The Prevalent Misuse of Fisher’s Partial Eta Squared Formula

The recording of an estimate of effect size is an essential tool for empirical science because it allows for statistical power. In addition, it enables researchers to replicate studies because it assists in choosing subject amounts effectively. A popular measure of effect size is partial eta squared and is often calculated using Fisher's formula. Despite the positive impact that partial eta provides to empirical researchers, it comes with two problems. One is that researchers are misusing this formula because it was initially made for between-subject designs. When measuring the effect size via partial eta squared in a between-subject design, it measures by means of the ratio of variance related to an effect, and that effect added to its associated error variance. It works specifically for between-subjects because the values are independent of any other aspects of the design. However, researchers have been using it for the past decade on repeated measures designs, which do not have independent values. In this thesis, I will examine how often studies report effect size, which measurement used to estimate effect size, and which subject design they apply it to. I analyze various articles from psychological journals in their latest December 2019 edition to see how many researchers continue to use Fisher’s partial eta squared.