ABSTRACT
The one-factor analysis of covariance (ANCOVA) model can be extended to more complex models in the same way as we expanded the one-factor ANOVA model. The example procedures briefly described here are easily translated from the ANOVA context into the ANCOVA context. Given a fixed sample size, ANCOVA is more powerful than ANOVA, and this has been demonstrated. Approaching power from a slightly different angle, a smaller sample size is needed in ANCOVA to obtain the same power in ANOVA. The introduction of a covariate requires several assumptions beyond the traditional ANOVA assumptions. Generating power analysis for ANCOVA models follows similarly to that for ANOVA and factorial ANOVA. In particular, if there is more than one independent variable, we must test for main effects and interactions separately. Because we have only one independent variable for our ANCOVA model, our illustration assumes only one main effect.
