Using Joint Criteria for Class Enumeration in Growth Mixture Modeling

• In Event: Roundtable Session 67
  In Roundtable Session: Mixture and Growth Mixture Models

Abstract

Growth mixture modeling has been widely applied in the social and behavioral sciences, but class enumeration in such models remains a challenging issue and relies on a number of individual model fit indices. Previous studies have tried to find the most efficient single index to determine the number of latent classes. The current simulation study proposed using joint model fit indices for this purpose. Preliminary results indicate some of the joint-indices could have higher rate of accuracy than those single indices. More conditions and combinations of model fit indices will to be explored in this research.

Authors

• Min Liu, University of Hawaii - Manoa

• Gregory R. Hancock, University of Maryland