Which Effect Size Calculation is the Best to Estimate the Population Effect Size in the Welch T Test?

Abstract

The purpose of this study is to use Monte Carlo method to detect the most precise and least biased effect sizes calculations in a variety of conditions. The results show that there is no big difference to obtain effect sizes of using mean difference or trimmed mean difference as denominator. Cohen’s dA proves to be the less unbiased but more precise across all the conditions in Welch t test. It is worthwhile to notice that Hedges’ g remains the same as Cohen’s dP across all the conditions of Welch t test. When group sample sizes are equal, no matter which population effect size formula are applied, Cohen’s dA, Cohen’s dP, and Hedges’ g are the same estimates given the bias statistics.