A Bayesian solution for a statistical auditing problem

• 成果类型: Article
• 署名作者: Meeden, G
• 署名单位: University of Minnesota System; University of Minnesota Twin Cities
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.1198/016214503000000648
• 发表日期: 2003
• 页码: 735-740
• 关键词: finite population; bootstrap

摘要

Auditors often consider a stratified finite population where each unit is classified as either acceptable or in error. Based on a random sample, the auditor may be required to give an upper confidence bound for the number of units in the population that are in error. In other cases the auditor may need to give a p value for the hypothesis that at least 5% of the units in the population are in error. Frequentist methods for these problems are not straightforward and can be difficult to compute. Here we give a noninformative Bayesian solution for these problems. This approach is easy to implement and is shown to have good firequentist properties.