Reconciling Bayesian and Frequentist Evidence in the One-Sided Testing Problem

摘要: Abstract For the one-sided hypothesis testing problem it is shown that possible to reconcile Bayesian evidence against H 0, expressed in terms of posterior probability 0 true, with frequentist p value. In fact, for many classes prior distributions infimum equal value; other cases less than The results are contrast recent work Berger and Sellke (1987) two-sided (point null) case, where was found value much smaller infimum. Some comments on point null also given.