The Effects of Decision-Aid Design on Auditor Performance in Internal Control Evaluation Tasks

Independent auditors evaluate internal controls (IC) using internal control questionnaires (ICQs). The Public Company Accounting Oversight Board (PCAOB) recommends that rather than using generic ICQs, auditors should consider using ICQs customized for each audit client. Based on Silver's (1988, 1990) decision-aid design principles and dilution effect research, we hypothesize that the presence of irrelevant task cues in an ICQ decision aid will negatively affect auditors' task performance and that task performance will be positively affected by an ICQ decision aid in which task cues are presented one at a time (directed search) rather than all together (non-directed). We conduct an experiment with practicing auditors as participants who complete an ICQ decision-aid with either few or many irrelevant questions presented all at once or a question at a time. We find partial support for the PCAOB's recommendation, in that participants were more accurate with a directed versus non-directed search ICQ decision aid. Irrelevant cues did not affect accuracy. Although the participants with directed search ICQ decision-aids were more accurate evaluating IC, they were less accurate assessing the overall strength of the IC system and recommending additional testing of IC and transaction details. These findings have important implications for the design and use of decision aids.