Nonlinear Structural Equation Mixture Models in a Multilevel Framework: An Empirical Study

• In Event: 44.078 - Advanced Issues in Structural Equation Modeling

Abstract

Recently, Kelava and Brandt (2014) proposed a general nonlinear multilevel structural equation mixture modeling (GNM-SEMM) framework that accommodates non-normally distributed latent variables and nonlinear effects at both within- and between-cluster levels, and at the same time allows for each of these facets to be modeled separately. Despite its modeling flexibility, to date there has not been any methodological investigation of models within this broad class. This study aims to fill this void via a Monte Carlo simulation.

Authors

• Jeffrey R. Harring, University of Maryland

• Jinwang Zou, University of Maryland - College Park

• Holger Brandt, University of Kansas

• Augustin Kelava, Eberhard Karls Universität Tübingen