Blending Bayesian and frequentist methods according to the precision of prior information with applications to hypothesis testing

DOI10.1007/S10260-015-0299-6zbMATH Open1327.62132DBLPjournals/sma/Bickel15arXiv1107.2353OpenAlexW2163956611WikidataQ57826594 ScholiaQ57826594MaRDI QIDQ897846FDOQ897846

Authors: David R. Bickel 

Publication date: 8 December 2015

Published in: Statistical Methods and Applications (Search for Journal in Brave)

Abstract: The following zero-sum game between nature and a statistician blends Bayesian methods with frequentist methods such as p-values and confidence intervals. Nature chooses a posterior distribution consistent with a set of possible priors. At the same time, the statistician selects a parameter distribution for inference with the goal of maximizing the minimum Kullback-Leibler information gained over a confidence distribution or other benchmark distribution. An application to testing a simple null hypothesis leads the statistician to report a posterior probability of the hypothesis that is informed by both Bayesian and frequentist methodology, each weighted according how well the prior is known. Since neither the Bayesian approach nor the frequentist approach is entirely satisfactory in situations involving partial knowledge of the prior distribution, the proposed procedure reduces to a Bayesian method given complete knowledge of the prior, to a frequentist method given complete ignorance about the prior, and to a blend between the two methods given partial knowledge of the prior. The blended approach resembles the Bayesian method rather than the frequentist method to the precise extent that the prior is known. The problem of testing a point null hypothesis illustrates the proposed framework. The blended probability that the null hypothesis is true is equal to the p-value or a lower bound of an unknown Bayesian posterior probability, whichever is greater. Thus, given total ignorance represented by a lower bound of 0, the p-value is used instead of any Bayesian posterior probability. At the opposite extreme of a known prior, the p-value is ignored. In the intermediate case, the possible Bayesian posterior probability that is closest to the p-value is used for inference. Thus, both the Bayesian method and the frequentist method influence the inferences made.

Full work available at URL: https://arxiv.org/abs/1107.2353

Recommendations

• Bayesian frequentist hybrid inference
• The multivariate point null testing problem: a Bayesian discussion
• Reconciling Bayesian and frequentist evidence in the point null testing problem
• scientific article; zbMATH DE number 1302157
• \(\varepsilon\)-contaminated priors in testing point null hypothesis: A procedure to determine the prior probability

zbMATH Keywords

multiple hypothesis testingfiducial inferenceconfidence distributionmaximum entropyimprecise probabilityrobust Bayesian analysismultiple comparison procedure

Mathematics Subject Classification ID

Bayesian inference (62F15) Foundations and philosophical topics in statistics (62A01)

Cites Work

• Elements of Information Theory
• That BLUP is a good thing: The estimation of random effects. With comments and a rejoinder by the author
• Title not available (Why is that?)
• Bayes factors: Prior sensitivity and model generalizability
• A STATISTICAL PARADOX
• Testing a Point Null Hypothesis: The Irreconcilability of P Values and Evidence
• Calibration of \(\rho \) values for testing precise null hypotheses
• Principles of Statistical Inference
• Rectangular Confidence Regions for the Means of Multivariate Normal Distributions
• Confidence and Likelihood^*
• Title not available (Why is that?)
• The Selection of Prior Distributions by Formal Rules
• Statistical decision theory and Bayesian analysis. 2nd ed
• Estimating the null distribution to adjust observed confidence levels for genome-scale screening
• Large-scale inference. Empirical Bayes methods for estimation, testing, and prediction
• Title not available (Why is that?)
• The formal definition of reference priors
• Maxmin expected utility with non-unique prior
• Why least squares and maximum entropy? An axiomatic approach to inference for linear inverse problems
• Stochastic independence in a coherent setting
• The theory of interval-probability as a unifying concept for uncertainty
• Robust Bayesian analysis: sensitivity to the prior
• Ancillaries and conditional inference (with comments and rejoinder)
• Probability Theory
• Title not available (Why is that?)
• Title not available (Why is that?)
• The functional-model basis of fiducial inference
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Game theory, maximum entropy, minimum discrepancy and robust Bayesian decision theory
• In defense of the maximum entropy inference process
• In defence of objective Bayesianism
• The Uncertain Reasoner's Companion
• Combining probability distributions: A critique and an annotated bibliography
• A unified conditional frequentist and Bayesian test for fixed and sequential simple hypothesis testing
• Bayesian and robust Bayesian analysis under a general class of balanced loss functions
• Maximum entropy fundamentals
• Title not available (Why is that?)
• Γ-Minimax: A Paradigm for Conservative Robust Bayesians
• Unreliable probabilities, risk taking, and decision making
• Inference rules and inferential distributions
• Controlling the degree of caution in statistical inference with the Bayesian and frequentist approaches as opposite extremes
• Mathematical foundations for a theory of confidence structures
• Simple estimators of false discovery rates given as few as one or two \(p\)-values without strong parametric assumptions
• Abstract methods in information theory
• Estimators of the local false discovery rate designed for small numbers of tests
• Coherent frequentism: a decision theory based on confidence sets
• A prior-free framework of coherent inference and its derivation of simple shrinkage estimators
• Game-theoretic probability combination with applications to resolving conflicts between statistical methods
• Title not available (Why is that?)
• Elementary fundamentals of a more general probability theory. I: Interval probability as comprehensive conception. In cooperation with T. Augustin and A. Wallner
• A predictive approach to measuring the strength of statistical evidence for single and multiple comparisons
• Title not available (Why is that?)
• A Bayesian perspective on the Bonferroni adjustment
• Combining marginal probability distributions via minimization of weighted sum of Kullback-Leibler divergences
• Expected utility within a generalized concept of probability -- a comprehensive framework for decision making under ambiguity
• Bayesian frequentist hybrid inference
• Optimal decisions under complex uncertainty – basic notions and a general algorithm for data‐based decision making with partial prior knowledge described by interval probability
• Confidence distributions: a review
• Sensitivity in Bayesian Statistics: The Prior and the Likelihood
• Small-scale inference: empirical Bayes and confidence methods for as few as a single comparison
• Comment
• Rejoinder
• Observed Confidence Levels

Cited In (6)

• Bayesian frequentist hybrid inference
• Departing from Bayesian inference toward minimaxity to the extent that the posterior distribution is unreliable
• Model fusion and multiple testing in the likelihood paradigm: shrinkage and evidence supporting a point null hypothesis
• Maximum entropy derived and generalized under idempotent probability to address Bayes-frequentist uncertainty and model revision uncertainty: an information-theoretic semantics for possibility theory
• Bayesian neural tree models for nonparametric regression
• Inference after checking multiple Bayesian models for data conflict and applications to mitigating the influence of rejected priors

This page was built for publication: Blending Bayesian and frequentist methods according to the precision of prior information with applications to hypothesis testing

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q897846)