Colloquium

The Department of Statistics
presents

Anthony Rossini

Statistics Department
University of South Carolina

Variance and Variability for Parameter Estimates with Semiparametric Regression Models for Interval Censored Failure Time Data

Abstract

Much work has recently been done to demonstrate the efficiency of various methods for estimating regression parameters of semiparametric models with interval censored data. These methods are interesting, in the sense that the underlying nuisance parameter can not be estimated at the standard root-n rate. We discuss some issues with using these estimators for data analysis, including some progress related to open problems in link selection and variance estimates.