In theory what would the $e_i$ (propensity scores) be, for $n_i$ observations already randomized into various treatment groups ?

I know $e_i$ (propensity scores) are calculated for $n_i$ observations in observation studies and RCT studies are assumed to be already randomized.

However I like to know what would be the hypothetical $e_i$ (propensity scores) values for observations in studies that already randomize patients into various treatment groups ? Could the propensity score values for everybody in an already randomized study would be 1 ?


1 Answer 1


"Propensity may be defined as an individual's probability of being treated with the intervention of interest given the complete set of all information about that individual."

According to this definition, the propensity is the probability that patient being randomized to the treatment patient received, and generally it is 1/k, where k is # of treatment arms in that RCT.

  • $\begingroup$ Thanks 158565, so what I understand is that the propensity score in theory for all the individuals in already randomized RCT study is 1/k. In theory. Is that correct ? $\endgroup$
    – Science11
    Dec 5, 2018 at 2:57
  • $\begingroup$ Not "in theory", it is in practice. $\endgroup$
    – user158565
    Dec 5, 2018 at 2:58
  • $\begingroup$ Excellent, can you please provide a citation for this 158565 ? Thanks. $\endgroup$
    – Science11
    Dec 5, 2018 at 3:16
  • 1
    $\begingroup$ Maybe not exact, but the meaning comes from Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983;70:41–55. $\endgroup$
    – user158565
    Dec 5, 2018 at 3:23
  • $\begingroup$ Thanks 158565, I will go through that article. $\endgroup$
    – Science11
    Dec 5, 2018 at 3:25

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.