in der Mathematik: Probleme und Lösungsansätze
Münster: WTM-Verlag.
Chen, H.-T. (1990). Theory-driven evaluations. Newbury Park: SAGE Publications.
Frank, A., & Kaduk, S. (2015). Lehrveranstaltungsevaluation als Ausgangspunkt für
Reflexion und Veränderung. Teaching Analysis Poll (TAP) und Bielefelder
Lernzielorientierte Evaluation (BiLOE). QM-Systeme in Entwicklung: Change
(or) Management?, 39.
Gueudet, G. (2008). Investigating the secondary–tertiary transition. Educational
studies in mathematics, 67(3), 237-254.
Hoppenbrock, A., Biehler, R., Hochmuth, R., & Rück, H.-G. (2016). Lehren und
Lernen von Mathematik in der Studieneingangsphase. Wiesbaden: Springer
Spektrum.
Liebendörfer, M., Hochmuth, R., Biehler, R., Schaper, N., Kuklinski, C., Khellaf, S.,
.Colberg, C., Scürmann, M. , Rothe, L. (2017). A framework for goal
dimensions of mathematics learning support in universities. In T. Dooley & G.
Guedet (Eds.), Proceedings of the Tenth Congress of the European Society for
Research in Mathematics Education (CERME10, February 1 – 5, 2017) (pp.
2177-2184). Dublin, Ireland: DCU Institute of Education and ERME.
Liebendörfer, M., Hochmuth, R., Schreiber, S., Göller, R., Kolter, J., Biehler, R., . . .
Ostsieker, L. (2014). Vorstellung eines Fragebogens zur Erfassung von
Lernstrategien in mathematikhaltigen Studiengängen. In J. Roth & J. Ames
(Eds.), Beiträge zum Mathematikunterricht 2014 (Vol. 1, pp. 739-742).
Münster: WTM-Verlag.
Rakoczy, K., Buff, A., & Lipowsky, F. (2005). Dokumentation der Erhebungs- und
Auswertungsinstrumente zur schweizerisch-deutschen Videostudie.
"Unterrichtsqualität, Lernverhalten und mathematisches Verständnis", 1.
Befragungsinstrumente. Frankfurt, Main: GFPF u. a.
Thumser-Dauth, K. (2007). Evaluation hochschuldidaktischer Weiterbildung:
Entwicklung, Bewertung und Umsetzung des 3P-Modells. Hamburg: Kovač.

476 [Link]:indrum2018:174093
 From single to multi-variable Calculus: a transition?
Hadas Brandes1 and Nadia Hardy2
1
Concordia University in Montreal, Department of Mathematics and Statistics,
Canada, [Link]@[Link]; 2Concordia University in Montreal,
Department of Mathematics and Statistics, Canada
We recently used the notion of praxeology from the Anthropological Theory of the
Didactic to model the knowledge that is necessary for students to learn in order to
succeed in an undergraduate multivariable Calculus course. We considered the
presence and absence of elements of the knowledge to be taught, as proposed by
curricular documents, in the knowledge to be learned, as indicated by final exams.
Our results indicate that the mathematical activities expected of students at this level
align with the activities observed in differential and integral Calculus, where
exercise-driven assessments set students’ work mainly in the recognition of types of
tasks and recollection of appropriate techniques.
Keywords: transition to and across university mathematics, assessment practices in
university mathematics education, teaching and learning of analysis and calculus,
Anthropological Theory of the Didactic, praxeology.

INTRODUCTION
So far, research on the teaching and learning of Calculus has focused on single-
variable Calculus. Cognitive and epistemological obstacles have been illustrated
against students’ learning of Calculus (Tall & Vinner, 1981; Sierpinska, 1994) and
an institutional perspective has also been taken to study the influence of institutional
practices on students’ learning of Calculus (Barbé, Bosch, Espinoza, & Gascón,
2005; Hardy, 2009). There’s a pattern that indicates Calculus students mostly engage
in procedural work that requires only a superficial grasp of the underlying concepts
(Hardy, 2009; Lithner, 2004; Selden, Selden, Hauk, & Mason, 1999).
We recently undertook a study that shifts the focus to multivariable Calculus courses
(Brandes, 2017). Our goal was to determine the knowledge that is essential for
students to learn in order to provide acceptable solutions on the final exam of an
undergraduate multivariable Calculus course. To this end, we used the notion of
praxeology from the Anthropological Theory of the Didactic (Chevallard, 2002) to
model the knowledge students are expected to learn and the knowledge to be taught.
We present our operationalization of this concept in the first part of this paper. In
second stage, we discuss a partial result of our study that places this multivariable
Calculus course along the transitions that university mathematics students undergo in
their engagement with mathematics (Winsløw, Barquero, de Vleeschouwer, &
Hardy, 2014).

477 [Link]:indrum2018:174202
THE EDUCATIONAL SYSTEM
We studied a ‘Multivariable Calculus I’ course offered to students in two
mathematics programs at a large North-American university. One of the programs is
for those who plan to join the workforce after graduation; the other aims to prepare
students for graduate studies in mathematics. Students in either stream will have
completed one-variable differential and integral Calculus and an introductory Linear
Algebra course on matrix and vector algebra. The multivariable Calculus course and
its sequel (‘Multivariable Calculus II’) are prerequisite to most of the courses in the
program geared towards graduate studies. Students usually complete Multivariable
Calculus I and II within the first year of their degree.
In any given term, the course is split into two sections per program, with about 70
students per section. The course is heavily coordinated across sections and terms
through a strict curriculum, course examiner, and common assessments. The course
outline specifies what to teach every week along with exercises from the textbook.
The course examiner writes common assessments for students in all sections. A
student’s grade is obtained from the highest of the following: 10% assignments, 30%
midterm, 60% final exam, or 10% assignments and 90% final exam. Finals exams are
therefore the crux of a student’s performance; in turn, the exams are consistent from
term to term in both format and content. Past exams are readily available to students,
and concern with their reactions prevents changes being made to the final exams.

ROUTINE PROBLEMS IN SINGLE-VARIABLE CALCULUS

We are interested in the mathematical activities with which students of a
multivariable Calculus course are expected to engage. We focus on the types of
problems that typify the learning of multivariable Calculus; a wealth of studies do so
for single-variable Calculus (Hardy, 2009; Lithner, 2004; Selden et al., 1999). These
studies emphasize the exercise-driven quality of the course assessments, in the sense
of Selden et al.’s (1999) routine problems, which “mimic sample problems found in
the text or lectures, except for minor changes in wording, notation, coefficients,
constants, or functions” and “can be solved by well-practiced methods” (p.18).
The exercise-driven quality of the course assessments extends to elements of the
curricula (Lithner, 2004). Calculus textbooks traditionally adhere to a definition-
theorem-example-exercise format, wherein the exercises repeat the problematics of
the examples and algorithms outlined in the text. Lithner (2004) measured the extent
to which intrinsic mathematical properties play a role in the minimal reasoning
required to solve routine tasks in traditional Calculus textbooks. Lithner’s
classification of reasoning types runs along a scale of how big a role is played by the
mathematical properties intrinsic to the problem versus the reapplication of known
algorithms; this scale runs parallel to Selden et al.’s (1999) spectrum of problems
from very routine to very non-routine, which vary based on how familiar the solver

478 [Link]:indrum2018:174202
is with the given problem. The more routine the problem, the less interaction is
required of the solver with the mathematics specific to that problem.
Assessments in North-American Calculus courses are largely drawn from the course
textbook, which Lithner (2004) showed to be steeped in routine problems.
Accordingly, he found students’ strategies to be anchored in what they recall
superficially rather than in the mathematics specific to a problem. This correlates
with Calculus students’ failure to complete non-routine problems (Selden et al.,
1999; Hardy, 2009). If textbook exercises can mostly be solved by identifying
superficial similarities with a known example (Lithner, 2004), then students’ non-
reliance on intrinsic mathematical properties and over-reliance on the recall of
algorithms may have roots in their learning environment. We follow this view by
framing our study within the Anthropological Theory of the Didactic and focusing
on elements of students’ learning environment: curricular and assessment documents.

ANTHROPOLOGICAL THEORY OF THE DIDACTIC (ATD)

Framework
From the perspective of the ATD, knowledge does not exist in a vacuum, rather, it is
bound to the institution in which it is shared and somehow connected to the
knowledge shared in related institutions; such connection is called transposition and
is of a didactic nature in the context of educational institutions (Chevallard, 1985).
Didactic transpositions take place along a spectrum of knowledge in which scholarly
mathematics (the knowledge developed, shared, and used by the experts – the
mathematicians) is transposed into knowledge to be taught in a given institution, up
to a transposition into knowledge actually learned by the students. This
transformation of mathematical knowledge takes form in several stages: scholarly
knowledge, knowledge to be taught, knowledge actually taught, knowledge to be
learned, and knowledge actually learned.
An essential feature of the ATD is an epistemological model called praxeology. It
allows the researcher to model knowledge at any stage of a didactic transposition.
The notion of praxeology is based in the assumption that any human activity consists
of a practical block (praxis) and a theoretical block (logos). The praxis is made up of
tasks to be accomplished and techniques with which to accomplish them; the
logos is the discourse that produces, justifies, and explains the techniques in the
practical block. Chevallard (1999) specifies two components of a theoretical block:
technology , the discourse that produces and justifies the techniques in the practical
block, and theory that justifies the technology.
In light of these theoretical considerations, and given our goal of finding the minimal
core of knowledge that students must learn in order to succeed in their multivariable
Calculus course, we treated three instances of didactic transposition. We created a
model of the knowledge to be learned, as determined by the final examinations; to

479 [Link]:indrum2018:174202
this end, we needed a model of the knowledge to be taught, as indicated by the
curricular documents. In order to familiarize ourselves with the mathematics prior to
these two instances of didactic transposition, we also created a reference model
based on the scholarly multivariable Calculus knowledge to be transposed. Before
we present our praxeological models of the knowledge to be taught and to be
learned, we review some of the literature about mathematics students’ praxeologies.
Transitions in students’ praxeologies
Winsløw et al. (2014) explain that students, at the pre-university level and in some
cases at the university level, tend to have a praxeology defined mostly by practice.
This is especially the case in differential and integral Calculus courses where
assessment is concerned mostly with the practical block and does not address the
ways in which the theoretical maintains the practical. This may have a precedent in
the way knowledge is taught in the classroom, as teachers may not have time to
justify tasks and techniques, given often-hefty curricula to deliver. Students, for their
part, tacitly accept the existence of a theoretical discourse supporting the practical
without concerning themselves with it (Hardy, 2009; Winsløw et al., 2014). Their
work is mainly in recognizing types of tasks and identifying a suitable technique
(Hardy, 2009; Winsløw et al., 2014), much as in Lithner’s identification of
similarities reasoning (2004) and Selden et al.’s routine problems (1999).
As students progress in university mathematics, they undergo two transitions. Where
once they might have ignored theoretical blocks and worked exclusively within the
practical block of a praxeology, they increasingly have to engage with theory and
technology in their completion of tasks. Winsløw et al. (2014) call the transition
from praxeologies that are purely practical to praxeologies that include a theoretical
and a practical block a first transition of university mathematical praxeologies
(p.101). For example, prior to the first transition, students complete tasks such as
using derivative rules to find the derivative of a function. Here, differentiability is an
always-met condition of the functions upon which students act in the tasks they do.
Prior to the first transition, it is sufficient for students to attend only to the practical
block of the mathematical knowledge; at the other end of this transition, students are
required to acknowledge the theoretical block as the justification for the techniques
they use for accomplishing a task. For instance, students may have to address the
differentiability condition of a function before engaging in finding its derivative.
A second transition occurs when students reach courses whose curricula and
assessment prioritize what once may have been the theoretical block of a praxeology;
as students transition into proof-making and validating, theoretical blocks of the past
become their practical blocks. For instance, the second transition will have occurred
in a student who knows to use the definition or theorems about continuity to prove
that, if a function is continuous, then some property of that function is true. The
characteristics of a second transition are that students explicitly acknowledge and use
the theoretical block to generate a technique for achieving a task.

480 [Link]:indrum2018:174202
KNOWLEDGE TO BE TAUGHT
The textbook of the multivariable Calculus course is typical of those used in North-
American Calculus courses and follows the usual definition-theorem-example-
exercise format. The course outline lists the textbook sections to be covered each
week and a choice of end-of-section exercises. By knowledge to be taught (KT) we
mean the mathematical knowledge in the sections and exercises listed on the outline.
To model the KT, we identified the praxeologies of which it consists.
In the case of the knowledge to be delivered in this course, we found that technology
and theory can be taken as one. There is no clear distinction between the two in the
textbook; the discourse throughout is set in the geometry and algebra of three-
dimensional space organized in the Cartesian system, and at times in Euclidean
metric spaces. However, the theory is not made explicit and tends to be woven into
the technology. Further, we found that the focus of the KT is mainly in the practical
blocks. For the purpose of this study, then, it was sufficient to compile a list of items
(definitions, theorems, etc.) that form the theoretical blocks of the praxeologies of
KT without distinguishing theory from technology.
This tended to the theoretical block of the praxeology that modelled each section of
the textbook on the course outline. To identify the tasks to be accomplished, we
considered the examples and the end-of-section exercises listed in the outline. To
describe the associated techniques, we consulted the examples and discussion
portions (theorems, explanations) of the text. To account for the build-up of
knowledge between sections (e.g. the notion of derivative of a vector function is
defined in one section and reused in later sections), we cross-referenced across
theoretical blocks and across and within practical blocks.

KNOWLEDGE TO BE LEARNED
In an operational sense, we define knowledge to be learned (KL) as the subset of the
KT which students need to know in order to provide solutions on final exams. This
operationalization was necessary from a methodological perspective: while the
questions in the final exams indicate the tasks to be accomplished, in most cases
there is no indication as to the expected technique or theoretical justifications. The
model of KT was therefore necessary to identify these elements of a mathematical
activity. In this sense, the main purpose of the model of KT was to model the KL.
Our operationalization, although useful to describe and characterize the KL, does not
properly reflect the fact that a transposition takes place and that some of the
praxeological elements (likely, elements of the theoretical block) are more likely ill-
defined than well-defined subsets of the praxeological elements of the KT. While the
KL may borrow elements of the KT praxeologies, the discourse that unifies the two
blocks of a praxeology might be distorted in the transposition.
Our model of KL is based on twelve final exams given recently within a span of
three years. We described the solution to each exam question in terms of KT task-

481 [Link]:indrum2018:174202
technique pairs that occur in the solution. Here is an instance of this work. Consider
the following item from one of the exams:
Find the tangent plane that touches at , where the surface is given by

We recorded this as “to find the tangent plane to a surface at a point.” This task
corresponds identically to task from the KT model; in turn, the technique for
this task requires the completion of : to find the value of the partial derivative of
a function at a point. Thus, we associated to this task the KT sequence
This particular exam task corresponded identically to a
KT task; this was not always the case. Nevertheless, apart from a handful of cases,
we were able to identify sequences of task-technique pairs that would form complete
solutions to the exam questions; this methodological affordance may attest to the
routine quality (Selden et al., 1999) of the tasks students are expected to accomplish.
Next, we grouped tasks of the same type so as to reflect praxeologies that occur in
the KT. For example, the following tasks come up in solutions to exam questions:
To find the first partial derivatives of a function
To find the first partial derivatives of a two-variable function defined implicitly

To verify that a two-variable function satisfies a partial differential equation

This cluster of tasks is drawn from the praxeology of KT specific to partial

derivatives. Altogether, we partitioned the model that captures KL about partial
derivatives and surfaces into groups of tasks that match up with these praxeologies
of KT: the above cluster specific to partial derivatives, along with tasks that draw
from KT praxeologies specific to functions of several variables, the chain rule,
tangent planes and linear approximations, directional derivatives and the gradient
vector, extreme values, and Lagrange multipliers. Organizing the model of KL in
parallel to the model of KT facilitated our analysis of the structure of the KL.

STRUCTURE OF THE KNOWLEDGE TO BE LEARNED

The KL has to do with partial derivatives and surfaces; space curves and vector
functions; equations of lines and planes and distance in ; limits of rational
functions; polar curves; and Taylor Series. Let’s call ‘ideal student’ one who has the
requisite knowledge to write acceptable solutions in a final exam. How might we
characterize the praxeologies of an ideal student in this course? Below, we consider
which parts of the KT praxeologies are to be learned and characterize them in the
language of Lithner (2004) and Selden et al. (1999).
Knowledge from all KT praxeologies occurs as knowledge to be learned. Thus, the
KL is not necessarily a subset of KT in the sense that some praxeologies are to be

482 [Link]:indrum2018:174202
learned while others are not. Rather, we found that the KL is a subset in the sense of
what’s left of the KT praxeologies after the didactic transposition of KT into KL.
First, the practical blocks of the KT praxeologies are downsized in this transposition.
For instance, consider the praxeology of KT about polar coordinates. The ideal
student can convert polar equations into Cartesian equations and sketch the curve –
given the following curves (up to a change in constants and functions sine or cosine):
;
The algebraic manipulations specific to converting these types of polar equations
into Cartesian equations are in examples from the textbook, as is the technique for
sketching them. The ideal student’s topos (‘action space’) (Chevallard, 2002) does
not need to extend beyond the point praxeology (a praxeology of knowledge that is
particular to a single type of task) specific to these functions. We found many of the
practical blocks of KT praxeologies to be reduced in this way to point praxeologies.
Most of the praxeologies of KT are downsized in another sense: their theoretical
block is removed following the transposition from KT to KL. For instance, consider
the praxeologies that constitute the knowledge to be taught about partial derivatives.
We found that the practical blocks are reduced to computational tasks where the
ideal student needs to apply the appropriate differentiation algorithm; the geometric
interpretation of partial derivatives as slopes is unneeded and the ideal student does
not need to know any of the theory or technology at the backbone of the procedures.
The ideal student does not need to know the limit-based definition of partial
derivative nor the definitions and roles of limits, continuity, and differentiability in
the concepts of derivative, gradient, and extrema of a function. The theoretical
blocks of these praxeologies vanish in the transposition from KT to KL. In general, it
seems that the ideal student needs to be fluent in the algorithms prescribed by
praxeologies of the KT but doesn’t need to justify or explain them.
The absence of theoretical blocks in the ideal student’s praxeology is manifested in
several ways: first, the student needn’t justify the validity or choice of technique (e.g.
by verifying or stating that the chain rule is applicable, since the functions in the
exams are always differentiable); second, the exam questions do not require students
to interpret any results (e.g. by making a sketch of a surface near a point where some
geometric properties of the surface were computed); and finally, it suffices to have a
superficial grasp of the concepts in the theoretical blocks in order to accomplish the
types of tasks in the final exams. We expand on this point.
In general, the ideal student can recognize task types and identify the appropriate
technique, in reasoning similar to Lithner’s (2004) identification of similarities (IS),
whereby a strategy for tackling a problem is chosen based on the similarities of
certain surface properties between the new problem and a known problem (e.g. given
a limit-finding problem, note whether the limit is taken at a numerical value or
infinity and identify the type of function involved). For instance, the exam questions

483 [Link]:indrum2018:174202
specific to limits of multivariable functions are as follows: find the limit of a
function at the origin, if it exists, or show that it does not exist. The function
given in the exams is either an odd rational function with no limit at the origin or it
involves a trigonometric component which could be rid of to reduce to a rational
function in the process of an argument (in these cases, the exam functions
invariably have limit ). This task occurs in examples in the textbook and exercises
in these students’ assignments. In general, tasks required by the exam questions were
similar to those in the KT, so that students could rely on IS reasoning rather than on
the underlying mathematics in order to choose the appropriate technique. This course
is therefore in line with students’ pre-university mathematics, where much of their
responsibility is in recognizing types of tasks and choosing an appropriate known
technique (Winsløw et al., 2004).
IS reasoning is characterized as requiring little reflection on the intrinsic
mathematical properties of the problem at hand (Lithner, 2004). To successfully
implement IS, the ideal student needs to recognize terms in the question statements
(arc length, curvature, normal plane, binormal vector…) and the formulas for
deriving them. But the ideal student is not tested on the meaning of these quantities
and geometric properties as they relate to a curve at a point (e.g. a student might need
to find the equation for an osculating plane, but does not need to explain what the
osculating plane describes). The irrelevance of intrinsic mathematical properties to
the tasks students need to achieve suggests that the theoretical block of KT
praxeologies need not be present in the ideal student of either course.
Theoretical blocks are missing from the ideal student’s topos in a few senses: the
student is not required to justify or explain the techniques chosen to complete a task,
and at times is even told which technique to use (e.g. via instructions to ‘use
Lagrange Multipliers’ or ‘use the chain rule’). The ideal student is not required to
interpret the numerical or algebraic results of their calculus in any way; and it
suffices to learn the components of the theoretical blocks only superficially. In all,
this multivariable Calculus course seems to follow in the pre-university mathematics
tradition whereby students need not link the practical and theoretical blocks of a
praxeology (Winsløw et al., 2014). Further, the components of the practical blocks
themselves are discrete, as the ideal student does not need to combine tasks in any
way – for instance, the ideal student must know how to find invariant quantities of a
curve, but needn’t provide a local description of a curve based on its invariant
quantities. This may be called the “compartmentalization of knowledge in calculus
courses” described by Winsløw et al. (2014, p.104).
On the whole, it appears that only a surface version of the KT theoretical blocks is
essential for the ideal student to learn: they need to know terms and associated
formulas, in some cases have some intuitive image of certain concepts, and be fluent
in the algorithms described by the technologies. This surface acquisition of the
theoretical block serves to recognize routine tasks and identify a suitable technique.

484 [Link]:indrum2018:174202
In light of the absence of theoretical blocks in the minimal core of knowledge that is
essential for students to learn in order to provide solutions to exam questions, we
conclude that the KL cannot be described by actual praxeologies (made up of a
praxis and a logos). Rather, the KL is an amalgamation of practical blocks. This
places this university-level multivariable Calculus course in the stage prior to the
first transition in university mathematics education previously discussed:

Figure 1. Transitions in university mathematics education (Winsløw et al., 2014, p.101)

where refers to the practical block of a praxeology and to its theoretical block.
Winsløw et al. (2014) explain that this first transition occurs when students no
longer work strictly within the practical block of a praxeology and begin to
incorporate a theoretical block, perhaps by using it to justify or produce a technique;
a second transition occurs when students’ past theoretical blocks turn into their
current practical blocks, as when they start making and validating proofs in Analysis.

CONCLUSIONS
The aim of our study was to determine the minimal core of knowledge that is
necessary for students to learn in a multivariable Calculus course in order to provide
acceptable solutions on their final exam. We found that the exercise-driven quality of
the course assessments makes it essential for students to recognize certain types of
tasks and to identify the appropriate technique, but does not require students to learn
the theoretical block that maintains these tasks and techniques.
Historically, the studied educational system introduced the multivariable Calculus
course as a prerequisite to Analysis in an effort to help students adapt to university
mathematics in the first year of their studies. It seems, however, that the
mathematical activities expected of students in this bridge between pre-university
and university courses are of the type expected in past Calculus courses: students’
action space is fully within the practical block of the praxeologies that model the
knowledge to be taught. As a result, students are no more required to engage with the
theoretical in this Calculus course than they previously were. Meanwhile, the
mathematical activities in Analysis courses are two steps ahead, after the second
transition described by Winsløw et al. (2014), where students must work within what
once were the theoretical blocks that backed the practical of Calculus. The question
therefore remains: what course would make it essential for students to incorporate
the theoretical blocks of a praxeology into the work they do in a practical block?

485 [Link]:indrum2018:174202
REFERENCES
Barbé, J., Bosch, M., Espinoza, L., & Gascón, J. (2005). Didactic Restrictions on the
Teacher’s Practice: The Case of Limits of Functions in Spanish High Schools. In
C. Laborde, M.-J. Perrin-Glorian, A. Sierpinska (Eds.), Beyond the Apparent
Banality of the Mathematics Classroom (pp. 235-268). Boston, MA: Springer.
Brandes, H. (2017). The bare necessities for doing undergraduate multivariable
Calculus (Unpublished master's thesis). Concordia University, Montreal, Quebec,
Canada.
Chevallard, Y. (1985). La transposition didactique, du savoir savant au savoir
enseigné [The didactic transposition, from the learned knowledge to the taught
knowledge]. Grenoble: La Pensée Sauvage.
Chevallard, Y. (1999). L’analyse des pratiques enseignantes en théorie
anthropologique du didactique [The analysis of teaching practice in the
anthropological theory of the didactic]. Recherches en Didactique des
Mathématiques, 19, 221–266.
Chevallard, Y. (2002). Organiser l’étude 1. Structures & fonctions [To organise the
study, 1]. In J.-L. Dorier, M. Artaud, M. Artigue, R. Berthelot, & R. Floris (Eds.),
Actes de la XIe école d’été de didactique des mathématiques [Proceedings of the
11th summer school on didactics of mathematics] (pp. 3–32). Grenoble: La Pensée
Sauvage.
Hardy, N. (2009). Students’ perceptions of institutional practices: The case of limits
of functions in college level Calculus courses. Educational Studies in
Mathematics, 72, 341–358. doi:10.1007/s10649-009-9199-8
Lithner, J. (2004). Mathematical reasoning in Calculus textbook exercises. The
Journal of Mathematical Behavior, 23(4), 405-427.
doi:10.1016/[Link].2004.09.003
Selden, A., Selden, J., Hauk, S., & Mason, A. (1999). Do Calculus students
eventually learn to solve non-routine problems? (Report No. 1999-5). Cookeville,
TN: Tennessee Technological University.
Sierpinska, A. (1994). Understanding in Mathematics. Bristol, PA: The Falmer
Press, Taylor & Francis Inc.
Tall, D., & Vinner, S. (1981). Concept image and concept definition in mathematics
with particular reference to limits and continuity. Educational Studies in
Mathematics, 12(2), 151-169. doi:10.1007/bf00305619
Winsløw, C., Barquero, B., Vleeschouwer, M. D., & Hardy, N. (2014). An
institutional approach to university mathematics education: from dual vector
spaces to questioning the world. Research in Mathematics Education, 16(2), 95-
111. doi:10.1080/14794802.2014.918345

486 [Link]:indrum2018:174202
 A study of transitions in an undergraduate mathematics program
Laura Broley and Nadia Hardy
Concordia University in Montreal, Canada, l_brole@[Link]
In this paper, we introduce an in-progress study of the transitions students face as
they advance in their mathematics courses. Previous work has discussed the changes
that occur in the transition from high school to university. With regards to the
knowledge students are expected to learn, however, significant similarities have
been noted: to do well in introductory university courses, students can learn to solve
a particular subset of tasks through routinized techniques, with limited awareness of
the supporting mathematical theory. In contrast, students in advanced courses are
required to work with and on that theory. The first stage of our project aims to better
understand this transition by building praxeological models of the knowledge to be
learned in a succession of two introductory analysis courses.
Keywords: Transition to and across university mathematics, teachers’ and students’
practices at university level, teaching and learning of analysis and calculus.

INTRODUCTION
Several studies have discussed the specific knowledge taught and learned in pre-
calculus, calculus, and analysis courses, from different perspectives: for example,
concept image and concept definition (e.g., O’Shea, 2016), APOS theory (e.g.,
Martínez-Planell, Trigueros Gaisman, & Mcgee, 2016), and the Anthropological
Theory of the Didactic (ATD; e.g., Bergé, 2016). Our starting point is the general
and relatively vague question of when in an undergraduate degree in mathematics
does a student need (need in the sense of to succeed in the course) to engage in
mathematical activities that may substantially, or meaningfully, lead to developing
mathematical practices. We consider and frame this question within the ATD
(Chevallard, 1999), which provides theoretical tools for modelling any human
activity or practice. The semantic distinction between these two words is essential to
us. Our hypothesis is that the kinds of didactic constructs to which professors and
students are exposed are decisive in fostering the emergence of practices out of
collections of local, particular, and relatively short-lived activities. From the
theoretical stance we take, this means the development of mathematical knowledge
out of local, particular, and relatively short-lived mathematical activities.
Previous research has found that the activities proposed to students in introductory
calculus courses do not necessarily encourage the development of mathematical
practices. Lithner’s (2004) study of the exercises in undergraduate calculus
textbooks used in Sweden led to the conclusion that the majority of tasks students
encounter can be solved by mathematically superficial techniques such as finding
and copying a similar solution outlined somewhere in the same section of the book.
When working in Spanish high school calculus classes, Barbé, Bosch, Espinoza, and

487 [Link]:indrum2018:174227
Gascón (2005) observed teachers implementing mathematically incomplete
practices: they solved numerous tasks in hopes of guiding students in developing
solid mathematical techniques, but struggled to introduce any lasting rational
discourse (i.e., theoretical block) that produced or explained the techniques. Hardy
(2009), who conducted task-based interviews with students in North American
college calculus courses, showed that in the absence of such a theoretical block,
students construct non-mathematical reasoning to support the highly routinized
practical block they develop (i.e., the techniques and corresponding types of tasks).
To describe the kinds of transitions students are expected to go through as they
progress in their university mathematics coursework, Winsløw (2008) introduced the
model depicted in Figure 1 below. The conjecture is that students encounter at least
two types of transitions in the practices they are supposed to develop. The first
requires them to gain some level of awareness of the theoretical block that was once
absent from their exclusively practical work; the second occurs when elements of
that theoretical block become part of the practical block with which they must
engage autonomously. Think, for instance, of how some early university courses
spend a significant amount of time in lectures elaborating previously scarce
definitions, theorems, and proofs, which students may be expected to understand
enough to quote in assignments or reproduce on exams. In contrast, more advanced
coursework requires students to develop their own proofs, often involving the more
abstract objects that were part of the theoretical block constructed in earlier courses.

Figure 1: Transitions in university mathematics coursework (from Winsløw, 2008)

A recent study suggests that in the context of undergraduate multivariable calculus
courses, students are not yet required to go through Transition 1: the models of the
knowledge to be learned in these courses show that students are exposed to a limited
practical block (Π), with no need to work with or on the corresponding theoretical
block (Λ; Brandes, 2017). This said, students are indeed expected to work with and
on mathematical theory when they take advanced courses later on.
A few questions arise from this:
1. What does this “work with and on a theoretical block” look like in comparison
to the routinized, principally practical activity in which students seem to be
engaging in introductory courses?

488 [Link]:indrum2018:174227
 2. If the practices students develop in advanced mathematics courses can be
modelled by the third stage shown in Figure 1, when, if ever, do students’
practices reflect the second stage, and what are the mathematical activities
proposed to them in such contexts?
The purpose of our study is to contribute to addressing these questions, and
therefore, to the discussion of the transitions students face. To do so, we propose to
model the knowledge at different stages in the didactic transposition process in two
courses contained in what we will call the “analysis path” in a typical undergraduate
mathematics program in North America (US and Canada). Ultimately, the goal is to
reflect on the general question mentioned above: Can the activities in which students
are obliged to engage lead to the development of mathematical practices (i.e.,
mathematical knowledge)?

THEORETICAL FRAMEWORK
“Activity and practice”
As mentioned above, we have come to see the semantic difference between activity
and practice as pertinent to our work. The ATD’s notion of praxeology provides a
fundamental model for defining mathematical practice, which, in the context of the
theory, is equated to mathematical knowledge. According to the model, any practice
(or piece of knowledge) can be represented by a quadruplet [T, τ, θ, Θ] involving
four interconnected components: a type of tasks T, which generates the practice, the
corresponding collection of techniques τ developed to accomplish T, the discourse
used to describe, justify, explain, and produce the techniques (i.e., their technologies
θ), and the underlying theories Θ that serve as a foundation of the technological
discourse. As students progress in their studies of mathematics, they engage in
numerous activities, which progressively determine the practices they develop.
As a strictly hypothetical example, we could imagine students in an introductory
calculus course being asked to engage in the following activities, inspired by a
commonly used calculus textbook (Stewart, 2008):
a1: Estimate the area under the graph of from to using
four approximating rectangles and right endpoints. Sketch the graph and the
rectangles. Is your estimate an underestimate or an overestimate? What happens if
you repeat the exercise with left endpoints? (Areas and Distances, Section 5.1)
a2: Evaluate . (The Substitution Rule, Section 5.5)

a3: Determine if is convergent or divergent. If it is convergent,

evaluate it. (Improper Integrals, Section 7.8)
If these were the first activities completed by students in the corresponding sections,
we could expect their actions to be localized and particular. In other words, the

489 [Link]:indrum2018:174227
solutions students produce would likely be the result of their engagement in a
relatively isolated act of figuring out how to solve the specific given problem. As the
students participate in more activities, however, they may be exposed to tasks of the
same type, and may consequently begin to develop a related practice. By the end of a
calculus course, for example, students will have typically solved a large number of
problems involving the calculation of definite integrals by way of various integration
techniques. From this, they may have learned to recognize other activities (e.g.,
or ) as forming a type of task with a2, and therefore as
requiring the same technique: making a substitution (not forgetting to change the
bounds!), determining the anti-derivative of the new function, and calculating the
difference of this anti-derivative evaluated at the bounds. In comparison, certain
activities may be encountered by students only in insignificant (e.g., unevaluated),
rare, and/or disconnected situations. The action of accomplishing those tasks may
hence remain isolated and particular, never contributing to the development of
practices. Activities like a1 or a3, for instance, might never be encountered beyond a
few recommended exercises at isolated, unique moments in the course.
Research confirms that the collection of activities given (and not given) to students
play a crucial role in determining the kinds of practices they develop (and do not
develop). Although students may seem to be learning mathematical practices, they
may in fact be engaging in isolated activities or developing practices of a non-
mathematical nature. In her research, Hardy (2009) noticed that when first-year
calculus students are given activities related to slightly non-routine tasks, they often
apply techniques in a mathematically unjust way. For example, when asked to
compute , 20 out of 28 students factored, seven of which did direct
substitution first. Her analysis of students’ discourse during task-based interviews
led her to conclude that the students tended to justify their techniques through
perceived norms. She specifies, for example, that “it seems that students were doing
substitution not to find the limit or to characterize an indetermination, but because
that is ‘what you do first’” (p. 351). To explain her observations, Hardy (2009)
discusses how the kinds of activities to which the students were exposed led them to
develop such practices, composed of a limited practical block and non-mathematical
technologies. The activities in which students participated did not only relate to sets
of highly routinized tasks, they also required no form of mathematical justification.
Engaging in such activities, students observed patterns that led to the construction of
techniques based on arbitrary lists of steps that just seemed to work; at least enough
to do well on assignments and exams.
In a similar sense, we could imagine a student in our hypothetical example justifying
their solution to the activity “integrate over ” by
saying something like: “first, find the antiderivative, then find the difference between
the value at 1 and the value at -1, because that’s how we always do it!” An activity

490 [Link]:indrum2018:174227
such as “integrate over ”, might therefore elicit the following
erroneous response:

Unless of course the isolated activities, a1 and a3, were eventually, substantially, and
meaningfully incorporated into developing the above non-mathematical practice into
a practice more mathematical in nature.
“Undergraduate mathematics coursework”
As illustrated in the previous section, an anthropological perspective does not
interpret students’ non-mathematical practices (or knowledge) as reflecting a
common misconception inspired by difficulties inherent to a given mathematical
concept. Rather, it sees such practices as resulting from a concrete situation within
which the student finds themselves, under the influence of institutions (Douglas,
1986). In the ATD, the word “institution” is taken in a wide sense. For example,
mathematicians work within an overarching institution that we could call
Mathematical Research (MR), where their praxeologies are shaped by various shared
criteria (concerning consistency, beauty, explanatory power, efficiency, etc.), but
survive only if they follow the strict rules of mathematical reasoning. The students of
interest to us, in contrast, are subjects of the institution Undergraduate Mathematics
Coursework (UMC), which was in large part created to train potential participants of
MR. This said, various conditions and constraints within UMC can require and
enable a network of praxeologies that is fundamentally different from that built and
recognized by MR. The non-mathematical praxeologies described in the previous
section provide some examples.
To capture the transposition of knowledge as it moves from MR into UMC,
Chevallard, and others (e.g., Bosch, Chevallard, & Gascón, 2005), have introduced a
distinction between different types of knowledge (i.e. practice):
 Scholarly Knowledge, produced and used by mathematicians;
 Knowledge to be Taught, as determined by curricula, textbooks, and
professors’ teaching plans;
 Knowledge Actually Taught, according to professors’ actual interactions with
students, e.g., in lectures;
 Knowledge to be Learned, i.e., the knowledge students are expected to
develop, which is often a transposed subset of the knowledge to be taught and
actually taught, with the minimal core indicated by assessment tools;
 Knowledge Actually Learned, which can only be predicted, through analyses
of student work, in-class observations of students, or other specially-designed
interactions with students, such as interviews or problem-solving situations.

491 [Link]:indrum2018:174227
Although a lot can happen in university lectures, the minimal knowledge students are
obliged to learn to pass their courses is determined by their assignments and exams.
It is not surprising that the knowledge actually learned by students is often only a
transposed subset of this minimal core. Hence, if we want to know what kind of
knowledge students are or could possibly be developing in UMC, then we cannot
restrict our exploration to curricula, textbooks, and teachers’ lecturing practices: we
need to pay careful attention to the way in which students are assessed.
Students’ learned knowledge in UMC may also be characterized as a progression
through various sub-institutions: from secondary school to early university courses
(e.g., in single and multivariable calculus), through to more advanced university
courses (e.g., in real analysis, metric spaces, measure theory, and functional
analysis), which may eventually lead to graduate studies and beyond. Programs can
vary from school to school and from country to country. However, a common
phenomenon in secondary schools seems to be that assessments focus solely on the
practical block of mathematical knowledge. The teacher may be expected to know
the theoretical block for explaining the material to students; but the students are
typically not obliged or even invited to develop an awareness of the technology or
theory, let alone how it is linked to the practical block (Barbé et al., 2005; Winsløw,
Barquero, De Vleeschouwer, & Hardy, 2014). One observed result is that many
students interpret mathematical knowledge (practice) as equivalent to identifying a
type of task and applying the corresponding technique (Bergqvist, Lithner, &
Sumpter, 2008). Several studies confirm that this same kind of situation can arise in
early university coursework (e.g., Lithner, 2003; Hardy, 2009; Brandes, 2017).
Indeed, over multiple years of coursework, students not only gain a particular view
of what mathematical knowledge is, but they also develop knowledge that, when
judged against the scholarly knowledge produced and used by mathematicians, is
evidently non-mathematical – from the strategies they develop to identify tasks, to
the discourses they use to justify these strategies and the techniques they choose.
Nevertheless, as conjectured in the schema shown in Figure 1, a transition is
expected to occur at some point: students are eventually required to develop
knowledge that is completely and coherently mathematical. These circumstances lead
Winsløw et al. (2014) to wonder about how teachers could help students accomplish
such transitions. In parallel, we are inspired to validate, specify, and extend these
researchers’ claims by constructing praxeological models of how the different kinds
of knowledge produced in the didactic transposition process progress throughout an
entire undergraduate degree. In other words, we are inspired to investigate more
closely the nature of the mathematical training being received by future
mathematicians in the progression of their undergraduate coursework.
Of course, developing praxeological models to represent the knowledge (to be)
taught and (to be) learned throughout an entire undergraduate degree is a hefty task.
Within the context of our PhD project, we propose to accomplish a first stage, based

492 [Link]:indrum2018:174227
on a subset of courses in one coursework path. Like in the Mathematical Research
institution, Undergraduate Mathematics Coursework is divided into several sub-
institutions according to domain – e.g., algebra, geometry and topology, analysis,
statistics, mathematical physics, or probability – each of which contain a grouping of
courses, which can be placed in some chronological order according to their
prerequisites. Having already carried out research in the early courses of an “analysis
path”, this is the context that seemed most appropriate for our work.

METHODOLOGY
Although our project aims at modelling different levels of knowledge that can be
identified in the didactic transposition process, in this paper we discuss only the
modelling of the knowledge to be learned (KTL).
Our research is conducted at a large, urban, Canadian university. The mandatory
courses in the analysis path of an Honours Bachelor of Science in Mathematics
include multivariable calculus (MVC) I and II, and mathematical analysis (MA) I, II,
and III. Since previous work (Brandes, 2017) suggests that the KTL in MVC I and II
is similar in nature to the KTL in calculus and pre-calculus courses, we decided to
start by focussing our attention on the two courses that come next and are likely
candidates for housing the transitions of interest to us: MA I and II.
As mentioned above, the KTL represents the knowledge that students are expected to
develop, which can be gleaned from the various activities in which they are invited
to engage (lectures, assignments, and exams), as well as the materials that frame and
support the activities (course outlines and textbooks). Since the minimal core of the
KTL is represented in the assessment activities students must complete on their own,
we have decided to ignore what happens in lectures and focus on the activities that
comprise assignments and (practice) exams.
MA I and II are institutions in themselves in that they enjoy some sort of stability.
For various reasons, course outlines and assigned textbooks tend to remain the same
from year to year. The courses also maintain the same assessment structure: students
complete assignments on a regular basis during the term, a midterm exam halfway
through, and a final exam, with most of their mark (90% or more) concentrated in the
examinations. This said, the actual activities proposed on assignments, midterms, and
finals have less stability in that they can reflect personal choices of the professors
assigned to teach the course in a given term. On top of this, our approach to
modelling the knowledge actually learned will involve task-based interviews with
students after they have passed MA I or II. Hence, we have collected the assignments
and (practice) exams proposed only by the professor(s) who would be teaching those
students. For instance, from the two MA I professors teaching in Fall 2017, we
collected eleven weekly assignments, seven practice midterms, six practice final
exams, and the actual examinations they gave to their students (these professors
worked together in that they gave the same set of activities to their students).

493 [Link]:indrum2018:174227
To analyse such a collection of activities, we think about whether each activity is
“isolated” or part of a “path to a practice”. An “isolated” activity may occur only
once in the sense that no other activities engage students in accomplishing the same
type of task. Since such activities are unlikely to contribute directly to the
development of a practice, we reflect on why they are proposed. In comparison, the
activities that belong to a “path to a practice” typically combine with other activities
to expose students to a type of task. Our goal in studying these activities is to extract
a theoretical model of the praxeologies that the ideal student (i.e., the student that
receives a good passing grade) is expected to develop in the course. We start by
constructing punctual praxeologies related to groups of non-isolated activities.
Looking at the problem statements, we can establish the types of tasks (T) that
generate the praxeologies. Determining the technologico-theoretical blocks ([τ, θ,
Θ]), however, requires more data. We rely on the solutions a professor makes
available to students to uncover the intended techniques, as well as portions of the
expected theoretical blocks; and we complete the latter by checking the course
outline and reading the relevant textbook chapters. The resulting collection of
punctual praxeologies then becomes part of our data, which we use to construct more
generalized praxeologies, think about how they are related to one another, and reflect
on the nature of the KTL.
Eventually, we plan to put the models of the KTL for MA I and II together and
compare our results with what previous researchers have found in calculus and pre-
calculus courses. This, we hope, will allow us to discuss how the ideal student is
expected to progress in the early stages of the analysis path. At the time of the
INDRUM 2018 conference, we will have completed this initial theoretical stage of
our project and will thus be able to share our results.

SUMMARY AND EXPECTATIONS

For a long time, mathematics students survive their courses based on developing a
transformed version of a practical block, where they learn to recognize routine tasks
and apply techniques to solve them in a sort of mechanical, naturalized, or
normalized way, void of a mathematical theoretical block. At some point throughout
an undergraduate degree in mathematics, however, the conditions for students’
survivability change dramatically and possibly abruptly: they are faced with
activities that require them not only to fill the void of a mathematical theoretical
block, but also to develop techniques for accomplishing tasks (e.g., proofs) that
involve the abstract theoretical objects that have come out of hiding. Through
modelling the knowledge the ideal student is expected to develop, as well as,
eventually, the knowledge students actually develop in early analysis courses, we
expect our project to bring about a more detailed and concrete understanding of
praxeological “transitions” that have been theorized to occur, and give us some
insight into how (or if) students adapt to them.

494 [Link]:indrum2018:174227
Returning to the vague and general question that originally inspired our project, we
ultimately hope to learn more about what students are actually learning throughout
an undergraduate degree in mathematics. The empirical data we collect will be a
contribution to largely anecdotal discussions about when (if at all) students’
knowledge is invited to become, and actually becomes, coherently, completely, and
complexly mathematical, just like the scholarly knowledge produced and used in the
institution of Mathematical Research. The significant difference between elementary
and advanced courses, as professors gain more freedom and teach topics more
closely related to their field of study, leads us to predict that students are eventually
required to develop mathematical practices. After all, in spite of the apparent
disconnection that is often observed between university mathematics courses and
mathematical research (cf. Broley, Caron, & Saint-Aubin, 2017), the field of
mathematics continues to live on, with new mathematicians emerging from the
coursework that made up their mandatory professional education. In any case,
through studying the principal conditions that currently shape the activities in which
undergraduate mathematics students engage, we feel that we will be in a better
position to discuss realistic and meaningful ways of encouraging these students to
develop practices that are truly “mathematical”, within the confines of educational
institutions. This, we hope, will serve as complementary to the recent surge of
studies (cf. Barquero, Serrano, & Ruiz-Munzon, 2016) aiming to explore innovative
teaching approaches that question not the nature of the knowledge developed, but the
dynamics of the knowledge development.

REFERENCES
Barquero, B., Serrano, L., Ruiz-Munzon, N. (2016). A bridge between inquiry and
transmission: The study and research paths at university level. First Conference of
International Network for Didactic Research in University Mathematics, March
2016, Montpellier, France.
Barbé, J., Bosch, M., Espinoza, L., & Gascón, J. (2005). Didactic restrictions on the
teacher’s practice: The case of limits of functions in Spanish high schools.
Educational Studies in Mathematics, 59, 235–268.
Bergé, A. (2016). Le rôle de la borne supérieure (ou supremum) dans l’apprentissage
du système des nombres réels. First Conference of International Network for
Didactic Research in University Mathematics, March 2016, Montpellier, France.
Bergqvist, T., Lithner, J., & Sumpter, L. (2008). Upper secondary students’ task
reasoning. International Journal of Mathematical Education in Science and
Technology, 39(1), 1–12.
Bosch, M., Chevallard, Y., & Gascón, J. (2005). ‘Science or Magic?’ The use of
models and theories in didactics of mathematics. 4th Congress of the European
Society for Research in Mathematics Education, February 2005, Sant Feliu de
Guixols, Spain.

495 [Link]:indrum2018:174227
Brandes, H. (2017). The bare necessities for doing undergraduate mutlivariable
calculus (Master’s thesis). Concordia University, Montréal.
Broley, L., Caron, F., & Saint-Aubin, Y. (2017). Levels of programming in
mathematical research and university mathematics education. International
Journal of Research in Undergraduate Mathematics Education.
[Link]
Chevallard, Y. (1999). L’analyse des pratiques enseignantes en théorie
anthropologique du didactique. Recherches en Didactique des Mathématiques, 19,
221–266.
Douglas, M. (1986). How institutions think. Syracuse, NY: Syracuse University
Press.
Hardy, N. (2009). Students’ perceptions of institutional practices: The case of limits
of functions in college level Calculus courses. Educational Studies in
Mathematics, 72, 341-358.
Lithner, J. (2003). Students’ mathematical reasoning in university textbook
exercises. Educational Studies in Mathematics, 52, 29–55.
Lithner, J. (2004). Mathematical reasoning in calculus textbook exercises. Journal of
Mathematical Behavior, 23, 389–404.
Martínez-Planell, R., Trigueros Gaisman, M., & Mcgee, D. (2016). Student
understanding of the relation between tangent plane and the differential. First
Conference of International Network for Didactic Research in University
Mathematics, March 2016, Montpellier, France.
O’Shea, A. (2016). Concept images of open sets in metric spaces. First Conference
of International Network for Didactic Research in University Mathematics, March
2016, Montpellier, France.
Stewart, J. (2008). Calculus: Early transcendentals (6th ed.). Belmont, CA:
Thomson Learning, Inc.
Winsløw, C. (2008). Transformer la théorie en tâches : la transition du concret à
l’abstrait en analyse réelle. In A. Rouchier & I. Bloch (Eds.), Actes de la XIIIème
école d’été en didactiue des mathématiques (cd-rom). Grenoble: La Pensée
Sauvage.
Winsløw, C., Barquero, B., De Vleeschouwer, M., & Hardy, N. (2014). An
institutional approach to university mathematics education: from dual vector
spaces to questioning the world. Research in Mathematics Education, 16(2), 95–
111.

496 [Link]:indrum2018:174227
Bridging probability and calculus: the case of continuous distributions
and integrals at the secondary-tertiary transition
Charlotte Derouet1, Gaetan Planchon², Thomas Hausberger², and Reinhard
Hochmuth3
1
LISEC team AP2E, University of Strasbourg, France,
[Link]@[Link]; ²Institut Montpelliérain Alexander Grothendieck,
CNRS, University of Montpellier, France; 3Leibniz University of Hannover,
Germany

This paper focuses on two mathematical topics, namely continuous probability

distributions (CPD) and integral calculus (IC). These two sectors that are linked by
𝑏
the formula 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) = ∫𝑎 𝑓(𝑥)𝑑𝑥 are quite compartmented in teaching classes
in France. The main objective is to study whether French students can mobilize the
sector of IC to solve tasks in CPD and vice versa at the transition from high school to
higher education. Applying the theoretical framework of the Anthropological Theory
of the Didactic (ATD), we describe a reference epistemological model (REM) and use
it to elaborate a questionnaire in order to test the capacity of students to bridge CPD
and IC at the onset of university. The analysis of the data essentially confirms the
compartmentalisation of CPD and IC.
Keywords: Transition to and across university mathematics; Teaching and learning of
analysis and calculus; Teaching and learning of probability; Anthropological Theory
of the Didactic.
INTRODUCTION
Continuous probability distributions (CPD) and integral calculus (IC) are two topics
that are taught in France during the last year of high school (grade 12 of the scientific
track). They constitute two sectors (in the sense of the Anthropological Theory of the
Didactics, ATD) that belong to the two different but closely related mathematical
domains of probability theory and calculus respectively. Indeed, the continuous
probability of an event and the definite integral with respect to a non-negative function
are both defined as areas of suitable two-dimensional domains in the syllabus (in
𝑏
France), and the formula 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) = ∫𝑎 𝑓(𝑥)𝑑𝑥 is the key for solving several
standard tasks in CPD, where 𝑋 represents a random variable and 𝑓 its associated
density function.
IC is the focus of extended studies in mathematics education: for instance, research
(Schneider, 1992; Tran Luong, Bessot & Dorier, 2010; Haddad, 2013) was conducted
in the context of Belgian, French, Vietnamese and Tunisian secondary education,
including the secondary-tertiary transition (Haddad, 2013). By contrast, there is hardly

497 [Link]:indrum2018:172495
any literature on CPD and, to make things worse, the available studies mostly put the
emphasis on normal distributions (Batanero, Tauber & Meyer, 1999; Wilensky, 1997;
Batanero, Tauber & Sánchez, 2004; Pfannkuch & Reading, 2006). Therefore, the
teaching-learning phenomena generated by the interrelationship between CPD and IC
are still to be investigated.
A first stone was laid by Derouet and Parzysz (2016; see also Derouet, 2016), who
studied possible ways to introduce the density function at grade 12 so that students may
construct this concept starting from considerations regarding histograms and therefore
might relate continuous probability to the integral. By an analysis of textbooks,
Derouet could show that the two sectors CPD and IC are very much compartmentalised
in the French curriculum (Derouet, 2016, pp. 127-190). For instance, the above formula
is seldom justified by a thorough discussion of the definitions involving areas, which
certainly hinders the bridging of the two sectors by students.
In this paper, we regard this “compartmentalisation” of knowledge as an institutional
phenomenon and therefore use ATD as the theoretical framework (see below). Our
goal is to study the impact of this compartmentalisation on the learning of mathematics:
are French students able to mobilize the sector of IC to solve tasks in the sector of CPD
and vice versa at the transition from high school to higher education?
After the presentation of theoretical constructs from ATD used in this research, we will
describe the reference epistemological model (REM) that we elaborated for the types
of tasks in CPD and IC with regard to studying interrelations of the two sectors. We
will then describe our methodology that builds on the elaboration of a questionnaire,
based on the REM, that has been submitted to students at the entrance of university.
We finally present results of a primary analysis of the data from the questionnaire and
draw some conclusions and perspectives opened up through this study.
THEORETICAL CONSTRUCTS
ATD “postulates that any activity related to the production, diffusion or acquisition of
knowledge should be interpreted as an ordinary human activity, and thus proposes a
general model of human activity built on the key notion of praxeology” (Bosch &
Gascon, 2014). The praxeology 𝛱 is represented by a quadruple [𝑇/𝜏 /𝜃/𝛩]: its praxis
part (or know-how) consists of a type of tasks 𝑇 together with a corresponding
technique 𝜏 (useful to carry out the tasks 𝑡 ∈ 𝑇 in the scope of 𝜏). The logos part (or
know-why) includes two levels of description and justification: the technology 𝜃, i.e.
a discourse on the technique, and the theory 𝛩, which often unifies several
technologies.
The elaboration of a reference epistemological model (Florensa, Bosch, & Gascon,
2015) as sequences of praxeologies, for a given body of knowledge, is an important
step in any research carried out in the ATD framework. It is the tool that will be used
by the researcher to describe, analyse, put in question or design the specific contents

498 [Link]:indrum2018:172495
that are at the core of a teaching and learning process. In order to build such a model,
“mathematical praxeologies are described using data from the different institutions
participating in the didactic transposition process, thus including historical, semiotic
and sociological research, assuming the institutionalized and socially articulated nature
of praxeologies” (loc. cit. p. 2637).
Our study relies on an overview of standard textbooks used at grade 12 in France, as
well as the official syllabus, in order to identify the standard praxeologies in CPD and
IC that may be related and the nature of this relationship at the praxeological level. An
epistemological investigation of the historical development and the interrelation of
both domains have previously been carried out in (Derouet, 2016, pp. 67-85). In order
to test the effect of the institutional compartmentalisation of knowledge on the learning
of mathematics, we need to check the availability of the identified praxeologies in the
praxeological equipment of students, and then submit tasks to students which need to
bridge CPD and IT as mutually interdependent sectors that share techniques or
technologies, borrow them from or lend them to each other. Special care must be taken
in the phrasing of these bridging tasks, taking into account the effect of ostensives
(Bosch & Chevallard, 1999), that is to say the role of signs. Indeed, ostensives
contribute to the activation of the specific sectors to which they belong and therefore
direct students toward specific techniques.
REFERENCE EPISTEMOLOGICAL MODEL
Even though problems of quadratures arose in ancient Greece, IC finds its roots as a
systematic method in the 17th century. The emergence of continuous probability may
be situated in the 18th century with the theory of errors in physical measurements.
Various functions were introduced to model the distribution of errors and the area
under the curve permitted to evaluate the “theoretical frequency” (so the probability)
of the deviation from the “true” value. CPD was thus naturally connected to IC in its
historical roots. The gaussian distribution was proposed later by Gauss in 1809.
To identify the different praxeologies, we have analysed 12 textbooks of the grade 12
of the scientific track (edition 2012). We focused on the exercises with a given solution
in the textbooks to have access to the usual techniques for the different tasks.
In our study, we will focus on two main types of mathematical tasks 𝑇𝐼 and 𝑇𝑃 , which
are related to the mathematical domains of integral calculus and the continuous
probability respectively:
𝑏
● 𝑇𝐼 : compute a value for an integral ∫𝑎 𝑓(𝑥)𝑑𝑥 for a positive continuous function
𝑓;
● 𝑇𝑃 : determine the probability 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) for a random variable 𝑋 endowed
with a density function 𝑓.
The type of tasks 𝑇𝐼 may further be split into two subtypes of tasks, depending on the
expected result: an exact value (𝑇𝐼,𝑒𝑥𝑎𝑐𝑡 ) or an approximation (𝑇𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 ).

499 [Link]:indrum2018:172495
The most useful technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 to solve 𝑇𝐼,𝑒𝑥𝑎𝑐𝑡 is to compute a primitive of the
function and apply the fundamental theorem of calculus. The corresponding
technology 𝜃𝐼,𝑒𝑥𝑎𝑐𝑡 is given by the fundamental theorem of calculus that relates
𝑏
integrals and primitives: ∫𝑎 𝑓(𝑥)𝑑𝑥 = 𝐹(𝑏) − 𝐹(𝑎) with 𝐹′ = 𝑓. The theory 𝛩𝐼,𝑒𝑥𝑎𝑐𝑡
includes the definition of the definite integral for a continuous positive function as an
area and properties of areas that may be formalised into a local axiomatic theory1.
The technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 thus resorts to praxeologies dedicated to the computation of
primitives. The standard technique at high school level is to use the “tabular of
primitives” (deduced from the tabular of derivatives). The technology comprises the
properties of the derivative and the theory is that of differential calculus. In the case of
piecewise affine functions, an alternative technique to 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 is to interpret the integral
as the area of an elementary surface (or a union of these).
The type of tasks 𝑇𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 may be solved using two main techniques: using a
calculator (or software), more or less a blackbox, or applying the “rectangle method”.
The latter technique 𝜏𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 consists in considering the integral as an area, taking a
subdivision of the interval of integration and computing the sum of rectangular areas.
The technology 𝜃𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 comprises the definition of the integral and properties of
areas. A further theoretical level 𝛩𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 is mainly non-existent at high school level
(cf. endnote 1).
Regarding the type of tasks 𝑇𝑃 , two cases need to be distinguished, depending on
whether a primitive of the density function f is known (𝑇𝑃,𝑝𝑟𝑖𝑚 ) or not. The latter case
is reduced to that of the normal distribution (𝑇𝑃,𝑛𝑜𝑟𝑚 ), which is dealt with using the
implementation of such a distribution in a calculator or software. Computer scientific
tools are mainly used as a blackbox by students, which hinders the possibility for
students to make connections with IC.
𝑏
The generic technique 𝜏𝑃,𝑝𝑟𝑖𝑚 for 𝑇𝑃,𝑝𝑟𝑖𝑚 is to compute ∫𝑎 𝑓(𝑥)𝑑𝑥, in other words to
resort to the praxeology 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 . The technology 𝜃𝑃,𝑝𝑟𝑖𝑚 is given by the formula
𝑏
𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) = ∫𝑎 𝑓(𝑥)𝑑𝑥, and the theory 𝛩𝑃,𝑝𝑟𝑖𝑚 comprises the definition of a
continuous probability (as the area of the corresponding domain) and the definition of
a probability density function. At high school, two particular cases are emphasised and
lead to local techniques, as concrete formulas are available for 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) in the
case of the exponential and uniform distributions. For instance, the technique 𝜏𝑃,𝑒𝑥𝑝 in
the case of the exponential may be reduced to computing 𝑒 𝜆𝑎 − 𝑒 𝜆𝑏 with the
technological argument 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) = 𝑒 𝜆𝑎 − 𝑒 𝜆𝑏 .
Let us recall that our model is based on the study of standard textbooks used in grade
12 classes in France and is dedicated to the description of the teaching-learning of CPD
and IC as it actually is (we are not planning task-design at this stage of the research).

500 [Link]:indrum2018:172495
In this model, we note the following links between CPD and IC: at the level of the
theoretical blocks, praxeologies in both sectors are anchored on the empirical notion of
area. At the level of the technique, 𝛱𝑃,𝑝𝑟𝑖𝑚 uses 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 , so that, from an ecological
point of view (Bosch & Gascon, 2014, p. 72), CPD contributes to the thriving of such
IC praxeologies. By contrast, 𝛱𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 does not seem to be reinvested in CPD
(whereas the computation of a probability of the Gaussian distribution could be an
opportunity to mobilize 𝛱𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 ). Conversely, we did not detect elements of the
praxis of CPD in the IC sector. This isn’t a surprise: IC is regarded as a prerequisite to
CPD and precedes the teaching of CPD in all textbooks. Nevertheless, the normal
distribution is a prototypical example of a function whose primitive cannot be
expressed in terms of available elementary functions. This fact explains the choice of
techniques in 𝛱𝑃,𝑛𝑜𝑟𝑚 and contributes also to the logos of 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 (by complementing
the statement that every continuous function admits a primitive). What about the type
2 2
𝑏 𝑒 −(𝑥−𝑚) /2𝜎
of tasks 𝑇𝐼,𝑛𝑜𝑟𝑚 : compute ∫𝑎 𝑑𝑥 ? It could appear in the IC sector through
𝜎 √2𝜋
𝛱𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 only but we didn’t find it in any textbook and it is never stated as such in the
CPD sector. The most efficient technique requires to use the formula 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) =
𝑏
∫𝑎 𝑓(𝑥)𝑑𝑥 from right to left: although the equality is symmetrical as an equivalence
relation, it isn’t symmetrical as a sign which denotes a succession of operations in
performing a computation. How would students react to such a task that asks to bridge
CPD and IC in an unusual way? This question came to us as a starting point for the
elaboration of our questionnaire dedicated to the investigation of the educational
effects of the institutional phenomenon of compartmentalisation of knowledge, in the
case of CPD and IC.
THE QUESTIONNAIRE
Our main goal is to test whether students are able or not to connect CPD and IC, and
especially mobilize the CPD sector to solve an IC task when ostensives do not indicate
explicitly the probability domain. To do so, we have elaborated a questionnaire both to
check the availability of standard praxeologies of CPD and IC in the praxeological
equipment of students and the capacity of students to complete such bridging tasks.
Bridging tasks appear at the very end of the questionnaire and are stated as follows:

Question 6: Expliquer toutes les méthodes que vous pouvez utiliser pour déterminer une valeur exacte et/ou approchée
2
1 𝑒 −𝑥 /2
de l’intégrale suivante : 𝐼 = ∫−0,5 𝑑𝑥. On pourra se limiter à donner une idée de la méthode si sa mise en oeuvre
√2𝜋
est trop compliquée.
𝜆
Question 7: Soit A la fonction définie par 𝐴(𝜆) = ∫0 𝑓(𝑥)𝑑𝑥 avec 𝑓(𝑥) = 𝑥𝑒 −𝑥 , pour tout  ∈ [0; +∞[. On peut
𝜆
démontrer que lim ∫0 𝑥𝑒 −𝑥 𝑑𝑥 = 1. D’après ce résultat, expliciter tout ce que vous pouvez dire sur la fonction A et
𝜆→+∞
la fonction f.
Translation:

501 [Link]:indrum2018:172495
 Question 6: Explain all the methods that can be used to determine an exact and/or approximate value of the following
2
1 𝑒 −𝑥 /2
integral: 𝐼 = ∫−0,5 𝑑𝑥 . It suffices to give an idea of the method if its implementation is too complicated.
√2𝜋
𝜆
Question 7: Let A be the function defined by 𝐴(𝜆) = ∫0 𝑓(𝑥)𝑑𝑥 with 𝑓(𝑥) = 𝑥𝑒 −𝑥 , for all  ∈ [0; +∞[. It can be
𝜆
proved that lim ∫0 𝑥𝑒 −𝑥 𝑑𝑥 = 1. According to this result, note everything that you can say about the function A and
𝜆→+∞
the function 𝑓.

Figure 1: bridging tasks submitted to students

Question 6 contains an instance of the type of tasks 𝑇𝐼,𝑛𝑜𝑟𝑚 discussed in the REM. The
task is stated in an opened way, asking for every method that students may know to
compute an exact or approximate value for the Gaussian integral. Praxeologies 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡
and 𝛱𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 should therefore also show up. In question 7, we intend to check whether
students can say that A is both a primitive for f and a probability associated with the
density function f, or restrict to the IC sector with an interpretation in terms of areas.
Previous questions intend to “activate” both sectors CPD and IC equally. In this
respect, question 1 offers a routine task of type 𝑇𝐼,𝑒𝑥𝑎𝑐𝑡 in the case of a straightforward
exponential function. Analogously, the first part of question 4 is a routine task of type
𝑇𝑃,𝑒𝑥𝑝 (compute 𝑃(1 ≤ 𝑋 ≤ 5) when X has an exponential distribution of parameter
3). In its second part, students are asked for a graphical interpretation of the probability
𝑃(1 ≤ 𝑋 ≤ 5) that is in fact defined as an area in the high-school syllabus, as well as
the integral: this interpretation is therefore essential in order to link the logos of 𝛱𝑃 and
that of 𝛱𝐼 .
Question 2 activates the CPD sector by soliciting an element of the logos of 𝛱𝑃 , namely
the properties that define a density function. This logos is crucial in the bridging
question 7, which is stated in the IC sector without any reference to CPD. Question 3
tests if students are able to retrieve the definitions of both the exponential and normal
distributions by specifically asking for those in the case of simple parameters (the
reduced centered gaussian law). The latter is an element of the logos of 𝛱𝑃,𝑛𝑜𝑟𝑚 : we
wish to check if students are able to recognize the normal distribution in the statement
of the bridging question 6, while taking care not to direct them towards a specific
technique (hence the order of questions).
In question 5, we rather activate the IC sector, more precisely elements of the logos of
𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 (primitives), but we have in mind the praxeology 𝛱𝑃,𝑛𝑜𝑟𝑚 in relation to the
bridging question 6: students are asked to provide an example of a continuous function,
if it exists (or justify the impossibility), that a) doesn’t possess primitives b) admits
primitives but expressions for these are not “explicitly known” (expected answer: the
density function of the normal law).
Summarizing, by the questions 1 to 4 we want to investigate whether the students
master techniques 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 in a calculus context and 𝜏𝑃,𝑝𝑟𝑖𝑚 in a probability context and

502 [Link]:indrum2018:172495
whether they know technologies related to IC and CPD. Then, by questions 6 and 7,
we focus on relationships between CPD and IC and the previous questions: we analyse
links between the questions 3, 4 (second part), 5 and 6, on the one hand, and links
between the questions 2 and 7, on the other hand.
DATA ANALYSIS AND RESULTS
The questionnaire was used at the beginning of September 2017 (the first week of
classes) in two classes of first year CPGE (French engineers school preparatory classes)
students, which is in fact at the transition between secondary and tertiary levels. The
first class (called class N) is a class of MPSI (Mathematics, Physics and Engineering
Science) and the second class (called class R) is a class of PCSI (Physics, Chemistry
and Engineering Science) of a rather prestigious establishment. The students working
on the questionnaire are in selective classes, so we can assume that they are “good”
scientific students, and in particular, if they meet difficulties then these are shared by
the other students. We only analysed answers from students who studied in French high
school during the past year because we constructed the questionnaire taking into
account the context of the French high school institution. We retrieved 82
questionnaires (40 of the class N and 42 of the class R). Except for a few students (less
than 5), the students didn’t use a calculator during the test.
From the 82 students, only one does not mobilize the technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 to resolve the
routine task concerning IC (question 1). 85% find a correct expression of the primitive
and 78% obtain the correct result, which means that this technique is mastered quite
well by students. Regarding the computation of the probability for an exponential
distribution (question 4), 63% of the students get a correct result. 82% of the students
use the technique 𝜏𝑃,𝑝𝑟𝑖𝑚 and the other students directly apply a formula. More than
70% of the students could identify the probability as an integral. Summarizing, except
for some errors regarding the primitive or the computation, the majority of students are
able to pass from a probability to an integral and, moreover, know the fundamental
theorem of calculus. In praxeological terms, they are able to mobilize the technique
𝜏𝑃,𝑝𝑟𝑖𝑚 that implies the use of the technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 in the case of an exponential
distribution.
Regarding CPD and neglecting formulation and formalization issues, only 39% of the
students know the definition of the density function (question 2) and only 27% recollect
the density function of the normal distribution (question 3b). In view of question 5,
only 32% mobilize the theorem claiming that all continuous functions on an interval
admit a primitive. 52 % give an example of a function for which they don’t explicitly
know a primitive (although it might exist and be expressed in terms of standard
functions2, for instance 𝑙𝑛(𝑥)). Among these, 15% mention the density function of the
Gaussian distribution (or a function of the type 𝑒 −𝑥² ). Of the 22 students who know the
density function of the normal distribution, half propose it as an example in question 5
(13% of all the students).

503 [Link]:indrum2018:172495
9% of the students do not answer question 6. The method most often proposed is
technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 (46%). Less than 20% (16 students) mention the normal distribution
and 16% propose the rectangle method (𝜏𝐼,𝑎𝑝𝑝𝑟𝑜𝑥 ). A few students propose the
technique “integration by parts”, which is beyond the curriculum in grade 12.
Moreover, only 23% of the students who propose the rectangle method are able to
illustrate the method by drawing the graph of the Gaussian curve (the other students
draw a wrong curve or do not consider any graph). Only 11 of the 16 students (69%)
who mention the normal distribution in this question write that the integral I is equal
to the probability 𝑃(−0,5 ≤ 𝑋 ≤ 1) with 𝑋 a random variable of a reduced normal
distribution and 7 of them state more precisely that they have to use the calculator to
evaluate this probability (𝜏𝑃,𝑛𝑜𝑟𝑚 ). So, our results indicate that the ostensive
2
1 𝑒 −𝑥 /2
∫−0,5 √2𝜋 𝑑𝑥 without indication invite students to stay in the IC sector and even more
particularly in the praxeology 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 (even if it is not possible here), which reflects
that 𝛱𝐼,𝑒𝑥𝑎𝑐𝑡 is the praxeology most developed in grade 12. To pass from an integral to
a probability and to change the sector in this direction does not seem to be natural for
students. Moreover, of the 22 students who know the expression for the density
function of the reduced normal distribution (question 3b), only 10 (45%) recognize the
density function of the normal distribution is this context. Of the 15 students who say
that the Gaussian function does not admit primitives expressed by using standard
elementary functions (question 5b), around 27% proposes to use the technique 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡
nevertheless and only 53% recognize the normal distribution in question 6. This means
that most of the students were not able to mobilize 𝛱𝑃,𝑛𝑜𝑟𝑚 in the IC-context of
question 6, that is 𝛱𝐼,𝑛𝑜𝑟𝑚 . In particular the application of 𝜏𝐼,𝑒𝑥𝑎𝑐𝑡 , although question
5b is answered correctly, demonstrates the strong compartmentalisation between CPD
and IC.
Regarding the answers to question 7, we notice that 38% of the students (31) state that
the function 𝑓 is a density function. 42% of them (13 students) could justify that it is a
density function including 4 who forget to mention the positivity of the function
(because they don’t write this condition in their definition in question 2). Moreover, 12
of the 31 students as well as one additional one identify 𝐴(𝜆) as 𝑃(0 ≤ 𝑋 ≤ 𝜆). 8
students think that the distribution in question 7 is an exponential distribution of
parameter 𝑥 and 7 students state that 𝐴 is a density function. Finally, more than 46%
of the students mention “probability”, which means that they manage to identify at
least some link between the IC embedding of question 7 and CPD. Probably, the
questionnaire itself influenced students and the percentage would be lower otherwise,
i.e. if question 7 was asked independently. Overall correct results with justification are
rare and of the 32 students who master the definition of a probability density function
(one element of the technology 𝜃𝑃,𝑝𝑟𝑖𝑚 tested in question 2), 31% (10 students) do not
identify 𝑓 as a density function. 16% of the students do not at all answer question 7.

504 [Link]:indrum2018:172495
Again, we observe also with respect to question 7 that CPD praxeologies, although
they are in principle available, often cannot be mobilized in the IC contextualization.
Summarizing, the data analysis shows that techniques related to one sector are
available for the majority of students only when ostensives related to this sector are
provided. Additionally, related technologies are much less mastered by students.
Perhaps, this could be an explanation why it is not natural for students to mobilize
praxeologies of the CPD sector for a task in the IC sector, in addition to the fact that
these tasks are not taught in the classroom. The data analysis by all means shows a
strong compartmentalisation between CPD and IC.
CONCLUSION AND PERSPECTIVES
The results of our primary data analysis clearly demonstrate a strong
compartmentalisation between CPD and IC. In particular, techniques from CPD,
although available in a CPD task, could not be mobilized in an IC-contextualized task.
A next step in our research will be more detailed data analyses looking for correlations
and interdependencies between techniques and technologies of CPD and IC. We further
observed that available techniques were not accompanied by related technologies. One
could claim that more elaborated technologies might support the transfer of techniques.
More generally, we think about studies investigating the impact of changes in the
institutional setting, i.e. establishing innovative teaching sequences with less
compartmentalisation. A teaching sequence articulating CPD and IC is proposed in
Derouet (2016). The effect of this teaching on the answers of students to the
questionnaire could be analysed by comparing the latter with the present results.
NOTES
1. This axiomatic remains implicit at the secondary level; it may be related to Measure Theory at university level.
2. We realised a posteriori that our question was not phrased properly: “a function that admits a primitive whose
expression is not explicitly known to you” may be interpreted as a lack of techniques to actually compute the primitive
and not the impossibility to provide an expression (in terms of elementary functions).

REFERENCES
Batanero, C., Tauber, L., & Meyer, R. (1999). From data analysis to inference: a
research project on the teaching of normal distributions. In Bulletin of the
International Statistical Institute: Proceedings of the Fifty-Second Session of the
International Statistical Institute (Tome LVIII, Book 1) (pp. 57–58). Helsinki.
Batanero, C., Tauber, L. M., & Sánchez, V. (2004). Students’ reasoning about the
normal distribution. In B.-Z. D. & J. Garfield (Eds.), The Challenge of Developing
Statistical Literacy, Reasoning and Thinking (pp. 257–276). Dordrecht: Springer.
Bosch, M. & Chevallard, Y. (1999). La sensibilité de l’activité mathématique aux
ostensifs. Recherche en Didactique des Mathématiques, 19(1), 77-124.

505 [Link]:indrum2018:172495
Bosch, M., & Gascon, J. (2014). Introduction to the Anthropological Theory of the
Didactic. In A. Bikner-Ahsbahs & S. Prediger (Eds.), Networking Theories as a
Research Practice in Mathematics Education (p. 67-83). Springer.
Derouet, C., & Parzysz, B. (2016). How can histograms be useful for introducing
continuous probability distributions? ZDM - Mathematics Education, 48(6), 757–
773.
Derouet, C. (2016). La fonction de densité au carrefour entre probabilités et analyse
en terminale S. Etude de la conception et de la mise en œuvre de tâches
d’introduction articulant lois à densité et calcul intégral. Thèse de doctorat.
Université Paris Sorbonne Cité, Université Paris Diderot.
Florensa, I., Bosch, M., & Gascon, J. (2015). The epistemological dimension in
didactics: Two problematic issues. In K. Krainer & N. Vondrova (Eds.),
Proceedings of the ninth congress of the european society for research in
mathematics education (p. 2635-2641). Prague: Faculty of Education, Charles
University.

Haddad, S. (2013). Que retiennent les nouveaux bacheliers de la notion d’intégrale

enseignée au lycée ? Petit x, 92, 7–32.

Pfannkuch, M., & Reading, C. (2006). Reasoning about distribution: A complex

process. Statistics Education Research Journal, 5(2), 4–9.

Schneider, M. (1992). Des objets mentaux “aire” et “volume” aux calculs de primitives.
In L’enseignement de l’analyse au lycée. Pourquoi et comment ? Actes de la
journée de Nice du 18 mai 1991 (pp. 1–32). Nice: IREM de Nice.

Tran Luong, C. K., Bessot, A., & Dorier, J.-L. (2010). dx et ∫ ostensifs pour questionner
l’enseignement des intégrales. Montpellier : IUFM de Montplellier, 2010. In G. C.
et C. L. A. Bronner, M. Larguier, M. Artaud, M. Bosch, Y. Chevallard (Ed.),
Diffuser les mathématiques (et les autres savoirs) comme outils de connaissance
et d’action. (Apports de la théorie anthropologique du didactique) (pp. 501–526).
Montpellier: IUFM de Montpellier.

Wilensky, U. (1997). What is normal anyway? Therapy for epistemological anxiety.

Educational Studies in Mathematics, 33(2), 171–202.

506 [Link]:indrum2018:172495
 Affective variables in the transition from school to university
mathematics
Sebastian Geisler, Katrin Rolka
Ruhr-University Bochum, Faculty of Mathematics, Germany,
[Link]@[Link]
The transition from school to university mathematics is a difficult step for students,
which many of them do not succeed to manage immediately. In this contribution we
use questionnaires, which measure mathematics-related affective variables as well as
subject-unspecific affective variables and students’ achievement during the semester
to predict the outcome of the exam at the end of the first semester (as a first indicator
for success in their studies) and the students’ attendance in this exam (as an indicator
for early dropout). We are interested in whether the mathematics-related or the
“general” affective variables are more suitable to predict the students’ exam
attendance and the exam outcome. The students’ achievements during the semester
turned out to be the best predictor for the exam outcome, whereas the students’
attendance was best predicted by their interest in mathematics.
Keywords: transition to and across university mathematics, teaching and learning of
analysis and calculus, dropout, study success, affect
INTRODUCTION
Dropout is a big problem for German universities, especially in mathematics. Nearly
80% of all mathematics students drop out or change their subjects – most of them
during their first year at university (Dieter & Törner, 2012). However, this is not a
typical German phenomenon: Chen (2013) reports similar figures of dropout and
subject change for the United States. In this paper, we do not distinguish between
students who drop out and leave the university system without examination and those
who “just” leave mathematics and change to another subject.
At Ruhr-University Bochum, where our study takes place, students begin their studies
of mathematics with two lecture courses in the first semester – calculus I and
linear algebra I. Two so called “mini-tests” are written during the semester to prepare
the students for their first “real” exam at university. These “mini-tests”, designed by
the lecturer, cover conceptual and procedural knowledge about definitions and proof
which have been discussed in the lecture before. 73% of the students at Ruhr-
University (who attended the exams) failed their calculus exam in 2017. Furthermore,
25% of the students did not even attend the exam. Following the goal to improve the
support of students who are at risk to fail their exam and/or drop out, we are
interested in which way students at risk differ from those who succeed.
Due to data protection regulations, it was not possible for us to identify which
students really dropped out from mathematics at Ruhr-University in 2017. Instead,
we could match the results from our questionnaires with the results of the exam at the
end of the first semester. We therefore could identify which students were successful

507 [Link]:indrum2018:171801
in their exam, which failed and which did not attend the exam. Students who did not
attend the exams might have dropped out before or may be at risk to drop out soon.
Baars and Arnold (2014) found that students who do not attend their exams in the
first semester have a high risk to drop out.
THEORETICAL BACKGROUND
Both, dropout and study success, are considered to be influenced by multiple factors,
which are often called predictors, such as the socio-economic and school background,
personal psychological prerequisites, learning behaviour and study conditions (Tinto,
1975; Heublein et al., 2009; Thiel et al., 2008).
The predictors, which are listed by the students for their decision to drop out, are
called dropout reasons. In Germany, most dropped out mathematics students name
the course requirements (e.g. failed exams, work-overload) (33%) and low motivation
(25%) as their main reasons to quit their studies of mathematics. Other reasons such
as study-conditions (13%) and reasons related to health- or financial problems (12%)
and personal reasons like family problems (10%), were less important – specially for
early dropout (Heublein et al., 2009).
Given the fact, that most dropped out mathematics students in Germany name the
requirements at university and their lack of motivation as crucial for their decision to
drop out, we want to shed light on the following affective variables, which are
considered to influence the students’ motivation and academic achievements:
mathematical self-concept, interest in mathematics, beliefs concerning the nature of
mathematics, basic needs and general self-efficacy. These variables are briefly
discussed in the following.
Mathematical self-concept
The self-concept can be seen as the mental model of one’s personal competences,
abilities and properties, or “in very broad terms, self-concept is a person’s perception
of himself” (Shavelson, Hubner, and Stanton, 1976 as cited in Bong and Skaalvik,
2002). The self-concept is influenced by the students’ former experiences and
achievements and can itself influence students’ motivation (Bong and Skaalvik,
2002). The self-concept is considered to be domain specific. Rach and Heinze (2016)
found that the mathematical self-concept is a significant predictor for dropout but not
for students’ success in the first semester.
Interest in mathematics
The interest in mathematics is considered to have a positive impact on the learning of
mathematics. Schiefele et al. (1993) define interest as a specific relation between a
person and an object. The interest in the subject that one is studying is rather stable,
since it has been developed over a longer time through different experiences.
Due to contradictory results in various studies the impact of interest in mathematics
on students’ performance and success in their studies of mathematics remains
uncertain. Rach and Heinze (2016) found no significant influence of the interest in

508 [Link]:indrum2018:171801
mathematics on students’ success during the first semester or on their risk to drop out.
However, Blömeke (2009) found significant correlations between the interest in
mathematics and the students’ intention to drop out.
Beliefs concerning the nature of mathematics
It has been widely discussed that the nature of mathematics changes with the
transition from school to university (e.g. Rach and Heinze, 2016). Mathematics in
German schools is often focused on applying mathematical techniques to solve real
world problems (modelling, problem solving). New mathematical contents are
regularly presented more intuitively with examples and illustrations and yield on an
intuitive or practical understanding of the concepts. Mathematics at university is
more theoretically and proof oriented. New concepts are presented in a rather formal
and abstract way and therefore less illustrated than in school. The focus often lies on
encouraging logical and abstract thinking. The students have to develop
understanding for deductive argumentations and proof – applying the theory is less
important than at school. This change from a practical to a theoretical approach is not
easy for most students. Many of them feel a big gap between mathematics at school
and university (Geisler, 2017). This feeling might be a result of unfulfilled
expectations and incongruences between the mathematical “reality” at university and
their established beliefs concerning the nature of mathematics, which are based on
their school experiences. Daskalogianni and Simpson (2001) call this phenomenon
“belief overhang”. Andrà, Magnano and Morselli (2011) found hints that students’
beliefs concerning the nature of mathematics can influence their decision to drop out
or to stay. Traditionally we distinguish between a static view, where mathematics is
viewed as a summary of (unconnected) rules, facts and techniques, and a dynamic
view, where mathematics is considered as a process and a creative field of research
(Grigutsch and Törner, 1998). However, it is yet unclear which beliefs are beneficial
for a successful transition from school to university.
Basic Needs
Following the framework of self-determination theory (Ryan & Deci, 2000), there are
three basic psychological needs that are important for the well-being of humans and
to generate motivation: social relatedness, competence and autonomy. In the special
situation of the transition from school to university mathematics, many students do
not experience autonomy and competence (Liebendörfer and Hochmuth, 2013). This
is problematic since Faye and Sharp (2008) found that especially the feeling of
competence is strongly associated with motivation in university. In an explorative
case-study, we found hints for the impact of social relatedness and competence on the
decision to drop out (Geisler, 2017).
General self-efficacy
The general self-efficacy is the strength of a persons’ belief to be able to reach certain
goals and to solve problems by his or her own competences and abilities
(Luszczynska et al., 2005). This general belief is not limited to a special domain like

509 [Link]:indrum2018:171801
mathematics or special academic settings. In contrast to the self-concept, self-efficacy
is more focused on the consequences of one’s own competences and abilities than on
the competences and abilities themselves. That’s why self-concept is rather past
oriented whereas the self-efficacy focuses on the future (Bong and Skaalvik, 2002).
Besides, self-concept is considered to be the more stable variable. Self-efficacy can
influence the students’ motivation in the sense that students who believe that they are
able to succeed in their studies of mathematics are more motivated to put effort in
their learning than those who believe that they have no chance in the exams. Self-
efficacy is therefore associated with academic achievement (Luszczynska et al.,
2005; Bong and Skaalvik, 2002). Students with lower self-efficacy have a higher risk
to drop out than those with higher self-efficacy (Krieger, 2011).
Achievement during the semester
The students’ achievement is an important factor for success and dropout. In Tinto’s
(1975) framework, achievement, as a part of the academic integration, is important
for the decision to drop out or to stay. In an explorative case-study, Geisler (2017)
found hints that students who are not satisfied with their achievement during the first
semester sometimes drop out, even if they are successful in their exams at the end of
the semester. Though achievement is closely connected with the perceived feeling of
competence.
RESEARCH QUESTIONS
In order to support students who are at risk to fail their exam or even to drop out we
want to know in which way these students differ from those who succeed. Following
the theoretical background described above, we decided to focus on the students’
achievements during the first semester and on affective variables which are likely to
influence students’ motivation. Since the dropout rate in mathematics is high
compared to other subjects, it seems plausible that mathematics related variables have
an important impact on dropout and success. We therefore distinguish between
mathematics-related affective variables (mathematical self-concept, interest in
mathematics, beliefs concerning the nature of mathematics) and more “general”
affective variables (basic needs, self-efficacy). We are interested in whether the
mathematics-related or the “general” affective variables are more suitable to predict
students’ exam attendance and their exam outcome. This leads to the following
research questions:
Differences between the three groups of students
1) Which differences in the affective variables and the achievements can be found
between students who do not attend the exam, students who fail in the exam
and those who succeed?
Prediction of the students’ exam attendance
2.1) In which way can the mathematics-related affective variables predict the
students’ attendance for the exam at the end of the first semester?

510 [Link]:indrum2018:171801
2.2) In which way can the “general” affective variables and the students’
achievements approve this prediction?
Prediction of the exam outcome
3.1) In which way can the mathematics-related affective variables predict the
outcome of the exam at the end of the first semester?
3.2) In which way can the “general” affective variables and the students’
achievements approve this prediction?
METHODOLOGY
209 students in the calculus lecture in wintersemester 2016/17 voluntarily
participated in our study. Undergraduate mathematics students as well as pre-service
teachers in mathematics usually attend this lecture during their first year at university.
The questionnaires were filled out during the lecture in the mid of the first semester,
taking into account that students cannot rate their satisfaction of the basic needs at the
begin of the semester. Due to incomplete datasets, only N=193 cases could be
included in our analysis. The instruments used in our questionnaire can be found in
Table 1.
construct source No. of items / Item-example
Crobach’s α
Interest Schiefele et al. 12 / 0.82 “It is personally important for me
2007 that I can study mathematics.”
Self-Concept Kauper et al. 4 / 0.82 “I am very good in mathematics.”
2012
Beliefs: static Laschke & 6 / 0.67 „Mathematics means learning,
Blömeke 2013 remembering and applying.“
Beliefs: dynamic Laschke & 6 / 0.73 “Mathematics involves creativity
Blömeke 2013 and new ideas.”
Social Relatedness Kauper et al. 6 / 0.78 “I feel comfortable with the other
2012 students.”
Competence Kauper et al. 3 / 0.66 “I get clear and detailed feedback
2012 on my achievements.”
Autonomy Kauper et al. 3 / 0.58 “I can do tasks in my way.”
2012
Self-Efficacy Beierlein et al. 3 / 0.88 “I can solve most problems on my
2012 own”
Table 1: Instruments with numbers of items, reliability and item-example
All reliabilities (Cronbachs α) were at least sufficient – except for the reliability of
the autonomy-subscale. All items were answered on a five-point Likert scale
(1=totally disagree; 5=totally agree). To measure the students’ achievement during
the semester, we used their results in the first “mini-test” (1 to 12 points).

511 [Link]:indrum2018:171801
The results of the questionnaires were analysed using a MANOVA (to prevent the
accumulation of the α-error compared with t-tests) to answer research question 1. To
answer the other research questions, we used linear and binary logistic regressions.
RESULTS
Students who do not attend the exam at the end of the first semester differ
significantly from those who fail the exam and those who succeed in almost all
affective variables (research question 1), except for the static beliefs (Table 2).
Focussing on the mathematics-related affective variables, the biggest difference
between the three groups of students can be found in the interest in mathematics,
which can explain 13% of the variance (η2=0.13***). Regarding the “general”
affective variables, the self-efficacy turned out to explain the most variance between
the three groups of students (η2=0.1***). Taking into account all measured variables,
the biggest difference between the three groups of students can be found in their
achievements in the “mini-test” (η2=0.19***).
No Attendance Failed Succeeded
N = 54 N = 101 N = 38
M SD M SD M SD F η2
Interest 2.99 0.66 3.43 0.7 3.7 0.56 13.89 0.13***
Self-Concept 2.64 0.83 2.93 0.67 3.11 0.58 5.45 0.05**
Beliefs: static 3.77 0.66 3.79 0.58 3.68 0.5 0.46
Beliefs: dynamic 3.07 0.68 3.46 0.65 3.56 0.62 8.22 0.08***
Self-Efficacy 2.46 0.96 2.78 0.78 3.25 0.69 10.18 0.1***
Social Related. 3.77 0.68 4.07 0.55 4.23 0.57 7.69 0.08**
Competence 2.94 0.87 3.23 0.86 3.43 0.81 3.91 0.04*
Autonomy 2.89 0.72 3.28 0.79 3.42 0.66 6.83 0.07**
Achievement 6.19 3.9 7.81 3.16 10.89 2.75 22.74 0.19***
Table 2: Means, standard deviations and results of the variance analysis
*p<0.05 **p<0.01 ***p<0.001
To answer research questions 2.1 and 2.2, three different logistic regression models
were tested (Table 3). Model 1 only contains the mathematics-related affective
variables. The only significant predictor for the students’ attendance in this model
was interest in mathematics. Model 1 can explain 19% of the variance and is able to
classify 71% of the students correctly as attending or not attending. Model 2
additionally contains the affective variables basic needs and self-efficacy. None of
these variables has a significant influence on the students’ attendance and they do not
improve the students’ classification. In contrast, the students’ achievements are a
(weak) significant predictor for the students’ attendance (Model 3). The inclusion of
the students’ achievement can improve the classification of the students (75.1%

512 [Link]:indrum2018:171801
correct) and increases the explained variance (Nagelkerke’s R2=0.26). Note that the
interest in mathematics is still the most significant predictor in Model 3.
Model 1 Model 2 Model 3
** **
Interest 2.43 2.5 2.42**
Self-Concept 1.22 0.89 0.8
Beliefs: static 1.14 1.3 1.29
Beliefs: dynamic 1.66 1.3 1.24
Self-Efficacy 1.33 1.18
Social Relatedness 1.3 1.14
Competence 0.93 1.04
Autonomy 1.4 1.39
Achievement 1.13*
Nagelkerke’s R2 0.19 0.22 0.26
Correct classification 71 % 71 % 75.1 %
Table 3: Results (coefficients Exp(B)) of the logistic regression to predict the exam
attendance - *p<0.05 **p<0.01
We used three linear regression models to answer research questions 3.1 and 3.2
(Table 4). In Model 1 only the mathematics-related affective variables were included.
Model 1 Model 2 Model 3
Interest 0.08 0.06 0.01
Self-Concept 0.17 -0.03 0.04
Beliefs: static -0.09 -0.01 0.02
Beliefs: dynamic 0.05 -0.02 -0.05
Self-Efficacy 0.41*** 0.2*
Social Relatedness 0.11 0.07
Competence -0.08 0.02
Autonomy -0.03 -0.07
Achievement 0.54***
R2 0.05 0.13 0.37
Table 4: Results (standardized beta coefficients) of the linear regression to predict the
exam outcome - *p<0.05 ***p<0.001
None of them can predict the exam outcome significantly. All these variables
together only explain 5% of the variance in the exam outcome. The “general”
affective variables, specially the self-efficacy, seem to have a bigger impact on the
exam outcome (Model 2). The inclusion of these variables improves the variance that
can be explained (R2=0.13). The self-efficacy is a highly significant predictor. The

513 [Link]:indrum2018:171801
most important (highly significant) predictor for the exam outcome is the students’
achievement in the “mini-test” (Model 3). The inclusion of the achievement increases
the explained variance to 37%. Multicollinearity of the variables in our models is at
least tolerable (tolerance>0.48, VIF<2) and should not have a big impact on results.
CONCLUSION
The interest in mathematics is the most important predictor for exam attendance in
our study, whereas the mathematical self-concept has no significant influence. This is
contradictory to Rach’s and Heinze’s (2016) findings, where only the mathematical
self-concept was able to predict the attendance. However, both studies have in
common that a mathematics-related affective variable is the most important predictor
for the attendance, whereas the “general” affective variables have no influence.
The mathematics-related affective variables do not predict the exam outcome. The
only affective variable that significantly predicts the exam outcome is the general
self-efficacy. This is rather surprising, taking into account that self-efficacy and self-
concept are (at least theoretically) closely connected variables and it seemed plausible
that the mathematics-related variable provides more insights. However, in contrast to
the mathematical self-concept, the general self-efficacy is not only focussed on one’s
competences and achievements in mathematics but also takes into account subject
unspecific competences which could be beneficial at university, too.
All in all, it turned out that the students’ achievements can predict both, exam
outcome and exam attendance. Interest in mathematics is suitable to predict the exam
attendance, whereas self-efficacy can predict the exam outcome. Firstly, this finding
shows that, mathematics related as well as general affective variables play an
important role in the transition from school to university. Secondly, it suggests that
success and dropout should not necessarily be viewed as two sides of a coin.
Our study has some limitations. We conducted data from only one university, which
could lead to cohort specialities. Furthermore, questionnaires were filled out in the
mid of the semester during the lecture. Students who do not (regularly) attend
lectures or have dropped out before have not been captured by our study. Some
results might be different if we could capture those students, too.
Our on-going research will now focus on a more detailed characterisation of dropped
out students, taking into account cognitive and metacognitive variables (e.g. students
learning behaviour) as well. In addition, it seems to be useful to identify different
types of dropped out students in mathematics. This might help to design and evaluate
more individualized supporting programs for students in the transition.
REFERENCES
Andrà, C., Magnano, G. & Morselli, F. (2011). Drop-out undergraduate students in
mathematics: an exploratory study. In Roesken, B. & Casper, M. (Eds.), Current
State of Research on Mathematical Beliefs XVII. Proceedings of the MAVI-17
Conference September 17-20, 2011, Bochum, Germany 13–22.

514 [Link]:indrum2018:171801
Baars, G. J. A., & Arnold, I. J. M. (2014). Early Identification and Characterization
of Students Who Drop Out in the First Year at University. Journal of College
Student Retention: Research, Theory & Practice, 16(1), 95-109.
Beierlein, C., Kovaleva, A., Kemper, C. J., & Rammstedt, B. (2012). Ein Messinstru-
ment zur Erfassung subjektiver Kompetenzerwartungen (An Instrument to
mearsure subjective Competenceexpectations). GESIS – Leibniz-Institut für
Sozialwissenschaften. Retrieved from [Link]
Manual_2012.PDF.
Blömeke, S. (2009). Ausbildungs- und Berufserfolg im Lehramtsstudium im
Vergleich zum Diplomstudium – Zur prognostischen Validität kognitiver und
psycho-motivationaler Auswahlkriterien (Success in pre service Teachers
Studyprograms in contrast to Diplom Studyprograms – Regarding the prognostic
validity of cognictive and psycho-cognitive criteria). Zeitschrift Für
Erziehungswissenschaft, 12(1), 82–110.
Bong, M., & Skaalvik, E. M. (2002). Academic Self-Concept and Self-Efficacy: How
Different Are They Really? Educational Psychology Review, 15(1), 1–40.
Chen, X. (2013). STEM Attrition: College Students’ Paths Into and Out of STEM
Fields. Washington, DC: National Center for Education Statistics, Institute of
Education Sciences, U.S. Department of Education. Retrieved from
[Link]
Daskalogianni, K., & Simpson, A. (2001). Beliefs Overhang: The Transition from
School to University. In Winter, J. (Ed.), Proceedings of the British Society for
Research into Learning Mathematics Vol. 2, 97-108.
Dieter, M. & Törner, G. (2012). Vier von fünf geben auf (Four out of five are giving
up). Forschung und Lehre, 19(10) 826–827.
Faye, C. & Sharpe, D. (2008). Academic Motivation in University: The Role of Basic
Psychological Needs and Identity Formation. Canadian Journal of Behavioural
Science, 40(4), 189–199.
Geisler, S. (2017). Dropout & Persistence in University Mathematics. In Kaur, B.,
Ho, W.K., Toh, T.L., & Choy, B.H. (Eds.). Proceedings of the 41st Conference of
the International Group for the Psychology of Mathematics Education, Vol. 1, p.
197. Singapore: PME.
Grigutsch, S. & Törner, G. (1998). World views of mathematics held by university
teachers of mathematics science. Retrieved from [Link]
[Link]/servlets/DerivateServlet/Derivate-5249/[Link].
Heublein, U., Hutzsch, C., Schreiber, J., Sommer, D. & Besuch, G. (2009). Ursachen
des Studienabbruchs in Bachelor- und in herkömmlichen Studiengängen (Reasons
for Dropout in Bachelor- and conventional Studyprogramms). HIS: Projektbericht,
Dezember 2009. Retrieved from [Link]

515 [Link]:indrum2018:171801
Kauper, T., Retelsdorf, J., Bauer, J., Rösler, L., Möller, J. & Prenzel, M. (2012).
PaLea - Panel zum Lehramtsstudium: Skalendokumentation und
Häufigkeitsauszählungen des BMBF-Projektes (PaLea – Panel of pre service
Teacher Studyprograms: Documentation of the Instruments). Retrieved from
[Link]
content/uploads/2012/04/PaLea%20Skalendokumentation%201_%[Link].
Krieger, A. (2011). Determinanten des Studienabbruchs in naturwissenschaftlich
orientierten Studiengängen (Determinants of Dropout in Naturesciences). Johannes
Gutenberg-Universität Mainz.
Laschke, C. & Blömecke, S. (2013). Teacher Education and Development Study:
Learning to Teach Mathematics (TEDS–M). Dokumentation der
Erhebungsinstrumente. Münster: Waxmann.
Liebendörfer, M., & Hochmuth, R. (2013). Interest in Mathematics and the first Steps
at University. In Ubuz, B., Haser, C., & Mariotti, M. A. (Eds.) (2013). Proceedings
of the Eighth Congress of the European Society for Research in Mathematics
Education (CERME 8, February 6 - 10, 2013). Ankara, Turkey: Middle East
Technical University and ERME.
Luszczynska, A., Gutié Rrez-Doñ, B., & Schwarzer, R. (2005). General self-efficacy
in various domains of human functioning: Evidence from five countries.
International Journal of Psychology, 40(2), 80–89.
Rach, S., & Heinze, A. (2016). The Transition from School to University in
Mathematics: Which Influence Do School-Related Variables Have? International
Journal of Science and Mathematics Education. 15(7), 1343–1363.
Ryan, R. M. & Deci, E. L. (2000). Self-Determination Theory and the facilitation of
Intrinsic Motivation, Social Development, and Well-Being. American
Psychologist, 55(1) 68–78.
Schiefele, U., Krapp, A., Wild, K.-P. & Winteler, A. (1993). Der Fragebogen zum
Studieninteresse (FSI) (The Studyinterest-Questionnaire). Diagnostica, 39, 335–
351.
Shavelson, R. J., Hubner, J. J., and Stanton, G. C. (1976). Self-concept: Validation of
construct interpretations. Review of Educational Research 46, 407–441.
Thiel, F., Veit, S., Blüthmann, I., Lepa, S. & Ficzko, M. (2008). Ergebnisse der
Befragung der Studierenden in den Bachelorstudiengängen an der Freien
Universität Berlin - Sommersemester 2008 (Results of the Survey among Students
in Bachelorprograms at Freie Universität Berlin). Retrieved from [Link]
[Link]/einrichtungen/arbeitsbereiche/lehr_studienqualitaet/zentrale-
evaluation/bachelorbefragung/[Link].
Tinto, V. (1975). Dropout from Higher Education: A Theoretical Synthesis of Recent
Research. Review of Educational Research, 5(1), 8–9.

516 [Link]:indrum2018:171801
 Discussing Mathematical Learning and Mathematical Praxeologies
from a Subject Scientific Perspective
Reinhard Hochmuth
Leibniz University Hanover, Germany, hochmuth@[Link]
This programmatic contribution discusses the link between concepts from
Anthropological Theory of Didactics (ATD) and the “subject-scientific point of view”
according to Holzkamp (1985, 1993). The main common concern of ATD and the
subject-scientific approach is to conceptualize and analyse “objects” like
“institutionalized mathematical knowledge” and “university” not as conditions that
cause reactions but essentially as meanings in the sense of generalized societal
reified action possibilities. The link of both approaches is illustrated by the issue of
“real numbers” in the transition from school to university: Hypotheses are derived
for further actual-empirical research, which intrinsically incorporate content- and
subject related perspectives as well as societal and school-related findings.
Keywords: Curricular and institutional issues concerning the teaching of
mathematics at university level, transition to and across university mathematics,
subject scientific approach, mathematical praxeologies, real numbers.
INTRODUCTION
This paper contributes to an ongoing major research project that describes and ana-
lyses form and content of advanced mathematics and its teaching and learning from a
subject scientific point of view. This approach is grounded in “Critical Psychology”,
framed by Holzkamp (1985) (see Tolman (1991) and Schraube & Højholt (2015) for
English written introductions). Recently this theory becomes internationally more
known in the mathematics education community due to Roth & Radford (2011), who
assessed “German Critical Psychology” as a further development of the culture-
historical activity approaches by Leontjev (1978) and Vygotsky (1978). It’s beyond
the scope of this paper to describe and analyse in which respects “German Critical
Psychology” differs and goes beyond culture-historical activity theory. Instead the
paper intends to point out its compatibility and (partial) complementarity with ATD
as well as to illustrate its potential relevance for further research concerning
university mathematics education.
Main features of “Critical Psychology” and its subject-scientific point of view are
well elaborated psychological categories (roughly: basic notions, see for details
(Holzkamp, 1985, pp. 28)) for describing and analysing cognitive and emotional-
motivational dimensions of subject [individual] related experiences, in particular
thoughts, actions and learning, in a way that major societal aspects are inherently be
incorporated. It aims (besides others) to provide individuals with analytic tools for
their self-reflection of problematic experiences and situations to reveal their inherent
dependencies and circumstances, thus allowing individuals to achieve a more
reflective learning. Within this framework, there is so far a lack of research that

517 [Link]:indrum2018:174703
relates to mathematical learning in general and in university in particular. “Critical
Psychology” provides points of contact for incorporating research results concerning
the societal and historical genesis of knowledge and reference structures as well as
institutionally framed (e.g. school, university, study courses) external and internal
transposition processes (Chevallard, 1991).
ATD’s praxeological analyses could principally inform any psychological or socio-
logical theory considering teaching and learning. Already in Castela (2015) a link
between ATD and cultural historical activity theory is discussed, in particular
between Roth’s concept of crossing boundaries between different socio-cultural
contexts and the issue of inter-institutional transitions. This basic idea is in the
following taken up in a broader sense.
In view of the ongoing major research project, this contribution has the status of an
intermediate step presenting programmatic ideas about linking ATD with the subject
scientific approach. This combination might in particular be fruitful for research
connecting detailed analyses of mathematical practices with a complex vision of
learners and teachers in a way that (with respect to both sides) their intrinsic societal
mediatedness is systematically incorporated. Though this rather pretentious goal can
easily be formulated (at least by using abstract notions, which would require a lot of
pages to be embedded in a coherent theory and to be explained in detail (Holzkamp,
1985, 1993)), the following lines also indicate that there is still a way to go
combining both approaches in actual-empirical research (Holzkamp, 1985, pp. 509).
The structure of the paper is organized as follows: In the first two sections we intro-
duce some notions from ATD and the subject scientific approach. Then we discuss
the link of both approaches. Finally we illustrate the link and some of its aspects and
opportunities considering the issue of “real numbers” in the transition from school to
university: After an ATD-orientated overview about the nowadays typical treatment
of real numbers in German secondary schools and considering various options of
extending this discourse in the transition from school to university , we discuss
subject scientific related aspects taking into account societal and school-related
findings and how they might contribute to validate mathematical and didactical
practices. This theoretical analysis exemplarily demonstrates how the link
intrinsically integrates content- and subject-related perspectives and leads to
hypotheses for actual-empirical research projects.
SOME NOTIONS FROM ATD
ATD (Chevallard, 1992; Winslow, Barquero, Vleeschouwer & Hardy 2014) aims at a
precise description of knowledge and its epistemic constitution. Its concepts make
possible to explicate institutional specificities of knowledge and related practices. An
underlying conviction of this approach is that cognitive-oriented accesses tend to
misinterpret contextual or institutional aspects of practices as personal dispositions. A
basic concept of ATD are praxeologies represented in so called “4T-models
(T,τ,θ,Θ)” consisting of a practical and a theoretical or logos block. The practical

518 [Link]:indrum2018:174703
block (know-how, “doing math“) includes the type of task (T) and the relevant
solving techniques (τ). The logos block (knowledge block, discourse necessary for
interpreting and justifying the practical block, “spoken surround“) covers the
technology (θ) explaining and justifying the used technique and the theory (Θ)
justifying the underlying technology. Praxeologies give descriptions of mathematics
by reference models that are activity oriented (techniques, technologies). The
interconnectedness of knowledge is modelled in ATD by means of local and regional
mathematical organizations that allow contrasting and integrating practical and
epistemological aspects in relation to different institutional contexts. Therefore ATD
is in particular helpful for analysing institutional realizations of mathematical
knowledge within different learning contexts, e.g. the use of mathematics in signal
theory (Hochmuth & Schreiber, 2015).
More than 15 years ago Chevallard has introduced the additional notion of “scale of
levels of codeterminations” that in the meanwhile has become rather important in
ATD analyses (Bosch & Gascón, 2006). The hierarchical sequence of levels covers
civilisation, society, school, pedagogy, discipline, domain, sector, theme and subject
(in the sense of topic). Each level provides some kind of framework, within among
others actions on lower levels are possible, supported or hindered and praxeologies
are in a certain sense embedded. In Barbé et al. (2005) is shown, for example, how
general didactic restrictions for teaching mathematical topics in school can affect
teachers’ practices and their established praxeologies, in particular the shaping of the
practical and the logos block and the relations between them. To mention one further
example, Job & Schneider (2014) argue that smoothing the transition-gap from
calculus to analysis shows at least the tendency to blur the distinction between the
different discourses in school and university, which tends to reinforce an empirical
positivist attitude by students as an epistemological obstacle to learning (ditto; p.
641). Generalizing their arguments, one might say that there are issues relating to
general world views (society) that affect institutionally settled praxeologies.
Moreover, Chevallard (1991) introduced the notion indicating the relation of
a position x (a typical position of an individual) within an institution I to a praxeology
o. The “scale of levels of codeterminations” underlines that the institution and, with
the institution, the position x and the praxeology o have to be considered as being
societal situated, i.e., that in their analyses specific emphasize also has to be put to
societal assignments that are related to societal mediation processes. The subject
scientific perspective, which is introduced and discussed in the next both sections,
allows to further specify positions x keeping the significance of societal and
institutional mediatedness (in a materialistic sense, see for example (Arndt, 2013)).
SOME NOTIONS CONCERNING THE SUBJECT SCIENTIFIC APPROACH
(“CRITICAL PSYCHOLOGY”)
“Critical Psychology” claims to present a scientific discussable and criticisable elabo-
ration of basic psychological concepts (categories). The starting point is a historical-

519 [Link]:indrum2018:174703
empirical investigation of general historical-specific characteristics of relations
between societal and individual reproduction as well as its dialectic mediatedness.
One of the central subject related categories is “action potence”, which is the potence
to ensure the disposal about the subject’s individual living conditions together with
others (Holzkamp, 1985, pp. 239).
Within the context of this paper there are three important points to notice: First, the
actual historical-specific form of subjectivity is characterized by the “possibility rela-
tion” regarding the societal reality, which includes in particular the basic experience
of intentionality. Second and connected to the first, the specific modality of
subjective experiences comprises the discourse form “reasoning discourse”: “I” speak
about my “own” actions in terms of subjective reasonable (not necessarily “rational”)
activities and of premises in the light of “my” life interests. A third crucial point is
that the “human’s relationship to the environment is almost always mediated. […]
Categories of psychology like learning, emotion, motivation and cognition cannot fail
to be significantly altered by the fact of our existence’s social mediatedness. The
most important mediation category is meaning.” (Tolman, 1991, pp. 14-15)
These three interrelated issues are combined in the assertion that conditions are given
to “me” in terms of meanings in the sense of generalized societal action possibilities
and that reality aspects, which are relevant for “me”, denoting again the generalized
subject standpoint, become premises for “me” in the light of “my” life interests.
Therefore, subject scientific considerations are essentially given by meaning-
premises-reasons-relations, which a priori situate experience and activities of the
(individual) subject “within the world” Accordingly, Holzkamp (1985, pp. 342)
figured out the level of subjective action reasoning as the main subject specific level:
It represents the level with respect to which individual experiences and activities (e.g.
learning) has to be reconstructed and analysed.
Via the specific notion of meaning, human activities, like teaching and learning, are
intrinsically thought as societal mediated. This implies that any analysis of subjective
actions requires the reconstruction of subjectively relevant conditions in the sense of
generalized action possibilities and the consideration of their societal mediatedness.
Since meanings appear (via objective-subjective premises) as the medium within
which subjects’ reasoning discourses are grounded, their study is a prerequisite for
describing and analysing related cognitive, motivational and emotional processes as
aspects of subjects’ activities. But, although meanings in the indicated sense are
rather relevant for acting and thinking, they do not determine them. Instead, they
represent action possibilities that might become premises in the light of subjectively
perceived “life interests”.
THE LINK BETWEEN ATD AND THE SUBJECT SCIENTIFIC APPROACH
The position x within an institution can be (re-)interpreted as the “position” and/or
“situation of life” from the subject point of view, which includes intentionality, the
modus of reasoning discourse and societal mediatedness. As an “element” (a

520 [Link]:indrum2018:174703
“position x”) of an institution a subject is typically not confronted with the whole
world but only with a local “situated” section represented by those meanings
typically produced and reproduced within the institutional context. Hence, insti-
tutional contexts provide specific frameworks for premise-reasoning-patterns. In this
sense ATD’s praxeological analyses contribute to concretisations of meaning-
premise-reasoning-patterns that are typical within an institution at the position x. In
particular the concept of praxeologies allows capturing substantial aspects of
mathematical practices in such a way, that they can be injected as facets of action
related meanings, i.e., they can be (re-)interpreted as generalized societal action
possibilities, which were potentially reflected in subject related reasoning schemes as
premises and/or reasons. In this sense praxeological analyses can be seen as one non-
trivial first step within subject scientific research projects: They might inform
microanalyses of task solution processes by exploring institutional established
practices. They are relevant for describing and analysing related activities, since they
appear as institutionalized medium, within subjective action reasoning grounds. With
respect to premises-reasoning-patterns the technological dimension of praxeologies,
i.e. the justification and validation of techniques, is of specific importance. But, see
above, praxeologies do not determine subjects’ activities, since there is an
unconscious-conscious step by subjects of selecting, neglecting or highlighting facets
of praxeologies in view of their evaluation of “life interests” and how they are
perceived by them at “position x” in the “institution I” in view of all prospects
addressed by “the scale of level of codeterminations”. Thus, the latter is rather
relevant for both, the analysis of meanings (essentially by ATD) and the analyses of
premises-reasons-relations (essentially by the subject scientific approach). In fact,
both strands can’t be seen as totally separated but as dialectically interrelated, since
institutionalized practices live through subjects’ [individual] activities.
THE ISSUE OF “REAL NUMBERS”
In this section we give first a short overview of the nowadays typical treatment of real
numbers in German secondary schools in grade eight or nine. Because of the space
limitation a detailed praxeological analysis can’t be presented.
The Treatment of Real Numbers in German Schools
The treatment of real numbers in German schools presumes that rational numbers are
known and can be represented by ratios, decimal fractions and points on the number
line. Moreover it is presumed that students are able to switch between those
representations. In particular basic calculations should be understood and can be
executed with respect to the different representations. The typical starting point for
the introduction of real numbers is the observation or proof (sometimes!) that there
are quadratic equations like without rational solutions. Next it is observed
(but typically not proven) that one can find approximations by proper decimal
fractions that fulfil those equations up to an arbitrary chosen error. On the other hand
it is (geometrically) argued that there is a magnitude x, the length of the square

521 [Link]:indrum2018:174703
diagonal, satisfying that corresponds to a certain point on the number line.
The intuitive conviction about the existence of those points on the number line
supports the idea, that (somehow converging) infinite sequences of approximating
proper decimal fractions give a (unique!) final finite result, a number, that can be
represented by a non-terminating decimal fraction. This type of discourse justifying
the existence of infinite-finite objects (i.e., infinite processes giving in a certain sense
a finite result) has in particular been considered in Lakoff & Nūñes (2002) as basic
for the whole analysis and denoted as “basic metaphor of infinity”. The new objects
of non-terminating and non-periodic decimal fractions are called irrational numbers
and build, together with the already known rational numbers, the set of real numbers.
Moreover, the calculation rules that are known for rational numbers are assumed to
be also true for all real numbers. Whereas in former years one can at least find
Descartes’ geometrical arguments for explaining multiplication and division for
general real numbers such arguments are nowadays missing. Corresponding to the
sketched treatment of real numbers there are nearly no tasks that are related to
structural aspects of real numbers or that enforce to reflect arguments of the discourse
concerning limits or the existence of points or numbers respectively. Instead the tasks
focus on various isolated techniques that are locally established, for example
approximation techniques like interval bisections and the Heron algorithm or the use
of calculation rules.
In terms of the 4T-model the established mathematical praxeologies can be
characterized as essentially punctual (or at most local) with isolated types of tasks
and corresponding isolated techniques, where the tasks can be solved without
referring to superordinated technological aspects, i.e., there are praxeologies
with technologies having in particular weak connections to
for .The technological and theoretical discourse remains (so far it is
represented at all) mostly implicit and essentially incomplete. These observations
blend with those presented by González’s et al. (2013) institutional analysis and with
results from a qualitative study by Bauer, Rolka & Törner (2005). For corresponding
results considering prospective secondary mathematics teachers we refer to (Sirotic &
Zazkis, 2007) and for an actual rather detailed study, which problematizes in
particular the use of the number line and investigates the knowledge of fresh French
university students see (Durand-Guerrier, 2016).
Real Numbers in University Education and Potential Foci for Transitions
First of all it is interesting to notice that mostly the treatment of real numbers in
university is either done axiomatically (courses for math majors) or (more or less)
skipped (courses for engineers and natural scientists), but does generally not intend to
connect or complete school praxeologies, i.e. for example: showing the one-to-one
correspondence of number line and decimal fraction views (Kirsch, 1966); discussing
geometrically the completeness of the number line (Artmann, 1983); showing how to
add and multiply non-terminating and non-periodic decimal fractions.

522 [Link]:indrum2018:174703
For improving this situation one might think of transition measures that are adjusted
to study courses for, e.g., math majors, prospective grammar school teachers or
engineering students. A general scheme suitable for describing and analysing
desirable transitions is given as follows (Hochmuth, 2018):

where o represents a praxeology and s a student within the institution S (school), ω a

praxeology and  a student within the institution U (university) in relation to one or
several blocks of the praxeology and is the power set symbol. The
scheme works as a heuristic tool and allows to express that techniques, technologies
or theories of o might be differently relevant for the relation of S within U to a
(perhaps new) praxeology (see (Biehler & Hochmuth, 2017) for a slightly more
restricted scheme).
Applying the scheme we illustrate next various transition foci: At first the focus
might be on techniques and technologies of o such that related skills are
improved, but tasks, techniques and technologies are only slightly extended, for
example: ordering of square roots and decimal fractions; applying calculation rules
for simplifying terms. This might be important for all above mentioned study courses.
Secondly the focus might be on technologies and their further development
(possibly also their theoretical embedding), for example: knowing, that square roots
like or can only be approximated by finite decimal fractions; justifying real
exponentials and powers (Winsløw & Grønbæk, 2014). This might in particular be
important for prospective grammar school teachers. Thirdly the focus might be on
techniques and the replacement of technologies by technologies that are
strongly and systematically embedded in real analysis, for example: constructing the
set of real numbers by Dedekind cuts or Cauchy sequences; starting with axioms for
ℝ and identifying natural and rational numbers within this new set of objects. This
might in particular be important for math majors. Within the ATD-framework each
case could be analysed, explored and specified in greater details and in view of the
scale of level of codeterminations. We skip further details in this paper.
Links to the Subject Scientific Approach
From the subject scientific point of view the analysed praxeologies represent
meanings in the sense of institutionalized action possibilities. Students’ reasoning and
activities ground in those praxeologies but select, neglect or highlight them in view of
an evaluation of their “life interests” and how they are perceived in view of the
institutions S and U, corresponding to positions s and , and in particular aspects
related to levels of codetermination. It is well-known that in the transition a lot of
issues play a role, see e.g. Gueudet (2008). In the following we will discuss two
different but complementing issues that are specifically linked to “Critical
Psychology”.

523 [Link]:indrum2018:174703
The first issue refers to the level “society” and in particular to Dowling’s sociological
analysis of myths, exemplarily the myth of “reference” (Dowling, 2002). Within our
frameworks Dowling’s myths can be (re-)interpreted as technological aspects of
mathematical praxeologies in school that are related to the societal significance of
educational processes. With respect to the latter Dowling differentiates between as-
signments concerning general societal aspects and those concerning their historic
specificity. Considering the historic specificity the exchange-value aspect comes into
play, which somehow undermines the use-value aspect and establishes a problematic
mixture of both. This fits to the observation that actual school introductions of real
numbers typically refer to “real world” problems like doubling the area of a quadratic
piece of chocolate. Such references dominate the justification of introducing real
numbers although treating the “real world” problems does not require “exact” solu-
tions and, moreover, algebraic extensions would be sufficient to resolve this issue.
Following Dowling, the myth of “reference” is not only a didactical issue that relates
to illustrative introductions, but possibly leads to problematic technological and
theoretical ideas, which do not disappear by establishing new, for example, axiomatic
praxeologies, instead they possibly survive and constitute a strong epistemological
(and motivational) obstacle (similar to the epistemological obstacle considered in
(Job et al., 2014)) for students’ learning of university mathematics, in particular for
future grammar school teachers. This could in particular happen, since Dowling’s
myths might dominate students’ general view of their “situation of life” and therefore
their accentuating of meanings.
The second issue relates to the organization of learning in school and to the “school-
and-exam system” mentioned for example in Chevallard (2013). Partly because of
this issue Holzkamp (1993) introduced the notion of defensive learning, a learning
which primarily intends to prevent negative consequences. An important aspect of
this notion is the opposition between ostensive and conceptual thinking that repre-
sents, according to (Holzkamp, 1985), the historic-specific societal concretization of
the cognitive aspect of human activities: Ostensive thinking is essentially
characterized by taking things as they appear to be and, in terms of the 4T-model, by
strongly focusing on locally situated technical and technological issues, which blend
with the above described praxeological organisation of “real numbers” in school.
Again, corresponding “ostensive” students’ views on their “situation of life” and
related meanings-premises-reasoning-patterns might let transition measures’
intentions fail and in particular the incorporation of technological school-blocks
within technological university-blocks, which results in new isolated praxelogies with
new but still weak technologies. It is an empirical open question how this tendency is
amplified by actual initiatives aiming to reduce transition problems by establishing a
“university-and-exam system”.
FINAL REMARK
The hypotheses derived in the last section illustrate the necessity that an analysis of
measures supporting students in the transition from school to university have

524 [Link]:indrum2018:174703
systematically to take into account both approaches, the praxeological and the subject
scientific as well as, with respect to both, the scale of level of codeterminations. The
established link between ATD and the subject scientific approach facilitates
theoretical and actual-empirical studies factoring in systematically aspects, which are
intrinsically connected to the institutional and societal level and have impact both on
institutionalized praxelogies and subjects’ meaning-premises-reasoning-patterns.
REFERENCES
Arndt, A. (2013): Unmittelbarkeit. Owl of Minerva Press.
Artmann, B. (1983). Der Zahlbegriff. Göttingen: Vandenhock & Ruprecht.
Barbé, J., Bosch, M., Espinoza, L. & Gascón J. (2005). Didactic Restrictions on the
Teacher’s Practice: The Case of Limits of Functions in Spanish High Schools.
Educational Studies in Mathematics 59 (1-3), 235-268.
Bauer, L., Rolka, K., & Törner, G. (2005). Mentale Repräsentationen von
Irrationalzahlen – eine Analyse von Schülerinnenaufsätzen. JMD, 26(1), 3-27.
Biehler, R., & Hochmuth, R. (2017). Relating different mathematical praxeologies as
a challenge for designing mathematical content for bridging courses. In: Didactics
of Mathematics in Higher Education as a Scientific Discipline. khdm-Report 17-05
(pp. 14-20). Kassel: Universität Kassel.
Bosch, M. & Gascón, J. (2006). Twenty-Five Years of the Didactic Transposition.
ICMI Bulletin No. 58, 51-65.
Castela, C. (2015). When Praxeologies Move from an Institution to Another One:
The Transpositive Effects. In: 23rd annual meeting of the Southern African asso-
ciation for research in mathematics, science and technology, SAARMSTE, 6-19.
Chevallard, Y. (1991). La transposition didactique. Du savoir savant au savoir
enseigné, 2nd edition. Grenoble: La Pensée Sauvage Éditions.
Chevallard, Y. (1992). Fundamental Concepts in Didactics: Perspectives Provided by
an Anthropological Approach. Recherches en didactique des mathématiques,
Selected Papers. La Pensée Sauvage, Grenoble, 131–167.
Chevallard, Y. (2013). Teaching mathematics in tomorrow’s society: A case for an
oncoming counter paradigm. Regular lecture at ICME-12 (Seoul, 8–15 July 2012).
Dowling, P. (2002). The sociology of mathematics education: Mathematical
myth/pedagogical texts. London: Routledge.
Durand-Guerrier, V. (2016). Conceptualization of the Continuum, an Educational
Challenge for Undergraduate Students. International Journal of Research in
Undergraduate Mathematics Education, 2(3), 338-361.
Gueudet, G. (2008). Investigating the Secondary–Tertiary Transition. Educational
Studies in Mathematics, 67(3), 237-254.

525 [Link]:indrum2018:174703
González-Martin, A.S., Giraldo, V. & Souto, A.M. (2013). The introduction of real
numbers in secondary education – an institutional analysis of textbooks. Research
in Mathematics Education 13(3), 230-248.
Hochmuth, R. (2018). A General Scheme for a Heterogeneous Manifold of
Transitions. Poster, accepted for citad6 (6th International Congress of
Anthropological Theory of Didactics), Grenoble (22nd -26th Jan. 2018).
Hochmuth, R., & Schreiber, S. (2015). Überlegungen zur Konzeptualisierung
mathematischer Kompetenzen im fortgeschrittenen Ingenieurwissenschaftsstudium
am Beispiel der Signaltheorie. In: Lehren und Lernen von Mathematik in der
Studieneingangsphase. Wiesbaden: Springer, 549-566.
Holzkamp, K. (1985). Grundlegung der Psychologie. Frankfurt/M.: Campus.
Holzkamp, K. (1993). Lernen: Subjektwissenschaftliche Grundlegung. Frankfurt/M.:
Campus.
Job, P., & Schneider, M. (2014). Empirical positivism, an epistemological obstacle in
the learning of calculus. ZDM, 46(4), 635–646.
Kirsch, A. (1966). Zur Behandlung der reellen Zahlen im Oberstufenunterricht. Der
Mathematikunterricht im Gymnasium, 215-227.
Lakoff, G., & Núñez, R. E. (2000). Where Mathematics comes from: How the
embodied mind brings mathematics into being. Basic books.
Leontjev, A.N. (1978). Activitiy, consciousness and personality (M.J. Hall Trans.).
Englewoold Cliffs, NJ: Prentice Hall.
Roth, W.-M., & Radford, L. (2011). A cultural-historical perspective on mathematics
teaching and learning. Rotterdam: Sense.
Schraube, E., & Højholt, C. (Eds.). (2015). Psychology and the conduct of everyday
life. Routledge.
Sirotic, N., & Zazkis, A. (2007). Irrational numbers: The gap between formal and
intuitive knowledge. Educational Studies in Mathematics, 65(1), 49-76.
Tolman, C. W. (1991). Critical Psychology: An Overview. In C. W. Tolman & W.
Maiers (Eds.), Critical Psychology: Contributions to an historical science of the
subject (pp. 1-22). Cambridge etc.: Cambridge Univ. Press.
Vygotsky, L.S. (1978). Mind in Society: The development of Higher Psychological
Processes. Cambridge, M.A.: Harvard University Press.
Winsløw, C., Barquero, B., Vleeschouwer, M. de, & Hardy, N. (2014). An institu-
tional approach to university mathematics education: from dual vector spaces to
questioning the world. Research in Mathematics Education, 16(2), 95–111.
Winsløw, C., & Grønbæk, N. (2014). Klein's double discontinuity revisited:
contemporary challenges for universities preparing teachers to teach calculus.
Recherches en didactique des mathématiques, 34(1), 59-86.

526 [Link]:indrum2018:174703
 Evaluating Innovative Measures in University Mathematics – The
Case of Affective Outcomes in a Lecture focused on Problem-Solving
Christiane Kuklinski1, Elena Leis1, Michael Liebendörfer1, Reinhard Hochmuth1,
Rolf Biehler², Elisa Lankeit², Silke Neuhaus², Niclas Schaper², and Mirko
Schürmann²
1
Leibniz Universität Hannover, Institut der Didaktik für Mathematik und Physik,
Germany, kuklinski@[Link] ; ²Universität Paderborn, Germany
The transition from high school to university mathematics has proven to be difficult
for many students but especially for pre-service secondary teachers. To support these
students at mastering this transition, various universities have introduced support
measures of various kinds. The WiGeMath project developed a taxonomy that makes
it possible to describe and compare these measures concerning their goals as well as
their frame characteristics. We exemplify the use of the taxonomy in the description of
one specific innovative measure that was part of the WiGeMath evaluations.
Moreover, we present first results concerning the goal-fulfilment of this measure
concerning affective characteristics of the student cohort and their predominant
beliefs.
Keywords: Transition to and across university mathematics, Novel approaches to
teaching, Teacher education, Motivational developments, Beliefs.

BACKGROUND
In German mathematics teacher education, pre-service teachers first study at
university before they enter a practical training. In this first phase, there is a strong
focus on mathematical content, in particular in higher secondary teacher education
where students mostly attend the same courses as mathematics major students. In
these shared lectures, many internationally known problems of the secondary-tertiary
transition arise (Gueudet, 2008), in particular, motivational problems and drop-out
are often reported. There is a substantial decline in students’ mathematical interest in
the first semester with Cohen’s d around 0.4 (Rach und Heinze, 2013, 2016), a
decline in their mathematical self-concept with Cohen’s d ranging from 0.5 to 0.7
(Rach und Heinze, 2013, 2016), and a strong dominance of controlled motivation
over autonomous motivation (Liebendörfer, in press) in the terminology of Ryan and
Deci’s (2017) self-determination theory. Consequently, many pre-service students
experience their university courses as a necessary evil rather than a helpful
qualification towards their aspired job (Kalesse, 1997; Liebendörfer, in press).

THE WIGEMATH PROJECT

To counteract the negative effects which for many students seem to occur at the
transition between school and university mathematics, many universities have
introduced support measures of various kinds. Even though university internal
evaluations of these measures mostly exist, a framework that helps to facilitate the

527 [Link]:indrum2018:174570
comparison of design and outcomes of different measures has until recently been
lacking. The WiGeMath project (Wirkung und Gelingensbedingungen von
Unterstützungsmaßnahmen für mathematikbezogenes Lernen in der
Studieneingangsphase; Effects and success conditions of mathematics learning
support in the introductory study phase) [1], which is a joint research project of the
Universities of Hannover and Paderborn (Colberg et al., 2016) led by Biehler,
Hochmuth and Schaper, has developed a framework for goal dimensions and frame
conditions of mathematics learning support in universities (Liebendörfer et al., in
press) that aims at such a comparison. Moreover, the project has used the framework
in first evaluations of various support measures at different universities in Germany.
Some exemplary results for one representative of the category of redesigned lectures,
which is one type of support measure that was evaluated in the project, is presented
below.

THE WIGEMATH TAXONOMY

The aim of the WiGeMath project is to develop and exemplify in use a taxonomy that
categorizes features and goals of Projects of Mathematics Learning Support (PMLS)
and to use this taxonomy to evaluate different support measures at German
universities. All projects that fall under the category of PMLS have in common that
they try to support students at the beginning of their university studies in mastering
the critical transition to university mathematics. They are innovative insofar as they
deviate from the standard format of lectures and tutorials that is encountered in
university mathematics even though the way how they do this differs. In the
WiGeMath project, different PMLS are subsumed under one of four categories,
namely bridging courses, mathematics support centres, support measures that parallel
courses and redesigned lectures. Due to space limits, this text focuses only on
redesigned lectures; a description of the other types of PMLS is given in
(Liebendörfer et al., in press).
Redesigned lectures are lectures that offer particular support to students that have
been shown to have higher risks at failing mathematics courses or focus on very
specific learning goals in a non-traditional way. We examined both redesigned
lectures that address preservice secondary teachers, who often show the greatest
problems with the transition from school mathematics to more abstract mathematical
content, and redesigned lectures which address engineering students who had already
failed a compulsory mathematics test of some kind. All redesigned lectures have in
common that new mathematical content is not the focus of teaching.
Different PMLS have different aims some of which are explicit but some of which
stay hidden even to the teaching staff until they are inquired about them by an
outsider. WiGeMath aimed at evaluating different PMLS based on their own
assumptions following a program evaluation approach (Chen, 1990) as well as
comparing them on grounds of an encompassing taxonomy that should categorize

528 [Link]:indrum2018:174570
descriptive (non-normative) goals in the sense of criterions that the PMLSs set out to
meet in their conception, features and conditions of PMLSs.
The taxonomy was constructed in a two-fold process. A first draft was developed by
the project members by means of a document analysis, taking into account documents
provided by project partners. The WiGeMath project collaborates with 11 partner
universities in Germany at which PMLS have been implemented. The draft for the
taxonomy was then used as a guiding thread for guided interviews with teaching staff
of eight PMLS, two of each category. The interviews were taped and transcribed and
afterwards the draft for the taxonomy was tested by trying to fit mentioned goals into
the draft’s categories. This led to minor refinements and reformulations of categories
and yielded the final WiGeMath taxonomy.
This final WiGeMath taxonomy consists of three main categories, namely frame
conditions, measure categories/ characteristics and goal categories. The frame
conditions include various sub-categories, which help to characterize the student
cohort addressed by a PMLS, the way it came about and developed, its embeddedness
in the university course system, organisational characteristics that may pertain to it,
characteristics of the room where it is held, financial and staff conditions and lastly
characteristics of the learning culture. Measure categories/ characteristics serve to
describe certain elements that characterize the PMLS in its structure, its didactical
elements and its teaching staff. Finally, goal categories encompass various sub-
categories of goals that either regard the individual learner or goals that the university
may have in implementing the PMLS as a broader organization as well as a sub-
category that allows to describe the quality of the goals. Not every sub-category has
to be relevant in the description of every PMLS and some aspects of a PMLS may
pertain to more than one category but the use of these categories aims to give an all-
encompassing description of a PMLS’s characteristics.
In the following, the use of the WiGeMath taxonomy shall be exemplified by
applying it to one of the redesigned lectures.

CHARACTERIZATION OF REDESIGNED LECTURES

All in all, six redesigned lectures at five German universities were evaluated in the
WiGeMath project. Out of these, four addressed preservice secondary teachers and
two addressed engineering students. To reach a characterization of each lecture, an
interview guided by the WiGeMath taxonomy was held with the teacher before the
start of the semester. It was audiotaped and transcribed and the transcript was then
used to name the measure’s characteristics and sort them into the right categories of
the WiGeMath taxonomy. What follows is the exemplary description of one of the
evaluated lectures along the lines of the taxonomy’s categories and sub-categories.
Frame conditions
Concerning the characteristics of the student cohort, those students that attended
the given redesigned lecture were preservice secondary teachers in their first semester

529 [Link]:indrum2018:174570
meaning that most of them were less than 20 years old and had just graduated from
high school. The lecture has had round about 200 participants in each of its turns.
The development of the lecture officially started in 2011. Since then, there has not
been a strict script, which is followed each year, but the different lecturers who have
been responsible focused on different aspects. Nonetheless, the basis of the lecture
always is the book by Grieser (2013) which deals with problem-solving strategies and
proofs.
As to the embeddedness of the lecture in the wider system of university lectures, it
is compulsory for preservice secondary teachers and voluntary for mathematics
majors in their third semester. For preservice secondary teachers the lecture has
substituted Linear Algebra as a first semester lecture though they still have to attend
Linear Algebra in their second semester.
Staff conditions have been marked recently by problems to find qualified tutors to
give the tutorials that support the lecture. As noted before, though there is only one
lecturer per semester, it is not always the same one.
Finally, the learning culture is characterized by a strong focus on the students being
active in their learning. They are supposed to try new methods and solve tasks during
the lectures as well as during the tutorials. The concept of a “thinking pause” is very
much enforced in the problem solving process. The lecturer gives some new input at
the beginning of each class and collects and discusses results after the students have
worked on problems or proofs.
Measure categories/characteristics
As to the structural characteristics, the measure consists of a lecture of two times
90 minutes per week with a tutorial of 90 minutes per week. One cycle of the
measure starts at the beginning of each winter semester and finishes at its end
(October through January).
The didactical elements include weekly homework and tutorial work of three to four
exercises. All exercises may be worked on in groups. The solutions are discussed in
the tutorials but no exemplary solutions are handed out. During the lecture there are
phases of teacher talk, partner work and individual work. The book by Grieser (2013)
is named as a reference text and can be accessed online on campus. At the end of the
semester, a written exam concludes the course.
Concerning the characteristics of the teaching staff, the lecturer has his PhD in
mathematics and is responsible for the contents of the lecture as well as the tutorials,
the exercises and the final exam. The tutorials are given by six tutors who are
students in higher semesters. These same tutors also have to correct the exercises
which are handed in by students. As mentioned before, the selection of tutors proved
difficult due to a small number of qualified applicants.

530 [Link]:indrum2018:174570
Goal categories
In the category of individual learning goals the measure focuses on activity-oriented
rather than on knowledge-related goals: Both mathematical working strategies, like
problem-solving strategies or use of examples and counter-examples, and learning
strategies shall be improved. Attitudinal goals play a major role, as well: The measure
aims at strengthening process beliefs and weakening toolbox beliefs in the sense of
Grigutsch and Törner (1998) and wants to introduce the students into the
mathematical professional community. Affective characteristics are to be influenced
insofar as anxieties shall be lessened, interest and motivation shall be strengthened
and the students shall gain a higher mathematical self-efficacy. Moreover, the
measure wishes to let students recognize the relevance of its contents for further
university studies.
The system-related goals include the preparation of the participants for their further
university studies and the decrease of the number of dropouts. Besides, the measure
wants to increase the quality of the feedback that students receive during their
studies.
The quality of the goals as understood in the WiGeMath taxonomy is not to be
understood in a normative sense but rather as a description of their substantiality.
With this aim in mind, goals are examined concerning how specific, measurable,
accepted, realistic and time-phased they are. Such a description of the quality of a
goal would be done for every goal individually in a thorough analysis but due to
space restrictions we will only focus on one specific goal in this paper to illustrate the
point: One of the measure’s goals is to improve affective characteristics of the
participants, i.e. to lessen maths anxiety, increase motivation and interest and improve
the participants’ mathematical self-efficacy, in the course of the semester. This goal is
specific to the point that it explicates what shall be achieved when by whom. It
remains unspecific in naming why the goal is important, who holds the responsibility
to reach it and which preconditions or limitations possibly exist. The goal is indirectly
measurable through a survey directed at the students and it is accepted as it was
named by the lecturer as a goal he wants to achieve rather than a goal he has to
achieve due to orders given from above. The goal seems to be realistic to the point
that the affective characteristics of the students seem to change for worse quite fast at
the beginning of their university studies so it seems plausible that they may be
changed for better even within a single semester. Still, this has to be checked as will
be shown below. Finally, the goal is time-phased as it shall be reached within a
limited time, namely the duration of the measure.
In this paper we will evaluate the following research question: To what extent was the
described redesigned lecture successful in achieving the last-mentioned goal of
influencing affective characteristics of the student cohort and in how far was the
lecture successful in changing beliefs away from toolbox beliefs towards process
beliefs?

531 [Link]:indrum2018:174570
METHODS
To measure the extent to which the above mentioned goals were reached, two
questionnaire surveys were conducted with the participants of the described lecture.
The first survey (t1) took place in the second week of the winter semester 2016 and
the second one (t2) was conducted in the second to last week of the same semester.
For each survey a questionnaire was developed, each one laid out to take about thirty
minutes to complete. These questionnaires were handed out at the end of a lecture so
that only those who were present that day could participate and participation was
voluntary which the students were informed about. Moreover participation was
anonymous but students used an individually constructed code so that the results of
each participant in the first survey could be compared and contrasted to the results in
the second survey. 163 students participated in the first survey and 103 in the second.
We analyze the data of the 76 participants who answered both questionnaires.
We used adopted versions of the scale of Schiefele, Krapp, Wild and Winteler (1993)
to measure interest, the scale of Schöne, Dickhäuser, Spinath, & Stiensmeier-Pelster,
(2002) to measure mathematical self-concept, a translation of the instrument of
Longo, Gunz, Curtis and Farsides (2014) to measure the experience of competence,
autonomy and social relatedness, an adopted version of the PISA 2000 instrument for
self-efficacy (Kunter et al., 2002), a shortened version of the scales by Grigutsch and
Törner (1998) to measure beliefs (application, process, system and toolbox) and a
translated version of the academic motivation scale (Vallerand, Pelletier, Blais,
Briere, Senecal and Vallieres, 1992) to measure different types of motivational
regulation (intrinsic, identified, introjected and extrinsic). We used Likert scales
ranging from 1 to 4 for self-concept, self-efficacy and beliefs, from 1 to 5 for
motivational regulation, from 1 to 6 for interest and from 1 to 7 for the experience of
competence, autonomy and social relatedness. All reliability coefficients (Cronbach’s
alpha) are acceptable or better, compare Table 1.
For each scale and each survey, a descriptive data analysis was conducted in order to
get an overview of the results. Though the entirety of scales included more than the
ones mentioned above, we concentrate on these only as our focus is to check to what
extent affective characteristics of the student cohort, their experience of competence
and their attitude towards different beliefs was changed in the course of the semester.

RESULTS
Table 1 shows the changes in mean values during the semester and effect sizes
(Cohen’s d) as well as p-values of paired t-tests.
Scale Number Cronbach’s α Mean value Cohen’s p-
of items t1 t2 t1 t2 d value

Interest for mathematics 9 .83 .83 4.23 3.96 0.32 .001

Mathematical self-concept 3 .81 .81 3.03 2.96 0.12 .149

532 [Link]:indrum2018:174570
 Experience of competence 6 .80 .81 4.63 4.25 0.40 < .001
Experience of social 6 .85 .89 5.39 5.40 0.01 .900
relatedness
Experience of autonomy 6 .73 .77 4.81 4.61 0.20 .132
Mathematical self-efficacy 4 .83 .87 2.72 2.66 0.10 .353
Application beliefs 4 .80 .88 3.01 3.02 0.02 .889
Process beliefs 4 .67 .85 3.26 3.18 0.12 .306
System beliefs 7 .79 .84 2.97 2.93 0.07 .534
Toolbox beliefs 5 .66 .74 2.75 2.56 0.34 .002
Intrinsic regulation 5 .88 .88 3.82 3.55 0.33 .001
Identified regulation 4 .72 .78 4.01 3.81 0.24 .026
Introjected regulation 4 .73 .78 2.04 2.19 0.18 .097
Extrinsic regulation 4 .64 .72 1.78 1.88 0.12 .278

Table 1: Scales and their Cronbach’s alphas, means, effect sizes of changes between
the two surveys and p-values for a significant change.
We see a substantial decline in interest and in the experience of competence, whereas
students’ mathematical self-concept and self-efficacy did not change significantly. We
can also see that the objective of reducing toolbox beliefs was clearly achieved, but
not the objective of strengthening process beliefs. The mean values of motivational
regulation show that intrinsic and identified regulation are dominating although they
are decreasing in the course of the semester.

DISCUSSION
The observation of a decline in interest is similar to the results showing a decline in
traditional courses; however, student’s mathematical self-concept does not change
significantly, which is a major difference (Rach und Heinze, 2013, 2016). Although in
our tests we were not able to show that the course could raise student’s self-efficacy,
it did not reduce it significantly either, which may still be an achievement. The
dominance of intrinsic and identified motivation is a positive result as a study in
traditional courses found extrinsic and introjected regulations to be dominant
(Liebendörfer, in press). Thus, although students’ interest in university mathematics
and their intrinsic motivation may reduce, they do not seem to develop a stronger
feeling of being inadequate for studying mathematics in the newly designed lecture.
The decline in interest as well as intrinsic and identified regulation may be explained
by a decline in the experience of competence. The change in students’ toolbox beliefs
is remarkable as beliefs are rather stable by definition and toolbox beliefs did not
change in other studies in the first year of lower secondary or primary mathematics

533 [Link]:indrum2018:174570
teacher education (Kolter, Liebendörfer & Schukajlow, 2016; Liebendörfer &
Schukajlow, 2017).
These results show that a specifically designed lecture may reduce problems of the
secondary-tertiary transition in mathematics. Nonetheless, the question remains
whether such lectures prepare the students for their further studies just as well as
traditional teaching does, considering that the course covered fewer mathematical
topics.
Moreover, the analysis that has been done to this point cannot ensure that the results
obtained were produced by the innovative measure alone. First of all, the lecturer’s
personality has an influence on the measure’s outcomes that could not be separated
from the outcomes of the measure itself in our study. A possible further effect may be
caused by the change in order of other lectures, in explicit the postponement of
Linear Algebra to a later semester: Whereas students usually experience their low
competency in both Analysis and Linear Algebra in the first semester, in this case it is
only one lecture.
In order to test this hypothesis, a next step in the WiGeMath project will be to
distribute the same questionnaire that was used in the investigation described above
to a different innovative measure at a university where different courses are attended
simultaneously, as well as to a traditional lecture. This will make comparisons more
explicit.
As to the taxonomy that was developed by the WiGeMath project, this in part
resembles the objectives of other taxonomies (Krathwohl, 2002) though with a
different focus. Whereas other taxonomies are mostly concerned with individual
learning outcomes, the WiGeMath taxonomy targets a description and ensuing
comparison of innovative measures as a whole. Though other taxonomies exist which
classify systems of higher education institutions (for example the Carnegie
Classification of Institutions of Higher Education, described in Bartelse & Vught,
2009), the perspective taken by WiGeMath to interpret such characteristics as goals is
a new one.
So far, the taxonomy is only laid out to serve innovative measures and even in this
area will have to be adapted as measures develop and improve. Traditional lectures
have not been taken into consideration so far but we propose that these would also
benefit from a similar taxonomy in terms of communicating frame conditions and
learning goals. In our interviews with lecturers, we found that often even to them
goals remained implicit until they were asked about them specifically. This might
even more be the case in traditional lectures that have “worked” for a long time.
As mentioned above, in many cases goals stay hidden until a framework like the one
developed by the WiGeMath project provides a common language to talk about them.
Even though lecturers have specific intentions when they design a course with
specific learning goals that a student cohort with certain characteristics shall achieve
in a setting framed by staff conditions, learning culture, etc., they often lack

534 [Link]:indrum2018:174570
guidelines to arrange these in a way that is comprehensible for others. Yet, only if
they can explicate their ideas, can an evaluation be successful and show strengths as
well as possible weaknesses of the designed course. In our example, the lecturer had
his PhD in mathematics and had hardly been in contact with didactical theories and
frameworks until the point of the WiGeMath evaluation. Hence, he had would not
speak in terms of mathematical beliefs, for example. When the concept was explained
to him, though, he clearly saw that one intention of the lecture was to change
students’ beliefs but to that point he simply lacked the vocabulary to explain this
intention.
Our taxonomy will help to communicate goals between universities, staff and
students as it provides a frame of reference and a common language as has been
shown for one example in this text.

NOTES
1. The WiGeMath project is funded by the German Federal Ministry of Education and Research (BMBF), grant
identifiers 01PB14015A and 01PB14015B.

REFERENCES
Bartelse, J., & Vught, F. van. (2009). The European Higher Education Classification:
Objectives and Concepts. In F. van Vught (Ed.), Mapping the Higher Education
Landscape (pp. 57–69). Springer Netherlands.
Chen, H.-T. (1990). Theory-driven evaluations. Newbury Park, California: Sage.
Colberg, C., Biehler, R., Hochmuth, R., Schaper, N., Liebendörfer, M., &
Schürmann, M. (2016). Wirkung und Gelingensbedingungen von
Unterstützungsmaßnahmen für mathematikbezogenes Lernen in der
Studieneingangsphase. In Beiträge zum Mathematikunterricht 2016 (pp. 213–216).
Heidelberg: WTM-Verlag für wissenschaftliche Texte und Medien.
Grieser, D. (2013). Mathematisches Problemlösen und Beweisen. Wiesbaden:
Springer.
Grigutsch, S., & Törner, G. (1998). World views of mathematics held by university
teachers of mathematics science (Preprint Series No. 420). Duisburg: University of
Duisburg, Department of Mathematics. Retrieved from [Link]
[Link]/go/duett-05272002-102811
Gueudet, G. (2008). Investigating the secondary–tertiary transition. Educational
Studies in Mathematics, 67(3), 237–254.
Kalesse, I. (1997). Change in Mathematical Views of First Year University Students
II. In M. Hannula (Ed.), Current State of Research on Mathematical Beliefs V (pp.
30–35). Helsinki: University of Helsinki, Dept. of Teacher Education.
Kolter, J., Liebendörfer, M., & Schukajlow, S. (2016). Mathe – nein danke?
Interesse, Beliefs und Lernstrategien im Mathematikstudium bei

535 [Link]:indrum2018:174570
 Grundschullehramtsstudierenden mit Pflichtfach. In A. Hoppenbrock, R. Biehler,
R. Hochmuth, & H.-G. Rück (Eds.), Lehren und Lernen von Mathematik in der
Studieneingangsphase (pp. 567–583). Wiesbaden: Springer Fachmedien
Wiesbaden.
Krathwohl, D. R. (2002). A Revision of Bloom’s Taxonomy: An Overview. Theory
Into Practice, 41(4), 212–218.
Kunter, M., Schümer, G., Artelt, C., Baumert, J., Klieme, E., Neubrand, M., … Weiß,
M. (2002). PISA 2000: Dokumentation der Erhebungsinstrumente. Berlin: Max-
Planck-Inst. für Bildungsforschung.
Liebendörfer, M., Hochmuth, R., Biehler, R., Schaper, N., Kuklinski, C., Khellaf, S.,
… Rothe, L. (2017). A framework for goal dimensions of mathematics learning
support in universities. In T. Dooley & G. Gueudet (Eds.), Proceedings of the
Tenth Congress of the European Society for Research in Mathematics Education
(CERME10, February 1 – 5, 2017). Dublin: DCU Institute of Education & ERME.
Liebendörfer, M. (in press). Motivationsentwicklung im Mathematikstudium.
Wiesbaden: Springer.
Liebendörfer, M., & Schukajlow, S. (2017). Interest development during the first year
at university: do mathematical beliefs predict interest in mathematics? ZDM, 49(3),
355–366.
Longo, Y., Gunz, A., Curtis, G. J., & Farsides, T. (2014). Measuring Need
Satisfaction and Frustration in Educational and Work Contexts: The Need
Satisfaction and Frustration Scale (NSFS). Journal of Happiness Studies, 17(1),
295–317.
Rach, S., & Heinze, A. (2013). Welche Studierenden sind im ersten Semester
erfolgreich? Journal Für Mathematik-Didaktik, 34(1), 121–147.
Rach, S., & Heinze, A. (2016). The Transition from School to University in
Mathematics: Which Influence Do School-Related Variables Have? International
Journal of Science and Mathematics Education, 1–21.
Ryan, R. M., & Deci, E. L. (2017). Identity Development, Self-Esteem, and
Authenticity. In Self-Determination Theory: Basic Psychological Needs in
Motivation, Development, and Wellness. New York: Guilford Publications.
Schiefele, U., Krapp, A., Wild, K. P., & Winteler, A. (1993). Der “Fragebogen zum
Studieninteresse” (FSI). Diagnostica 39(4), 335–351.
Schöne, C., Dickhäuser, O., Spinath, B., & Stiensmeier-Pelster, J. (2002). Skalen zur
Erfassung des schulischen Selbstkonzepts: SESSKO. Göttingen: Hogrefe.
Vallerand, R. J., Pelletier, L. G., Blais, M. R., Briere, N. M., Senecal, C., & Vallieres,
E. F. (1992). The Academic Motivation Scale: A measure of intrinsic, extrinsic, and
amotivation in education. Educational and Psychological Measurement, 52(4),
1003–1017.

536 [Link]:indrum2018:174570
 Defining as discursive practice in transition – Upper secondary
students reinvent the formal definition of convergent sequences
Alexander Schüler-Meyer
TU Dortmund University, Germany
[Link]-meyer@[Link]
The investigation of limits is at the heart of analysis at university. Accordingly, it is a
worthwhile topic for transition courses. The present study engages upper secondary
students in reinventing the definition of convergent sequences. Using a commognitive
framework, the central development stages of the definition from experiential to
abstract are empirically investigated in terms of activated secondary school
discourses. The students’ familiarity with secondary school discourses is critical, as
it allows them to transition from grasping processes with metaphors towards
grasping them as formal and abstract objects. For this, school objects act as
intermediate steps. Further studies of transition courses should explicitly address the
role of students’ secondary notions as resources for reifying processes into abstract
objects.
Keywords: transition; limits; convergence; practice of defining; design research.
INTRODUCTION
The investigation of limits is at the heart of analysis on the university level (Cornu,
1991). Accordingly, limits are a worthwhile object of investigation in a transition
course. In the German context, students also have previous knowledge about limits in
the domain of derivatives from upper secondary school. At school, the students’
notions are usually not developed into a formal understanding of limits, as it is not
expected in the German curriculum. Students transitioning to university have to
develop their formal notion of limits in the regular lecture, which can be difficult.
Thus, transition courses located at school have a huge potential for supporting
students in developing a more compatible formal and abstract understanding of
limits.
Objects in school have an experiential basis, while objects in tertiary mathematics
“are specified by formal definitions and their properties reconstructed through logical
deductions” (de Guzmán, Hodgson, Robert, & Villani, 1998, p. 753). For different
reasons, students in transition often do not use definitions as starting points for their
reasoning about objects, as would be expected in tertiary mathematics (Edwards &
Ward, 2008, Vinner, 1991). Engaging students in activities of defining objects on a
trajectory from experiential to formal approaches might help alleviate these issues.
In this paper, five students in a transition course from secondary to tertiary education
in Germany are asked to create a definition of convergent sequences, after having
studied them with the model of epsilon strips the week before. The paper illustrates
how students’ progress from their experiential understanding of convergence towards

537 [Link]:indrum2018:196204
a formal definition based on deductions. The qualitative analysis also unfolds how the
students use their previous school knowledge for this.
DEFINING AS A MATHEMATICAL PRACTICE IN TRANSITION
Students’ intuitions about limits
The transition from secondary to tertiary mathematics has been extensively addressed
in terms of changes and obstacles (Thomas et al., 2015). A specific issue of transition
in regard to the understanding of limits are the students’ intuitions (for typical
metaphors: Oehrtmann, 2009). The students’ intuitions about limits are not
surprising: Historically, mathematicians used their intuitions to think about infinity
based on infinitesimals, and this kind of reasoning still permeates modern analysis
despite not being accepted as adequate (Cornu, 1991). Students’ intuitions are a
fruitful starting point for reasoning about limits. Activities of using an epsilon-strip
help students to understand the convergence of sequence in terms of neighborhoods,
in which are “almost all terms of the sequence” (Przenioslo, 2005, p. 88). They can
also help students to understand the logical relations of  and N (Roh, 2010).
Another issue is the role of pre-formal notions about limits. In upper secondary
classrooms in Germany, pre-formal notions of limit are encouraged by teachers and
textbooks, as for example n is commonly referred to as “tends to”. By using and
extending metaphors of “how many terms make a party” (p. 335) into “an infinite
amount of terms will be in that epsilon neighborhood and a finite amount of terms
will be outside”, this understanding can be developed into a formal understanding of
convergence (Dawkins, 2012). It has been illustrated that students can reinvent the
formal definition (Swinyard, 2011), but not how students’ intuitions and previous
knowledge from secondary schools can systematically be activated in this process.
Hence, carefully guided activities of investigating sequences with epsilon strips might
help students make experiences that connect to their intuitions, but might at the same
time be rich enough for students to develop a formal and abstract understanding of
limits. While the first aspect has been investigated (see above), the latter aspect of
transitioning to the formal and abstract while connecting to previous knowledge and
intuitions has not yet been investigated.
The practice of defining
Definitions are the central means for grasping objects on the tertiary level, and the
starting point for mathematical reasoning about these objects (Vinner, 1991; Alcock
& Simpson, 2002). However, in nearly all mathematical domains, students in
transition often rely on their informal understanding of objects, instead of definitions
(overview in Thomas et al., 2015). There are several reasons why students have
difficulties with defining and definitions. Definitions describe objects in arbitrary
ways (Vinner, 1991), i.e. students do not have an experiential basis with the defined
objects. Furthermore, students are usually not engaged in practices of defining.
Instead, students encounter definitions in processes of proofing or validating
(Swinyard & Larsen, 2012). When students can encounter definitions as product of

538 [Link]:indrum2018:196204
their own making, it might help them to become aware of the nature of definitions at
the university level (Swinyard, 2011).
In sum, students can and should be engaged in practices of defining limits. From a
transitionary standpoint, practices of defining should be rooted in familiar secondary
activities with empirical investigations to create an experiential basis (de Guzmán et
al., 1998). Then, practices of defining in a transition course should start with an
experiential basis, and let students reinvent an own ‘arbitrary’ but viable definition,
and later progress to a more formal and abstract definition.
THEORETICAL BASIS: COMMOGNITION AND GUIDED REINVENTION
Commognitive Perspective on Learning and tertiary mathematics
Within the framework of commognition, students’ participation in discourses is in-
vestigated in terms of changes in the ways students discursively realize mathematical
objects in their utterances, where such changes constitute learning (Sfard, 2008).
Mathematical discourses can be distinguished from each other by their use of
keywords (e.g. “tends to”), of visual mediators (e.g. graphs and symbols), of practices
(“routines”, patterned activities like defining or proving) and of narratives (like
definitions) (Sfard, 2008, p. 134). Of special interest in the present study is the
students’ use of visual mediators and of narratives. Visual mediators are central
means for grasping convergent sequences with epsilon strips and with mathematical
symbols. Narratives are the means for students to discursively realize relations
between objects, or between the facets of the object convergent sequence, like  or N.
The development of a definition (as a narrative) and its associated mathematical can
be characterized through the ways it is discursively realized over time. This develop-
ment of the definition can be visualized as a “realization tree”. The branches of the
realization tree represent the distinct ways in which a definition, in this case about the
object convergent sequence, is discursively realized by the students (Sfard, 2008, p.
153), up to a certain point in time in the discourse.
Research program of design research and framework of guided reinvention
The present study follows a guided reinvention approach based on realistic mathe-
matics education. It engages students in context problems in which a mathematical
object is the result of own experiences and activities (Gravemeijer & Doorman,
1999). This is a fruitful approach for introducing tertiary mathematical objects (e.g.
Dawkins, 2012; and many others). It is located in the research program of design
research (Prediger, Gravemeijer, & Confrey, 2015).
The underlying hypothetical learning trajectory builds on the research on the learning
of limits. Building on the epsilon-strip activity in the previous session, students are
asked to document all facets which are relevant for convergence, for example the
height of the strip. Afterwards, the students have to bring these previously found
facets into a logical relationship in the form of a narrative. At the same time, the
students have to formalize the activity, including the process of finding a limit

539 [Link]:indrum2018:196204
candidate. A critical step in this trajectory is the need to progress from an x-first
perspective, in which students focus on inputs (x-values) and their respective outputs
(y-values), to a y-first perspective. The x-first perspective emphasizes finding a
candidate for a limit (Swinyard & Larsen, 2012). In regard to continuity, the y-first
perspective and finding a limit candidate is implied in the epsilon-strip activity, as
students first choose a strip with a certain height and afterwards arrange it on the
sequence. Nevertheless, previous experiences with functions might still guide
students to consider an x-first-perspective, so that the y-first perspective needs to be
stabilized.
Research questions
In the present study, the students’ progression from their experiential notion of
convergence with epsilon-strips towards a more the formal, abstract notion of
convergence along the outlined trajectory is investigated. The students are engaged in
practices of defining, in line with the previous considerations. This encompasses,
among others, the following activities: 1. Identifying central elements that should
constitute a definition, and baptizing them, and 2. exploring or deducing relations of
these elements. Of special interest are the students’ ways of building on and
connecting with their previous secondary school discourses. The study focuses on the
emergent practice of defining, as students’ preceding exploration of sequences with
epsilon strips has already been studied in other studies (Przenioslo, 2005).
The following research questions will be investigated in the present study:
Q1. What are the crucial steps in the students’ learning trajectory from experiential
investigations towards a formal definition of convergence?
Q2. How do students activate their previous knowledge from school, and what role
does this knowledge play to proceed to a formal, abstract definition?
METHODOLOGY
Participants and implementation
The design research project encompasses three design experiment cycles, the data
analyzed in this paper stem from the third cycle in which five students participated
(Ludwig, Lawrence, Dominic, Leif and Tanja). These students are highly proficient
eleventh-graders, in their penultimate year of upper secondary education. They
participate voluntarily in a one-year long transition course, designed by the author for
preparing for university STEM-studies. From their regular mathematics classrooms,
the students are familiar with an informal understanding of continuity (“Drawing
without lifting the pen”) and with limits as “tends to”. The teaching unit comprised 5
sessions of 90 minutes each, three sessions on convergent sequences, two sessions on
continuity. The teaching unit was taught by a Master student with tutoring
experiences at university level. All sessions were videotaped and transcribed.
The analysis focuses on Task 2 of Session 2, in which the students attempt to find a
formal definition for convergence by drawing upon their activities with the epsilon-

540 [Link]:indrum2018:196204
strips (Session 1). Before Task 2, the students recapitulated the activity of using the
epsilon-strips with a given convergent sequence. Furthermore, the students have
generated a ‘knowledge storage’ in the first session, where they documented the
relevant mathematical facets of convergence, and visualized them. Session 2 finishes
with the completion of the here investigated task.
Analysis of data
The transcripts of Task 2 in Session 2 were analyzed qualitatively in the framework
of commognition (Sfard, 2008). The central steps of the analysis are:
To answer Q1, the collective ways to realize the object convergent sequences over
time are analyzed and depicted in form of a realization tree. For that, the transcript is
segmented according to the qualitatively different ways in which the students realized
the object convergent sequence. As illustrated, changes in the ways of realizing
objects with keywords, visual mediators and their relations with narratives in the
discourse are indicative of learning. Each branch in the tree hence represents one
distinct way of grasping convergence discursively. The progression from one branch
of the realization tree to the next one marks a crucial step in students’ learning
trajectory, as the students change the discourse about convergence in some
substantial regard.
To answer Q2, the transcript is analyzed in regard to episodes in which the students’
discourse builds on school objects, which is indicated by keywords like “function”,
“y-value”, and narratives, e.g. the metaphor of “tends to”. These episodes are
analyzed in terms of how the students proceed from secondary utterances about
school objects towards more tertiary abstract and formal utterances. An object
becomes abstract, when its narratives do not refer to empirical phenomena, but to
other (previous) mathematical narratives. An object is considered formal, if it is
typically realized with a symbol as visual mediator (Sfard, 2008).
RESULTS
Different Realizations of convergent sequences and their progression
After having investigated a convergent sequence for 20 min, the students engage in
finding a definition for convergent sequences in Task 2 for the remainder of Session
2. For that, they can build on their “knowledge storage”, in which they documented
the central facets of convergence together with their respective graphical
representations. The realization tree in Figure 1 illustrates how students progressively
realize convergent sequences, from left to right, and achieve an abstract, formal
definition (fourth branch). The small rectangles denote the predominant discursive
means used. The turn numbers “Tx” under each oval localize the realization branches
in the classroom conversation. The conversation lasts 393 Turns.
In regard to the question of proceeding from an experiential notion of convergence
towards an abstract definition (Q1), this realization tree reveals several interesting
features of the developing practice of defining in transition. At two points in the pro-

541 [Link]:indrum2018:196204
cess students specifically go back to the experiential basis of working with the
epsilon strip, in the first and third branch of the realization tree. In these instances, the
students rely on metaphors to make sense of their experiences. In interaction with the
teacher, who explicitly states the rule to not use the metaphor “tends to” after #237,
the students adopt the phrase “it exists” from the teacher to form a new narrative
about A in terms of functional relationships, leading to the notion of A(m)/ N.
In the second and fourth branch of the tree, the students activate their previous school
knowledge in order to grasp their experiences in a more formal and objectified way.
This will be investigated in more detail in the next section.
Activation of secondary discourses
The starting point for the students to define convergence are narratives about limits
grounded in “tends to”, resulting in a narrative about two interconnected limit
processes: “m tends to 0 if A tends to infinity. And we need a target value” (#171-
172). In this first branch of the realization tree (Fig. 1), the students rely heavily on
the familiar school narratives of “tends to”. Now, the teacher establishes the dis-
cursive rule that the students should avoid using “tends to” (#213-230). The students
attempt to follow this rule, which leads to a change in the discourse (second branch in
Fig. 1). The following conversation occurs right in the beginning of these attempts:
237 Lawrence Yes, A is getting smaller and smaller, I mean, bigger, if m is getting
smaller and smaller. But we should avoid the tends to.
238 Leif One could try somehow proportional, like proportional-technique to
plot it.
239 Ludwig Dependent from A. Hence, writing in the index.
240 Lawrence A in dependence of m.
241 Leif I would have said, like, somehow m is proportional to 1 divided by A.
In this episode, Lawrence summarizes the result of the previous discussions in terms
of “tends to”. Leif proposes to think about the relations in terms of proportionality,
and Ludwig and Lawrence pick this up in terms of functional relations.
The students can give up narratives with “tends to” by first replacing the “if…then”
relationship with a functional relationship, as indicated by keywords of ‘dependence’.
These keywords show that the students’ ideas are rooted in familiar secondary
discourses about functional relationships. It seems that these secondary notions are
brought into the discourse associatively. Accordingly, the viability of these notions is
up to debate, and competing narratives are uttered (#240, 241). This guides Leif to
specify his proposal into “m is proportional to 1/A”. In this episode, school notions
allow the students to collaboratively develop new narratives. They help students to
engage in a new discourse in which the discursive rules have changed.
In the following third episode, the students try to decide about the nature of the
hypothesized functional relationship.

542 [Link]:indrum2018:196204
Figure 1. Realizations of the object convergent sequence

543 [Link]:indrum2018:196204
 279 teacher Yes. You already said, what else is dependent on m?
280 Lawrence Yes, A.
281 Leif We have the dependence of m and A.
282 Lawrence I write this down now like in short as tends to. [Lawrence writes
m0, A]
283 teacher What is dependent on what? […]
287 Ludwig m is the y-value, thus it is always dependent from the x-value.
288 Lawrence You take, like, the strip and look then, from which value on all of
them are in it.
289 Leif Yes, but you, you #
290 Lawrence We have never taken a value and then looked at how high the strip
needs to be. […]
293 Leif Yes, OK, then A is dependent from m.
294 Lawrence If m gets smaller, A gets bigger.
While the students still document their ideas in terms of the narrative of “tends to”
(#282), the teacher engages them in thinking again about the functional relationship.
Ludwig treats sequences in terms of properties of functions, namely in terms of y-
and x-values (#287). Based on this, he concludes that, as with functions, m has to be
dependent from A, taking a x-first perspective. Lawrence goes back to activities of
using the epsilon strips, and how to choose a strip (#288), proposing an y-first pers-
pective as implied by the epsilon strip activity. Hence, Ludwig treats sequences as a
variation of the familiar secondary objects of functions, but his functional perspective
is contested by Lawrence, who enforces his narratives about the epsilon strips and
their height and placement (#288, 290). In the end, both narratives are merged as “A
is dependent from m” (#293, 294), resulting in an endorsed narrative about A(m).
Here, the students’ previous school knowledge about functions and functional
relationships is the link between the students’ experiences with epsilon strips and the
formal notion of A(m) as an object. It provides the narrative that A and m can be
brought together as a combined object A(m) under a y-first perspective, and the visual
mediator/symbol of A(m). However, as Ludwig’s utterance illustrates, treating se-
quences as functions is at the same time misleading, as it reinforces an x-first per-
spective. Above that, there is still an echo of the metaphor “tends to” in A(m), as the
students use this narrative to summarize their ideas about the functional relationships
(#282, #294), suggesting that A(m) (analogue to N) is an evolution of this narrative.
Synthesis and Summary
In answer to Q2, secondary discourses seem to have at least two functions in the
students’ progression to more abstract discourses (branches of the realization tree in
Fig. 1), where abstract means that students more and more endorse narratives about
narratives instead of narratives about experiences. First, they are the source for the
students to develop new discourses with a new set of narratives and visual mediators
– in this case about A(m) [N] – but similar keywords, after the teacher establishes the

544 [Link]:indrum2018:196204
discursive rules that “tends to” is not to be used (episode 1). Second, secondary
discourses unlock objectified narratives. They allow to transform the narratives about
experiences with “tends to” or with “being in the strip” into narratives about abstract
objects of A(m) [n] (episode 2) and about d [|an – a|] and “d ≤ m” (Fig. 1). This is an
example of saming, where “proportionality” (#238) is imported into narratives about
experiences with the epsilon strips (Sfard, 2008, p. 170). This step is critical in
progressing from an experiential process-notion of relations, which is suggested by
the activities of using the epsilon strips, towards understanding these relations as
abstract objects. Nevertheless, the teacher is needed to provide the phrases and
scaffolds by which the students can engage in new discourses with new narratives,
e.g. about logical relations. This is expected, as the students neither have previous
knowledge about the typical phrases and narratives from the tertiary level, nor about
the practice of defining.
DISCUSSION
The here presented study contributes to the ongoing investigation of teaching inter-
ventions that help students to understand the abstract definition of limits. It especially
highlights the role of the students’ previous experiences with secondary mathematical
discourses in practices of defining: Students from an upper secondary classroom
connect to secondary narratives about limits and other objects (functions) in order to
make sense of their experiences with the epsilon strips, and to formulate a definition.
The students’ secondary school notions are critical for transposing secondary
narratives with metaphors (“tends to”) into more abstract narratives about narratives
that grasp logical relations. As relatively formal objects, school objects mediate
between informal experiences of relationships and a more tertiary, abstract
understanding of these relationships as objects, by allowing saming (Sfard, 2008).
Investigating students’ practices in terms of familiar secondary discourses and of
distinct steps of realizations has proven highly insightful to understand students’
resources in transitioning to tertiary mathematical discourses. The results of the
present study call for investigating students’ transition to tertiary mathematics not
only in terms of difficulties, but also in terms of students’ (secondary school)
resources and how these resources are situationally activated when the discursive
rules change.
REFERENCES
Alcock, L., & Simpson, A. (2002). Definitions: Dealing with categories
mathematically. For the Learning of Mathematics, 22(2), 28-34.
Cornu, B. (1991). Limits. In D. Tall (Ed.), Advanced mathematical thinking (pp. 153-
167). Dordrecht, The Netherlands: Kluwer Academic.
Dawkins, P. C. (2012). Metaphor as a possible pathway to more formal
understanding of the definition of sequence convergence. The Journal of
Mathematical Behavior, 31(3), 331-343.

545 [Link]:indrum2018:196204
De Guzmán, M., Hodgson, B. R., Robert, A., & Villani, V. (1998). Difficulties in the
passage from secondary to tertiary education. In G. Fischer & U. Rehmann (Eds.),
Proceedings of the international Congress of Mathematicians (pp. 747-762).
Berlin: Documenta Mathematica.
Edwards, B., & Ward, M. B. (2008). The role of mathematical definitions in
mathematics and in undergraduate mathematics courses. In M. P. Carlson & C.
Rasmussen (Eds.), Making the connection: Research and teaching in
undergraduate mathematics education (pp. 223-232). Washington, DC:
Mathematical Association of America.
Gravemeijer, K., & Doorman, M. (1999). Context problems in realistic mathematics
education: A calculus course as an example. Educational Studies in Mathematics,
39(1-3), 111-129.
Oehrtmann, M. (2009). Collapsing Dimensions, Physical Limitation,
and Other
Student Metaphors for Limit Concepts. Journal for Research in Mathematics
Education, 40(4), 396-426.
Prediger, S., Gravemeijer, K., & Confrey, J. (2015). Design research with a focus on
learning processes: an overview on achievements and challenges. ZDM
Mathematics Education, 47(6), 877-891.
Przenioslo, M. (2005). Introducing the Concept of Convergence of a Sequence in
Secondary School. Educational Studies in Mathematics, 60(1), 71-93.
Roh, K. H. (2010). An empirical study of students’ understanding of a logical
structure in the definition of limit via the ε-strip activity. Educational Studies in
Mathematics, 73(3), 263-279.
Sfard, A. (2008). Thinking as communicating: Human development, the growth of
discourses, and mathematizing. New York, NY: Cambridge University Press.
Swinyard, C. (2011). Reinventing the formal definition of limit: The case of Amy and
Mike. The Journal of Mathematical Behavior, 30(2), 93-114.
Swinyard, C., & Larsen, S. (2012). Coming to understand the formal definition of
Limit: Insights gained from engaging students in reinvention. Journal for Research
in Mathematics Education, 43(4), 465.
Thomas, M. O. J., de Freitas Druck, I., Huillet, D., Ju, M.-K., Nardi, E., Rasmussen,
C., & Xie, J. (2015). Key Mathematical Concepts in the Transition from Secondary
School to University. In S. J. Cho (Ed.), The Proceedings of the 12th International
Congress on Mathematical Education (pp. 265-284). Cham: Springer.
Vinner, S. (1991). The role of definitions in the teaching and learning of
mathematics. In D. Tall (Ed.), Advanced mathematical thinking (pp. 65-81). New
York, NY et al.: Kluwer Academic.

546 [Link]:indrum2018:196204
 Fostering Heuristic Strategies in Mathematics Teacher Education
Peter Stender and Ann Sophie Stuhlmann
University of Hamburg, Germany; [Link]@[Link],

The “double discontinuity” stated by Felix Klein 1908 is still relevant in the
mathematics teacher education at German universities. We are developing a course
approach, which is intended to bridge the double discontinuity in a didactic
dimension. We offer additional learning opportunities for teacher students, which
are characterized by clarifications of mathematical methods that are fundamental to
the regular lecture contents. We focus especially on reflections on heuristic
strategies, which are the core part of mathematical methods according to George
Pólya. Feedback shows that students consider our approach as helpful in the
transition from school to university.

Keywords: Novel approaches to teaching, teaching and learning of analysis and

linear algebra, mathematical methods, heuristic strategies, teacher education,
Double discontinuity.

INTRODUCTION
In Germany, the first year at university is often associated with great problems for
mathematics teacher students. The dropout rates in the study entrance phase are
extremely high in mathematics compared to other study programs. (Dieter et al.,
2008). Klein’s “double discontinuity” is still an issue in the mathematics teacher
education at German universities (Hefendehl-Hebeker, 2013). On the one hand, the
double discontinuity can be understood as a discontinuity of content, but Klein also
assigns a didactic dimension to it (Allmendiger, 2016). While mathematics teaching
at school can be described as intuitive and problem-based, mathematics at the
university is characterized by a deductive structure (Klein, 1905). The ideas and
strategies behind the findings of mathematics are rarely presented at the university.
The mathematics teacher students take part in the same mathematics lectures as
regular students who study mathematics as a major subject. Therefore, bridging the
double discontinuity is only possible by offering additional learning opportunities for
teacher students. A few years ago, we implemented additional seminars for teacher
students where we showed connections between school mathematics and university
mathematics. Feedback of such seminars showed that students did not consider our
approach as helpful for bridging the discontinuity. They said that the regular lectures
were challenging and time consuming enough so they did not want to think about
additional issues in our seminars. For this reason, we have developed an approach for
additional learning opportunities for teacher students, which is characterized by
page 1 of 10

547 [Link]:indrum2018:171181
reflections on intuitions and strategies that are fundamental to the regular lecture
contents. Our approach is intended to bridge the didactic discontinuity. On the one
hand, we want to support teacher students in the transition from school to university,
and on the other hand, knowledge about strategies and intuitions is important in the
future teaching activity at school. In the following report, we describe first
experiences we have made with our seminars.

THEORETICAL FRAMEWORK
Subjects and methods in mathematics
In the philosophy of science, it is generally accepted that each science has its specific
subject and its specific methodi. As mathematical subjects in university mathematics,
we deem the mathematical definitions, theorems and their proofs including
mathematical algorithms. The mathematical methods describe how the subjects of
mathematics are created or investigated by a mathematicianii. Which methods does a
mathematician use, when formulating a new definition or a theorem? Furthermore:
how does the mathematician come to the ideas when developing a proof? Sometimes
the steps of a proof are easy in the sense that a beginner could do them but the
decision to do these steps needs a lot of experience. In fact, producing a proof that
does not already exist is an act of problem solving according to the definition of
Schoenfeld (1985, p.74) or the idea Pólya (1973, p. 3) describes by mentioning
“problems to proof”. This point of view corresponds also with the definition of the
concept “problem” the German cognitive psychologist Dörner gave in 1976 (p. 10 iii,
own translation):
What a problem is, is easy to define: an individual faces a problem if he or she is in an
inner state that is unwanted by any reason, but the individual has not the means in the
moment to transform the unwanted situation into the desired one. Three aspects
characterize a problem:
1. An unwanted start situation sα.
2. A desired final situation sω.
3. A barrier that prevents the transformation from sα to sω at the moment.
“We distinguish problems from tasks. Tasks are intellectual challenges for which the
methodsiv are known. […]. Tasks only need reproductive thinking while in problem
solving you have to produce something new.” (1976, p. 10, own translation)
Dörner (1976, S. 10, own translation) emphasised: “Whether a question is a problem
or a task depends on the experience of the individual.” If a student just learned
induction “Prove the Bernoulli Inequality by induction!” as the first exercise might
still be a problem but having done a series of similar proofs, it should be a task.

page 2 of 10

548 [Link]:indrum2018:171181
Based on this definition of a problem Dörner (1976) defines heuristic strategies as
„the structure that organises and controls a problem-solving process” (Dörner, 1976,
p. 273, own translation), which covers a very wide range of methods and is not
bound to mathematics. Following this concept, heuristic strategies cover much of
what is meant by the concept of “method” in the philosophy of science once applied
to mathematics. Pólya (1973) already underlined this as the subtitle of “How to solve
it” is “A new aspect of mathematical methods”.
The following list of heuristic strategies was formulated based on the work of Pólya
and others relying on the broad definition by Dörner. These strategies are arranged in
groups for a better overview:
 organise your material / understand the problem: change the representation of the
situation if useful, trial and error, use simulations with or without computers,
discretize situations,
 use the working memory effectively: combine complex items to supersigns,
which represent the concept of ‘chunks’, use symmetry, break down your problem
into sub-problems,
 think big: do not think inside dispensable borders, generalise the situation,
 use what you know: use analogies from other problems, trace back new problems
to familiar ones, combine particular cases to solve the general case, use
algorithms where possible,
 functional aspects: analyse special cases or extreme cases, in order to optimise
you have to vary the input quantity, discretize the situation,
 organise the work: work backwards and forwards, keep your approach – change
your approach – both at the right moment.
Examples and detailed descriptions for each strategy are shown in Stender (2017).
These strategies do not cover all the methods in the work of mathematicians but a
broad range due to our literature analysis. Many of these strategies are also
applicable beyond the field of mathematicsv and this way of importance as general
problem solving strategies. We would call a list of heuristic strategies that are used
only in mathematics as they rely on a formal language “proof strategies”. This list
contains for example mathematical induction, proof by contradiction, proof by
exhaustion, the Invariance Principle (Engel, 1998, p. 1), Cantor's diagonal argument,
the box principle (Engel, 1998, p. 59) but also very simple ideas like “use an
adequate substitution” or “add zero in an appropriate way.” We do not claim that any
of these lists is complete. A third aspect is the use of mathematical language like
reading a mathematical text appropriately (for example see Hodds, 2014). We
visualised these aspects in figure 1.

page 3 of 10

549 [Link]:indrum2018:171181
Fig. 1: Mathematical Content and Methods
The importance of the methods in mathematics
The discussion on problem solving since Pólya (1973) shows the importance of
heuristic strategies in mathematics and thus (if one follows the argument above) the
importance of mathematical methods in the process of learning mathematics. But
mathematical methods are not emphasized often in (German) lectures in the
mathematics departments. Therefore, we formulate an additional argument here.
We look at a special case of mathematics competence. Imagine a person that knows
many mathematical definitions, theorems and even proofs by heart, but has hardly
any mathematical methods available. This person would not be able to solve any
mathematical problem.
The other extreme would be a person with little declarative knowledge of
mathematical facts but competent in applying mathematical methods. He or she just
has forgotten most of the mathematical content he or she learned. This person would
be able to understand new mathematical content very fast and could fill gaps in the
mathematical knowledge quickly if necessary.
This thought experiment shows that a lack of mathematical methods is a much more
severe problem while doing mathematics than a lack of knowledge about
mathematical content.
A mathematician often uses mathematical methods unconsciously. This is
appropriate, as this is faster in the most cases than channelling everything through
the restricted working space (see Miller, 1956) in the prefrontal cortex. For a future
teacher it is much more important to be aware of his own methods as he or she shall
page 4 of 10

550 [Link]:indrum2018:171181
later explain his actions to children. An explicit reference to his own methods is
essential in these explanations.
Furthermore, the methods mentioned are the same in university mathematics and in
school mathematics while the content shows little similarities as Klein (1945/1908)
pointed out. This way the methods are an important link between school mathematics
and university mathematics and give studying abstract mathematics a meaning for
students who want to become a teacher.
These considerations lead us to the following goal for our research: we want to
implement and evaluate teaching and learning of heuristic strategies in the
mathematics teacher education in a close connection to the mathematics content
trying to show that learning abstract mathematics means learning mathematical
methods which is crucial for teaching mathematics in school.

IMPLEMENTED INTERVENTIONS
The approach we have developed for teacher students includes different additional
learning opportunities:
 We offer special tutorials for mathematics teacher students connected to the
Linear Algebra and mathematical Analysis lectures.
 Within the scope of Linear Algebra and Analysis lectures students must do
weekly exercises. We have implemented a weekly workshop where students get
support in creating and formulating proofs.
 All students may choose a preparation lecture before beginning their studies. Part
of this lecture is a tutorial. We offer special tutorials for mathematics teacher
students too.
These learning opportunities have in common that they do not intervene in the
regular lectures and are not obligatory. This way we are free to realise our teaching
approach, but this also leads to smaller groups so a standard pre-test-post-test design
to evaluate the tutorials is not appropriate. We offered the tutorial over four
semesters with 20 participants. This number was quite stable while participants
changed: some gave up studying mathematics, some new students joined during the
semesters when they heard about the tutorial. Due to the individual timetable of
students, some of them couldn’t participate over all four semesters. The tutorial was
open to students with a major in math too (parallel traditional tutorials exists that
only focus on mathematical content) so overall five teacher students participated
over the whole time.
As the tutorials fulfil different functions in the learning process, we offer the
following aspects for the students:

page 5 of 10

551 [Link]:indrum2018:171181
Among other aspects as answering questions related to the lectures, we point out the
appearance of heuristic strategies in the lecture content. Besides students solve small
exercises so we can exemplify the application of heuristic strategies. We discuss the
application of heuristic strategies in school by means of examples. Furthermore, we
give a structured overview over the lecture content several times per semester to help
students to network their knowledge. The overview is closely connected to some
heuristic strategies as using analogies and generalise the situation.
An example for the use of heuristic strategies is the use of equivalence classes in
linear algebra for example using quotient rings. Quotient rings were presented in the
lecture using a formal definition. The lecturer gave ℤ /mℤ as an example but this
structure was not familiar to the students. In the tutorial we showed the following
ideas to the students:
 Using a special example, we presented the ring ℤ /4ℤ .
 For this structure, we gave several different representations – a number line
where the numbers of the same equivalence classes were coloured in the same
way, the four classes as sets, a representation on a circle and a table that
showed the calculations in this structure.
 The importance of the concept of equivalence classes was emphasised: a
number of objects is chunked together to one new object according to Miller
(1965) so an element of a quotient ring is what we call a supersign according
to Kießwetter (1983).
 We pointed out, that similar structures occur in school when dealing with
rational numbers, which are classes of pairs of natural numbers (supersigns)
and where several representations are necessary while working with fractions.
Even if fractions are not implemented in this abstract description in school,
students must deal with this mathematical structure.
This way similarities in school mathematics and university mathematics are brought
into the focus of the students and showed that students in school dealing with
fractions might experience the same difficulties as students in university dealing with
quotient rings.
A second example from the tutorial is a short exercise we presented to the students:
In the tutorial matrices were used in vector spaces with bases. In the tutorial, we
showed how the matrix describing a rotation (ℝ2→ℝ2, angle α) could be build. Then
we asked the students to develop the matrix for a reflection through an axis that has
the angel α with the x-axis. As this group was already used to heuristic strategies, we
gave the following help:
 Change the representation (figure).
 Look for special cases (basis).
page 6 of 10

552 [Link]:indrum2018:171181
  Trace back new problems to familiar ones (use the knowledge from the
rotation).
 Break down your problem into sub-problems (work on one element of the
basis at a time).
 Use symmetry (to work on sin(2α-180°).
This example could also be situated in high school, so it breaks down university
mathematics to school mathematics and shows in addition how heuristic strategies
can be used supporting students work.
In the weekly workshops, students do exercises on their own. Mostly students have
to prove small theorems related to the lecture content. They work in groups of three
or four and a tutor supports them in developing and formulating proofs. In case
students are not able to overcome barriers in proving processes by themselves, the
tutor promotes the application of an appropriate heuristic strategy. Following a
proving process, students reflect on the strategies they have applied in the process.

EVALUATION RESULTS
As said before the number of students participating in the additional learning
opportunities is too small for quantitative designs. We are in the process of design-
based research as our interventions are an early approach for the use of heuristic
strategies in mathematics education. This means results of efficiency do not make
sense in this phase of the research.
We have conducted interviews with five participants of the tutorials and the
workshops at the end of each semester. We have analysed these interviews using
qualitative content analysis according to Mayring (2010). In the interviews, the
students were asked what they remembered from the tutorial and what they think was
helpful (not discriminating helpful for studying mathematics or helpful for the later
profession). If whole aspects were not mentioned the interviewer stimulated an
answer using a key word. The two kinds of answers (with or without stimulation)
were separated in the analysis.
Over the four semesters of this process, a slightly increasing awareness of the
heuristic strategies in mathematics can be observed in the responses of the students.
Over the time, the students gave more examples of the use of heuristic strategies
during the interviews. The students gave a positive feedback according to the
heuristic strategies as a link between school and university. They said that it
delivered a meaning to the fact that the content of the mathematical studies had
hardly any connect to school mathematics. One student mentioned that he used
heuristic strategies to explain mathematics when he was working in school as a
substitute teacher. The students regard the workshops as very helpful for their
learning progress. They emphasize the importance of developing proofs by their own
page 7 of 10

553 [Link]:indrum2018:171181
to understand the lecture content. Therefore, they consider it beneficial that the tutor
gave only a few strategic hints in case of barriers in the proving process rather than
presenting solutions. Especially the heuristic strategy of changing the representation
helps students to develop new ideas.
We think that it can be expected that the heuristic strategies integrated in the lecture
as metacognition, added to the mathematical content would be far more effective for
learning mathematical methods.

CONCLUSIONS AND LOOKING AHEAD

Overall, we can observe that there are effects of the metacognitive input according
the heuristic strategies as mathematical methods, but the learning process is very
slow. It needs many different examples for each strategy to be able to realise the
importance of these strategies and even more to identify them in your own
mathematical work or use them explicitly to solve problems. Pólya (1973, p. 208)
quotes a fictional traditional professor “A method is a device which you use twice.”
Obviously “twice” is far not enough in the learning process – and not enough to
make a device so important that it should be taught.
The teaching concept to support students by stressing the mathematical methods
particularly by stressing the heuristic strategies seems to have a positive impact but is
not realised easily. This corresponds with the difficulty of teaching problem solving
and heuristic strategies (Schoenfeld1985, p. 95). The examples we gave to the
students during our tutorials and workshops had an impact on realising heuristic
strategies in mathematics on the long run. However, we do not know if they were not
enough to let the student’s use these strategies in their mathematical work explicitly.
We redesign the tutorial if necessary and expand the number of learning situations
where we implement metacognition on the use heuristic strategies. A long-time goal
is to include metacognition in the lectures itself but that means to convince
mathematics lecturer to change their lectures.
Our next project is developing a tutor training for master students tutoring beginners.
In the tutor training, we will use former homework tasks and analyse the solutions
according to the use of heuristic strategies. This will be presented to the students in
few examples and then done on further tasks by the students working in small
groups. In the next step, we will demonstrate how these strategies can be used
explaining mathematics and supporting students to overcome barriers during
problem solving. The students shall use their own results from the first step to
practise this.

REFERENCES
Allmendiger, H. (2016). Die Didaktik in Felix Kleins „Elementarmathematik vom
höheren Standpunkt aus“. Journal für Mathematikdidaktik, 37 (1), 209-238.
page 8 of 10

554 [Link]:indrum2018:171181
Dieter, M., Brugger, P., Schnelle, D., Törner, G. (2008). Zahlen rund um das
Mathematikstudium – Teil 3. Mitteilungen der Deutschen Mathematiker
Vereinigung, 16(3), 176-182.
Dörner, D. (1976). Problemlösen als Informationsverarbeitung. Kohlhammer,
Stuttgart.
Engel, A. (1998). Problem Solving Strategies. Springer, New York.
Hefendehl-Hebeker, L. (2013). Doppelte Diskontinuität oder die Chance der
Brückenschläge. In C. Ableitinger, J. Kramer & S. Prediger (Hrsg.), Zur doppelten
Diskontinuität in der Gymnasiallehrerausbildung. Ansätze zur Verknüpfung der
fachinhaltlichen Ausbildung mit schulischen Vorerfahrungen und Erfordernissen
(S. 1-16). Springer, Wiesbaden.
Hodds, M., Alcock, L., Inglis, M. (2014). Self-Explanation Training Improves Proof
Comprehension. Journal for Research in Mathematics Education, 45(1), 62-101.
doi:10.5951/jresematheduc.45.1.0062
Miller, G. A. (1956). The Magical Number Seven, Plus or Minus Two: Some Limits
on our Capacity for Processing Information. Psychological Review, 63, 81–97
Kießwetter, K. (1983). Modellierung von Problemlöseprozessen. Der
Mathematikunterricht, 29(3), 71-101.
Klein, F. (1908). Elementarmathematik vom höheren Standpunkt. Dover
Publications, New York.
Klein, F. (1945). Elementary mathematics from an advanced standpoint. Springer,
Berlin.
Mayring, P. (2010). Qualitative Inhaltsanalyse: Grundlagen und Techniken. Beltz
Pädagogik.: Beltz, Weinheim.
Miller, G. A. (1956). The Magical Number Seven, Plus or Minus Two: Some Limits
on our Capacity for Processing Information. Psychological Review, 63, 81-97.
Pólya, G. (1973). How to solve it. A new aspect of mathematical methods, 2nd edition.
Princeton University Press.
Stender, P. (2017). The use of heuristic strategies in modelling activities. ZDM
Mathematics Education. [Link]
Schoenfeld, A. H. (1985). Mathematical Problem Solving. Academic Press, Orlando.

i
The concept of “method” is used in the philosophy of science in a much broader way than used by
e.g. Schoenfeld (1985) when he mentions “This way a method becomes a strategy”. In this paper
method is mostly used in this broader meaning of the philosophy of science.
page 9 of 10

555 [Link]:indrum2018:171181
ii
There is a broad discussion whether mathematics is produced by humans or discovered and for
both standpoints and for the idea, that mathematics is a mixture of both there are good arguments.
For this paper it makes no difference if one changes the verb “producing” by “discovering”.
iii
The original is in German language, this is an own translation.
iv
Dörner uses the word “method” not in the same way “scientific method” is used in this paper. The
word method has more the meaning of an algorithm here.
v
Pólya (1973) gave examples from mathematics for the heuristic strategies he mentioned but also
examples from outside mathematics for most of the heuristic strategies shown.

page 10 of 10

556 [Link]:indrum2018:171181
INDRUM 2018 was organised by the University of Agder (Department of Mathematical Sciences),
with the financial support of MatRIC, Centre for Research, Innovation and Coordination of
Mathematics Teaching.

INDRUM 2018 was an ERME Topic Conference.

ERME Topic Conferences (ETC) are conferences organised on a specific research theme or themes
related to the work of ERME as presented in associated working groups at CERME conferences.
Their aim is to extend the work of the group or groups in specific directions with clear value to the
mathematics education research community.
[Link]

List of participants: ABDALLAH Elise (France), ABLEITINGER Christoph (Austria), ALZUBAIDI

Ibrahim (United Kingdom), ANGER Astrid (Austria), BAHADUR POUDEL Amrit (Norway),
BARBERO Marta (Italy), BARQUERO Berta (Spain), BIEHLER Rolf (Germany), BJØRKESTØL
Kirsten (Norway), BOSCH Marianna (Spain), BRANCHETTI Laura (Italy), BRANDES Hadas
(Canada), BREEN Sinead (Ireland), BREKKE Morten (Norway), BROLEY Laura (Canada),
BROWN Stacy (United States), CHANG Beng (China), CHELLOUGUI Faïza (Tunisia), CHORLAY
Renaud (France), COUSSEMENT Els (Belgium), DALLAS Markos (Norway), DELGADO
REBOLLEDO Rosa (Chile), DREYFUS Tommy (Israel), DURAND-GUERRIER Viviane (France),
EGELANDSAA Lillian (Norway), EICHLER Andreas (Germany), FEUDEL Frank (Germany),
FLEISCHMANN Yael (Germany), FREDRIKSEN Helge (Norway), FUGLESTAD Anne Berit
(Norway), GASPAR MARTINS Sandra (Portugal), GEISLER Sebastian (Germany), GHEDAMSI
Imène (Tunisia), GIBARA Ryan (Canada), GLASSMEYER David (United States), GONZÁLEZ-
MARTÍN Alejandro (Canada), GOODCHILD Simon (Norway), GRADWOHL Julia (Germany),
GUEUDET Ghislaine (France), HADJERROUIT Said (Norway), HAMANN Tanja (Germany),
HANKE Erik (Germany), HAUSBERGER Thomas (France), HECK André (The Netherlands),
HILLESUND Eivind (Norway), HOCHMUTH Reinhard (Germany), HOGSTAD Ninni Marie
(Norway), HOWARD Emma (Ireland), HUGET Judith (Germany), HÄSÄ Jokke (Finland),
JAWORSKI Barbara (United Kingdom), KANWAL Shaista (Norway), KAWAZOE Mitsuru (Japan),
KIRSTEN Katharina (Germany), KNUTSON Inger Johanne (Norway), KOCK Zeger-Jan (the
Netherlands), KONTOROVICH Igor' (New Zealand), KOUROUNIOTIS Christos (Greece),
KUKLINSKI Christiane (Germany), LAHDENPERÄ Juulia (Finland), LALAUDE-LABAYLE
Marc (France), LANDGÂRDS Ida (Norway), LANKEIT Elisa (Germany), LANSZWEERT Ilse
(Belgium), LARSON Niclas (Norway), LAWSON Duncan (United Kingdom), LECORRE Thomas
(France), LOCKWOOD Elise (United States), MARTINEZ-PLANELL Rafael (Puerto Rico),
MATHIEU-SOUCY Sarah (Canada), MAYENDE Godfrey (Norway), MEEHAN Maria (Ireland),
MEYER Antoine (France), NARDI Elena (United Kingdom), NEUHAUS Silke (Germany),
NICOLÁS Pedro (Spain), NIEMINEN Juuso Henrik (Finland), NOSSUM Rolf Thomas (Norway),

557
O’SHEA Ann (Ireland), OUVRIER-BUFFET Cécile (France), PAPADAKI Evi (Greece),
PETTERSSON Kerstin (Sweden), RACH Stefanie (Germany), RASMUSSEN Chris (United States),
RASMUSSEN Elisabeth (Norway), REED Zackery (Unit ed States), ROGOVCHENKO Svitlana
(Norway), ROGOVCHENKO Yuriy (Norway), ROSICH Núr ia (Spain), RÄMÖ Johanna (Finland),
RØNNING Frode (Norway), SCHMIDT Karsten (Denmark), SCHMIDT-THIEME Barbara
(Germany), SCHÜLER-MEYER Alexander (Germany), SCHÄFER Ingolf (Germany), STENDER
Peter (Germany), STRØMSKAG Heidi (Norway), STUHLMANN Ann Sophie (Germany),
SYRDALEN Erlend (Norway), TABCHI Theresia (France), TETAJ Floridona (Norway), THOMA
Athina (United Kingdom), TRIGUEROS María (Mexico), UF ER Stefan (Germany),
VANDEBROUCK Fabrice (France), VIIRMAN Olov (Sweden), VOS Pauli ne (Norway), WAWRO
Megan (United States), WINSLØW Carl (Denmark), WOLD Margrethe (Norway), ZANDIEH
Michelle (United States).

558
Author Index

Abdallah Elise, 194–203 Geisler Sebastian, 507–516

Alzubaidi Ibrahim, 452, 453 Ghedamsi Imène, 34–43
Anastasakis Marinos, 316–325 Gibel Patrick, 234–243
Apkarian Naneh, 286–295 Gjesteland Thomas, 175–184
Glassmeyer David, 454, 455
Barbero Marta, 204–213 Gómez-Chacón Inés M., 204–213
Barquero Berta, 85–94, 155–164 González-Martín Alejandro S., 115–124
Biehler Rolf, 12, 224–233, 467–476, 527–536 Gradwohl Julia, 125–134
Björklund Johan, 464, 465 Gueudet Ghislaine, 135–144
Bloch Isabelle, 234–243
Bosch Marianna, 373–382 Hamann Tanja, 456, 457
Boustedt Jonas, 464, 465 Hanke Erik, 44–63
Branchetti Laura, 14–23 Hao Jie, 454, 455
Brandes Hadas, 477–486 Hardy Nadia, 403–411, 477–496
Breen Sinéad, 383–392 Hausberger Thomas, 497–506
Broley Laura, 487–496 Hernandes-Gomes Gisela, 115–124
Brown Stacy, 214–223 Hewitt Dave, 316–325
Hillesund Eivind, 185, 186
Chorlay Renaud, 24–33 Hochmuth Reinhard, 467–476, 497–506,
Conniffe Dan, 316–325 517–536
Copeland Cillian, 306–315 Howard Emma, 306–315
Corriveau Claudia, 403–411 Huget Judith, 458, 459
Cortes Kimberly, 454, 455 Häsä Jokke, 366, 367
Croft Tony, 10
Cuadros Jordi, 373–382 Jaworski Barbara, 12, 316–325
Jones Keith, 452, 453
Dar Ayesha, 316–325
Delgado Rebolledo Rosa, 393–402 Kanwal Shaista, 145–154
Derouet Charlotte, 497–506 Kirsten Katharina, 326–335
Dreyfus Tommy, 286–295 Kock Zeger-Jan, 336–345
Durand-Guerrier Viviane, 14–23 Kosiol Timo, 356–365
Kourouniotis Christos, 265–274
Eichler Andreas, 125–134 Kuklinski Christiane, 467–476, 527–536

Feeney Matt, 316–325 Lahdenperä Juulia, 346–355

Feudel Frank, 95–104 Lalaude-Labayle Marc, 234–243
Fleischmann Yael, 224–233 Landgärds Ida, 187, 188
Florensa Ignasi, 373–382 Lankeit Elisa, 467–476, 527–536
Fredriksen Helge, 296–305 Larson Niclas, 368, 369
Lawson Duncan, 10
Gascón Josep, 373–382, 412–421 Lecorre Thomas, 34–43
Gaspar Martins Sandra, 105–114 Leis Elena, 467–476, 527–536

559
Lévi Laurent, 234–243 Rämö Johanna, 366, 367
Liebendörfer Michael, 467–476, 527–536 Rønning Frode, 12
Lockwood Elise, 244–254
Salem Béchir Sghaier, 74–83
Martı́nez-Planell Rafael, 64–73 Schaper Niclas, 467–476, 527–536
Mathieu-Soucy Sarah, 403–411 Schmidt Karsten, 165–174
Meehan Maria, 306–315, 383–392 Schmidt-Thieme Barbara, 456, 457, 460,
Meyer Antoine, 255–264 461
Modeste Simon, 255–264 Schüler-Meyer Alexander, 537–546
Monreal Galán Nacho, 155–164 Schürmann Mirko, 467–476, 527–536
Schäfer Ingolf, 54–63
Neuhaus Silke, 467–476, 527–536
Shrish-Thapa Damma, 316–325
Nicolás Pedro, 412–421
Stender Peter, 547–556
Nieminen Juuso Henrik, 366, 367
Strømskag Heidi, 422–431
O’Shea Ann, 383–392 Stuhlmann Ann Sophie, 547–556
Ouvrier-Buffet Cécile, 255–264 Syrdalen Erlend, 462, 463

Papadaki Evi, 265–274 Tabach Michal, 286–295

Parnell Andrew, 306–315 Tabchi Theresia, 432–441
Pauna Matti, 366, 367 Tetaj Floridona, 191, 192
Pepin Birgit, 336–345 Treffert-Thomas Stephanie, 189, 190,
Pesonen Henri, 366, 367 316–325
Pettersson Irina, 464, 465 Trigueros Marı́a, 64–73
Pettersson Kerstin, 368, 369
Planchon Gaëtan, 497–506 Ufer Stefan, 356–365

Quéré Pierre-Vincent, 135–144 Vandebrouck Fabrice, 74–83

Viirman Olov, 442–451, 464, 465
Rach Stefanie, 356–365 Vlaseros Nikolaos, 316–325
Rasmussen Chris, 286–295 Vos Pauline, 175–184
Ray Herman, 454, 455
Reed Zackery, 244–254 Watson Kevin, 275–284
Rogovchenko Svitlana, 189, 190 Wawro Megan, 12, 275–284
Rogovchenko Yuriy, 189, 190 Winsløw Carl, 12, 165–174
Rolka Katrin, 507–516 Wold Margrethe, 175–184, 370, 371
Rowland Tim, 383–392
Ruiz-Munzón Noemı́, 155–164 Zakaryan Diana, 393–402
Rushton Gregory, 454, 455 Zandieh Michelle, 275–284

560