During last week’s East Users Group Meeting, Michael Proschan of the NIH and NIAID, gave a presentation on ‘Blinded Adaptations, Permutations and t-tests.’ Given the close connection between adaptive t-tests and adaptive permutation tests, Proschan argued that it is possible to determine the validity of an adaptive t-test from the validity of an adaptive permutation test. Proschan concludes that in adaptive settings, “permutation tests are often valid in adaptive settings, even if blind is maintained.”
Michael's slides are available below. You can also find two of his recent papers on blinded adaptations in the 'Related Items of Interest' section.
Related Items of Interest
Posch, Martin, and Michael A. Proschan. "Unplanned adaptations before breaking the blind." Statistics in medicine 31.30 (2012): 4146-4153.
