STEVE INSKEEP, HOST:
Judges, baseball umpires and loan officers constantly have to make judgments - does the person performing qualify for asylum? Is the pitch a ball or a strike? Should I approve this loan application? Since all these people are human, they make mistakes, and our social science correspondent Shankar Vedantam has come across a mistake that all make, and some research about that.
Hi Shankar.
SHANKAR VEDANTAM, BYLINE: Hi Steve.
INSKEEP: And the mistake is what?
VEDANTAM: Well, the mistake lies in seeing patterns where none exist. It's actually a common error, Steve. We often want to impose a sense of order on things that are actually random. We see this in casinos all the time. In fact, the phenomenon is called the gambler's fallacy. If you toss a coin up five times and it comes down tails five times in a row, you have a feeling that the next coin flip has to come down heads.
INSKEEP: It must be heads this time - put all my money on heads, sure.
VEDANTAM: Exactly. Now, in reality of course, there's a 50-50 chance the next coin flip is going to be heads. Each coin flip is completely independent and strings of heads or strings of tails are not surprising, but we're often surprised when they do occur.
INSKEEP: OK so how does that apply to this situation where people are judging balls and strikes or judging human beings in a courtroom?
VEDANTAM: Well, many analyses of the gambler's fallacy have typically been lab experiments. Kelly Shue, at the University of Chicago, along with her colleagues Daniel Chen and Toby Moskowitz, they analyzed 150,000 rulings by judges in immigration courts. So these were judges hearing asylum cases. The researchers analyzed, how often does an asylum approval follow a rejection or vice versa? Now, each case obviously has no bearing on the next case. So how judges rule on one case should not affect how they rule on the next case. Here's Kelly Shue.
KELLY SHUE: What we find is that judges tend to deny asylum if they granted asylum to the previous case and the opposite. So they tend to grant asylum if they just denied asylum to the previous case.
INSKEEP: Whoa. So it's like that person with the coin, flipping it heads or tails, they're thinking on some level, sometimes if I just had tails, the next one maybe ought to come up heads.
VEDANTAM: That's right. Now, actually, if you ask the judges, they will probably tell you that each case has no bearing on the next. But I think the gambler's fallacy often operates outside of people's conscious awareness. So very subtly, it's biasing people into making judgments. The researchers find the same thing in a separate analysis of nearly one million pitchers in major league baseball. Umpires have to call a ball or a strike depending on where the ball crosses home plate. And again this is especially true with close calls. Umpires don't like to call a series of balls or a series of strikes in a row - gambler's fallacy at work again.
INSKEEP: And this is all very subtle. It happens some of the time.
VEDANTAM: That's right. It happens very rarely, but because you're magnifying this effect over a large number of cases, the net effect can be potentially very large.
INSKEEP: You could have games decided or human lives changed in disastrous ways. So what is a solution to this bias then?
VEDANTAM: Well, as with most biases, Steve, the first step is just to be aware of the bias. But in one experiment that Shue and her colleagues conducted with loan officers in India, they found something that did work. This was a lab experiment where they had loan officers come and study loan applications and assess whether to offer a loan. Now, these were loans that had already been issued so the researchers knew which ones had turned out well and which ones had turned out badly so they could assess the accuracy of the loan officers' decisions. Researchers gave the loan officers a financial incentive for each correct decision. Here's Kelly Shue.
SHUE: We find that when the loan officers are given stronger incentives to be accurate that the magnitude of the gambler's fallacy significantly decreases.
VEDANTAM: So the thinking here, Steve, is that the financial incentives causes the officers to slow down, think hard and make a deliberate judgment rather than make a quick intuitive call.
INSKEEP: You're saying when some money is put on the line, we take the brain off autopilot and actually think.
VEDANTAM: So I don't think you can say this applies in every single case. For example, you can't offer financial incentives to judges deciding asylum cases. But in that case, there might be a slightly different solution. The researchers say that varying the kind of cases that the judges see consecutively could help. If the judges feel that each case is sufficiently different from the next, they might be less prone to gambler's fallacy.
INSKEEP: Shankar, thanks very much.
VEDANTAM: Thank you Steve.
INSKEEP: That's NPR's Shankar Vedantam who joins us regularly to talk about social science research and to explore his different ideas about the science of human behavior on his new podcast, Hidden Brain. Transcript provided by NPR, Copyright NPR.
