Unresolved Issues in
Classical Audit Sample Evaluations
Donald R. Nichols
Texas Christian University
Rajendra P. Srivastava
The University of Kansas
Bart H. Ward
The University of Oklahoma
Classical variables techniques can be usefully employed in certain audit situations. They may be useful, for example, when auditing high error rate populations or accounts with numerous negative balances or when the auditor is concerned about both over arid under-statement errors. Classical variables techniques may also be useful when the auditor is concerned with assessing the reasonableness of proposed adjustments in light of statistical test results. This paper reviews several issues associated with the evaluation of classical statistical hypothesis testing results in auditing. Though presented in terms relevant to classical statistical testing, some of the issues reviewed may be germane to other statistical or non-statistical approaches to audit sampling as well.
Some of these issues have been isolated and examined in greater detail by other studies. This paper mainly deals with the comparison and reconciliation of certain alternative evaluation strategies which can be employed when achieved allowances for sampling risk differ from planned levels. This situation can occur when the apparent achieved efficiency of a sample estimator is different from the level on which the auditor based the audit sampling plan.
Comparative Evaluation Strategies
Several strategies are available for use in evaluating the results of a classical variables hypothesis test. Conclusions drawn from the evaluation of sample results may vary depending upon which strategy is employed. Three of these strategies are explained and compared in this paper.
No one of the three strategies is uniformly dominant or necessarily superior to the others in all situations. However, they can lead to different conclusions. Therefore, it is important to understand how they differ. In this respect, the selection of an appropriate evaluation strategy is similar to the dilemma