Prior Sensitivity of Bayesian Structural Equation Modeling Fit Indices With Model Misspecification

• In Event: Bayesian Techniques in Structural Equation Modeling

Abstract

The purpose of this study is to systematically evaluate the sensitivity of Bayesian structural equation modeling (BSEM) fit indices to prior choices and model misspecification. The Bayesian fit indices evaluated include BRMSEA_ppmc, BRMSEA_DevM, BCFI, BTLI, BNFI, Bgamma_hat, and BM_C. Models are misspecified by omitting cross-loadings or constraining latent factor correlations. Various prior specifications are considered that include diffuse, aligned, and divergent priors. As prior sensitivity of these fit measures with misspecified models has not received adequate attention, this study aims to fill this research gap.

Authors

• Ejike J. Edeh, University of Arkansas
  Presenting Author

• Xinya Liang, University of Arkansas
  Presenting Author

• Chunhua Cao, University of Alabama
  Presenting Author