ABSTRACT

This chapter introduces a major step into the sort of significance testing that tends to dominate experimental research. Researchers rarely use just two samples in a test of difference. ANOVA (analysis of variance) is introduced for the analysis of more than two conditions. ANOVA is a powerful parametric procedure for testing the differences between several means and (later, in Chapter 21) several independent variables. It avoids the problem of producing Type I errors that is incurred if we make multiple comparisons between pairs of means using several t tests. ANOVA allows us to compare all means as a group without raising our chance of a Type I error above .05.