I am using the Matching R package to compute Average Treatment Effect within Propensity Score Matching (PSM).

I have two groups: treatment and control. In addition to assessing the statistical difference between these groups on standard continuous variables, I need to verify if the two survival curves are the same between the two groups.

Before knowing PSM I had used the classical log - rank test. How can I update my approach taking into account PSM (possibly using the Matching R package)?

  • $\begingroup$ Would it to simple to say: Do as you always do, only use the matched sample in place of the "raw" one? This, at least, is the main point from the authors of MatchIT in R (Ho, King, Garry and Imai). "Matching does not add addtional parametric assumptions, but does not reduce them either". $\endgroup$ – Repmat May 25 '15 at 21:49
  • $\begingroup$ Can you expand on your query "I need to verify if the two survival curves are the same within groups"? Not sure what the within-group element here refers to... $\endgroup$ – James Stanley May 25 '15 at 22:00
  • $\begingroup$ @James... you are right. I have edided. $\endgroup$ – Giorgio Spedicato May 26 '15 at 12:51

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.