Modeling auditor judgment in nonstatistical sampling

Abstract

Since its issuance in June 1981, Statement on Auditing Standards (SAS) No. 39, "Audit Sampling," has been the center of much controversy. Practitioners are voicing their concerns as they anticipate difficulties in designing, selecting, and evaluating a nonstatistical sampling procedure in accordance with SAS 39. This proposed exploratory study seeks to identify those factors that underlie the auditor's judgment with respect to nonstatistical sample size decisions in substantive tests. The research will utilize Egon Brunswik's Lens Model to provide mathematical representations of the auditor's judgment process. Correlational statistics will be used to assess judgment accuracy, agreement (consensus) , and auditor "self-insight" into his decision process. The study will provide empirical insight, into whether the auditor's determination of the appropriate extent of testing is consistent with his judgment as to the assurance level needed from his sampling application, or conversely, the degree of risk he is willing to accept. The ability of auditors to formulate their sample size decisions properly is crucial because of their impact on audit effectiveness and efficiency. In addition, this project should provide additional evidence bearing upon the arguments of both proponents and opponents of SAS 39.