Comparison between Multivariate Fixed-Effects and Mixed-Effects Meta-Analytic Approaches

The use of different multivariate meta-analytic approaches to quantitatively combine and aggregate effect sizes from different primary studies can have different important implications for the cumulative research findings and research-based conclusions. Multivariate fixed and mixed effects (multivariate random-effects regression) models are examples for such approaches. Gaining knowledge of the differences between these various approaches help meta-analysts to accurately choose between them based on the statistical and sampling assumptions of the two models. Therefore, the purpose of this study is to compare the multivariate parameter estimates and their standard errors from applying these two multivariate approaches. The comparison findings show that the multivariate fixed-effects model results yielded smaller standard errors for multivariate mean effect sizes and the moderator variables than the multivariate mixed-effects. Also, multivariate fixed-effects yielded much smaller effect (regression coefficient) of the moderator variable.