Comparing Effect Sizes and their Confidence Intervals: A Primer on Equivalence Testing in Educational Research

Abstract

This paper’s objective was to teach the Equivalence Testing applied to Educational Research to emphasize recommendations and to increase quality of research. Equivalence Testing is a technique used to compare effect sizes or means of two different studies to ascertain if they would be statistically equivalent. For making accessible Equivalence Testing, this technique was explained with two examples by conducting manual calculations, using an online calculator, the software R, the software SPSS, and a t table. Furthermore, the software R with an Equivalence Testing code was used, and its results were graphed and discussed with details. Among other recommendations given, Equivalence Testing can be a useful tool for comparing means and effects within certain bounds that could hopefully imply a practical significance to provide meaning to findings. The results of Equivalence Testing can indicate that two treatments´ effects are statistically equivalent or not. Thus, the Equivalence Testing can be a channel to replicate studies and observe if there is a possible pattern regarding the appearance of a phenomenon.

Keywords

• confidence interval
• cohen’s d
• effect size
• equivalence testing
• practical significance