Estimating a Three-Level Latent Variable Regression Model With Cross-Classified Multiple Membership Data

• In Event: 65.060 - Multilevel Modeling: Nonnested Multilevel Data, Measurement Invariance, and Bayesian Approaches

Abstract

This study proposes a new model, the cross-classified multiple membership latent variable regression (CCMM-LVR) model for handling student mobility across schools that can enhance the three-level latent variable regression model (HM3-LVR). Using a large-scale longitudinal dataset containing mobile students, CCMM-LVR model estimates were compared with estimates using the HM3-LVR that ignores mobility by only recognizing the first school attended. The impact of ignoring mobility was investigated by comparing parameter estimates, standard error estimates, and model fit indices for the CCMM-LVR and HM3-LVR.

Authors

• Audrey J. Leroux, Georgia State University

• Susan Natasha Beretvas, The University of Texas - Austin