Abstract
A confidence interval for effect sizes provides a range of plausible population effect sizes (ES) that are consistent with data. This article defines an ES as a standardized linear contrast of means. The noncentral method, Bonett’s method, and the bias-corrected and accelerated bootstrap method are illustrated for constructing the confidence interval for such an effect size. Results obtained from the three methods are discussed and interpretations of results are offered.
DOI
10.22237/jmasm/1383278640
Included in
Applied Statistics Commons, Social and Behavioral Sciences Commons, Statistical Theory Commons
