A Problem in Data Analysis: Implications for Organizational Behavior Research

Abstract

Research testing the relative predictive accuracy of various combinations of variables in three areas of organizational behavior research is reviewed. Failures to obtain clear results favoring specific combinations in the three areas may not be separate problems, but rather manifestations of a single, more general one—the inability of widely used research methods and analytical techniques to detect differences in predictive accuracies under conditions common in organizational behavior research.