The Effects of Estimation Methods in Multilevel Structural Equation Modeling With Unequal Family Groups and Small Samples

• In Event: Estimation Accuracy and Conditions for Complex Latent Variable Models

Abstract

Family data are common in social and educational research. Multilevel structural equation modeling (MSEM) has become a widespread analytic technique for nested data analysis. Approaches such like MUML, ML and Robust-ML can be used for MSEM of family data. However, practices associated with the employing of MSEM with small unequal family groups using different methods are problematic due to the lack of empirical research. Thus, a Monte Carlo simulation was employed here to investigate the effects of estimation methods in multilevel SEM with unequal family groups. Results indicates the within-group model poses no problems, however, for the between-group model, the estimation methods, number of family groups, ICC and family size all have clear effects on the performance of factor models.

Authors

• Yu Su, University of Florida

• Walter L. Leite, University of Florida