The use and interpretation of unstandardized and standardized effect sizes in psychology: current practices and challenges

Abstract Effect sizes quantify the outcome of empirical studies and thus provide the crucial answer to the research question. Unstandardized effect sizes do so in absolute terms, i.e., in terms of the values of the scales used, and are therefore generally easy to interpret. Standardized effect sizes, on the other hand, are a relative piece of information with the advantage that they are free from arbitrary units. However, since many commonly used effect sizes take the variance into account, they carry the risk of being misinterpreted or overinterpreted. In Study 1, the typical use of effect sizes in psychological research was analyzed. Results indicate that although effect sizes are almost always reported, they are most often reported primarily in standardized form, and conventions are most often used to interpret them, most commonly those of Cohen (1988). In Study 2, potential challenges in interpreting standardized effect sizes were analyzed using the example of a mean difference and typical research scenarios for which Cohen’s d was reported. The results suggest that the majority of researchers in psychology attribute more meaning to the standardized effect size than it actually conveys. Therefore, in line with the existing literature, it is recommended that more emphasis be placed on non-standardized effects when interpreting empirical results. In addition, existing conventions for standardized effect sizes should be revised and adapted to psychological subdisciplines or specific research topics.