摘要：This paper describes current practices relating to statistical sampling used by auditors performing compliance audits. An example audit case is presented where a compliance audit is performed using statistical sampling. Fines for non-compliance are based upon the sample results. While the practice is quite common and generally accepted, there are significant ramifications associated with the use of a single-point estimator. Approximately half of the time, an auditee will be charged more than is actually deserved. We explore the shortcomings associated with the use of a single-point estimator and the appropriateness of the use of a one-tailed confidence interval to provide greater assurance that fines are appropriate and reasonable.
