What Moderates Moderators? A Meta-Analysis of Interactions in Management Research

Continuous moderators (i.e., interactions) are some of most frequently hypothesized and tested relationships in organizational behavior human resource management and applied psychology, but there is relatively little information on their average magnitudes and conditions under which they vary. We meta-analyzed 15 years of continuous moderators published in four of our leading journals and found that moderator effects are relatively small (f 2 = .030) and subsequently, the majority of studies examining continuous moderators are statistically underpowered. We found that a number of factors including level of analysis, outcomes examined, and number of variables in the model influence moderator effect sizes. In addition, the strong and negative correlation between sample size and effect size (r = -.490) and the strange behavior of p-values around the .05 threshold of statistical significance are suggestive of a small sample bias. Finally, we provide a 2- step process that allows researchers, reviewers, and ed...