“Confirmation bias” refers to a psychological tendency to search out information that confirms a belief, rather than evidence that belies it. It’s a common behavior for individuals, applying to a range of cognitive behaviors from something as simple as picking out the same style of clothing over and over, to defending political beliefs.
Prior research shows that companies experience their own confirmation bias, particularly when facing the difficult decision to issue a restatement. But David N. Ricchiute, Deloitte Professor of Accountancy, discovered in his research that this bias doesn’t always rule. Further, the type of information available to an audit team may influence the restatement decision they recommend.
Restatements occur when an error comes to light after a publicly traded company has issued its audited financial statements. The statements are routinely accompanied by an independent auditor’s “unqualified” opinion, which means the financial statements comply with accountancy rules.
So a restatement amounts to a second-guessing of the auditor’s opinion.
In the event of a potential restatement, two independent investigative teams review the evidence supporting the original opinion. One team represents the board of directors; the other team is sent in by the auditors who issued the opinion. Existing accountancy research would predict that professional auditors would look for confirming evidence. “We have a rich literature that tells us that when auditors are being second-guessed, they search for evidence that is consistent, that confirms their prior opinion. They want to make their case stronger,” said Ricchiute.
But in reality, he observed something else was happening, especially in cases when the stakes were very high. “Contrary to prior confirmation-bias research in auditing, the audit firm’s investigation team searched for disconfirming evidence,” said Ricchiute. “They wanted evidence that presented the worst-case against the firm’s prior opinion.”
In order to explain this discrepancy, Ricchiute examined what type of information the investigative team was analyzing—summarized reports, or actual evidence in the form of contracts, invoices, memoranda and so on. According to another psychological theory called evidence complexity, he predicted that investigative teams using summarized or simpler reports would look for evidence confirming their prior statement, while teams looking at the stacks of receipts and other source documents would look for disconfirming evidence.
And indeed, that’s what he found in his study using two pools of audit partners and managers. The group with summarized information looked for confirming evidence, while the group given complex information looked for disconfirming. “And so research has a disconnect from what I’ve observed in practice,” he said.
Ricchiute’s research has important implications for audit practice in the event of a possible restatement.
“If the task you’re studying is complex, you need to use complex actual evidence in order to reach conclusions that translate readily to practice,” said Ricchiute. “An audit partner who is making a decision about whether to recommend restatement would not rely on summarized information.”
Ricchiute’s study, “Evidence Complexity and Information Search in the Decision to Restate Prior-Period Financial Statements,” was published in 2010 in The Journal of Accounting Research.