JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 520
Type: Contributed
Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
Sponsor: Section on Statistics in Epidemiology
Abstract - #305629
Title: A Robust Bayesian Approach to Assessing Average Indirect (Mediation) Effect in Multilevel Models, with Application to a Smoking Study
Author(s): Yisheng Li*+ and Ying Yuan
Companies: MD Anderson Cancer Center and MD Anderson Cancer Center
Address: Department of Biostatistics, Houston, TX, 77030, United States
Keywords: mediation analysis ; indirect effect ; multilevel model ; mixed-effects model ; Dirichlet process ; nonparametric Bayes
Abstract:

Mediation analysis is commonly conducted in the behavioral and social sciences with a goal of assessing whether the effect of an independent variable X on an outcome variable Y is through an intermediate variable M. Two regression models are fit, one being M on X (path a) and the other being Y on both X and M (path b). When X, Y and M all involve repeated measurements within clusters, paths a and b are typically multilevel (mixed-effects) models in which the effects of X on M and M on Y controlling for X are assumed cluster-specific and normally distributed. The normality assumption may result in poor performance when the random effects are not normally distributed (Bauer et al, 2006). We propose a Bayesian nonparametric approach to testing the average indirect effect using a Dirichlet process prior for the distribution of the random effects in both paths a and b. This approach makes use of recent results on the calculation of the moments of the random moments of the nonparametric random effect distribution with a Dirichlet process prior (Li et al., 2011). We evaluate the performance of the proposed approach using simulations, and illustrate the method with a smoking dataset.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program




2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.