Alternative Estimation Procedures for Event-History Analysis: A Monte Carlo Study

Abstract

The authors compare alternative procedures for estimating parameters of event-history models: ordinary least squares, Kaplan-Meier least squares, maximum likelihood, and partial likelihood. They report results of simulations comparing maximum likelihood and partial likelihood estimators for small samples in terms of several potential sources of error. A related paper is Tuma, Hannan, and Groeneveld (1979).

Description

Keywords

event history models

Citation