Hypotheses with nominal data are tested with a Chi Square. The type of Chi Square depends on the type of hypothesis being tested. The descriptive statistics for nominal data are frequency distributions, in which the frequency of participants in each category is computed. If there are two or more groups being compared, the frequency distribution is called a cross-tabulation, in which the frequency of participants in each combination of conditions is reported. Both of these are described in Chapter 5.
One there is only one group, the null hypothesis is that the distribution of frequencies for the categories fits a specified hypothesis, such as the frequencies in all categories are equal. The null hypothesis could be anything. For example, it could be that categories 1 and 2 each have 25% and category 3 has 50%. However, there should be a theoretically sound reason for the hypothesis. The Chi Square used in this situation is the Chi Square Goodness-of-Fit test. It is called that because it tests how well the distribution fits the hypothesized distribution.
Compute a Chi Square Goodness-of-Fit Test Manually |
Compute a Chi Square Goodness-of-Fit Test Using SPSS |
When comparing two or more groups on nominal data, the appropriate test is the Chi Square Test for Independence. This test compares the pattern of frequencies in the groups, testing the null hypothesis that the pattern is comparable in the populations from which the group samples were drawn.
Compute a Chi Square Test for Independence Manually |
Compute a Chi Square Test for Independence Using SPSS |