Ratio Estimation in Accounting Populations with Probabilities of Sample Selection Proportional to Size of Book Values

Abstract

It is generally recognized that accounting populations possess statistical characteristics which render inappropriate the routine application of sample survey techniques to auditing problems. Kaplan [1973] points out that inferences about accounting populations are probably best made using auxiliary information, such as recorded book values. However, since typical accounting populations are highly skewed and contain relatively few accounts in error, the standard distributional assumptions1 underlying auxiliary information estimators (e.g., ratio or regression estimators) are likely to be violated. This leads Kaplan to conclude that entirely new approaches may be required for statistical sampling in auditing. In this paper we study a modification of a standard unbiased ratio estimator2 based on a population sample selected with probability pro-