Journal of Data Science

Although hypothesis testing has been misused and abused, we argue that it remains an important method of inference. Requiring preregistration of the details of the inferences planned for a study is a major step to preventing abuse. But when doing hypothesis testing, in practice the null hypothesis is almost always taken to be a “point null”, that is, a hypothesis that a parameter is equal to a constant. One reason for this is that it makes the required computations easier, but with modern computer power this is no longer a compelling justification. In this note we explore the interval null hypothesis that the parameter lies in a fixed interval. We consider a specific example in detail.

Abstract: Existing methods on sample size calculations for right-censored data largely assume the failure times follow exponential distribution or the Cox proportional hazards model. Methods under the additive hazards model are scarce. Motivated by a well known example of right-censored failure time data which the additive hazards model fits better than the Cox model, we proposed a method for power and sample size calculation for a two-group comparison assuming the additive hazards model. This model allows the investigator to specify a group difference in terms of a hazard difference and choose increasing, constant or decreasing baseline hazards. The power computation is based on the Wald test. Extensive simulation studies are performed to demonstrate the performance of the proposed approach. Our simulation also shows substantially decreased power if the additive hazards models is misspecified as the Cox proportional hazards model.