Large Sample Multivariate Procedures for Comparing and Combining Effect Sizes Within a Single Study

Publisher Website
Google Scholar
Abstract
Educational research typically involves multiple measures on two or more independent groups of students. Results of such studies can be summarized by measures of treatment effects. The purpose of this paper is to present large sample multivariate methods for estimating and comparing effect sizes across independent samples within a single study. Procedures are described on how to pool treatment effects to determine an overall study effect size for treatments with similar effects. These proposed procedures also are recommended for use before undertaking a metaanalysis when some of the studies have multivariate measures of the same construct.