The statistical analysis of truncated and censored data under serial dependence

The subject of this work is estimation of two dimensional distribution function resp. integral w.r.t. such distribution function and the statistical analysis of this estimator in case when the data are incomplete. The variables of interest are lifetimes, where the first one is at risk of being truncated and the second can be censored. Additionally, the variables may be dependent. For example in the analysis of AIDS one often encounters data which are observed sequentially over time. In such a case the first variable can be the time elapsed between infection and seroconversion and the second may be the time between seroconversion and AIDS. A big part of the work deals with the Lynden-Bell estimator for truncated data and its proparties.