A random effects model for effect sizes.

This chapter presents an analogue to random effects analysis of variance for effect size analyses. This model assumes that the population values of the effect size are samples from a distribution of effect size parameters. Thus, the observed variability in sample estimates of effect size is partly because of the variability in the underlying population parameters and partly because of the sampling error of the estimator about the parameter value. The chapter explains a method for obtaining unbiased estimates of the variance of population effect sizes, that is, the parameter variance component. It also presents a statistical test of the hypothesis that the variance in population effect sizes is zero. The chapter discusses the estimation of the mean of the effect size distribution and empirical Bayes estimation procedures for random effects models.