Testing for predicted patterns: When interest in the whole is greater thanin some of its parts

This article provides a general rationale and step-by-step procedures for s tatistically assessing a theory-based or predicted pattern in a set of K > 2 means. A coherent hypothesis-testing strategy that incorporates Type I er ror and power considerations is an integral part of the process. Some speci fic advantages that are conferred by predicted pattern testing, relative to commonly adopted alternatives, include its (a) scientific parsimony; (b) p recision, with respect to getting to the heart of a researcher's hypothesis ; (c) sensitivity, as reflected by greater statistical power and which incl udes applicability to experimental factors associated with small "piecewise " effects; and (d) ability to lend varying degrees of statistical support t o a researcher's hypothesis. Both a hypothetical example and a published ap plication of predicted pattern testing are provided to demonstrate its use and usefulness.