ABSTRACT
The simplest relationship which involves both covariate and factor has parallel regression lines, with possibly different intercepts for different factor levels. The classical situation in analysis of covariance has the experimental factor and the covariate affecting response in ways that are readily separable. The analysis of variance assumes that the mean responses at different factor levels are constant but possibly different. The analysis of covariance, adjusting cell means by a covariate, can be readily extended to several factors. Pivot statistics and tests of hypotheses for the main inferential questions concerning the effect of treatments adjusted for the covariate and the effect of covariate adjusted for the treatments. Interaction plots for factorial arrangements, suggest the form of relationship between covariate and response for each level of the factor.
