Comparing Bayesian and Structural After-Measurement Estimation for Multilevel Structural Equation Models With Latent Interactions

• In Event: Fresh Perspectives on Estimating Complex Multilevel Models: Counts, Mixtures, Interactions, and Machine Learning

Abstract

A flurry of recent developments have expanded the ability of researchers to consider structural equation models with latent interactions. Two promising approaches are Bayesian estimation and structure after measurement (SAM) techniques. Both approaches have been successfully applied in models of varying complexity, in multilevel structural equation models, and with small to moderate sample sizes (e.g., fewer than 100 clusters). However, a comprehensive comparison of these approaches when latent interactions are included in multilevel structural equation models has yet to be completed. To address this gap, we conducted several simulation studies to compare estimator performance with MLSEMs when a within-, between-, or cross-level latent interaction is present. Results map out conditions in which each approach should be prioritized.

Authors

• Kyle T. Cox, University of North Carolina - Charlotte

• Benjamin Kelcey, University of Cincinnati