Concerns about drawing causal inferences from meta-analyses: An example in the study of gender differences in aggression.

The present findings indicate that the gender differences in aggression appear to be remarkably stable when changes in study characteristics over time are controlled.

Abstract

Meta-analysis has increasingly been used as an explanatory research tool. The present investigation was designed to illustrate the potential limitations of meta-analysis for making causal inferences. Several meta-analytic investigations have led others to conclude that gender differences are getting smaller over time, however, there has been little concern regarding changes in research methodology over time. The present findings indicate that the gender differences in aggression appear to be remarkably stable when changes in study characteristics over time are controlled. The authors discuss the implications for the use of meta-analytic procedures to make causal inferences and the implications for understanding the causes of gender differences.