Identifiability and Estimation in Generalized Linear Models with Nonignorable Missing Data

• Activity Number: 584
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Jiwei Zhao*+ and Jun Shao
• Companies: and University of Wisconsin
• Keywords: nonignorable ; identifiability ; pseudo-likelihood ; missing mechanism

Abstract

In this talk, we consider statistical models with nonignorable missing data. Without any further assumption, unknown parameters may not be identifiable when the missing mechanism is nonignorable. We develop a pseudo likelihood method under nonignorable missingness without specifying the form of the missing mechanism. For the identifiability issue, We derive explicit conditions for generalized linear models. We establish asymptotic theory for this pseudo likelihood based estimator and provide an efficient algorithm to handle the computational challenge. We illustrate the method through the analysis of HIV-CD4 data and some simulation studies. This is joint work with Dr. Jun Shao, University of Wisconsin, Madison.