Drug Prevention Programs: Unpacking their Mechanisms of Change
Mediation analysis is the statistical method used to investigate how a prevention program changes mediating constructs that are hypothesized to influence designated outcomes. This type of analysis contributes to program evaluation and improvement by identifying the causal intervening mechanism, determining effective and ineffective program components, checking manipulation, and refining the program accordingly.
This book chapter describes examples of mediation that form the backbone of drug prevention theory and quantitative mediation analysis in drug prevention research.
The chapter covers the following topics: estimation of the mediated effect, assumptions and effect size measures, multiple mediator models, multilevel mediation, longitudinal mediation models, person-centered approaches, and categorical outcomes. Finally, recent promising quantitative methods and experimental design approaches to address causal inference in mediation models are described. Future directions for mediation analysis are discussed.
Kisbu-Sakarya, Y., MacKinnon, D.P. & O’Rourke, H. (2015). Strategies for drug program evaluation: Statistical models of mediation for drug program evaluation. In L. M. Scheier (Ed.), Handbook of Drug Abuse Prevention Research, Intervention Strategies, and Practice. Washington, DC: American Psychological Association.