ABSTRACT
A meta-analysis involves identifying all the available studies which are relevant to the area being explored. These have to be classified according to their design and the DVs which they have employed. The decision has to be made as to how many meta-analyses will be necessary to describe the area fully. Each meta-analysis can have a combined effect size and a combined probability calculated for it. In addition, the heterogeneity of the effect sizes should be calculated. When heterogeneity of effect size is identified, either the analysis should follow that for a random model, a focused comparison of studies with different characteristics can be conducted or studies should be removed from the meta-analysis until a homogeneous set of studies has been identified. The combined probabilities for all the studies and for the homogeneous set should be reported, as should the combined effect size and confidence intervals for both the complete set of studies and the homogeneous set. All decisions about the inclusion and exclusion of studies should be made explicit in the report of the meta-analysis.
