STA4820/014/0/2024
Tutorial Letter 014/0/2024
Statistical Techniques for Science and
Engineering
STA4820
Year module
Department of Statistics
Assignment 4 questions.
STA4820/014/0/2024
ASSIGNMENT 04
Unique Nr.: 776500
Fixed closing date: 01 August 2024
Study chapters 16 to 17 of the text-book Statistics for Management and Economics by Keller
G, Gaciu N. (2020) before attempting these questions. Use Excel to complete this
assignment.
QUESTION 1 (10)
A chemist is calibrating a spectrophotometer that will be used to measure the concentration of carbon
monoxide (𝐶𝐶𝐶𝐶) in atmospheric samples. To check if the machine is in calibration, samples of known
concentration are measured. The true concentrations (𝑥𝑥) and the measured concentrations (𝑦𝑦) are
given in the following table. Because of random error, repeated measurements on the same sample
will vary. The machine is in calibration if its mean response is equal to the true concentration.
Table 1
True concentration Measured concentration
(ppm) (ppm)
0 1
10 11
20 21
30 28
40 37
50 48
60 56
70 68
80 75
90 86
100 96
To check if the machine is in calibration, the linear model 𝑦𝑦 = 𝛽𝛽0 + 𝛽𝛽1 𝑥𝑥 + 𝜀𝜀 is fitted to the data in
Table 1 above. Ideally, the value of 𝛽𝛽0 should be 0 and the value of 𝛽𝛽1 should be 1.
a. �0 and 𝛽𝛽
What are the least-squares estimates 𝛽𝛽 �1 ? (2)
b. Can you reject the null hypothesis 𝐻𝐻0 : 𝛽𝛽0 = 0 at the 0.05 level of significance? (2)
c. Can you reject the null hypothesis 𝐻𝐻0 : 𝛽𝛽1 = 1 at the 0.05 level of significance? (3)
2
STA4820/014/0/2024
d. Do the data provide sufficient evidence to conclude that the machine is out of calibration at the 0.05
level of significance? (3)
QUESTION 2 (14)
The article “The Selection of Yeast Strains to Produce Premium Quality South African Brandy Base
Products” (C. Steger and M. Lambrechts, Journal of Industrial Microbiology and Biotechnology,
2000:431–440) contains detailed information on the volatile compound composition of base wines
made from each of 16 selected yeast strains. Below are the results from the linear regression analysis
of the concentrations of total esters (𝑥𝑥) and total volatile acids (𝑦𝑦)(in mg/L).
Table 2
SUMMARYOUTPUT
Regression Statistics
Multiple R 0.3235
R Square 0.1046
Adjusted R Square 0.0407
Standard Error 96.7713
Observations 16
ANOVA
df SS MS F Significance F
Regression 1 15320.2728 15320.2728 1.6359 0.2217
Residual 14 131105.5746 9364.6839
Total 15 146425.8474
Coefficients Standard Error t Stat P-value Lower 95% Upper 95%
Intercept 352.3247 52.7742 6.6761 1.0506E-05 239.1352 465.5141
Esters -0.2495 0.1951 -1.2791 0.2217 -0.6679 0.1689
Figure 1 Figure 2
Esters Residual Plot Esters Line Fit Plot
200 500
Acids
Residuals
0 Acids
0 200 400 600 800 0
Predicted Acids
0 500 1000
-200
Esters Esters
a. From the above Figures 1 and 2, is there an outlier? (1)
3
STA4820/014/0/2024
b. Write down the coefficients of the least-squares line for predicting acid level (𝑦𝑦) from ester level
(𝑥𝑥), along with their estimated standard errors. (4)
c. Write down the P-value of the test of the null hypothesis 𝐻𝐻0 : 𝛽𝛽1 = 0. (1)
Table 3: Outlier deleted (for answering parts (d) to (f))
SUMMARYOUTPUT
Regression Statistics
Multiple R 0.1880
R Square 0.0353
Adjusted R Square -0.0389
Standard Error 100.3896
Observations 15
ANOVA
df SS MS F Significance F
Regression 1 4798.8365 4798.8365 0.4762 0.5023
Residual 13 131014.8078 10078.0621
Total 14 135813.6444
Coefficients Standard Error t Stat P-value Lower 95% Upper 95%
Intercept 359.2321 91.0763 3.9443 0.0017 162.4737 555.9905
Esters -0.2834 0.4108 -0.6905 0.5023 -1.1708 0.6040
d. Using Table 3, rewrite down the coefficients of the least-squares line, along with their estimated
standard deviations. (4)
e. Use Table 3 to rewrite down the P-value of the test of the null hypothesis 𝐻𝐻0 : 𝛽𝛽1 = 0 for the data
with the outlier deleted. (1)
f. Does a linear model appear to be useful for predicting acid concentration from ester
concentration? Explain. (3)
4
STA4820/014/0/2024
QUESTION 3 (24)
The Chief executive officer (CEO) of a company that manufactures drywall wants to analyse the
variables that affect demand for his product. Drywall is used to construct walls in houses and offices.
Consequently, the CEO decides to develop a regression model in which the dependent variable is
monthly sales of drywall (in hundreds of 4 × 8 sheets) and the independent variables are as follows:
Five-year bond rates (Bond in percentage points), Vacancy rate in apartments (A Vacancy in
percentage points), Vacancy rate in office buildings (O Vacancy in percentage points). To fit a multiple
regression model, he used monthly observations from the past 2 years (in Table 4).
Table 4
Drywall Permits Bond A Vacancy O Vacancy
328 49 8.35 2.98 13.43
376 79 8.08 5.6 14.51
373 79 7.9 2.25 14.24
144 50 7.69 4.26 14.3
194 37 7 2.6 11.64
220 53 7.32 2.97 10.61
126 22 8.4 5.35 18.45
301 69 8.28 3.13 18.52
54 21 8 5.6 10.29
252 46 8.95 4.81 11.91
381 79 8.21 5.88 17.75
173 30 7.24 2.98 18.16
152 38 7.35 5.69 17.14
351 73 7.27 4.86 16.11
233 55 7.08 5.68 18.54
35 12 7.76 4.46 19.46
290 62 8.21 2.23 19.26
5 12 7.76 5 17.28
335 60 7.2 2.42 15.15
280 49 7.57 3.25 19.94
101 14 8.44 3.61 15.47
297 66 8.43 2.13 12.75
309 62 8.14 4.35 12.24
233 40 8.81 2.31 18.65
a. Find the regression equation. (1)
b. What is the estimate sigma (square root of the error variance)? Can you use this statistic to
assess the model’s goodness of fit? If so, how? (2)
c. What is the coefficient of determination, and what does it tell you about the regression model’s
goodness of fit? (3)
d. Test the goodness of fit of the model? (3)
5
STA4820/014/0/2024
e. Conduct tests of hypotheses to determine whether or not each of the independent variables is
linearly related to drywall demand in this model. (12)
f. Predict next month’s drywall sales with 95% confidence if the number of building permits is 50,
the 5-year bond rate is 9.0%, the vacancy rate in apartments is 3.6% and the vacancy rate in
office buildings is 14.3%. (3)
QUESTION 4 (9)
A developer who specializes in summer Cottage properties is considering purchasing a large
tract of land adjoining a lake. The current owner of the tract has already subdivided the land into
separate building lots and has prepared the lots by removing some of the trees. The developer wants
to forecast the value of each plot. From previous experience, she knows that the most important
factors affecting the price (𝑦𝑦) of a plot are (𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠/ 𝑥𝑥1 ), number of mature trees (𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇/𝑥𝑥2 ), and
distance to the lake (𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷/ 𝑥𝑥3 ). From a nearby area, she gathers the relevant data for 60 recently
sold plots.
Table 5
SUMMARYOUTPUT
Regression Statistics
Multiple R 0.4924
R Square 0.2425
Adjusted R Square 0.2019
Standard Error 40.2435
Observations 60
ANOVA
df SS MS F Significance F
Regression 3 29029.7163 9676.5721 5.9749 0.0013
Residual 56 90694.3331 1619.5417
Total 59 119724.0493
Coefficients Standard Error t Stat P-value Lower 95% Upper 95%
Intercept 51.3912 23.5165 2.1853 0.0331 4.2820 98.5004
Plot size 0.6999 0.5589 1.2524 0.2156 -0.4196 1.8194
Trees 0.6788 0.2293 2.9603 0.0045 0.2195 1.1382
Distance -0.3784 0.1952 -1.9379 0.0577 -0.7695 0.0128
a. Find the regression equation (from the regression of 𝑦𝑦 on the independent variables). (1)
b. What is the estimate of sigma (square root of the error variance)? Interpret its value. (2)
c. What is the coefficient of determination? What does this statistic tell you? (3)
6
STA4820/014/0/2024
d. Test the goodness of fit of the model. What does the p-value of the regression model tell you?
(3)
Total Marks: [57]
