No abstract

The purpose of this paper is to describe the results of an effort to utilize casebased reasoning (CBR) to model a specific audit judgment task. To date, most efforts to develop computational models of audit judgment have used strictly rule-based representation methods. Some researchers have recently adopted more robust structures to model the auditor domain knowledge. Although these recent efforts to extend the representation methods appear to be more accurate descriptions of auditor reasoning and memory, they still lack a comprehensive theory to guide the development of the model. A commonly encountered phenomenon in audit judgment is for an auditor to compare the current case to similar previous experiences. Others have proposed a model for this type of judgment in other expert judgment domains. This model has become known as case-based reasoning (CBR). This study describes our initial efforts to utilize CBR to model a specific audit judgment task.

Decision-support systems can be improved by enabling them to use past decisions to assist in making present ones. Reasoning from relevant past cases is appealing because it corresponds to some of the processes an expert uses to solve problems quickly and accurately. All this depends on an effective method of organizing cases for retrieval. This paper investigates the use of inductive networks as a means for case organization and outlines an approach to determining the desired number of cases-or assessing the reliability of a given number. Our method is demonstrated by application to decision making on corporate tax audits.

Recent IS research has begun to explore behavioral filtering patterns associated with content and contextual cues on a network forum. Using eye-tracking technology, this work has shed light on the cues attended to during filtering (Meservy et al. 2014) and how the attentional switching patterns between these cues (e.g., evaluating all cues of a single solution versus comparing a single cue across multiple solutions) affects filtering accuracy (Fadel et al. 2015). In the present study, we extend this prior work while making note of two important observations. First, although these studies have shed light on the role of different types of cues in forum information filtering, they are limited with respect to their ability to elucidate the actual cognitive processes that underlie this filtering. Gaze data from an eye-tracker can prompt inferences about the types of information attended to during the filtering process, but it is silent on the neurocognitive processes that occur. This leaves several important questions for ongoing theory development. For example, are different types of cues (e.g., content versus contextual) processed by different cognitive centers in the brain, which, depending on their relative activation levels, could produce more or less accurate filtering decisions? Or do similar neural mechanisms underlie both content and context-based processing, and any difference lies only in the type of information evaluated? Moreover, which types of cues are most important when filtering solutions, and how do combinations of cues affect this filtering process on both a behavioral and a cognitive level? Second, prior studies have relied on dual process theories of cognition (Chaiken 1987; Petty and Cacioppo 1986) as a theoretical frame for examining information filtering on a network forum. Originating in the domain of persuasion psychology, dual process theories posit that persuasion can occur via two primary cognitive routes: the central (systematic) route, in which the arguments of the message itself are carefully evaluated, and the peripheral (heuristic) route, in which judgments are made primarily based on surrounding peripheral cues (Chaiken 1980; Petty et al. 2005; Petty and Cacioppo 1986). Applying this framing to the context of solutions on a network forum, central route processing would entail evaluation of solution content, and peripheral route processing would rely on evaluation of surrounding contextual cues such as source expertise and validation (Fadel et al. 2015; Meservy et al. 2014). We believe this conceptualization offers a useful lens for characterizing

No abstract