Semiparametric Models with Covariates for Lifetime Data under a General Censoring Scheme with an Application to Contingent Valuation

Abstract

We are interested in estimating the distribution of lifetimes, also called survival times, subject to a general censoring scheme called ``middle censoring'' (see Jammalamadaka and Mangalam (2003)). Both the Cox proportional hazards (Cox PH) and accelerated failure time (AFT) models are considered since each model has a baseline distribution function that is modified by the presence of covariates. The key contribution presented is the estimation of the effect of the covariate as well as the baseline distribution function. We conclude with an application to a contingent valuation study.