CHI-SQUARED EXAMPLE

A social scientist wants to see if there is a relationship between the divorce status (divorced or never divorced) among American men and their religious affiliation (or lack thereof). A random sample of 500 American men who were, or are, married is surveyed, and their divorce status and religious affiliation were recorded. These results are summarized in the table below.

A. Is there a relationship between divorce status and religious affiliation? Use a = .10.

B. For each of the religious affiliation classifications, calculate the percentage of men who have divorced and the percentage of men who haven’t been divorced.

C. Discuss the meaning of the percentages found in part A with respect to the relationship between divorce status and religious affiliation.

 

SOLUTION

A. It is desired to see if there is a relationship between two qualitative variables; divorce status and religious affiliation.

          H0: Divorce status is independent of religious affiliation

          Ha: Divorce status and religious affiliation are related

           Decision Rule: Accept Ha if the calculated p-value < .10.

            

          Calculations from StatCrunch: Χ2 = 11.36, p-value = 0.0228

Interpretation: At the .10 level of significance I conclude that there is a relationship between divorce status and religious affiliation.

B. The percentage of men who have divorced for each of the religious affiliation classifications is

Using these the following summary table is constructed.

 

      Religious Affiliation

 

Divorce Status

    A

    B

    C

    D

 None

Divorced

  27%

  31%

  21%

  29% 

  33%

Never Divorced

  73%

  69%

  79%

  71%

  67%

Totals

100%

100%

100%

100%

100%

C. For the religious affiliations classified as A, B, D, and None all of the percentages of divorced men are reasonably close to 30%, and of course those never divorced are close to 70%.  However, for the classification of C, the percentage of divorced men is much lower at 21% with 79% being never divorced. The differences in these percentages for classification C as compared to the other classifications is the main reason that it was concluded that a relationship exists between divorce status and religious affiliation.

 

COMMENTS ABOUT THE SOLUTION

            

Contingency table results:
Rows: Divorce Status
Columns: None

 

Cell format

Count
Expected count


 

A

B

C

D

None

Total

Divorced

15
13.66

25
19.52

36
51.48

28
23.91

18
13.42

122

Never Divorced

41
42.34

55
60.48

175
159.5

70
74.09

37
41.58

378

Total

56

80

211

98

55

500


Statistic

DF

Value

P-value

Chi-square

4

11.35965

0.0228

 

 

 

 

Prev ]

TABLE OF CONTENTS ]