CAVEAT-EMPTOR - RANK TRANSFORM METHODS AND INTERACTION

When distributional assumptions for analysis of variance are suspect, and nonparametric methods are unavailable, ecologists frequently emplo y rank transformation (RT) methods. The technique replaces observation s by their ranks, which are then analysed using standard parametric te sts. RT methods are widely recommended in statistics texts and in manu als for packages like SAS and IMSL. They are robust and powerful for t he analysis of additive factorial designs. Recently, however, RT metho ds have been found to be grossly inappropriate for use with non-additi ve models. This severe limitation remains largely unreported outside o f the theoretical statistics literature. Our goal is to explain this s hortcoming of RT methods.