DISCUSSION OF A Computational Model of Auditor Knowledge and Reasoning Processes in the Going-Concern Judgment.
讨论Biggs等人关于审计师在持续经营判断中知识与推理过程的计算模型研究，指出其四步研究策略中延迟测试可能带来的问题，并强调过程验证的重要性。
Abstract In their article, Biggs, Selfridge and Krupka (BSK) provide a status report on a very ambitious investigation of auditor knowledge and reasoning processes involved in going-concern judgments. The investigation takes the form of cognitive modeling in which the end product is a computer program that functions in much the same way that a human expert would in similar circumstances. The research strategy involves four steps: observe the behavior of experts, develop a theory or conceptual model of that behavior, implement that theory in a computer program and test the theory. The computational model becomes a tool for further development of both the conceptual model, or theory and the computational model. The stated intent of emphasizing iterations of the first three steps of a four-step strategy is problematic if one views the feedback from testing as a necessary guide to additional observation and modeling. If one accepts the idea of delaying formal tests of the model, it seems reasonable to demand some evidence of process validity along the way.