Expert Judgment in Auditing: An Expected Utility Approach

Abstract

With few exceptions, accounting research has provided evidence of significant individual differences in the professional judgments of auditors (see Joyce [1976]; Moriarity and Barron [1976]). If we accept consensus as a measure of the validity of professional judgments (Einhorn [1974]), this evidence provides a motivation for additional research into the nature of these differences. In this paper, I assume that audit decisions can be analyzed within an expected utility framework. This is not unreasonable, since the kinds of decision auditors must make fit into the expected utility framework very well. For example, Ward [1974] suggested a similar model for deciding which audit procedures to perform, and Kinney [1975] used the framework for developing optimal sampling plans. Within an expected utility framework, the analysis of the causes of differences in professional judgments can be simplified if we focus on a decision in which the feasible actions and states are completely specified. Under these circumstances, there are two sufficiency conditions for consensus. First, different auditors would have to arrive at the same probability distribution over the specified set of states. Second, auditors would have to possess homogeneous utilities for the outcomes or consequences of their decisions. Previous studies have generally failed to isolate these sufficiency conditions. Policy-capturing studies, for example, have attributed individual differences to differential cue weighting. A major limitation of this approach is that the dependent variable is a terminal judgment that is the result of a decision rule which combines probabilities and utilities. Different weightings, then, could be traced to either source, or both. The