July 2006

IZA DP No. 2198: Fixed and Mixed Effects Models in Meta-Analysis

published in: J. Osborne (ed.), Best Practices in Quantitative Methods. Sage, 2007

The last three decades the accumulation of quantitative research evidence has led to the development of systematic methods for combining information across samples of related studies. Although a few methods have been described for accumulating research evidence over time, meta-analysis is widely considered as the most appropriate statistical method for combining evidence across studies. This study reviews fixed and mixed effects models for univariate and multivariate meta-analysis. In addition, the study discusses specialized software that facilitates the statistical analysis of meta-analytic data.