THE STUDY OF EFFECTS IN MANOVA

A particular strategy for investigating effects resulting from a MANOVA [multivariate analysis of variance] is proposed. The strategy involves multiple 2-group multivariate analyses. The 2 groups result from considering multivariate pairwise group contrasts or multivariate complex group contrasts. Assuming a given 2-group analysis yields real effects, the resultant single linear discriminant function (LDF) may be studied. A rationale based on a transformation of LDF weights is recommended for assessing variable relative contribution. The analysis strategy is described in detail and illustrated with real data sets.