Using the stepwise Bayes technique to choose between experiments (Q578771)

Let \(\theta\) be the unknown state of nature and L(\(\theta\),d) a loss function. Before making the decision one of k possible experiments can be chosen. In this paper it is assumed that the set of available designs is a subclass of all designs. The decision consists of a design and a sequence of decision rules for the possible experiments. In Theorem 1 admissible decisions are characterized for the case of unique stepwise Bayes decision rules. Theorem 2 applies to finite population sampling; a sufficient condition for uniform admissibility is given.