2011•Wiley series in probability and statisticsRequires access
Randomized Pretest–Posttest Designs
Bradley E. Huitema
Open publisher page 0 citations
Abstract
Pretest–posttest designs designs are analyzed using one-factor ANOVA on the posttest. This chapter points out relationships among these methods, to explain why the two-factor ANOVA approach is often misleading, and to explain why ANCOVA is usually better than all the ANOVA alternatives. In the chapter the covariate is an early measurement on the variable to be used as the dependent variable rather than a separate variable measured in a different way. The analysis is carried out by using the pretest as the covariate and the posttest as the dependent variable. Frequently encountered analytic procedures include (1) one factor ANOVA on the posttest, (2) one-factor ANOVA on difference scores, and (3) two-factor ANOVA. These methods are less satisfactory than ANCOVA because they do not adjust for chance differences on the pretest and they generally have lower power. Controlled Vocabulary Terms analysis of covariance; analysis of variance