APPLIED STATISTICS (SQQS2013)

Exercise 4: CHI – SQUARE TEST

1. Give one measurement scales that can be analysed using Chi-square test?

Solution: nominal/ordinal data.

2. One guideline to ensure a good approximation to the Chi-square distribution is that

expected frequency for the ith category is at least 5 (ei≥ 5). If this were not possible, what
would be a possible solution?

Solution: Increase sample size/combine rows or column

3. Why the computed Chi-square test statistics is always positive?

Solution:
The difference between the observed frequency and the expected frequency is squared,
that is (O–E)2 and the denominator is the number expected which must also be positive.

4. Find the value of degree of freedom of the independence test/equal ratio test for following
situation:
a) A number of primary and secondary school children were asked which of three
elements they thought was most important to them, Grades, Popularity, or Sports.

b) Research studied the relationship between having the AIDS syndrome and gender.

c) A researcher selected 100 passengers from each of 3 airlines and asked them if the
airline had lost their luggage on their last flight.

Solution:
a) 2
b) 1
c) 2

5. Explain how the “test of independence” and the “test of homogeneity” are differ.

Solution:
The test of independence has one sample of data that is being cross-tabulated according
to the categories of two separate variables.

The test of homogeneity has multiple samples being compared side-by-side and together
these samples from the entire sample used in the contingency table.

1
6. A store owner wishes to see if people have a favourite day of the week to shop. A sample
of 400 people is selected, and each is asked his or her preference. The data are shown. At
5% significance level, test the claim that shoppers have a preference day of the week to
shop.

Day Sun. Mon. Tues. Wed. Thurs. Fri. Sat.

Number 28 16 20 26 74 96 140

Solution:
H0: The shoppers have no preference day of the week to shop.
H1: The shoppers have a preference day of the week to shop.

1
pi= =0.1429 E i=57.1429
7
2 2
χ =¿237.1398 χ 0.05 , 6=12.5916

Reject H 0
The shoppers have a preference day of the week to shop at α = 0.05.

7. Studies in the past have shown that 64%, 21%, 10% and 5% of road accidents are due to
human errors, mechanical failure, insufficient road signs and fate respectively. A data set
from a recent study had revealed the following information: 328, 120, 40 and 12 accidents
were due to human errors, mechanical failure, insufficient road signs and fate
respectively. Test at α = 0.01 level of significance if the current distribution of causes of
accidents differs from the past.

Solution:
H 0 : The current distribution of causes of accidents is same as the past.
H 1 : The current distribution of causes of accidents is differs from the past.

Causes of accidents Oi pi n pi=E i

Human errors 328 0.64 320
Mechanical failure 120 0.21 105
Insufficient road signs 40 0.10 50
Fate 12 0.05 25
Total 500 1 500
2 2
χ test =11.1029 χ 0.01 ,3 =11.345

Failed to reject H 0
The current distribution of causes of accidents is no differs from the past.

2
8. The number of letters received by an office in a day for 100 days is as shown in the
following table.

No. of letters, x 0 1 2 3 4 5
No. of days, f 9 11 24 25 19 12

Does the number of letters receive by this office per day follows a Poisson distribution at
significance level  = 0.01?

Solution:
H0: Number of letters received by this office per day follows a Poisson distribution.
H1: Number of letters received by this office per day does not follow a Poisson
distribution.

x̄=2 .7 , m = 1

No. of letters 0 1 2 3 4 5
No. of days, Oi 9 11 24 25 19 12
0.220
pi 0.0672 0.1815 0.2450 0.1488 0.1370
5
Ei 6.72 18.15 24.5 22.05 14.88 13.7
( Oi−E i )2 0.394
0.7736 2.8167 0.0102 1.1408 0.2109
Ei 7

2 2
test 0. 01,4
χ =5 .3469 χ =13. 277 ; df = 6 – 1 – 1 = 4

Fail to reject H0.

There is enough evidence to support that the number of letters received by this office per
day follows a Poisson distribution at significance level  = 0.01.

9. A company that hires workers with disabilities, such as blind and deaf, is carrying out a
study to determine whether workers’ performances and their disabilities are dependent.
The following data were collected to be analyzed.

Disabilities
Blind Deaf
Performance Below Average 18 25
Average 65 40
Above Average 25 17

Use a 2.5% level of significance to help the company conducting the study.

3
 Solution:
H 0 : The workers’ performances and their disabilities are independent (have no
relationship).
H 1 : The workers’ performances and their disabilities are dependent (have a relationship).

Disabilities
Blind Deaf
Performance Below Average Eij 24.4421 18.5579
Average Eij 59.6842 45.3158
Above Average Eij 23.8737 18.1263

2 2
χ test =5.1556 χ 0.025 ,2=7.3778

Failed to reject H 0
The workers’ performances and their disabilities are independent.

10. 300 and 200 households were chosen at random from two cities, Kuala Lumpur and Kota
Kinabalu respectively and their income rates are classified as follows:

Cities
Kuala Lumpur Kota Kinabalu
High Income 150 60
Medium Income 105 110
Low Income 45 30

Test at the level of significance α = 0.01 the hypothesis that the proportions of households
that belong to different income groups are the same in Kuala Lumpur and Kota Kinabalu.

Solution:
H 0 : The proportions of households that belong to different income groups are the same in
Kuala Lumpur and Kota Kinabalu.
H 1 : The proportions of households that belong to different income groups are different in
Kuala Lumpur and Kota Kinabalu.

Cities
Kuala Lumpur Kota Kinabalu
High Income Eij 126 84
Medium Income Eij 129 86
Low Income Eij 45 30

2 2
χ test =22.5913 χ 0.01 ,2=9.2103

Reject H 0
The proportions of households that belong to different income groups are difference in
Kuala Lumpur and Kota Kinabalu.

4
11. The following data shows the number of college students in a class according to the type
of school they attended and experience of homesick feeling when they first attended the
college.

Type of school * Homesick feeling Crosstabulation

Homesick feeling Total
Yes No
Type of school Daily Count - 11 46
Expected Count - B -
SBP Count 10 25 -
Expected Count - 17 -
MRSM Count C 32 -
Expected Count - - -
Total Count - - A
Expected Count - - -

Chi-Square Tests
Asymp. Sig.
Value Df (2-sided)
Pearson Chi-Square 27.285(a) D .000
Likelihood Ratio 28.268 2 .001
Linear-by-Linear
21.686 1 .005
Association
N of Valid Cases -
a 0 cells (.0%) have expected count less than 5. The minimum expected count is 15.96.

Based on output above;

a) Find the value of A, B, C and D.
b) Did the data provide sufficient evidence to conclude that the type of school attended is
related to the homesick feeling? Use  = 0.05.

Solution:
a) A=1 40
B=22 . 3429
C=27
D=2

b) H0: The type of school attended is not related to the homesick feeling.
H1: The type of school attended is related to the homesick feeling.

p-value = 0.000< = 0.05

Reject H0.

There is enough evidence to support that the type of school attended is related to the
homesick feeling at 5% significance level.