Appropriate Moderated Regression and Inappropriate Research Strategy: A Demonstration of Information Loss Due to Scale Coarseness

Paunonen and Jackson (1988) demonstrated stepwise moderated regression provides a test interaction effects that protects the nominal I error rate. However, the stepwise procedure also been characterized as failing to detect effects in empirical studies. This issue led to questions regarding the method’s power (Bobko, 1986; Zedeck, 1971) in research. It is demonstrated that, because a research strategy frequently used in empirical , the probability of Type II error in a true interaction effect is unknown. , the number of scale steps used in the dependent variable is shown to in a form of systematic error that can increase or decrease the expected effect of the interaction. The problem is also discussed the context of testing more complex models. for eliminating this problem in research designs are provided. Index terms: loss, interaction effects, Likert scales, regression, response transformation.