ABSTRACT Prior research documents that auditors fail to revise audit plans to effectively address identified fraud cues. While auditors may understand what evidence would address such cues, we propose that auditors fail to apply this understanding because they use implemental mindsets when making decisions for themselves (i.e., deciding). However, we also propose that auditors use deliberative mindsets when advising. To test our predictions, we assign auditors to a decider or an advisor role in a realistic case that contains seeded fraud cues and asks them to consider revising last year's plan. We also manipulate whether the case prompts auditors to revise the plan unconventionally. Results indicate decider‐condition auditors use implemental mindsets: Prompted deciders follow the unconventional plan without regard to underlying fraud risk and unprompted deciders stick with the same‐as‐last‐year plan. Advisor‐condition auditors use more deliberative mindsets: In the prompt and no prompt conditions, they identify plans that are strongly linked to their own fraud risk assessments and that better align with experts' recommended plan for effectively addressing the seeded fraud cues. Supplemental analyses suggest deciding and advising auditors both follow the experts' plan when they believe in its potential effectiveness but, after controlling for the influence of perceived effectiveness, deciding auditors follow it to a greater extent simply because they believe the PCAOB wants it. By contrast, advising auditors do not exhibit signs of excessive PCAOB influence. Our findings provide evidence that seeking informal advice (or thinking like an advisor) helps auditors to effectively revise audit plans in response to identified fraud risk—it helps when a prompt is present or not, suggesting it complements rather than merely substitutes for interventions meant to improve auditors' judgment and decision making.
Stored only when both papers are in the corpus — 25 citations exist upstream.
Imported from OpenAlex (CC0) — nothing here is generated.