Handling Nonorthogonal Analysis of Variance

Abstract

Nonorthogonal analysis of variance has received considerable attention in recent years resulting in numerous articles advocating a number of nonequivalent approaches. These approaches are briefly summarized along with the arguments for and against each. It is suggested that each technique is appropriate for certain situations, but that the comprehensive strategy presented by Appelbaum and Cramer (1974) is perhaps best for the majority of nonorthogonal analyses.