Nonparametric estimators of effect size in meta-analysis

When observations are normally distributed, the sample standardized mean difference is a consistent and asymptotically efficient estimator of effect size in meta-analysis. However, in some cases the observations are far from normally distributed, and a nonparametric index of effect magnitude is desirable. The logic of the estimator proposed by H. C. Kraemer and G. Andrews (see record 1982-11171-001) is extended to provide related nonparametric estimators of different parameters that may be appropriate under other experimental conditions. (21 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).