Abstract Competitive pressure is forcing the accounting profession to improve the efficiency of audit services. One method suggested in the literature for this purpose is the Bayesian statistical approach to sample size decisions and error likelihood revisions. This paper reports on the results of two interrelated case studies with 138 practicing auditors to investigate the efficiency implications of the Bayesian approach over the classical statistical approach in compliance testing tasks in auditing. The results provide evidence in support of the efficiency of the Bayesian approach: (1) Using the Bayesian approach, experienced auditors decided on smaller sample sizes than those using the classical approach for the same level of confidence (95 percent); (2) In two sample cases where auditors' behavior was conservative, once the samples were taken, the study indicates that the auditor's ret lance on the Bayesian revision model may result in higher efficiency than reliance on judgmental revision. While auditors acted in a manner consistent with the Bayes' model in another two-sample case, they showed excessive revision, beyond Bayes' model, in a fifth sample situation. Implications for audit practice and future research are discussed.
Mohammad J. Abdolmohammadi (Sat,) studied this question.