Modern approaches to statistical mediation analysis focus on estimation and inference about the indirect and direct effects of putative cause X on presumed effect Y through proposed intervening variable M. To date, all discussions of these approaches have assumed X is either dichotomous or continuous, even though investigators frequently are interested in testing mediation hypotheses involving a multicategorical causal agent (such as two or more experimental conditions relative to a control group). In this paper we describe the estimation of indirect and direct effects in statistical mediation analysis with a multicategorical X. We introduce the concepts of the relative indirect, relative direct, and relative total effect, show how they are estimated in OLS regression and structural equation modeling, and how they are interpreted as a function of how the groups are coded for analysis. We describe inferential tests for these relative and omnibus effects and provide example code and macros for Mplus, SPSS, and SAS that implements the tests. Discussion extends the method we describe to models with multiple intervening variables and models with latent variables.
CITATION STYLE
Hayes, A. F., & Preacher, K. J. (2011). Indirect and direct effects of a multicategorical causal agent in statistical mediation analysis. Manuscript Submitted for Publication, (July), 1–54. Retrieved from http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Indirect+and+Direct+Effects+of+a+Multicategorical+Causal+Agent+in+Statistical+Mediation+Analysis#0
Mendeley helps you to discover research relevant for your work.