A realist alternative to meta-analysis: Two papers

Abstract

Meta-analysis (the statistical combination of a set of studies in a given area, with the aim of establishing an overall or average effect of something) is increasingly common in work and organization studies. Critiques of meta-analysis are now common. There is also a well-known alternative based in realism. The purpose in bringing together the two papers by Nielsen and Miraglia and by Vincent and colleagues is not to rehearse the critiques or simply explain realism or realist evaluation. The two papers certainly perform these functions in setting out problems with meta-analysis and also identifying when and to what extent it remains valid. The goal, however, is to move forward by showing what a realist synthesis would look like and illustrating how it works. Vincent and colleagues lay out the principles, while Nielsen and Miraglia take the case of intervention studies to show how realist evaluation works.