Exercise Sheet 8 - Lecture 8

1. The following table gives the observed counts in 1-second intervals of alpha
particles emitted from a radioactive source. Use a Chi-squared test at the 5%
level to assess the goodness of fit to the Poisson distribution.

n Observed
0 5267
1 4436
2 1800
3 534
4 111
5+ 21

10187
λ̂ = 12169
= 0.8371

n Expected
0 5268.6
1 4410.5
2 1846.1
3 515.1
4 107.8
5+ 20.9

X 2 = 2.08 df = (6 - 1) - 1 = 4 Critical Region is X 2 < 9.49

⇒ No significant evidence against the null hypothesis at the 5% level.

2. Groups of six people are chosen at random and the number, x, of people
in each group who normally wear glasses is recorded. The results obtained
from 200 groups of six are shown in the table below. Fit a Binomial dis-
tribution to the data and assess the fit using a Chi-squared test at the 1%
level..

x 0 1 2 3 4 5 6
Counts 17 53 65 45 18 2 0

x̄ 2 1
n=6 p̂ = n
= 6
= 3

x 0 1 2 3 4 5 6
Expected Counts 17.6 52.7 65.8 43.9 16.7 3.3 0.3

Combining the last 3 cells together leads to X 2 = 0.0584

df = (5 - 1) - 1 = 3 Critical Region is X 2 > 11.34

⇒ No significant evidence against the null hypothesis at the 5% level.

1
3. Over a period of time a mens’ clothes shop sold 500 pairs of trousers. The
numbers sold in various waist sizes were:

Waist (cm) 70 75 80 85 90 95 100 105 110

Frequency 5 13 37 88 114 123 82 29 9

Estimate the mean and standard deviation of the waist measurements, and
test whether the figures support the theory that waist measurements fol-
low a normal distribution. Use a 5% level of significance. (Note that a
customer with, for example, a waist measurement between 80 and 85cm
requires trousers with an 85cm waistband.)
μ̂ = x̄ = 91.89 s2 = 60.9999

Waist (cm) 70 75 80 85 90 95 100 105 110

End point (u) 70 75 80 85 90 95 100 105 ∞
Standardized (z) -2.48 -1.84 -1.20 -0.56 0.08 0.72 1.36 2.00 ∞
P (Z < z) 0.007 0.033 0.115 0.287 0.531 0.764 0.913 0.977 1.0
P (a < Z < b) 0.007 0.026 0.082 0.172 0.244 0.233 0.149 0.064 0.019
E 3.26 13.09 40.96 86.20 122.05 116.29 74.57 32.17 9.33

Combining the first 2 cells together leads to X 2 = 2.56

df = (8 - 1) - 2 = 5 Critical Region is X 2 > 11.07

⇒ No significant evidence against the null hypothesis at the 5% level.

4. It is conventional wisdom in military squadrons that pilots tend to father
more girls than boys. Snyder (1961) gathered data for military fighter pi-
lots. The sex of the pilots offspring were tabulated for three kinds of flight
duty during the month of conception, as shown in the following table. Is
there any evidence of an association between father’s activity and offspring
gender? Use a 5% level of significance.

Female Offspring Male Offspring

Flying Fighters 51 38
Flying Transports 14 16
Not Flying 38 46

The expected cell counts are

Female Offspring Male Offspring

Flying Fighters 45.16 43.84
Flying Transports 15.22 14.78
Not Flying 42.62 41.38

X 2 = 2.75 df = (3 - 1)(2 - 1) = 2 Critical Region is X 2 > 5.99

⇒ No significant evidence against the null hypothesis at the 5% level.

2
 5. Grades in a statistics class were classified by students main subject of study.
Is there any relationship between grade and major? Use a 1% level of sig-
nificance.

Major
Grade Psychology Biology Other
A 8 15 13
B 14 19 15
C 15 4 7
D-F 3 1 4

The expected cell counts are

Major
Grade Psychology Biology Other
A 12.20 11.90 11.90
B 16.27 15.86 15.86
C 8.81 8.59 8.59
D-F 2.71 2.64 2.64

X 2 = 12.18 df = (4 - 1)(3 - 1) = 6 Critical Region is X 2 > 16.81

⇒ No significant evidence against the null hypothesis at the 5% level

Exam Paper Questions The table below lists past exam paper question numbers
that involve Chi-squared tests (useful for revision). NB. bracketed question num-
bers indicate that only part of the question involves Chi-squared tests.

Human Sciences Psychology

Year TT MT HT TT
2001 37 23 37
2000 348 28 34 348
1999 37 3 7 (4) 10 3 7
1998 36 36 36
1997 57 37 57