The Catholic University of Malawi

Faculty of Science

Mathematical Sciences Department

Generic & Weekend

STA4102N: Multivariate Methods

End of Semester Examination

Tuesday, June 8, 2021

08h30 – 12h30 – 4 Hours

Instructions:

1. Attempt all questions in this paper

2. Solutions to each question item MUST begin on a separate page

3. (a) .do files MUST be name after their corresponding question numbers e.g.
STA4102N_Question2a
(b) Place ALL .do files in a directory called STA4102N_ExamNumber_EndSem
(c) Compress this folder and submit it together with any work you want to be
graded to the invigilator when time has been called on

4. All program files/listings MUST have header comments and they should be
located at the TOP of the file

5. Among other things, the file header comments MUST have: The author; date;
type of assessment (End or Mid Semester examination); course code and title;
question number; a description of what the code in the file accomplishes; a list of
any modifications (bug fixes) to the file

6. Instructions 2 through 5 are part of the examination and are graded 7 marks

Page 1 of 7 – Multivariate Methods (STA4102N)

1. Twenty engineer apprentices and 20 pilots were given six tests (Travers 1939). The
variables were

y1 = intelligence,
y2 = form relations,
y3 = dynamometer,
y4 = dotting,
y5 = sensory motor coordination,
y6 = perseveration.

The data are given in Table 1

(a) Test H0 : µ1 = µ2 .
(b) If the T 2 -test in part 1a rejects H0 , carry out a t-test for each variable.
(c) Test the significance of y4 , y5 , y6 adjusted for y1 , y2 , y3 . [ 7, 11, 5 marks ]

Table 1: Comparison of Six Tests on Engineer Apprentices and Pilots

Engineer Apprentices Pilots
y1 y2 y3 y4 y5 y6 y1 y2 y3 y4 y5 y6
121 22 74 223 54 254 132 17 77 232 50 249
108 30 80 175 40 300 123 32 79 192 64 315
122 49 87 266 41 223 129 31 96 250 55 319
77 37 66 178 80 209 131 23 67 291 48 310
140 35 71 175 38 261 110 24 96 239 42 268
108 37 57 241 59 245 47 22 87 231 40 217
124 39 52 194 72 242 125 32 87 227 30 324
130 34 89 200 85 242 129 29 102 234 58 300
149 55 91 198 50 277 130 26 104 256 58 270
129 38 72 162 47 268 147 47 82 240 30 322
154 37 87 170 60 244 159 37 80 227 58 317
145 33 88 208 51 228 135 41 83 216 39 306
112 40 60 232 29 279 100 35 83 183 57 242
120 39 73 159 39 233 149 37 94 227 30 240
118 21 83 152 88 233 149 38 78 258 42 271
141 42 80 195 36 241 153 27 89 283 66 291
135 49 73 152 42 249 136 31 83 257 31 311
151 37 76 223 74 268 97 36 100 252 30 225
97 46 83 164 31 243 141 37 105 250 27 243
109 42 82 188 57 267 164 32 76 187 30 264

2. Baten, Tack, and Baeder (1958) compared judges’ scores on fish prepared by three
methods. Twelve fish were cooked by each method, and several judges tasted fish
samples and rated each on four variables: y1 = aroma, y2 = flavor, y3 = texture,

Page 2 of 7 – Multivariate Methods (STA4102N)

 and y4 = moisture. The data are in Table 2. Each entry is an average score for the
judges on that fish.
Compare the three methods using all four MANOVA tests. [ 11 marks ]

Table 2: Judges’ Scores on Fish Prepared by Three Methods

Method 1 Method 2 Method 3
y1 y2 y3 y4 y1 y2 y3 y4 y1 y2 y3 y4
5 6 6.3 6.7 5 5.3 5.3 6.5 4.8 5 6.5 7
5 6.2 6 5.8 4.8 4.9 4.2 5.6 5.4 5 6 6.4
6 5.9 6 7 3.9 4 4.4 5 4.9 5.1 5.9 6.5
5 5 4.9 5 4 5.1 4.8 5.8 5.7 5.2 6.4 6.4
5 5.7 5 6.5 5.6 5.4 5.1 6.2 4.2 4.6 5.3 6.3
6 6.1 6 6.6 6 5.5 5.7 6 6 5.3 5.8 6.4
6 6 5.8 6 5.2 4.8 5.4 6 5.1 5.2 6.2 6.5
4 5 4 5 5.3 5.1 5.8 6.4 4.8 4.6 5.7 5.7
6 5.4 4.9 5 5.9 6.1 5.7 6 5.3 5.4 6.8 6.6
6 5.2 5.4 5.8 6.1 6 6.1 6.2 4.6 4.4 5.7 5.6
6 6.1 5.2 6.4 6.2 5.7 5.9 6 4.5 4 5 5.9
5 5.9 5.8 6 5.1 4.9 5.3 4.8 4.4 4.2 5.6 5.5

3. Table 3, from Zerbe (1979a), compares 13 control and 20 obese patients on a glu-
cose tolerance test using plasma inorganic phosphate. Delete the observations cor-
responding to 12 and 1 12 hours so that the time points are equally spaced.

(a) For the control group, use orthogonal polynomials to find the degree of growth
curve.
(b) Repeat 3a for the obese group.
(c) Find the degree of growth curve for the combined groups, and compare the
growth curves of the two groups. [ 13, 13, 7 marks ]

4. Reaven and Miller (1979; see also Andrews and Herzberg 1985, pp. 215–219) mea-
sured five variables in a comparison of normal patients and diabetics. In Table 4
we give partial data for normal patients only. The three variables of major interest
were

x1 = glucose intolerance,
x2 = insulin response to oral glucose,
x3 = insulin resistance.

The two additional variables of minor interest were

y1 = relative weight,
y2 = fasting plasma glucose.

Page 3 of 7 – Multivariate Methods (STA4102N)

 (a) Find the canonical correlations between (y1 , y2 ) and (x1 , x2 , x3 ).
(b) Find the standardized coefficients for the canonical variates.
(c) Test the significance of each canonical correlation.

5. (a) Suppose y is N4 (µ, Σ), where


  
−2 11 −8 3 9
3 −8 9 −3 6
   
µ=  Σ2 = 
−1  3 −3 2 3

5 9 6 3 9

i. Find the distribution of z = 4y1 − 2y2 + y3 − 3y4 .

ii. Find the joint distribution of z1 = y1 + y2 + y3 + y4 and z2 = −2y1 + 3y2 +
y3 − 2y4 . [ 7, 7 marks ]
(b) Harry Roberts, a naturalist for the Alaska Fish and Game department, studies
grizzly bears with the goal of maintaining a healthy population. Measurements
on bears provided the following summary statistics:

Variable Weight Body length Neck Girth Head length Head width
(kg) (cm) (cm) (cm) (cm) (cm)
Sample mean 95.52 164.38 55.69 93.39 17.98 31.13
x

Covariance matrix
 
3266.46 1343.97 731.54 1175.5 162.68 238.37
1343.97 721.91 324.25 537.35 80.17 117.73
 
 731.54 324.25 179.28 281.17 39.15 56.8 
 
S=
 1175.5 537.35 281.17 474.98 63.73 94.85 

 162.68 80.17 39.15 63.73 9.95 13.88 
 
238.37 117.73 56.8 94.85 13.88 21.26

i. Obtain the large sample 95% simultaneous confidence intervals for the six
population mean body measurements.
ii. Obtain the 95% Bonferroni confidence intervals for the six means in Part
5(b)i [ 11, 11 marks ]

Page 4 of 7 – Multivariate Methods (STA4102N)

 Table 3: Plasma Inorganic Phosphate (mg/dl)
Hours after Glucose Challenge
1
Patient 0 2
1 1 12 2 3 4 5
Control
1 4.3 3.3 3.0 2.6 2.2 2.5 3.4 4.4
2 3.7 2.6 2.6 1.9 2.9 3.2 3.1 3.9
3 4.0 4.1 3.1 2.3 2.9 3.1 3.9 4.0
4 3.6 3.0 2.2 2.8 2.9 3.9 3.8 4.0
5 4.1 3.8 2.1 3.0 3.6 3.4 3.6 3.7
6 3.8 2.2 2.0 2.6 3.8 3.6 3.0 3.5
7 3.8 3.0 2.4 2.5 3.1 3.4 3.5 3.7
8 4.4 3.9 2.8 2.1 3.6 3.8 4.0 3.9
9 5.0 4.0 3.4 3.4 3.3 3.6 4.0 4.3
10 3.7 3.1 2.9 2.2 1.5 2.3 2.7 2.8
11 3.7 2.6 2.6 2.3 2.9 2.2 3.1 3.9
12 4.4 3.7 3.1 3.2 3.7 4.3 3.9 4.8
13 4.7 3.1 3.2 3.3 3.2 4.2 3.7 4.3
Obese
1 4.3 3.3 3.0 2.6 2.2 2.5 2.4 3.4
2 5.0 4.9 4.1 3.7 3.7 4.1 4.7 4.9
3 4.6 4.4 3.9 3.9 3.7 4.2 4.8 5.0
4 4.3 3.9 3.1 3.1 3.1 3.1 3.6 4.0
5 3.1 3.1 3.3 2.6 2.6 1.9 2.3 2.7
6 4.8 5.0 2.9 2.8 2.2 3.1 3.5 3.6
7 3.7 3.1 3.3 2.8 2.9 3.6 4.3 4.4
8 5.4 4.7 3.9 4.1 2.8 3.7 3.5 3.7
9 3.0 2.5 2.3 2.2 2.1 2.6 3.2 3.5
10 4.9 5.0 4.1 3.7 3.7 4.1 4.7 4.9
11 4.8 4.3 4.7 4.6 4.7 3.7 3.6 3.9
12 4.4 4.2 4.2 3.4 3.5 3.4 3.9 4.0
13 4.9 4.3 4.0 4.0 3.3 4.1 4.2 4.3
14 5.1 4.1 4.6 4.1 3.4 4.2 4.4 4.9
15 4.8 4.6 4.6 4.4 4.1 4.0 3.8 3.8
16 4.2 3.5 3.8 3.6 3.3 3.1 3.5 3.9
17 6.6 6.1 5.2 4.1 4.3 3.8 4.2 4.8
18 3.6 3.4 3.1 2.8 2.1 2.4 2.5 3.5
19 4.5 4.0 3.7 3.3 2.4 2.3 3.1 3.3
20 4.6 4.4 3.8 3.8 3.8 3.6 3.8 3.8

Page 5 of 7 – Multivariate Methods (STA4102N)

 Table 4: Relative Weight, Blood Glucose, and Insulin Levels
Patient Number y1 y2 x1 x2 x3
1 0.81 80.00 356.00 124.00 55.00
2 0.95 97.00 289.00 117.00 76.00
3 0.94 105.00 319.00 143.00 105.00
4 1.04 90.00 356.00 199.00 108.00
5 1.00 90.00 323.00 240.00 143.00
6 0.76 86.00 381.00 157.00 165.00
7 0.91 100.00 350.00 221.00 119.00
8 1.10 85.00 301.00 186.00 105.00
9 0.99 97.00 379.00 142.00 98.00
10 0.78 97.00 296.00 131.00 94.00
11 0.90 91.00 353.00 221.00 53.00
12 0.73 87.00 306.00 178.00 66.00
13 0.96 78.00 290.00 136.00 142.00
14 0.84 90.00 371.00 200.00 93.00
15 0.74 86.00 312.00 208.00 68.00
16 0.98 80.00 393.00 202.00 102.00
17 1.10 90.00 364.00 152.00 76.00
18 0.85 99.00 359.00 185.00 37.00
19 0.83 85.00 296.00 116.00 60.00
20 0.93 90.00 345.00 123.00 50.00
21 0.95 90.00 378.00 136.00 47.00
22 0.74 88.00 304.00 134.00 50.00
23 0.95 95.00 347.00 184.00 91.00
24 0.97 90.00 327.00 192.00 124.00
25 0.72 92.00 386.00 279.00 74.00
26 1.11 74.00 365.00 228.00 235.00
27 1.20 98.00 365.00 145.00 158.00
28 1.13 100.00 352.00 172.00 140.00
29 1.00 86.00 325.00 179.00 145.00
30 0.78 98.00 321.00 222.00 99.00
31 1.00 70.00 360.00 134.00 90.00
32 1.00 99.00 336.00 143.00 105.00
33 0.71 75.00 352.00 169.00 32.00
34 0.76 90.00 353.00 263.00 165.00
35 0.89 85.00 373.00 174.00 78.00
36 0.88 99.00 376.00 134.00 80.00
37 1.17 100.00 367.00 182.00 54.00
38 0.85 78.00 335.00 241.00 175.00
39 0.97 106.00 396.00 128.00 80.00
40 1.00 98.00 277.00 222.00 186.00
41 1.00 102.00 378.00 165.00 117.00

6. You are provided with a .txt file. The file is located in the Desktop directory of

Page 6 of 7 – Multivariate Methods (STA4102N)

 the computer you are using. It contains the problem data. This data comes from
a firm that surveyed a random sample of n = 50 of its employees in an attempt to
determine which factors influence sales performance. Two collections of variables
were measured:

• Sales Performance:
o Sales Growth
o Sales Profitability
o New Account Sales
• Test Scores as a Measure of Intelligence:
o Creativity
o Mechanical Reasoning
o Abstract Reasoning
o Mathematics

There are p = 3 variables in the first group relating to Sales Performance and q = 4
variables in the second group relating to Test Scores

(a) Test for relationship between canonical variate pairs

(b) Obtain estimates of canonical correlation
(c) Obtain the canonical coefficients
(d) Interpret each component

[ 5, 5, 5, 5 marks ]

7. Explain examples of typical questions that are answered by the MANOVA in the
following fields:

(a) Medicine
(b) Sociology
(c) Management Studies

[ 5, 5, 5, 5 marks ]

——— End of Examination ———

Page 7 of 7 – Multivariate Methods (STA4102N)