Meta-analysis has proved increasingly popular in management and organization studies as a way of combining existing empirical quantitative research to generate a statistical estimate of how strongly variables are associated. Whilst a number of studies identify technical, procedural and practical limitations of meta-analyses, none have yet tackled the meta-theoretical flaws in this approach. We deploy critical realist meta-theory to argue that the individual quantitative studies, upon which meta-analysis relies, lack explanatory power because they are rooted in quasi-empiricist meta-theory. This problem, we argue, is carried over in meta-analyses. We then propose a ‘critical realist synthesis’ as a potential alternative to the use of meta-analysis in organization studies and social science more widely.