A Semiparametric Estimation of Mean Function with Nonignorable Missing Data

• Activity Number: 658
• Type: Topic Contributed
• Date/Time: Thursday, August 5, 2010 : 10:30 AM to 12:20 PM
• Sponsor: IMS
• Author(s): Cindy Long Yu*+
• Companies: Iowa State University
• Keywords: Exponential tilting ; Not missing at random ; Nonparametric regression

Abstract

Parameter estimation with non-ignorable missing data is a challenging problem in statistics. Fully parametric approach for joint modeling of the response model and the population model can produce results that are very sensitive against the failure of the assumed model. In this paper, based on the exponential tilting model, we propose a semi-parametric estimation method of mean function with non-ignorable missing data. The parametric component is obtained by assuming a logistic regression model for the response probability and the non-parametric component is obtained by a nonparametric regression approach for missing data considered in Cheng (1994). By adopting a nonparametric part of the model, the estimation method can be made robust. Results from a limited simulation study and from an empirical case study are presented.