What question did this study set out to answer?

To validate Bayesian dollar-unit sampling bounds and compare their performance to non-Bayesian bounds.

March 18, 2026

A Note on a Comparison of Bayesian with Non-Bayesian Dollar-Unit Sampling Bounds for Overstatement Errors of Accounting Populations.

Key Points

To validate Bayesian dollar-unit sampling bounds and compare their performance to non-Bayesian bounds.
Analyzed Bayesian models using diffuse priors on typical tainting patterns.
Compared performance of Bayesian and non-Bayesian dollar-unit sampling bounds under audit conditions.
Some Bayesian models improved performance over non-Bayesian techniques.
Validation of Bayesian models showed robust outcomes even with diffuse priors.

Abstract

Abstract ABSTRACT: There has been a proliferation of dollar-unit sampling bounds over the past few years; however, in many cases there has been little guidance provided on the relative performances of these bounds under representative audit conditions. This is particularly true of the Bayesian bounds which are analyzed here. One purpose of this study, therefore, is to provide further validation of Bayesian models by assessing their robustness when using diffuse priors on the "typical tainting patterns" of Leitch, et al. 1982. A second purpose of the study is to compare Bayesian bounds with non-Bayesian bounds. The study found that even with diffuse priors, some Bayesian models show potential for improving on the performance of non-Bayesian dollar-unit sampling techniques.

Mark Helpful

Bookmark

Relay