Bayesian Sampling Procedures for Auditors: Computer-Assisted Instruction.

Abstract Bayesian statistical sampling procedures for auditors have the potential to improve audit efficiency by decreasing the sample size required to achieve a desired reliability for the auditor's statistical conclusions from the sample. The increased efficiency is obtained in the Bayesian approach by formally including the auditor's subjective prior estimates (e.g., error rates) in the sample evaluation. The Bayesian method in effect provides a formal analytic procedure for including much of the auditor's informal information about an audit population together with the information obtained from the sample. This thereby reduces the sampling required to achieve desired reliability for statistical statements about the audit population. The program is in conversational mode, so that students with little knowledge of computers can use it to understand the potential of the Bayesian approach in audit sampling. The introduction provided by this program can help the student learn that computer routines easily put complex mathematical methods within the reach of audit practice.