| The generally accepted goal of audit judgment research has been understand and improve audit decision making (Libby, 1981). The objective of this research is to describe a cognitive model of belief revision in audit risk assessment. That goal is to be achieved in the following steps: (1) to study the belief revision strategies used by experienced auditors in a real-world problem setting, (2) to analyze the data collected to identify behavior and strategies used by the auditors, (3) to propose a framework of belief revision in audit risk assessment to account for the observed behaviors, (4) to validate the model by applying it to a different data set to test its generalizability, and (5) to outline a plan for enhancing the model in the future.; The major contribution of this research is the way in which the model enhances our understanding of belief revision processes involved in updating evidence in audit judgments. Specifically, it provides added insight into the following areas: (1) knowledge organization--how is knowledge organized? (2) knowledge utilization--what reasoning processes and strategies do experienced auditors use in problem solving? (3) learning and improvement--how can we use the enhanced understanding to improve audit quality? Based on analytical results, nine propositions are posited and tested. Some of the propositions provide new insight into belief revision: e.g., contrary to the general belief, the data show that the primacy effect appeared more frequently in auditors' problem solving processes than the recency effect. Some findings have strengthened earlier research results with the application to a more complex and realistic problem situation.; Several strategies/biases auditors used in the assessment are also identified. These strategies include: polarization/scanning through incrementalism/muddling through, and elimination by aspects. Furthermore, with the combination of an anchor, an adjustment weight, and the direction of evidence, the model is able to predict the belief revision with acceptable accuracy. |
