No label defined (Q4512128)

The seminal paper by \textit{D. R. Cox} [J. R. Stat. Soc. Ser. B 34, 187-220 (1972; Zbl 0243.62041)]\ on hazard modelling that includes a baseline hazard, which is called henceforth hazard-based models, opened a new era of statistical methodology for studying time-to-event data. Since then much work has been done on improving hazard-based models to obtain more accurate inference for lifetime data. In this paper it is pointed out that some simple graphical procedures can be used to suggest the level of complexity which is required in a model for time-to-event data. The procedures allows easy model assessment and can be easily incorporated and used systematically in any exploratory analysis of time-to event setting. Emphasis is given to model building and comparison between hazard-based models rather than to estimation procedures, which were addressed by the author [Lifetime Data Anal. 3, 367-381 (1997; Zbl 0910.62095)]. It is verified whether there exits heteroscedasticity-across-individuals with different covariate levels in a particular data set, whether it is possible to describe it through a simple function of the covariates and the importance of taking it into account for model fit. In order to keep the procedure straightforward least-squares estimates are considered.