Assessing the Efficacy of Model Fit Indices in Multilevel SEM Using Standard Method (Poster 8)

• In Event: AERA Poster Session Saturday 3:20 p.m.
  In Poster Session: Multilevel Factor Modeling: On the Performance of Fit Indices

Abstract

The performance of model fit indices in multilevel structural equation modeling (MSEM) was assessed using a standard method. A Monte Carlo study evaluated common fit indices under varying conditions: sample size (SS), intraclass correlation coefficient (ICC), and specification model (SM). Results showed high convergence rates (CR) across all conditions, with a 100% rate for correctly specified (CSM), measurement misspecified (MMM), and structure misspecified (SMM) models. The χ^(2 )statistic effectively indicated the correctly specified CSM with a 6.2% rejection rate, while CFI and RMSEA retained the CSM with 100% accuracy. Both χ², CFI, and RMSEA accurately identified MMM with 100% rejection rates. However, while χ^2 correctly identified the misspecified SMM, CFI, and RMSEA inaccurately retained SMM with a 0% rejection rate.

Author

• Noor Alibrahim, University of Denver