In N individuals observed over respective Tn times, I have observed Kn events. Multiple events pro individual N can be observed Kni (where i is the index numder for the k event of the n individual).

With a set of X predictors (predictors on the individuals Xn and of each specific episode Xkni). I would like to model the length of the K events (Lnki), and try do find which of the Xn and Xkni influence L.

L are integers > 0

What would be the appropriate was to run such regression? Poisson? Maybe a mixed model (with N random intercepts?). Any help with links to R resources is greatly appreciated.

  • $\begingroup$ Is the jargon for what you are interested in "survival analysis", e.g., stats.stackexchange.com/questions/199744/… $\endgroup$ Oct 8, 2019 at 13:40
  • $\begingroup$ No, I am not interested in survival analysis. $\endgroup$
    – ECII
    Oct 8, 2019 at 14:56
  • $\begingroup$ Why not? The description of your problem sounds exactly like a survival analysis problem. $\endgroup$ Oct 8, 2019 at 18:30
  • $\begingroup$ I am not interested to see if the events occur or when the events occur but on the length on the events. I dont undestand how to formulate this as a survival analysis problem. $\endgroup$
    – ECII
    Oct 8, 2019 at 20:19


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.