Abstract
Consider the estimating equation proposed by Lai and Ying, based on thelinear rank test, we show the existence of local confounding between thehazard function and the covariates when the hazard function goes up and downalong the time scale. This phenomenon further indicates significant loss ofefficiency of Lai and Ying's estimating equation. To take care of this localconfounding without using nonparametric estimates of the derivative of thehazard function, we propose the empirical process approach (EPA) based on anempirical process constructed from Lai and Ying's linear rank estimatingequation by varying its trimming time point. The EPA will provide verytractable estimations of the regression parameters as well as Pearson'schi-square statistics for testing the model's assumptions.