On fitting Cox's proportional hazards models to survey data

摘要: SUMMARY Binder (1992) proposed a method of fitting Cox's proportional hazards models to survey data with complex sampling designs. He defined the regression parameter interest as solution partial likelihood score equation based on all values population under study, and developed heuristically procedure estimate corresponding variance. In this paper, we provide formal justification Binder's method. Furthermore, present an alternative approach which regards random sample from infinite universe accounts for randomness in statistical inference. Under approach, retains its original interpretation log hazard ratio, conclusion applies other populations. The related problem survival function estimation is also studied.