A Bayesian analysis of serial dilutions offers a worse positive bias than the MPN and proposes an inappropriate interval estimate

The recent article ‘Estimating bacterial density from tube dilution data by a Bayesian method', by Roussanov et al. (Food Microbiology, 1996 13, 1431–1437), inappropriately treats the likelihood function over a scalar (not logarithmic) axis as a Bayesian probability distribution over a universe...

Full description

Saved in:
Bibliographic Details
Published in:Food microbiology 1997-10, Vol.14 (5), p.515-517
Main Author: Garthright, W.E.
Format: Article
Language:English
Subjects:
bacterial counting
Bayesian theory
dilution
enumeration
food microbiology
most probable number
probability analysis
serial dilution tests
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The recent article ‘Estimating bacterial density from tube dilution data by a Bayesian method', by Roussanov et al. (Food Microbiology, 1996 13, 1431–1437), inappropriately treats the likelihood function over a scalar (not logarithmic) axis as a Bayesian probability distribution over a universe of possible concentrations of microbes. The authors recommend that the mean of that distribution serve as a point estimate, but they derive an interval estimate that is inconsistent with the use of the mean for the point estimate. Using the uniform prior distribution is unsatisfactory on Bayesian terms. Using the interval estimate is inappropriate. The point estimate derived from the scalar mean is much more positively biased than the MPN.
ISSN:0740-0020
1095-9998
DOI:10.1006/fmic.1997.0127