I am performing a survival analysis with cluster data cluster(id) using GEE in R (package:survival).

I was wondering I could correctly interpret the Robust value in the summary of the model output.

Robust = 14.65 p=0.4

thanks in advance

  • $\begingroup$ check the documentation for that package $\endgroup$
    – develarist
    Sep 5 '20 at 23:26
  • $\begingroup$ There is no information about that in the documentation $\endgroup$
    – Anjeline
    Sep 7 '20 at 4:58

For a robust model involving a cluster term, this is reported as an overall test of model significance along with the Likelihood ratio, Wald, and Score (log-rank) tests. It can be hard to find specific documentation on the Robust term, but its meaning is clear from the underlying code for survival:::print.summary.coxph and survival:::summary.coxph.

The "Robust" value represents a robust Score test value for the model as a whole. The associated p-value has the usual interpretation: the probability that a test value at least that large might be found by chance even if the null model holds. For this example, the interpretation would be that the model was not significantly different from the null model.


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.