Distribution Overlap to Estimate Effects in Meta-Analyses of Group Differences for Students with Learning Disabilities

Abstract

Listen

Translate article

ABSTRACT Between-group measured differences for students with learning disabilities (LD) versus students without LD have been frequently examined in meta-analytic research under the assumption that moderate or large effects can inform more accurate LD identification. An empirical synthesis of meta-analyses was conducted to examine the effect sizes from 19 meta-analyses of measured differences between students with and without LD on achievement, social-emotional/behavioral, cognitive, and physical measures. This study used the effect size estimates to compute the overlap in distributions for all measured variables and found that even with moderate-to-large effects, there was more than 50% overlap between the two distributions. In other words, students identified with LD scored the same as students without LD at least 50% of the time on commonly used cognitive, social-emotional/behavioral, and physical measures. Only measures of achievement (k = 12, 47.26%) resulted in less than 50% overlap. This study examined the moderating effect of disability area and measurement category and found in all cases that the overlap was high and suggested that percentage overlap should be considered when examining group differences between students with and without LD.