Efficiency of Sieve Sampling in Auditing

Abstract

Dollar-unit sampling is a popular method for sampling accounts by auditors. In this article, we analyze and compare two methods of obtaining dollar-unit samples, simple random sampling and sieve sampling. Two methods of assigning errors to the dollar units in the population are considered. For a variety of study populations based on actual accounting populations, it is found that sieve sampling, with either error assignment method, produces a more precise point estimator than simple random sampling of dollar units for sample sizes that are not excessively large for audit applications.