I've been working on putting together a survival analysis using Kaplan-Meier and the logrank test. I am doing the testing in R with survdiff(). Each plot has multiple groups/curves, and I've been testing for the difference between groups (where there are more than two) by doing individual pairwise tests, as has been suggested elsewhere.

The question arose about testing for significance between each group and the combined/pooled data for all the groups. To accomplish this, I've been doing pairwise testing between the data from all the groups, and the data from the individual groups- as if there was one curve for the combined data, and one curve for the single group. Is this appropriate? I just want to conform that I am not breaking assumptions or otherwise committing transgressions by approaching it this way.



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.