A comparison of the effect size estimators in meta-analysis

Abstract

The objective of a meta-analysis is usually to estimate the overall treatment effect and make inferences about the difference between the efffects of the two treatments. This article presents several forms of effect size estimators and compares these effect size estimators and the variance of overall treatment effect estimator within each group. as outcome measures, standarized difference is considered. Four modes of effect size estimators are discussed. Effect size estimators by Glass, Hedges, The Maximum Likelihood, and Shrunken Estimators of Effect Size are employed in this study. Finally, with the help of a software the results of these four effect size estimators are discussed. Estimators are illustrated using a comparison of the effectiveness of amlodipine and placebo on work capacity.