Inferences About Proportions

In Chapters 6 and 7, we considered testing inferences about means, ­rst for a single mean (Chapter 6) and then for two means (Chapter 7). The major concepts discussed in those two chapters included the following: types of hypotheses, types of decision errors, level of signi­cance, power, con­dence intervals (CIs), effect sizes, sampling distributions involving the mean, standard errors involving the mean, inferences about a single mean, inferences about the difference between two independent means, and inferences about the difference between two dependent means. In this chapter, we consider inferential tests involving proportions. We de­ne a proportion as the percentage of scores falling into particular categories. Thus, the tests described in this chapter deal with variables that are categorical in nature and thus are nominal or ordinal variables (see Chapter 1), or have been collapsed from higher-level variables into nominal or ordinal variables (e.g., high and low scorers on an achievement test).