Detecting Differential Effects Through Parallel Multiple-Mediator Mixture Models

• In Event: Advancements and Challenges in Cognitive Diagnostic Models and Mixture Modeling

Abstract

This simulation study examined the detection and estimation of differential effect pathways within a parallel mediator mixture model. The results showed that Bayesian Information Criterion (BIC) was superior to other model fit indices in correctly identifying the number of latent classes in the presence of heterogenous mediation. Parameter estimates were unbiased under most simulation conditions when sample size was sufficiently large.

Authors

• Min Liu, Baylor University

• Gregory R. Hancock, University of Maryland