Imbalanced Randomization (Rationales for)

Abstract Most randomized controlled trials use balanced randomization to achieve approximately or exactly equal sample sizes per group. However, in some settings imbalanced (unequal) randomization designs may be more powerful, less costly, and/or more ethically appealing to study participants. This article examines some important statistical rationales for unequal randomization designs. It covers both two‐arm and multi‐arm trial designs with either fixed or response‐adaptive randomization.