You are on page 1of 2

STAT3014/3914

Semester 2 Applied Statistics 2010

Exercise 2

Tutorial Exercise.
 
2 12
 8 9 
1. Consider the data matrix X = 
 6 9 .

8 10
 
2 7
Use Hotelling’s T to test the hypothesis H0 : µ = . What assumptions are
11
necessary to perform this test?

2. From the data on 9 girls living at high altitude


 
24.5000
x2 = (76.0, 58.4, 13.5)T and S 2 =  5.6375 1.9700 .
4.3125 1.4563 1.8125

(a) Write down an expression for the statistic used to test the hypothesis that the
population mean vector is (80, 60, 15)T . The T 2 statistic takes the value 13.37.
Carry out the test.
(b) Formulate a statistic to test the hypothesis that µ1 /5 = µ2 /4 = µ3 . Carry out the
test given that the test statistic takes the value 25.79.

3. (a) Measurements on cranial length (x1 ) and cranial breadth (x2 ) on a sample of 35
female frogs gave
 
T 17.178 19.710
x = (22.860, 24.397) and S X = .
23.710

Test the hypothesis H0 : µ1 = µ2 , stating any assumptions that you make.


(b) Similar measurements on 14 male frogs gave
 
T 17.159 17.731
y = (21.821, 22.843) and S Y = .
19.273

Test the hypothesis H0 : µ1 = µ2 .


(c) Test the hypothesis that the male and female frog populations have the same
mean cranial lengths and breadths (that is, µM = µF ). You may assume the
populations have the same covariance structure.
Computer Exercise.

Data on 9 mandible measurements (in mm) for samples drawn from different species of
dog are stored in R in canine1 (modern Thai dogs) and canine2 (golden jackals). You can
obtain the data using

canine1 = read.table(file =
url("http://www.maths.usyd.edu/u/UG/SM/STAT3014/r/Data/canine1.dat"))

and

canine2 = read.table(file =
url("http://www.maths.usyd.edu/u/UG/SM/STAT3014/r/Data/canine2.dat"))

The variables are


V1 Length of mandible
V2 Breadth of mandible below the first molar
V3 Breadth of articular condyle
V4 Height of mandible below first molar
V5 Length of first molar
V6 Breadth of first molar
V7 Length from first to third molars
V8 Length from first to 4th premolar
V9 Breadth of lower canine tooth.

1. Calculate the mean and covariance matrix for the two groups.

2. Test the hypothesis that the two dog populations have equal mean vectors.

3. Test that the two populations have common mean values for the subvector consisting
of variables V 5, V 6, V 7, and V 8.

You might also like