The Annals of Statistics

A Bayes but Not Classically Sufficient Statistic

D. Blackwell and R. V. Ramamoorthi

Full-text: Open access

Abstract

In a Borel setting, every classically sufficient statistic is Bayes sufficient, but not vice versa. The example is a hypothesis testing problem in which Bayesians, but not classicists, can achieve zero error probabilities.

Article information

Source
Ann. Statist. Volume 10, Number 3 (1982), 1025-1026.

Dates
First available in Project Euclid: 12 April 2007

Permanent link to this document
http://projecteuclid.org/euclid.aos/1176345895

Digital Object Identifier
doi:10.1214/aos/1176345895

Mathematical Reviews number (MathSciNet)
MR663456

Zentralblatt MATH identifier
0485.62004

JSTOR
links.jstor.org

Subjects
Primary: 62B20
Secondary: 62F15: Bayesian inference

Keywords
Sufficient statistic Bayes

Citation

Blackwell, D.; Ramamoorthi, R. V. A Bayes but Not Classically Sufficient Statistic. Ann. Statist. 10 (1982), no. 3, 1025--1026. doi:10.1214/aos/1176345895. http://projecteuclid.org/euclid.aos/1176345895.


Export citation