Robust Asymptotic Analysis for Mean and Covariance Structure Model

Mean and covariance structure model is widely applied in behavioral, educational, medical, social and psychological research. The classic maximum likelihood estimate is vulnerable to outliers and distributional deviation. In this paper, robust estimate based on minimizing the objective function is proposed, and M-ratio test based on the robust deviance is suggested to assess the model fit. Empirical results are illustrated by a real example.