Monotonicity of effect sizes: Questioning kappa-squared as mediation effect size measure.

Abstract

Mediation analysis is important for research in psychology and other social and behavioral sciences. Great progress has been made in testing mediation effects and in constructing their confidence intervals. Mediation effect sizes have also been considered. Preacher and Kelley (2011) proposed and recommended κ² as an effect size measure for a mediation effect. In this article, we argue that κ² is not an appropriate effect size measure for mediation models, because of its lack of the property of rank preservation (e.g., the magnitude of κ² may decrease when the mediation effect that κ² represents increases). Furthermore, κ² can lead to paradoxical results in multiple mediation models. We show that the problem of κ² is due to (a) the improper calculation of the maximum possible value of the indirect effect, and (b) mathematically, the maximum possible indirect effect is infinity, implying that the definition of κ² is mathematically incorrect. At this time, it appears that the traditional mediation effect size measure PM (the ratio of the indirect effect to the total effect), together with some other statistical information, should be preferred for basic mediation models. But for inconsistent mediation models where the indirect effect and the direct effect have opposite signs, the situation is less clear. Other considerations and suggestions for future research are also discussed.

A novel measure of effect size for mediation analysis.

“…For illustrative purposes and without loss of generality, variables will be assumed to be standardized with unit variance. Two of these limitations were identified by Wen and Fan (2015). The first is that 2  is not a monotonic function in raw or absolute value of the indirect effect.…”

Consumer Preferences for Local Food: Testing an Extended Norm Taxonomy

“…When it comes to reporting effect size, there is ambiguity in the scientific literature. Wen and Fan [67] stated that the traditional measure of the mediation effect size, P M (the ratio of the indirect effect to the total effect), together with some other statistical information, should be given for basic mediation models. However, Hayes and Rockwood [65] advised against this as it has no upper and lower bound; instead, they recommended the partially standardized indirect effect (defined as the indirect effect divided by the standard deviation of the independent variable).…”

Working mechanisms of a web-based self-management intervention for cancer survivors: a randomised controlled trial

“…Þ , as this measure of effect size is preferred for basic mediation models (Wen & Fan, 2015). In addition, simple mediation analyses were conducted to assess the influence of each hypothesised mediator on fatigue and depression separately.…”