The Mediation Myth
Abstract
The mediation myth is the false belief that mediation is actually estimated in the typical mediation analysis. This myth is based on a trifecta of shortcomings: (1) the typical mediation study relies on an inadequate design; (2) the researcher uses a flawed analysis strategy; and (3) there is scant attention to assumptions that are required when estimating mediation. These problems stem from overgeneralizing the classical product method for estimating mediation and overreliance on statistical significance testing as a decision criterion in mediation analysis. The goals of this article are to (1) raise awareness of these difficulties among researchers and (2) provide a roadmap about design and analysis options for a more rigorous and scientifically valid approach to mediation analysis.