© 1998 by Biometrika Trust
A note on nonparametric estimators of the bivariate survival function under univariate censoring
Division of Biostatistics, School of Public Health, Columbia University 600 West 168th Street, New York, New York 1032-3799, U.S.A.wyt{at}fertig.cpmc.columbia.edu
Fred Hutchinson Cancer Research Center 1124 Columbia Street, Seattle, Washington 98104, U.S.A.johnc{at}swog.fhcrc.org
A simple nonparametric estimator is proposed for the underlying bivariate distribution of paired failure times in the presence of univariate right-censoring. The estimator is shown to be a particular path-dependent estimator which is a generalisation of an estimator proposed by Campbell & FÖldes (1982). Under univariate censoring, this path-dependent estimator has the smallest variance among all path-dependent estimators, which includes the estimator proposed by Lin & Ying (1993).
Key Words: Asymptotic Bivariate survival function Nonparametric Univariate censoring