Moment-type estimators for the proportional likelihood ratio model with longitudinal data

Abstract

Luo & Tsai, Biometrika 99 , 211–22, 2012, proposed a proportional likelihood ratio model and discussed a maximum likelihood method for its parameter estimation. In this paper, we use this model as the marginal distribution to analyse longitudinal data, where the maximum likelihood method is not directly applicable because the joint distribution is not fully specified. We propose a moment-type method that is an extension of the generalized estimating equation method. The resulting estimators are consistent, asymptotically normal and perform well in our simulation study.