Uniformly most powerful Bayesian tests

DOI10.1214/13-AOS1123zbMATH Open1277.62084arXiv1309.4656OpenAlexW3105309374WikidataQ37649564 ScholiaQ37649564MaRDI QIDQ385759FDOQ385759

Authors: V. Johnson 

Publication date: 11 December 2013

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Uniformly most powerful tests are statistical hypothesis tests that provide the greatest power against a fixed null hypothesis among all tests of a given size. In this article, the notion of uniformly most powerful tests is extended to the Bayesian setting by defining uniformly most powerful Bayesian tests to be tests that maximize the probability that the Bayes factor, in favor of the alternative hypothesis, exceeds a specified threshold. Like their classical counterpart, uniformly most powerful Bayesian tests are most easily defined in one-parameter exponential family models, although extensions outside of this class are possible. The connection between uniformly most powerful tests and uniformly most powerful Bayesian tests can be used to provide an approximate calibration between p-values and Bayes factors. Finally, issues regarding the strong dependence of resulting Bayes factors and p-values on sample size are discussed.

Full work available at URL: https://arxiv.org/abs/1309.4656

Recommendations

• On the existence of uniformly most powerful Bayesian tests with application to non-central chi-squared tests
• Restricted most powerful Bayesian tests for linear models
• Bayesian (mean) most powerful tests
• Towards an uniformly most powerful binomial test
• Determination of uniformly most powerful tests in discrete sample spaces

zbMATH Keywords

Bayes factorsNeyman-Pearson lemmaHiggs bosonJeffreys-Lindley paradoxnon-local prior densitiesobjective Bayesone-parameter exponential family modeluniformly most powerful tests

Mathematics Subject Classification ID

Parametric hypothesis testing (62F03) Bayesian inference (62F15) Foundations and philosophical topics in statistics (62A01) Asymptotic properties of parametric tests (62F05) Applications of statistics to physics (62P35)

Cites Work

• Testing Statistical Hypotheses
• A STATISTICAL PARADOX
• Testing a Point Null Hypothesis: The Irreconcilability of P Values and Evidence
• Title not available (Why is that?)
• IX. On the problem of the most efficient tests of statistical hypotheses
• The Intrinsic Bayes Factor for Model Selection and Prediction
• Title not available (Why is that?)
• Bayesian model selection in high-dimensional settings
• Severe Testing as a Basic Concept in a Neyman–Pearson Philosophy of Induction
• Title not available (Why is that?)
• Bayes Factors Based on Test Statistics
• The case for objective Bayesian analysis
• Title not available (Why is that?)
• Title not available (Why is that?)
• Harold Jeffreys's \textit{Theory of probability} revisited
• Properties of Bayes Factors Based on Test Statistics
• Title not available (Why is that?)

Cited In (23)

• A Characterization of Most(More) Powerful Test Statistics with Simple Nonparametric Applications
• On biases in assessing replicability, statistical consistency and publication bias
• Prior distributions for objective Bayesian analysis
• A frequency-calibrated Bayesian search for new particles
• Testing un-separated hypotheses by estimating a distance
• Comparing Objective and Subjective Bayes Factors for the Two-Sample Comparison: The Classification Theorem in Action
• Maximum Average-Power (MAP) Tests
• Reply to Gelman, Gaudart, Pericchi: More reasons to revise standards for statistical evidence
• Introducing and analyzing the Bayesian power function as an alternative to the power function for a test
• A note on Type S/M errors in hypothesis testing
• Title not available (Why is that?)
• Restricted most powerful Bayesian tests for linear models
• On the existence of uniformly most powerful Bayesian tests with application to non-central chi-squared tests
• A sample size-dependent prior strategy for bridging the Bayesian-frequentist gap in point null hypothesis testing
• Abandon Statistical Significance
• The False Positive Risk: A Proposal Concerning What to Do About p-Values
• Bayesian t-tests for correlations and partial correlations
• Hypothesis Testing in Presence of Adversaries
• Uniformly most powerful unbiased test for conditional independence in Gaussian graphical model
• Bayesian (mean) most powerful tests
• Revised standards for statistical evidence
• A modified sequential probability ratio test
• A Powerful Bayesian Test for Equality of Means in High Dimensions

This page was built for publication: Uniformly most powerful Bayesian tests

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q385759)