As in the question - how do you decide, whether the raw variable x time vs. variable x log(time) - or any other function - should be taken in the survival analysis?

  • $\begingroup$ Do you mean "time-varying coefficients"? If so, please, edit the title accordingly. For time-varying covariates/predictors there is no function of time to specify, other than coding the data to indicate the period of time over which a set of covariates were in place. It would help if you illustrated the question with a specific situation that you face, as such decisions are based on the form of the lack of proportional hazards that you have. If it is the situation you describe here, consider whether a Cox model is actually a good choice. $\endgroup$
    – EdM
    Mar 29 '21 at 18:02

Your Answer

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

Browse other questions tagged or ask your own question.