Analogical Reasoning and Error Detection.
Garry Marchant · Deep Blue (University of Michigan) · 1987
The auditor is required to provide a satisfactory explanation for any exception that might indicate a possible material misstatement of the financial statements. Since errors are relatively infrequent events, even seasoned auditors may have no direct experience with many types of errors. This thesis addresses the question of how the auditor generates a potential explanation and suggests analogical reasoning is an effective means for dealing with these rare events. The first experiment examines the use of analogical reasoning in error detection. The second experiment examines whether two prior analogs are more effective than one prior analog. The transfer paradigm, based on Gick and Holyoak (1982), was used in both experiments. The detected error in the prior analog was the only difference between the two treatment conditions (timing and performance) and was for each condition analogous to only one of the alternative solutions. In the comparison between the timing and control conditions, experts mean probability estimate for the target error increased when a prior analog was available, indicating that experts were using the timing analogy. Novices showed no change across conditions. In the comparison between the performance and control conditions neither experts nor novices showed any change across conditions. In the second experiment, experts showed no change between the one prior analog and two prior analog conditions. Novices, with a high quality schema, showed an increase in the use of the analogy in the two prior analog condition. Experts may have been affected by the frequency of occurrence of the target error. The error contained in the timing condition is a high frequency error and the error contained in the performance and two prior analog condition is a low frequency error. When the target is a high frequency error experts will use the analogy since a high frequency target has a high activation strength. When the target is a low frequency error high frequency alternatives (with a higher activation strength) will interfere with the generation of the target. Expert auditors use of analogy is mediated by the strength of activation of the target error.