Nevertheless, plaintiffs’ predictions of thejudge’s award were substantially higher than those of the defendants
Defendants, on average, predicted that the judge would award the plaintiff $20,000. Plaintiffs expected the judge to specify a payment of $33,000.
Moreover, the degree of discrepancy between plaintiffs’ and defendants’ predictions of the judge was a strong predictor of whether they settled the case, as opposed to relying on the judge’s decision and paying penalties for not settling.
In a number of follow-up cases, we attempted to intervene in various ways to reduce the magnitude of the self-serving bias. In several studies, we paid participants for accurately predicting the judge’s ruling. This intervention had no measurable effects.
Participants consistently believed that the judge would come out on their side even when we paid them for estimating the judge’s rule accurately.
There was a strong tendency to view arguments supporting one’s own position as more convincing than those supporting the other side, suggesting that the bias operates by distorting one’s interpretation of evidence
We also attempted to reduce the magnitude of the self-serving bias by describing it to participants in some detail and administering a test to them to make sure they understood the description.
Also, just as our participants identified with either the plaintiff or defendant in the legal case, auditors are likely to identify with the firms they audit
The experimental intervention was successful insofar as participants became convinced that the negotiating opponent would be highly biased, but participants believed that they themselves would not succumb to the bias.
The fact that participants are unable to rid themselves of the bias when rewarded for doing so and when they’re told about the bias and their belief that they weren’t subject to the bias both demonstrate clearly that the self-serving bias is unconscious and not deliberate.
Participants were presented with eight arguments favoring the side they’d been assigned to; that is, either plaintiff or defendant, and eight arguments favoring the other side. They were asked to rate the importance of these arguments as would be perceived by a neutral third party.
Consistent with this interpretation, in another set of experiments, when the parties were presented with their role; that is, they were told that they’re either the plaintiff or the defendant, only after they read the case materials, the magnitude of the bias was substantially reduced, and almost all of the pairs reached rapid agreement on damages.
The task faced by auditors is, of course, differentfrom that faced by participants in our studies. However, there are many similarities. Like our participants, auditors need to sift through large amounts of information to arrive at a judgment.
In fact, auditors possess much greater amounts of information than participants in our experiment and there is, therefore, likely to be commensurately greater scope for biased interpretations of the information.
Needless to say, this is especially true if auditor firms are providing consulting services to the companies they audit or have a financial interest in the firm. In this case, they’ll have a strong desire to view the firm as a healthy profitable enterprise and will process the information accordingly.
Even in the absence of such explicit conflicts of interest, however, there are a variety of reasons why auditors may have an interest in taking a positive view of the firm they’re auditing.
First, the people who will be hurt by any misrepresentation are statistical. They cannot be identified at the time the decision is made. Researchers have found that people tend to be far less affected by imposing harm onstatistical victims than on known victims.
If a biased audit results in losses to investors, many people might lose a small amount of money, but it isn’t clear who will lose money. In contrast, the auditor is likely to be well-acquainted with the people who would be hurt by a negative opinion on the audit.