Reconciling Bayesian and Frequentist Evidence in the One-Sided Testing Program.
Abstract
For the one-sided hypothesis testing problem it is shown that it is possible to reconcile Bayesian evidence against H sub 0, expressed in terms of the posterior probability that H sub 0 is true, with frequentist evidence against h sub 0, expressed in terms of p-value. In fact, for many classes of prior distributions it is shown that the infimum of the Bayesian posteriors probability of H sub 0 is either equal to or bounded above by the p-value. The results are in direct contrast to recent work of Berger and Sellke (1985)in the two-sided (point null) case, where it was found that the p-value is much less than the Bayesian infimum. Some comments on the point null problem are given. Additional keywords: Bayes theorem. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1985
- Accession Number
- ADA155603
Entities
People
- G. Casella
- R. L. Berger
Organizations
- Florida State University