Chapter 12

Statistics for Educational
Researchers and Managers

Using Microsoft® Excel
5th Edition
Chapter 12
Chi Square Tests and Nonparametric
Tests
Statistics for Managers Using Microsoft Excel, 5e © 2008 Prentice-Hall, Inc. Chap 12-1
Learning Objectives
In this chapter, you learn:

 How and when to use the chi-square test for
contingency tables
 How to use the Marascuilo procedure for
determining pair-wise differences when
evaluating more than two proportions
 How and when to use the McNemar test
 How and when to use nonparametric tests
Contingency Tables
Contingency Tables
 Useful in situations involving multiple
population proportions
 Used to classify sample observations
according to two or more characteristics
 Also called a cross-classification table.
Contingency Table Example
Hand Preference vs. Gender

Dominant Hand: Left vs. Right
Gender: Male vs. Female
 2 categories for each variable, so the table

is called a 2 x 2 table
 Suppose we examine a sample of

300 college students
Sample results organized in a contingency table:
sample size = n = 300: Gender
120 Females, 12 were Hand

Female Male
left handed Preference
180 Males, 24 were Left 12 24 36
left handed
Right 108 156 264
120 180 300
H0: π1 = π2 (Proportion of females who are left
handed is equal to the proportion of
males who are left handed)
H1: π1 ≠ π2 (The two proportions are not the same –
Hand preference is not independent
of gender)
 If H0 is true, then the proportion of left-handed females
should be the same as the proportion of left-handed males.
 The two proportions above should be the same as the
proportion of left-handed people overall.
The Chi-Square Test Statistic
The Chi-square test statistic is:
2
(f  f )
χ2   o e
all cells fe
where:
fo = observed frequency in a particular cell
fe = expected frequency in a particular cell if H0 is true
2 for the 2 x 2 case has 1 degree of freedom
Assumed: each cell in the contingency table has

expected frequency of at least 5
The 2 test statistic approximately follows a chi-square

distribution with one degree of freedom
Decision Rule:
If 2 > 2U, reject H0,
otherwise, do not reject 
H0
0
Do not Reject H0 
reject H0
2U
Computing the
Average Proportion
The average X1  X 2 X
p 
proportion is: n1  n2 n
120 Females, 12 were  Here:

left handed
12  24 36
180 Males, 24 were p  0.12
left handed
120  180 300
The proportion of left handers overall is 0.12, that is, 12%
Finding Expected Frequencies
 To obtain the expected frequency for left handed females,
multiply the average proportion left handed (p) by the total
number of females
 To obtain the expected frequency for left handed males,
multiply the average proportion left handed (p) by the total
number of males
If the two proportions are equal, then

P(Left Handed | Female) = P(Left Handed | Male) = .12
i.e., we would expect (.12)(120) = 14.4 females to be left handed

(.12)(180) = 21.6 males to be left handed
Observed vs. Expected
Frequencies
Gender
Hand
Female Male
Preference
Observed = 12 Observed = 24
Left 36
Expected = 14.4 Expected = 21.6
Right 264
120 180 300
Gender
Hand
Female Male
Preference
Left 36
Right 264
The test statistic is: 120 180 300
( fo  fe )2
  
2
all cells fe
(12  14.4) 2 (108  105.6) 2 (24  21.6) 2 (156  158.4) 2
    0.7576
14.4 105.6 21.6 158.4
The test statistic is  2 0.7576 , U2 with 1 d.f. 3.841
Decision Rule:
If 2 > 3.841, reject H0, otherwise, do
not reject H0
Here,
=.05 2 = 0..7576 < 2U = 3.841,
so you do not reject H0 and
0 conclude that there is
Do not Reject H0 
reject H0 insufficient evidence that the
2U=3.841 two proportions are different.
2 Test for The Differences Among
More Than Two Proportions
 Extend the 2 test to the case with more than two

independent populations:
H0: π1 = π2 = … = πc
H1: Not all of the πj are equal (j = 1, 2, …, c)
2
( f  f )
2   o e
all cells fe
where:
 fo = observed frequency in a particular cell of the 2 x c table
 fe = expected frequency in a particular cell if H 0 is true
 2 for the 2 x c case has (2-1)(c-1) = c - 1 degrees of freedom
Assumed: each cell in the contingency table has expected frequency of at

least 1
Computing the
Overall Proportion
The overall X 1  X 2  ...  X c X
p 
proportion is: n1  n2  ...  nc n
 Expected cell frequencies for the c categories are

calculated as in the 2 x 2 case, and the decision rule
is the same:
Decision Rule: Where 2U is from the
If 2 > 2U, reject H0, chi-square distribution
otherwise, do not with c – 1 degrees of
reject H0 freedom
2 Test with More Than Two
Proportions: Example
 The sharing of patient records is a
controversial issue in health care. A survey
of 500 respondents asked whether they
objected to their records being shared by
insurance companies, by pharmacies, and by
medical researchers. The results are
summarized on the following table:
Organization
Object to Insurance Pharmacies Medical
Record Companies Researchers
Sharing
Yes 410 295 335

No 90 205 165
The overall p
X 1  X 2  ...  X c 410  295  335
 0.6933
proportion is: n1  n2  ...  nc 500  500  500
Organization
Sharing
Yes fo = 410 fo = 295 fo = 335

fe = 346.667 fe = 346.667 fe = 346.667
No fo = 90 fo = 205 fo = 165
fe = 153.333 fe = 153.333 fe = 153.333
Organization
Object Insurance Pharmacies Medical
to Companies Researchers
Record
Sharing
 fo  fe 
2
11 .571  f o  f e 2  f o  f e 2
7.700 0.3926
fe fe fe
Yes
 f o  f e 2  fo  fe 
2
17.409
 fo  fe 
2
0.888
26.159
fe fe fe
No
2
( f  f )
The Chi-square test statistic is:  2   o e 64.1196
all cells fe
H0: π1 = π2 = π3
H1: Not all of the πj are equal (j = 1, 2, 3)
Decision Rule: 2U = 5.991 is from the chi-

If 2 > 2U, reject H0, square distribution with 2
otherwise, do not reject H0 degrees of freedom.
Conclusion: Since 64.1196 > 5.991, you reject H0 and you

conclude that at least one proportion of respondents who object to
their records being shared is different across the three
organizations
The Marascuilo Procedure
The Marascuilo procedure enables you to

make comparisons between all pairs of
groups.
First, compute the observed differences pj - pj’
among all c(c-1)/2 pairs.
Second, compute the corresponding critical
range for the Marascuilo procedure.
 Critical Range for the Marascuilo Procedure:
p j (1  p j ) p j / (1  p j / )
Critical Range   2
U 
nj n j/
 Compute a different critical range for each

pair-wise comparison of sample proportions.
 Compare each of the c(c - 1)/2 pairs of
sample proportions against its corresponding
critical range.
 Declare a specific pair significantly different
if the absolute difference in the sample
proportions |pj – pj’| is greater than its critical
range.
Example
Organization
Sharing
Yes 410 295 335

P1 = 0.82 P2 = 0.59 P3 = 0.67
No 90 205 165
Example
MARASCUILO TABLE
Absolute
Proportions Differences Critical Range
| Group 1 - Group 2 | 0.23 0.06831808
| Group 1 - Group 3 | 0.15 0.0664689
| Group 2 - Group 3 | 0.08 0.074485617
Conclusion: Since each absolute difference is greater

than the critical range, you conclude that each
proportion is significantly different that the other two.
2 Test of Independence
 Similar to the 2 test for equality of more than two

proportions, but extends the concept to contingency
tables with r rows and c columns
H0: The two categorical variables are independent

(i.e., there is no relationship between them)
H1: The two categorical variables are dependent
(i.e., there is a relationship between them)
2 Test of Independence
2
( f  f )
2   o e
all cells fe
 where:
fo = observed frequency in a particular cell of the r x c table
fe = expected frequency in a particular cell if H0 is true
2 for the r x c case has (r-1)(c-1) degrees of freedom
Assumed: each cell in the contingency table has expected

frequency of at least 1)
Expected Cell Frequencies
 Expected cell frequencies:
row total column total

fe 
n
Where:
row total = sum of all frequencies in the row
column total = sum of all frequencies in the column
n = overall sample size
Decision Rule
 The decision rule is
If 2 > 2U, reject H0,
otherwise, do not reject H0
Where 2U is from the chi-square distribution with

(r – 1)(c – 1) degrees of freedom
Example: Test of Independence
 The meal plan selected by 200 students is shown below:
Number of meals per week

Class 20/week 10/week none Total
Standing
Fresh. 24 32 14 70
Soph. 22 26 12 60
Junior 10 14 6 30
Senior 14 16 10 40
Total 70 88 42 200
 The hypothesis to be tested is:
H0: Meal plan and class standing are independent

(i.e., there is no relationship between them)
H1: Meal plan and class standing are dependent
(i.e., there is a relationship between them)
Expected cell frequencies
if H0 is true:
Number of meals
Example for one cell: per week
Class 20/wk 10/wk none
row total x column total
fe  Standing Total
n
Fresh. 24.5 30.8 14.7 70
30 70
 10.5 Soph. 21.0 26.4 12.6 60
200
Junior 10.5 13.2 6.3 30
Senior 14.0 17.6 8.4 40
Total 70 88 42 200
 The test statistic value is:
( fo  fe )2
  
2
all cells fe
(24  24.5) 2 (32  30.8) 2 (10  8.4) 2
    0.709
24.5 30.8 8.4
2U = 12.592 for α = .05 from the chi-square

distribution with (4 – 1)(3 – 1) = 6 degrees of
freedom
Example: Test of
Independence
The test statistic is  2 0.709 , U2 with 6 d.f. 12.592
Decision Rule:
If 2 > 12.592, reject H0, otherwise,
do not reject H0
Here,
=0.05 2 = 0.709 < 2U = 12.592,
so do not reject H0
0 Conclusion: there is
Do not Reject H0  insufficient evidence that meal
reject H0
2U=12.592 plan and class standing are
related.
McNemar Test
 Used to test for the difference between two

proportions of related samples (not
independent)
 You need to use a test statistic that follows
the normal distribution
McNemar Test
Contingency Table
Condition (Group) 2
Condition (Group) 1 Yes No Totals
Yes A B A+B
No C D C+D
Totals A+C B+D n
Where A = number of respondents who answered yes to condition 1 and condition 2

B = number of respondents who answered yes to condition 1 and no to 2
C = number of respondents who answered no to condition 1 and yes to 2
D = number of respondents who answered no to condition 1 and condition 2
n = number of respondents in the sample
McNemar Test
Contingency Table
Condition (Group) 2
Yes A B A+B
No C D C+D
Totals A+C B+D n
The sample proportions are:
A B AC
p1  p2 
n n
McNemar Test
Contingency Table
Condition (Group) 2
Yes A B A+B
No C D C+D
Totals A+C B+D n
The population proportions are:

π1 = proportion of the population who answer yes to condition 1
π2 = proportion of the population who answer yes to condition 2
McNemar Test
Test Statistic
To test the hypothesis:
H0: π1 = π2
H1: π1 ≠ π2
Use the test statistic:

B C
Z
B C
McNemar Test
Example
 Suppose you survey 300 homeowners and ask
them if they are interested in refinancing their
home. In an effort to generate business, a
mortgage company improved their loan terms and
reduced closing costs. The same homeowners
were again surveyed. Determine if change in loan
terms was effective in generating business for the
mortgage company. The data are summarized as
follows:
McNemar Test
Example
Survey response after change
Survey response Yes No Totals
before change
Yes 118 2 120

No 22 158 180
Totals 140 160 300
Test the hypothesis (at the 0.05 level of significance):

H0: π1 ≥ π2: The change in loan terms was ineffective
H1: π1 < π2: The change in loan terms increased business
McNemar Test
Example
Survey response after The critical value (.05
Survey change significance) is Z = -1.96
response Yes No Totals
before change The test statistic is:
B C 2  22
Z   4.08
Yes 118 2 120 B C 2  22
No 22 158 180
Totals 140 160 300
Since Z = -4.08 < -1.96, you reject H0 and conclude that the
change in loan terms significantly increase business for the
mortgage company.
Wilcoxon Rank-Sum Test
 Test two independent population medians

 Populations need not be normally distributed
 Distribution free procedure
 Used when only rank data are available
 Can use when both n1 , n2 ≤ 10

 Assign ranks to the combined n1 + n2 sample values
 If unequal sample sizes, let n1 refer to smaller-sized
sample
 Smallest value rank = 1, largest value rank = n1 + n2
 Assign average rank for ties
 Sum the ranks for each sample: T1 and T2
 Obtain test statistic, T1 (from smaller sample)
Checking the Rankings
 The sum of the rankings must satisfy the formula

below
 Can use this to verify the sums T1 and T2
n(n  1)
T1 T2 
2
where n = n1 + n2
Hypothesis and Decision
M1 = median of population 1; M2 = median of population 2
Test statistic = T1 (Sum of ranks from smaller sample)
Two-Tail Test Left-Tail Test Right-Tail Test
H0: M1 = M2 H0: M1 ≥ M2 H0: M1 ≤ M2
H1: M1 ≠ M2 H1: M1 < M2 H1: M1 > M2
Reject Do Not Reject Reject Do Not Reject Do Not Reject Reject

Reject
T1L T1U T1L T1U
Reject H0 if T1 ≤ T1L Reject H0 if T1 ≤ T1L Reject H0 if T1 ≥ T1U
or if T1 ≥ T1U
Example with Small Samples
 Sample data are collected on the capacity
rates (% of capacity) for two factories.
 Are the median operating rates for two
factories the same?
 For factory A, the rates are 71, 82, 77, 94, 88
 For factory B, the rates are 85, 82, 92, 97

 Test for equality of the population medians
at the 0.05 significance level
Capacity Rank
Ranked
Factory A Factory B Factory A Factory B
Capacity
71 1
values:
77 2
Tie in 3rd and 82 3.5
4th places 82 3.5

85 5
88 6
92 7
94 8
97 9
Rank Sums: 20.5 24.5
Factory B has the smaller sample size, so the test
statistic is the sum of the Factory B ranks:
T1 = 24.5
The sample sizes are:

n1 = 4 (factory B)
n2 = 5 (factory A)
The level of significance is α = .05
 n1
 Lower and
Upper Critical n2 One- Two-
4 5
Values for T1 Tailed Tailed
from
4
Appendix
Table E.8: .05 .10 12, 28 19, 36
.025 .05 11, 29 17, 38
T1L = 11 and 5
.01 .02 10, 30 16, 39
T1U = 29
.005 .01 --, -- 15, 40
6
  = .05
Test Statistic (Sum of ranks
 n1 = 4 , n2 = 5
from smaller sample):
Two-Tail Test T1 = 24.5
H0: M1 = M2
H1: M1 ≠ M2
Decision:
Reject Do Not Reject Do not reject at a = 0.05
Reject
T1L=11 T1U=29 Conclusion:
There is not enough evidence to prove
Reject H0 if T1 ≤ T1L = 11
that the medians are not equal.
or if T1 ≥ T1U = 29
Using Normal Approximation
 For large samples, the test statistic T1 is approximately normal
with mean and standard deviation:
n1(n  1) n1 n2 (n  1)
μT1  σ T1 
2 12
 Must use the normal approximation if either n1 or n2 > 10
 Assign n1 to be the smaller of the two sample sizes
 Can use the normal approximation for small samples
 The Z test statistic is
T1  μ T1
Z
σT1
 Where Z approximately follows a standardized

normal distribution
Use the setting of the prior example:
The sample sizes were:
n1 = 4 (factory B)
n2 = 5 (factory A)
The level of significance was α = .05
The test statistic was T1 = 24.5
n1(n  1) 4(9  1)
μT1   20
2 2
n1 n2 (n  1) 4 (5) (9  1)
σT1   4.082
12 12
 The test statistic is
T1  μ T1 24.5  20
Z  1.10
σT1 4.082
 Z = 1.10 is less than the critical Z value of 1.96 (for α = .05)
so you do not reject H0 – there is not sufficient evidence that
the medians are not equal
Kruskal-Wallis Rank Test
 Tests the equality of more than 2 population medians
 Use when the normality assumption for one-way
ANOVA is violated
 Assumptions:
 The samples are random and independent
 variables have a continuous distribution
 the data can be ranked
 populations have the same variability
 populations have the same shape
 Obtain overall rankings for each value
 In event of tie, each of the tied values gets the average
rank
 Sum the rankings for data from each of the c
groups
 Compute the H test statistic
 The Kruskal-Wallis H test statistic:
(with c – 1 degrees of freedom)
 12 c T2
H   j
  3(n  1)
 n(n  1) j 1 n j 
where:
n = total number of values over the combined samples
c = Number of groups
Tj = Sum of ranks in the jth sample
nj = Size of the jth sample
 Complete the test by comparing the calculated H value
to a critical 2 value from the chi-square distribution
with c – 1 degrees of freedom
 Decision rule

 Reject H0 if test statistic H > 2U
0   Otherwise do not reject H0
Do not Reject H0
reject H0
2U
Example
 Do different branch offices have a different number of
employees?
Office size Office size Office size

(Chicago, C) (Denver, D) (Houston, H)
23 55 30
41 60 40
54 72 18
78 45 34
66 70 44
Example
 Do different branch offices have a different
number of employees?
Office size Office size Class size

Ranking Ranking Ranking
(Chicago, C) (Denver, D) (Houston, H)
23 2 55 10 30 3
41 6 60 11 40 5
54 9 72 14 18 1
78 15 45 8 34 4
66 12 70 13 44 7
 = 44  = 56  = 20
Example
H 0 : Median C Median D Median H
H A : Not all population Medians are equal
 The H statistic is
 12 c T2
H   j
  3(n  1)
 n(n  1) j 1 n j 
 12  44 2 56 2 20 2  
       3(15  1) 6.72
 15(15  1)  5 5 5 
Example
 Compare H = 6.72 to the critical value from the chi-square
distribution for 3 – 1 = 2 degrees of freedom and  = .05:
2
χ 5.991
U
2
Since H = 6.72 >U 5.991
χ ,
reject H0
There is evidence of a difference in the

population median number of employees in
the branch offices
Chapter Summary
In this chapter, we have
 Developed and applied the 2 test for the difference between
two proportions
 Developed and applied the 2 test for differences in more than
two proportions
 Examined the 2 test for independence
 Used the McNemar test for differences in two related
proportions
 Used the Wilcoxon rank sum test for two population medians
 Small Samples
 Large sample Z approximation
 Applied the Kruskal-Wallis H-test for multiple population
medians

Chapter 12

Uploaded by

Copyright:

Available Formats

Chapter 12

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 12

Uploaded by

Copyright:

Available Formats

Statistics for Educational

Researchers and Managers

In this chapter, you learn:

Hand Preference vs. Gender

 2 categories for each variable, so the table

 Suppose we examine a sample of

Sample results organized in a contingency table:

sample size = n = 300: Gender

120 Females, 12 were Hand

2 for the 2 x 2 case has 1 degree of freedom

Assumed: each cell in the contingency table has

The 2 test statistic approximately follows a chi-square

120 Females, 12 were  Here:

The proportion of left handers overall is 0.12, that is, 12%

If the two proportions are equal, then

i.e., we would expect (.12)(120) = 14.4 females to be left handed

120 180 300

 Extend the 2 test to the case with more than two

Assumed: each cell in the contingency table has expected frequency of at

 Expected cell frequencies for the c categories are

Yes 410 295 335

Yes fo = 410 fo = 295 fo = 335

Decision Rule: 2U = 5.991 is from the chi-

Conclusion: Since 64.1196 > 5.991, you reject H0 and you

The Marascuilo procedure enables you to

 Compute a different critical range for each

Yes 410 295 335

Conclusion: Since each absolute difference is greater

 Similar to the 2 test for equality of more than two

H0: The two categorical variables are independent

2 for the r x c case has (r-1)(c-1) degrees of freedom

Assumed: each cell in the contingency table has expected

row total column total

 The decision rule is

If 2 > 2U, reject H0,

otherwise, do not reject H0

Where 2U is from the chi-square distribution with

Number of meals per week

 The hypothesis to be tested is:

H0: Meal plan and class standing are independent

 The test statistic value is:

2U = 12.592 for α = .05 from the chi-square

 Used to test for the difference between two

Where A = number of respondents who answered yes to condition 1 and condition 2

The sample proportions are:

The population proportions are:

Use the test statistic:

Yes 118 2 120

Test the hypothesis (at the 0.05 level of significance):

 Test two independent population medians

 Can use when both n1 , n2 ≤ 10

 The sum of the rankings must satisfy the formula

Reject Do Not Reject Reject Do Not Reject Do Not Reject Reject

 For factory B, the rates are 85, 82, 92, 97

4th places 82 3.5

The sample sizes are:

 Where Z approximately follows a standardized

The level of significance was α = .05

The test statistic was T1 = 24.5

Office size Office size Office size