In R, to check if the proportional hazards assumption is met, we can run the cox.zph command to plot the scaled Schoenfeld residuals vs time. But what does zph actually stand for? I'd imagine that the ph stands for proportional hazards, but I am unsure what z stands for.

I want to know what it stands for because

  1. "checking scaled schoenfeld residuals vs time" is a mouth full to say
  2. "running zph" doesn't convey details to somebody unfamiliar with the model
  • $\begingroup$ "the cox.zph function calculates tests of the proportional-hazards assumption for each covariate, by correlating the corresponding set of scaled Schoenfeld residuals with a suitable transformation of time [the default is based on the Kaplan-Meier estimate of the survival function, K(t)]."` socserv.socsci.mcmaster.ca/jfox/Books/Companion/appendix/… $\endgroup$
    – hrbrmstr
    Sep 15 '15 at 1:05
  • $\begingroup$ @hrbrmstr This doesn't answer what the z actually stands for. $\endgroup$
    – RayVelcoro
    Sep 15 '15 at 1:14

My guess is that it comes from old SAS names. In Therneau & Grambsch book (Modeling Survival Data - Extending the Cox Model), chapter 7, it says:

A common test of proportional hazards, corresponding to the Z:ph test printed by the SAS phglm procedure (since retired) shows that proportional hazards is badly violated, but how?

edit: It is somewhat explained here in the SAS help

  • $\begingroup$ Totally unintuitive and makes no sense. I never would have guessed that. Sounds a lot like SAS. Thanks @theodor ! $\endgroup$
    – RayVelcoro
    Sep 16 '15 at 13:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.