Evaluation of Heterogeneity and Heterogeneity Interval Estimators in Random-Effects Meta-Analysis of Educational and Psychological Research

• In Event: 43.072 - Learning About Meta-Analysis and Meta-Analyses About Learning: Methods and Applications of Research Synthesis

Abstract

The random-effects meta-analysis model is based on the assumption that a distribution of true effects exists in the population, often assumed normal. The population variance, also called heterogeneity, and its interval can be estimated numerous ways. We compared 16 estimators of heterogeneity (Bayesian and non-Bayesian) with regard to bias and mean square error and associated heterogeneity interval estimators over a broad set of conditions found in educational and psychological meta-analyses. Three conditions were varied: (a) sample size per meta-analysis, (b) true heterogeneity, and (c) sample size per study within each meta-analysis. No estimator was uniformly superior, but without knowledge of heterogeneity the Paule and Mandel estimator with its interval estimated using the Jackson method are recommended for combined use.

Authors

• Peter Boedeker, Baylor College of Medicine

• Robin K. Henson, University of North Texas