Software Review Review: mediation Package in R Adam C. Sales University of Texas College of Education Causal mediation analysis is the study of mechanisms—variables measured between a treatment and an outcome that partially explain their causal rela- tionship. The past decade has seen an explosion of research in causal mediation analysis, resulting in both conceptual and methodological advancements. However, many of these methods have been out of reach for applied quantitative researchers, due to their complexity and the difficulty of implementing them in standard statistical software distributions. The mediation package in R provides a set of simple commands that execute some of the newer causal mediation methods. This article will summarize some of the recent advances in mediation analysis, critically review the mediation package, and demon- strate, by example, some of its capabilities. Keywords: causal inference; mediation; computational tools 1. Introduction The primary role of statistical causal inference in policy studies is to estimate the effects of interventions, treatments, and general causes. But esti- mating cause and effect does not satisfy the scientific mind and should not satisfy policy studies either. For both scientific and practical reasons, research- ers need to know how a treatment caused its effect. This is the realm of statis- tical mediation analysis. The past decade has seen an explosion into research in methods for statistical mediation analysis based on solid conceptual and statistical footing. In particular, methodologists such as VanderWeele (2008, 2014, 2015), Hong (2010, 2015), and a group of scholars including Imai, Keele, Tingley, and Yamamoto (2009) have adapted the Neyman–Rubin Causal Model (Rubin, 1978) to studies of mediation, and the result has included advances in conceptual clarity, a suite of new statistical methods, and a better understanding of what causal mediation analysis can and cannot do. The latter group has written mediation, a software package for the R statistical environment (R Development Core Team, 2011) that executes the many of the new methods they have developed. In this article, I will summarize some of the recent advances in mediation analysis and review the mediation package. I will demonstrate, by example, Journal of Educational and Behavioral Statistics Vol. XX, No. X, pp. 1–16 DOI: 10.3102/1076998616670371 # 2016 AERA. http://jebs.aera.net 1 as doi:10.3102/1076998616670371 JOURNAL OF EDUCATIONAL AND BEHAVIORAL STATISTICS OnlineFirst, published on November 2, 2016 at Harvard Libraries on November 12, 2016 http://jebs.aera.net Downloaded from