Natalia V. Kotchetova
University of Waterloo
William F. Messier
Georgia State University
Aasmund Eilifsen
Norwegian School of Economics and Business Administration
Abstract: We investigate whether auditors are more likely to make probabilistic judgments in accordance with Bayes rule and less likely to neglect base rates when they make their assessments in a frequency format versus a probability format. We test two hypotheses based on research in cognitive psychology by Gigerenzer and his colleagues (e.g., Gigerenzer et al. 1988; Gigerenzer 1991). Three factors are manipulated (base rate, response format, and case context) in a 2x2x2 mixed factorial design. Seventy-six auditors participated in the experiment. The results indicate that (1) consistent with prior research, auditors are more likely to attend to base rate when they are high, versus low, when both are based on realistic data, (2) for combinations of experimental materials containing auditing cases with high base rate of fraud, auditors' assessments of probabilities differ when evaluating an auditing case versus a non-auditing case, and most importantly, (3) auditors' responses using a frequency format are closer to the base rate and scaled Bayesian response than those auditors using a probability format.
Back to Program
