College of Professional and Continuing Education (CPCE), an affiliate of PolyU

THE HONG KONG POLYTECHNIC UNIVERSITY

HONG KONG COMMUNITY COLLEGE

Subject Title : Regression Analysis Subject Code : SEHH2313

Session : Semester One 2022/23 Time : 12:30 – 14:30

Date : 11 December 2022 Time Allowed : 2 Hours

Subject : Dr Elegance LAM

Examiner(s)

This question paper has a total of TWENTY-ONE pages (including this covering page).

Instructions to Candidates:

1. There are THREE sections in this paper.

Section A (20%) – Multiple-choice Questions. Answer ALL questions in this section on the
multiple-choice answer sheet provided. Each question carries 2 marks.
Section B (60%) – Compulsory Questions. Answer ALL questions in this section in the
answer script provided. Each question carries 20 marks.
Section C (20%) – Elective Questions. Answer any ONE out of the TWO questions in this
section in the answer script provided. Each question carries 20 marks.
If you answer more than one question, only the first attempted question
will be marked. Indicate in your answer script clearly which question
you are attempting.
2. Show all your workings clearly and neatly. Reasonable steps should be shown.
3. Correct your answers to FOUR decimal places, if necessary.
4. A list of formulae can be found on pages 13 and 14 of this paper.
5. The following statistical tables are attached in this paper:
Student’s t-Distribution (page 15)
F-Distribution (pages 16 and 17)
Chi-square Distribution (page 18)
Critical Values for Durbin-Watson Statistics (pages 19 and 20)
Standard Normal Distribution Table (page 21)
6. Electronic calculators, including programmable calculators, may be used provided that the
calculators are battery powered, silent in operation, with neither printout nor graphic /
word display facilities and do not use dot-matric technology in the main display. Other
electronic devices with graphic / word-display facilities (such as databank watches) are not
permitted.
7. Candidates are NOT allowed to retain this paper.
Authorised Materials:
YES NO
CALCULATOR [✓] [ ]
(All programmes stored should be cleared.)
SPECIFICALLY PERMITTED ITEMS [ ] [✓]

DO NOT TURN OVER THE PAGE UNTIL YOU ARE TOLD TO DO SO

Page 1 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Section B (60%) – Compulsory Questions

Answer ALL questions in this section in the answer script provided. Each question carries
20 marks.

Question B1

Mr. Chan is a president of XX Financial Services. He would like to find out the relationship between
the number of client contacts (𝑿) and the amount of sales (𝒀, HK dollars). A random sample of 12
clients was selected. The data are summarised as follows:

∑ 𝑿 = 𝟐𝟓𝟐, ∑ 𝒀 = 𝟑𝟖𝟒, ∑ 𝑿𝟐 = 𝟓𝟗𝟑𝟐 , ∑ 𝒀𝟐 = 𝟏𝟑𝟏𝟒𝟎 , ∑ 𝑿𝒀 = 𝟖𝟕𝟕𝟓

A simple linear regression model

𝒚𝒊 = 𝜷𝟎 + 𝜷𝟏 𝒙𝒊 + 𝜺𝒊 , 𝒊 = 𝟏, 𝟐, … , 𝟏𝟐

where 𝜺𝒊 is the normal random error with a zero mean and an unknown variance σ2 , is fitted to these
data.

Correct your final answers to 4 decimal places.

(a) Find the equation of the estimated regression line for the given data. (6 marks)

(b) Find an unbiased point estimate for 𝝈𝟐 . (4 marks)

(c) The president wants to carry out a hypothesis test to check whether the above data provide
sufficient evidence to indicate that a positive linear relationship exists between the number of
client contacts and the amount of sales. Test at a 5% significance level. (10 marks)

Page 7 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Question B2

Andy was interested in developing a regression model on health-related quality of life of hepatitis
patients ( 𝒀 ). He identified four independent variables ( 𝑿𝟏 , 𝑿𝟐 , 𝑿𝟑 , 𝑿𝟒 ) . Twelve sets of
observations on the variables were collected and an equation of the form was proposed as follows:

̂𝟎 + 𝜷
̂=𝜷
𝒀 ̂𝟏 𝑿𝟏 + 𝜷
̂𝟐 𝑿𝟐 + 𝜷
̂𝟑 𝑿𝟑 + 𝜷
̂𝟒 𝑿𝟒

The coefficient estimates and error sum of squares of the models built are listed as follows.

Number of
Predictor ̂0
𝛽 ̂1
𝛽 ̂2
𝛽 ̂3
𝛽 ̂4
𝛽 Error Sum of
Variable(s) Squares (SSE)
0 283.33333 –– –– –– –– 6274.66667
1 228.36364 0.90238 –– –– –– 3083.56706
1 132.17949 –– 5.75824 –– –– 4011.67766
1 270.18484 –– –– 0.14227 –– 6271.77376
1 263.13969 –– –– –– 0.19798 6247.54383
2 161.29646 0.69538 3.03529 –– –– 2622.68773
2 10.84794 1.06439 –– 2.24685 –– 2464.94313
2 191.06102 0.91355 –– –– 0.35904 2994.85276
2 102.63873 –– 5.7778 0.31409 –– 3997.60449
2 137.69443 –– 5.78605 –– -0.06122 4009.13649
2 260.42833 –– –– 0.03363 0.19409 6247.39266
3 -14.22208 0.87559 2.45085 1.94643 –– 2175.52042
3 144.47991 0.71185 2.88253 –– 0.19435 2597.86146
3 4.43168 1.06294 –– 2.1686 0.13467 2453.21246
3 106.48208 –– 5.83002 0.37531 -0.10658 3990.43796
4 -15.01777 0.87633 2.43857 1.9367 0.01934 2175.28584

Correct your final answers to 4 decimal places.

(a) Determine whether each of the following models is appropriate in terms of the 𝑪𝒑 statistics.
Explain briefly for each scenario.

(i) 𝒀 = 𝜷𝟎 + 𝜷𝟐 𝑿𝟐 + 𝜷𝟑 𝑿𝟑 + 𝜺 (3 marks)

(ii) 𝒀 = 𝜷𝟎 + 𝜷𝟏 𝑿𝟏 + 𝜷𝟑 𝑿𝟑 + 𝜷𝟒 𝑿𝟒 + 𝜺 (3 marks)

(b) Use the forward selection procedure and partial F-tests at a 5% significance level to select the
‘best’ subset of predictor variables. (14 marks)

Page 8 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Question B3

(a) A manager of manufacturing company was interested in whether there is a difference in the
number of units produced when the employees work on three different shifts including
morning (M), afternoon (A) and night (N). A random sample of 21 workers is selected and
their output recorded on each shift.

Shift Units produced

Morning (M) 33 36 29 30 27 33 36
Afternoon (A) 23 24 25 28 25 27 23
Night (N) 32 30 28 29 26 31 34

Assume that an one-way classification model is constructed.

(i) Do the data provide sufficient evidence to indicate that the three shifts differ
significantly with regard to the units produced? Test at 𝜶 = 𝟎. 𝟎𝟓. Correct the test
value to 4 decimal places. (7 marks)

(ii) Test, at the 5% level of significance, whether the mean number of units produced of
workers on afternoon and night shifts are different. Correct the test value to 4 decimal
places. (7 marks)

(b) A manager of a company wants to study whether or not the household income would affect the
decision to purchase an iPad. A random sample of recent 50 customers who were considering
purchasing an iPad was selected. The collected information was shown below.

Variables Description
Purchase 𝒀 Coded as 1 if the customer purchased an iPad;
Coded as 0 if the customer did not purchase an iPad.
Income 𝑿 Household income of the customer (in thousands of dollars)

Customer, 𝒊 1 2 3 ⋯ 48 49 50
𝑿𝒊 4.2 6.8 4.9 ⋯ 10.5 6.2 6.7
𝒀𝒊 0 1 0 ⋯ 1 0 0

Suppose the logistic regression model is fitted by using the data collected,

𝝅𝒊
𝐥𝐧 ( ) = 𝜷𝟎 + 𝜷𝟏 𝑿𝒊
𝟏 − 𝝅𝒊

where 𝝅𝒊 = 𝑷(𝒀𝒊 = 𝟏).

Page 9 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Question B3 (continued)

Some SAS outputs of the model are shown below.

Correct your final answers to 4 decimal places.

(i) What is the estimated probability that the customer will purchase an iPad if the
household income is $46 (in thousands of dollars)? (3 marks)

(ii) Find the 95% confidence interval for 𝜷𝟏 . (3 marks)

- End of Section B -

Page 10 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Section C (20%) – Elective Questions

Answer any ONE out of the TWO questions in this section on the answer script provided.
Each question carries 20 marks. If you answer more than one question, only the first
attempted question will be marked. Indicate in your answer script clearly which question
you are attempting.

Question C1

Maggie was interested in studying the impact of coffee consumption and time spent exercising on
anxiety of working adults. The weekly coffee consumption 𝑿𝟏 (in cups), weekly time spent
exercising 𝑿𝟐 (in hours) and anxiety score 𝒀 (on a scale of 1 to 100) for 20 working adults were
collected for data analysis. Higher score of anxiety indicates greater anxiety.

Weekly Coffee Weekly Time Spent

Working adults, i Anxiety Score, 𝒀𝒊
Consumption, 𝑿𝟏𝒊 Exercising, 𝑿𝟐𝒊
1 27 220 84
2 25 190 79
⋮ ⋮ ⋮ ⋮
19 2 90 35
20 1 92 33
∑ 𝑋1𝑖 = 232 ∑ 𝑋2𝑖 = 2912 ∑ 𝑌𝑖 = 1083

Consider the following model:

𝛀: 𝒀𝒊 = 𝜷𝟎 + 𝜷𝟏 𝑿𝟏𝒊 + 𝜷𝟐 𝑿𝟐𝒊 + 𝜺𝒊 , 𝒊 = 𝟏, 𝟐, … , 𝟐𝟎

where the 𝜺𝒊 ’s satisfy the usual normality assumptions, is formulated and fitted to the data. It is
given that
𝟏𝟕𝟒𝟏𝟕𝟏𝟕𝟕𝟏 𝟐𝟏𝟑𝟔𝟓𝟖𝟖 −𝟒𝟔𝟑𝟎𝟖𝟐
𝟕𝟐𝟐𝟎𝟗𝟗𝟒𝟖 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟎𝟖𝟑
𝟐𝟏𝟑𝟔𝟓𝟖𝟖 𝟏𝟓𝟏𝟗𝟑𝟒 −𝟐𝟔𝟕𝟕𝟗
(𝑿𝑻 𝑿)−𝟏 = , 𝑿𝑻 𝒀 = ( 𝟏𝟒𝟔𝟕𝟓 ), 𝒀𝑻 𝒀 = 𝟔𝟐𝟗𝟑𝟓 ,
𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕
−𝟒𝟔𝟑𝟎𝟖𝟐 −𝟐𝟔𝟕𝟕𝟗 𝟓𝟑𝟏𝟒 𝟏𝟔𝟖𝟓𝟖𝟔
( 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕 𝟏𝟖𝟎𝟓𝟐𝟒𝟖𝟕)

𝑺𝒚𝒚 = 𝟒𝟐𝟗𝟎. 𝟓𝟓

where X is the design matrix of the model 𝛀 and 𝒀 is a column vector containing observed values
of the dependent variable.

Correct your final answers to 4 decimal places.

(a) Find the least squares estimates for 𝜷𝟎 , 𝜷𝟏 and 𝜷𝟐 . (6 marks)

(b) Calculate the adjusted coefficient of multiple determination 𝑹𝟐𝒂 of the model 𝛀. (4 marks)

(c) Using the result of part (b), test the hypothesis that 𝜷𝟏 = 𝜷𝟐 = 𝟎 at the 5% level of
significance. Correct the test value to 4 decimal places. (10 marks)

Page 11 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Question C2

The following table shows the total number of travellers stayed per day and daily tourism
expenditure (in thousands dollars) in a small city for the past 15 days.

Day, 𝒕 Total number of travellers stayed per day, 𝒙𝒕 Daily tourism expenditure, 𝒚𝒕
1 135 4.8
2 180 4.5
3 200 4.8
4 155 6.1
5 135 4.5
6 72 2.1
7 115 3.6
8 95 3
9 155 5.5
10 95 2.8
11 115 3.7
12 135 4.6
13 87 2.5
14 115 3.3
15 125 4.5

A researcher assumed that the daily tourism expenditure are independent and normally distributed
with a common unknown variance. Correct your final answers to 4 decimal places.

(a) Do the above data provide sufficient evidence to indicate that there is a linear relationship
between total number of travellers stayed per day and daily tourism expenditure at a 5% level
of significance? (7 marks)

(b) Calculate the sum of squares due to pure error (SSPE). (3 marks)

(c) By carrying out a lack of fit test at a 5% level of significance, determine whether or not a
simple linear regression model (relating 𝒚𝒕 to 𝒙𝒕 ) is adequate for fitting the above data.
(10 marks)

- End of Section C -

Page 12 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Appendix: Useful Formulae for Reference

1. Simple Linear Regression

(a) A 100 (1 −  ) % confidence interval for 0

MSE  xi2
ˆ0  t
; n−2 nS xx
2

(b) A 100 (1 −  ) % confidence interval for 1

MSE
ˆ1  t
; n−2 S xx
2

(c) Test statistic for 0

ˆ0 − c
MSE  xi2
nS xx

(d) Test statistic for 1

ˆ1 − c
MSE
S xx

(e) A 100 (1 −  ) % confidence interval for E ( Y0 ) when X = x0

 1 ( x0 − x )2 
ˆ0 + ˆ1 x0  t MSE  + 
2
; n−2
 n S xx 


(f) A 100 (1 −  ) % prediction interval for Y0 when X = x0

 1 ( x0 − x )2 
ˆ0 + ˆ1 x0  t MSE 1 + + 
2
; n−2
 n S xx 


Page 13 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Appendix: Useful Formulae for Reference (continued)

2. Multiple Linear Regression

(a) A 100 (1 −  ) % confidence interval for  j

ˆ j  t MSE  a j +1, j +1
; n− p
2

(b) Test statistic for  j

ˆ j − c
MSE  a j +1, j +1

(c) A 100 (1 −  ) % confidence interval for E ( Y0 ) when x = x0

( )
−1
x0T ˆ  t MSE  x0T X T X x0
; n− p
2

(d) A 100 (1 −  ) % prediction interval for Y0 when x = x0

MSE 1 + x0T X T X ( ) x0 
−1
x0T ˆ  t
; n− p  
2

3. Variable Selection and Model Building

(Mallows’) C p statistic
SSE p
− (n − 2 p)
MSE F

4. Autocorrelation

Durbin-Watson statistic
n

 (e − e )
2
t t −1
t =2
n

e
t =1
2
t

Page 14 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 15 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 16 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 17 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 18 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 19 of 21
College of Professional and Continuing Education (CPCE), an affiliate of PolyU

Page 20 of 21
 College of Professional and Continuing Education (CPCE), an affiliate of PolyU

The Standard Normal Distribution Table

The table entry for z is the area to the left of z.

For example, P(Z  1.00) = 0.8413

z 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.0 .5000 .5040 .5080 .5120 .5160 .5199 .5239 .5279 .5319 .5359
0.1 .5398 .5438 .5478 .5517 .5557 .5596 .5636 .5675 .5714 .5753
0.2 .5793 .5832 .5871 .5910 .5948 .5987 .6026 .6064 .6103 .6141
0.3 .6179 .6217 .6255 .6293 .6331 .6368 .6406 .6443 .6480 .6517
0.4 .6554 .6591 .6628 .6664 .6700 .6736 .6772 .6808 .6844 .6879
0.5 .6915 .6950 .6985 .7019 .7054 .7088 .7123 .7157 .7190 .7224
0.6 .7257 .7291 .7324 .7357 .7389 .7422 .7454 .7486 .7517 .7549
0.7 .7580 .7611 .7642 .7673 .7703 .7734 .7764 .7793 .7823 .7852
0.8 .7881 .7910 .7939 .7967 .7995 .8023 .8051 .8078 .8106 .8133
0.9 .8159 .8186 .8212 .8238 .8264 .8289 .8315 .8340 .8365 .8389
1.0 .8413 .8438 .8461 .8485 .8508 .8531 .8554 .8577 .8599 .8621
1.1 .8643 .8665 .8686 .8708 .8729 .8749 .8770 .8790 .8810 .8830
1.2 .8849 .8869 .8888 .8907 .8925 .8944 .8962 .8980 .8997 .9015
1.3 .9032 .9049 .9066 .9082 .9099 .9115 .9131 .9147 .9162 .9177
1.4 .9192 .9207 .9222 .9236 .9251 .9265 .9279 .9292 .9306 .9319
1.5 .9332 .9345 .9357 .9370 .9382 .9394 .9406 .9418 .9429 .9441
1.6 .9452 .9463 .9474 .9484 .9495 .9505 .9515 .9525 .9535 .9545
1.7 .9554 .9564 .9573 .9582 .9591 .9599 .9608 .9616 .9625 .9633
1.8 .9641 .9649 .9656 .9664 .9671 .9678 .9686 .9693 .9699 .9706
1.9 .9713 .9719 .9726 .9732 .9738 .9744 .9750 .9756 .9761 .9767
2.0 .9772 .9778 .9783 .9788 .9793 .9798 .9803 .9808 .9812 .9817
2.1 .9821 .9826 .9830 .9834 .9838 .9842 .9846 .9850 .9854 .9857
2.2 .9861 .9864 .9868 .9871 .9875 .9878 .9881 .9884 .9887 .9890
2.3 .9893 .9896 .9898 .9901 .9904 .9906 .9909 .9911 .9913 .9916
2.4 .9918 .9920 .9922 .9925 .9927 .9929 .9931 .9932 .9934 .9936
2.5 .9938 .9940 .9941 .9943 .9945 .9946 .9948 .9949 .9951 .9952
2.6 .9953 .9955 .9956 .9957 .9959 .9960 .9961 .9962 .9963 .9964
2.7 .9965 .9966 .9967 .9968 .9969 .9970 .9971 .9972 .9973 .9974
2.8 .9974 .9975 .9976 .9977 .9977 .9978 .9979 .9979 .9980 .9981
2.9 .9981 .9982 .9982 .9983 .9984 .9984 .9985 .9985 .9986 .9986
3.0 .9987 .9987 .9987 .9988 .9988 .9989 .9989 .9989 .9990 .9990
3.1 .9990 .9991 .9991 .9991 .9992 .9992 .9992 .9992 .9993 .9993
3.2 .9993 .9993 .9994 .9994 .9994 .9994 .9994 .9995 .9995 .9995
3.3 .9995 .9995 .9995 .9996 .9996 .9996 .9996 .9996 .9996 .9997
3.4 .9997 .9997 .9997 .9997 .9997 .9997 .9997 .9997 .9997 .9998
3.5 .9998 .9998 .9998 .9998 .9998 .9998 .9998 .9998 .9998 .9998
3.6 .9998 .9998 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999
3.7 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999
3.8 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999 .9999

- END OF PAPER -

Page 21 of 21