ABSTRACT
In a situation that calls for Analysis of variance (ANOVA), the basis for choosing one technique over the other is simply convenience. Sometimes it is more convenient to state the problem in regression terms, whereas other times the ANOVA terminology is more useful. In fact, most computer programs perform ANOVA work by changing the problem into a regression problem, producing the regression results, and then reporting the results in ANOVA terms. The F test used in ANOVA to test for significant main effects and interactions is, of course, the same as that in regression. Although most textbooks and computer programs treat analysis of variance and regression as separate topics, ANOVA is itself a form of regression. The effect of a particular drug and the effect of being of a particular strain were represented separately, with the drug effect constant regardless of strain and the strain effect constant regardless of drug.
