Sufficient and appropriate evidence: auditors' use of stopping rules

Abstract

Auditing standards require auditors to collect 'sufficient appropriate' audit evidence, but fails to provide auditors with guidance on how to do so. Research in psychology calls the cognitive mechanisms to terminate the information search process "stopping rules". The limited literature in psychology identifies four main stopping rules used by decision-makers: magnitude threshold, difference threshold, mental model and mental list. The literature gives little guidance on which of these stopping rules would be used by auditors of financial statements. The purpose of this thesis is to identify the stopping rules used by auditors and to use behavioural decision theory to investigate the impact of the auditor’s knowledge, environment and motivation on auditor stopping rules and decision-making performance. Study One was a verbal protocol study using detailed case study to examine the identification of significant risks at the pre-planning stage of the audit. Participants conducted the task while thinking-aloud to capture their conscious cognitive processes. A checklist similar to those provided to audit staff was provided and used at their discretion. The results identified that auditors used multiple stopping rules during the large, complex and unstructured task. The mental model stopping rule was used to the greatest extent and supplemented by use of the other rules (most often the mental list). This is a significant extension to the stopping rule literature as previously it was reported that decision-makers use only a single stopping rule when determining to cease searching for information. It was also found that the stopping rule itself did not affect performance, but how the participant used the stopping rule; for example use of the mental model alone does not improve performance, but how an auditor uses the mental model. The second study was a behavioural experiment with the smaller, less complex task of identifying the cause of an unexpected gross margin fluctuation. Participants continued to use multiple stopping rules during the conduct of the task, though the magnitude threshold stopping rule was used to a greater extent. The level of risk impacted auditor performance, due to both increasing the accuracy of the participant’s initial cause selection and also the extent to which the participant used the mental model stopping rule. Time budget pressure was found to impact performance via its effect on selecting the correct initial cause.