Aggregation of Evidence in Auditing: A Likelihood Perspective.
在概率论框架下提出了一种聚合审计证据的通用方案，将审计证据视为变量网络，推导了针对单一变量和变量簇的证据组合表达式，并展示了如何在实际审计中应用以获得账户或财务报表公允表达的后验概率。
Abstract We have discussed a general scheme for aggregating audit evidence under the probability theory framework. We have argued that audit evidence forms a network of variables, variables being the accounts constituting the financial statements, the audit objectives of the accounts, and the financial statement as a whole. We have derived general expressions for combining evidence for two situations: (1) various items of evidence bearing on one variable, and (2) various items of evidence bearing on a cluster of variables. These general results are used to demonstrate how audit evidence can be aggregated on an actual audit to obtain posterior probability that the account or the financial statements are fairly presented. We have also discussed various types of dependencies encountered by the auditor and how these dependencies can be treated in the framework. Several limitations of the present study have been discussed and suggestions for further research are presented.