Interpreting and evaluating meta-analysis

This article offers some guidelines for interpreting and evaluating meta-analytic reviews of research. The fundamental goals of meta-analysis are to combine results across studies to yield an overall estimate of effect and to compare effects between studies in order to understand moderating factors. Suggestions are made for what readers should look for in a meta-analysis, and a discussion is provided of several issues that are not often explicitly addressed: choice of unit of analysis, fixed versus random effects, the meaning of heterogeneity, determination of when contrasts are appropriate, and the choice of measure of central tendency. We recommend that readers adopt a skeptical attitude about the results of meta-analysis, particularly when only complex analyses are reported.