Helps

Communicating Relationships of  Nominal Data Using Contingency Tables.

 The Chi Square test is usually used for nominal data. But the relationship is usually conveyed more simply with a contingency table. Percentages make the point even clearer.

The following examples of contingency tables were presented in Applied Statistics for Public Administration, Revised Edition, by Kenneth Meier and Jeffery Brudney (1987), pp 253-5.             

 

Contingency Table

 

 

Independent Variable

 

 

 

            Category 1

Category 2

Total

Dependent

Variable

Category 1

 

 

 

Category 2

 

 

 

 

Total

 

 

 

 

 

An example with data.

 

Contingency Table: Raw Data

 

Race

 

Civil Service Exam

            Black

White

Total

   No Pass

70

70

140

   Pass

60

135

195

   Total

130

205

335

 

 

The table is clearer when the raw data is converted to a percentage.

 

Contingency Table: Percentage Data

 

Race

Civil Service Exam

            Black

White

  No Pass

54%

34%

  Pass

46%

66%

  Total

100%

(n = 130)

100%

(n = 205)

 

 

Control Variable

Sometimes a third variable explains the difference in the dependent variable better than our original independent variable.  Here is an example of how to present a control variable from Applied Statistics for Public Administration, Revised Edition, by Kenneth Meier and Jeffery Brudney (1987), pp 234-5.             

 

Contingency Table

 

                      Control Variable 1                                                

 

            Control Variable 0

 

 

 

Independent Variable

 

 

Independent Variable   

 

 

 

Cat. 1

Cat. 2

Total

 

  Cat. 1

 Cat. 2

Total

Dep.

Var.

Cat. 1

 

 

 

Cat. 1

 

             

 

Cat. 2

             

 

 

Cat. 2

 

           

 

 

Total

 

 

 

Total

 

 

 

 

Contingency Table: Raw Data

College Graduates

 

Not College Graduates

 

 

Race

 

 

Race

 

Exam

Black

White

Total

Exam

  Black

 White

Total

  No Pass

10

40

50

  No Pass

60

30

90

  Pass

30

120

150

  Pass

30

15

45

  Total

40

160

200

  Total

90

45

135

 

Contingency Table: Percentage Data

College Graduates

 

Not College Graduates

 

 

Race

 

 

Race

 

Exam

Black

White

 

Exam

  Black

 White

 

  No Pass

25%

25%

 

  No Pass

67%

67%

 

  Pass

75%

75%

 

  Pass

33%

33%

 

  Total

100%

(n = 40)

100%

(n = 160)

 

  Total

100%

 (n = 90)

100%

(n = 45)

 

As we can see, it is not enough to just apply the statistical procedure.  An analyst also has to identify other variables that may be be affecting our dependent variable, and test them when possible.