Unified frequentist and Bayesian testing of a precise hypothesis

摘要: In this paper, we show that the conditional frequentist method of testing a precise hypothesis can be made virtually equivalent to Bayesian testing. The conditioning strategy proposed by Berger, Brown and Wolpert in 1994, for simple versus case, is generalized precise null composite alternative hypothesis. Using strategy, both Bayesian will report same error probabilities upon rejecting or accepting. This of considerable interest because it often perceived are incompatible in situation. That they are compatible, when conditional allowed, strong indication "wrong" tests currently being used postexperimental assessment accuracy. new unified procedure discussed illustrated several common situations.