David P. MacKinnon, Arizona State University

The investigation of mediating mechanisms is a measurement challenge. Progress is made as the mediating variable process is more precisely measured. In this view, researchers conduct studies to distill or uncover the most important facet of a construct that transmits the causal effect of an independent variable to a dependent variable. Examples in chemistry and genetics illustrate how improved measurement of unobservable processes helped researchers identify critical mediating mechanisms. In psychology, a variety of methods are used to measure mediating processes including self-report and biological measures. A strength of the application in psychology is that humans can report information about their feelings, states, and behaviors. A weakness is that humans are susceptible to response tendencies and other factors that increase error and bias in measurements. The goal of the presentation is to discuss measurement approaches for investigating mediating variables and to review the measurement literature on how quality of measurement affects mediation analysis. Topics include the influence of confounding variables, use of alternative measurement models, and measurement overlap on mediation analysis.  I discuss future methodological and substantive directions in the measurement of mediating processes including programs of research that include measurement of a mediating construct as a central goal.

about the speaker

David P. MacKinnon is a Regents Professor in the Psychology Department at Arizona State University. He earned a Ph.D. in measurement and psychometrics from UCLA in 1986.  He received the 2007 ASU Outstanding Graduate Mentor Award and the 2021 APA Division 5 Jacob Cohen Teaching and Mentorship award.  In 2011, he received the Nan Tobler Award from the Society for Prevention Research for his 2008 book on statistical mediation analysis. Dr. MacKinnon has been principal investigator on many federally funded grants and has had a National Institute on Drug Abuse grant to develop and evaluate methods to assess mediation since 1990. He received the MERIT award for this mediation research.  He is a Fellow of the Association for Psychological Science, APA Division 5, and Society for Prevention Research.  He is past president of APA Division on Quantitative and Qualitative methods and the Society for Multivariate Experimental Psychology. Dr. MacKinnon has wide-ranging interests in statistics and methodology, but his primary interest is in statistical methods to assess how prevention and treatment programs achieve their effects.