Evidential Reasoning for WebTrust Assurance Services

Abstract

This study looks at two aspects of assurance services. The first deals with the type(s) of evidential networks that will allow a professional accountant to provide assurance. Here, we develop an evidential network model for “WebTrust Assurance,” a service being provided by the American Institute of Certified Public Accountants (AICPA) and the Canadian Institute of Chartered Accountants (CICA). Our model augments the AICPA/CICA approach and provides goals, sub-goals and evidence relevant to the overall assurance to be provided. The aggregation of evidence and the resolution of uncertainties follow the belief-function approach of Srivastava and Shafer. Next we develop a decision theoretic model for the assurance-planning problem. Our approach is based on estimating the expected value of providing various levels of assurance and is illustrated with several different scenarios that may be faced in practice. We also consider the role of ambiguity in decision situations such as planning WebTrust engagements and calculate bounds in expected value based on whether auditors are conservative or not in their approach to risk.