Poster #32 - Beyond Testing Measurement Invariance: Investigating Sources of Heterogeneity With Three-Step Multilevel Factor Mixture Modeling

• In Event: 21.063 - MTCC Poster Session Three
  In Poster Session: 21.063-3 - Differential Item Functioning and Measurement Invariance

Abstract

Testing measurement invariance with a large number of groups is methodologically challenging and measurement invariance (MI) is often not supported in cross-national comparative studies. We propose the 3-step multilevel factor mixture modeling (ML FMM) to test MI across many groups and furthermore to model predictors of latent class membership that possibly induce measurement noninvariance. This Monte Carlo simulation study shows the adequacy of 3-step ML FMM regarding class assignment accuracy as well as the correct MI detection rates and class enumeration rates based on a new information criterion for large data. The performance of 3-step approach is also acceptable with well controlled Type I error and adequate power of the covariate effect.

Authors

• Eunsook Kim, University of South Florida

• Yan Wang, University of Massachusetts - Lowell