Unfair comparisons.

Suggests that when organizational researchers competitively test 2 or more theories, or compare the predictive or relational strength of 2 or more factors or variables, the results can lead them to conclude that one theory or variable is stronger than the other(s). It is argued that such conclusions can be insensitive to a potential alternative explanation of the results: The stronger theory, factor, or variable was favored by being more strongly operationalized, manipulated, or measured. This can render the comparative study into a test of a poorly framed empirical question, the results of which are at least partial artifacts of the procedures, manipulations, or measures used. Four illustrations of this alternative explanation of comparative results are drawn from recent manipulation-based (disadvantaged theories due to procedural nonequivalence and unclear comparisons of factors due to unassessed distributional equivalence) and measurement-based (disadvantaged variables due to procedural nonequivalence and to procedural and distributional nonequivalence) studies of organizational behavior. Generic procedures are described than can make for fairer comparisons. Some possible points of misunderstanding regarding the authors' arguments (e.g., all differential support is artifactual) are addressed. (23 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)