The Effect of Estimation Methods on Bifactor Models with Ordinal Responses

• In Event: Electronic Board Session IV

Abstract

This simulation study aims to systematically examine the statistical performance of estimators in bifactor models when ordered categorical responses are present. Results show that WLSMV outperformed ULSMV and MLR in statistical power, but MLR resulted in higher convergence rates and smaller bias.
Keywords: Bifactor model, WLSMV, ULSMV, MLR, ordinal data

Authors

• Xinchang Zhou, UIUC

• Bo Zhang, UIUC

• Yan Xia