The Comparison of One-stage and Two-stage Selection Rules in a Bayes Approach

The comparison of one-stage and two-stage selection rules was carried out using the Bayes look-ahead approach. Loss functions including subset size and costs of sampling were used in the work. The comparison was made under the assumptions of balanced models, reduction by sufficiency, permutation invariance and decreasing in transposition of distributions, permutation invariance, and selecting in favor of larger values of loss functions. The results were applied to normal models and binomial models. Everything was done with fixed sample sizes.