Discussion of Considerations in Choosing Statistical Sampling Procedures in Auditing

Abstract

The basic theme of the Loebbecke-Neter paper is that the choice of an appropriate statistical audit sampling procedure depends upon audit objectives, the audit environment, and audit procedures. Under certain conditions of each of these three factors, various statistical estimators perform poorly. Instances of this poor performance were studied in an extensive empirical study by the same two authors.' While the NeterLoebbecke empirical study is as yet unpublished, many of its conclusions are referred to in the paper under discussion, and the authors were good enough to make a draft copy of the study available to us for purposes of this discussion. Based on the study conclusions, Loebbecke and Neter have presented in the current paper a suggested decision flowchart to aid the auditor in selecting various sampling procedures which they feel are appropriate in various specified circumstances. In discussing this paper, it is impossible to avoid reference to the empirical study on which its principal conclusions are based. At the same time, it would not be proper in this forum, prior to the study's publication, to begin an exhaustive analysis of the study's numerical findings. We will confine ourselves, therefore, to such general comments about the study and its findings as are necessary to a meaningful consideration of the present paper. In our view, the most important contribution made by this paper, and by the study on which it is based, is the dramatic demonstration that many classical sampling plans are subject to significant unstated risks in certain