Xianzheng Huang

Department of Statistics
216 LeConte College
1523 Greene Street
University of South Carolina
Columbia, SC 29208

Office: LeConte 200F
Phone: (803) 777-5110
E-mail: huang@stat.sc.edu



Research Statement

Key phrases: Measurement error; Latent variables; Model misspecification; Nonparametric statistics


Data prone to measurement error arise in a wide range of applications. My longstanding research endeavor is studying effects of measurement error on statistical inference. This line of investigation motivates new strategies for drawing inference that adequately account for measurement error.


Measurement error models fall under the big umbrella of latent variables models, which are especially susceptible to model misspecification. I have studied properties of inference results associated with latent variable models in the presence of model misspecification. Findings here lead to diagnostic methods for detecting inadequate model assumptions.


The vulnerability of parametric inference to model misspecification motivates my interests in nonparametric statistics. In particular, I am interested in developing nonparametric methods for mean regression, mode regression, and density estimation in the presence or absence of measurement error.




Advanced Statistical Inference; Latent Variable Models; Linear Statistical Models; Nonlinear Statistical Models; Mathematical Statistics (I & II); Introduction to Statistical Theory (I & II, for Distance Learning); Statistical Methods (I); Statistics for Engineers; Elementary Statistics.