Abstract ABSTRACT: Before a Bayesian or decision-theoretic model can be utilized in actual audit practice, methods must be developed which allow auditors to make their beliefs explicit in the form of a probability distribution. This paper compares the probability distributions obtained from a group of audit seniors using the cumulative distribution function and equivalent prior sample elicitation techniques. Comparison of the probability distributions which resulted from use of the two elicitation techniques showed no statistically significant differences in either means or medians. Although measures of overall variability were not statistically different, some differences were found when the interquartile range was examined. Since this latter result is consistent with previous findings in the psychological literature, the failure to find differences in overall variability measures may be attributed to the unique nature of auditing populations.
Michael A Crosby (Wed,) studied this question.