Introduction to Statistical Mediation Analysis (Multivariate Applications Series)

This volume introduces the statistical, methodological, and conceptual aspects of mediation analysis. Applications from health, social, and developmental psychology, sociology, communication, exercise science, and epidemiology are emphasized throughout.

There are five parts to the book. Part 1 covers definitions, history, and applications for the mediation model. In Part II the conceptual model described in the first part of the book is quantified in the estimation of mediation in traditional single mediator models. This part of the book describes the estimation of mediation effects including assumptions of the methods, statistical tests, and construction of confidence limits for the mediated effect. Part III describes advanced mediation models with new chapters describing collider, confounder, and mediator third-variable models,the potential outcomes causal mediation model, and Bayesian mediation methods. Part IV describes the traditonal and then the potential outcomes approach for two or more mediators, categorical variables, structural equation modeling, measurement, multilvel models, and longitudinal mediation models. Part V, discusses general issues in the investigation of mediation including experimental designs, additional approaches to identifying mediating variable, guidelines for planning and conducting a mediation study, and possible future directions.

The primary audiences for this book are advanced undergraduate students, graduate students, and researchers from many substantive areas. The first type of reader will consist of persons skilled in a substantive area but less experienced with statistical methods. Researchers planning a study where mediating mechanisms are investigated also fall in this group. The second type of reader is sophisticated statistically but without much exposure to mediation analysis.

David P. MacKinnon is a Foundation and Regents Professor of psychology at Arizona State University. He is a globally recognized leader in quantitative psychology with primary substantive interest in prevention science and methodology for understanding how interventions achieve effects.