For more information on the published version, visit Science Direct's Website. https://www.sciencedirect.com/referencework/9780080970875/international-encyclopedia-of-the-social-and-behavioral-sciences

Xiaolu Wang is professor of International Business and Management at Dickinson College.

Wang, Xiaolu. Mediation, Statistical. In International Encyclopedia of the Social and Behavioral Sciences, 2nd ed., edited by James Wright, 75-80. Elsevier Science, 2015.

The goal of causal mediation analysis is to investigate the extent to which the causal effect of a ‘treatment’ variable on an outcome variable is propagated through a third (or more) variable – the mediator of interest. The focus is to partition the total effect into an indirect effect and a direct effect. After defining various relevant causal effects using potential outcomes, four approaches of causal mediation analysis are addressed, including linear structural equation models, linear counterfactual models with potential errors, marginal structural models, and principal stratification.