OPTIMAL BAYESIAN RANDOMIZATION

Scott M. Berry and Joseph B. Kadane

Abstract:

Randomization is a puzzle for Bayesians. The intuitive need for randomization is clear, but there is a standard result that Bayesians need not randomize. In this paper we propose a model in which randomization is a strictly optimal procedure. The most important aspect of our model is that there are several parties who make different decisions an observe different data. The result also sheds light on ethical considerations involving randomization in a clinical trial.

Here is the full postscript text for this technical report. It is 97208 bytes long.