Moderator (Sub-group) Analyses

Psychology studies contain numerous sources of bias, making it highly likely that the result of your overall meta-analysis will show signs of heterogeneity between studies. Consequently, meta-analysts often run moderator (sub-group) analyses to try and account for heterogeneity in results. Moderator analyses involve research teams coding methodological or sample characteristics that are potential sources of heterogeneity such as differences between studies in the gap between delivery of the intervention and follow-up measurement, the proportion of men or women in samples, or the type of screening (or other behaviour) results are based on. This chapter will run through some moderator analyses I have reported in my meta-analyses to introduce the issue of moderator analyses in meta-analysis.