ABSTRACT
One method, which may appear to be commonsense, is to analyze pretestposttest data using a repeated measures analysis of variance. After all, in a repeated measures experimental design, multiple measurements are made on the same subject and the test for treatment effects is based not on experimental error, but on within subject error. We shall see however that this approach is valid only when multiple posttest measurements are made on the same subject. When applied to a single pretest and single posttest, the output from a repeated measures analysis of variance has a unique interpretation.
