Randomization-based, Bayesian inference of causal effects

Abstract Bayesian causal inference in randomized experiments usually imposes model-based structure on potential outcomes. Yet inferences from are especially credible because they depend a known assignment process, not probability model of In this article, I derive randomization-based procedure for effects finite population setting. formally show that satisfies analogues unbiasedness and consistency under weak conditions prior distribution. Unlike existing methods inference, my supposes neither models generate outcomes nor independent identically distributed random sampling. does suppose discrete bounded. Consequently, researchers can reap the benefits without sacrificing properties make first place.