Of all the things to study, lying and cheating might not seem the most edifying of topics. And yet billions of people do these things every day, sometimes with catastrophic consequences. In a fascinating talk at this year’s South By South West (SXSW) festival, renowned behavioural economist Dan Ariely and Sarah Szalavitz revealed the results of their recent study on cheating.
Ariely, who authored the best selling Predictably Irrational, already has a lot of evidence that most people are willing to cheat a little some of the time. To prove this point he began by asking the audience who had not cheated or lied at all in the last month — perhaps 10% of people put up their hands. And in the last week? Perhaps 20% of hands this time. And today? Most of the hands now go up. “I suppose it’s still early,” chuckled Ariely.
“And yet how many of you, despite all this lying, consider yourselves a good and moral person?” Not a hand stays down this time. This is typical, explains Ariely, because although most humans have a strong sense of right and wrong, the lines can blur depending on circumstances. “We have a ‘fudge’ factor — a range of immorality that we’re willing to live with.”
Ariely illustrates his point with a joke: a little boy is sent home from school for stealing a pencil. His father berates him for his misdeed: “Timmy, you should never steal, stealing is wrong. Besides, if you’d needed a pencil you could just have asked me. I can bring you as many pencils you need from my office.”
Ariely points out that this is a perfect example of contextual cheating: the father would never take a dollar from petty cash, but a half a dozen pencils are fair game because they are a few steps removed from money. “This is what happens with mortgage backed securities — the moment something is multiple steps removed from money, people will cheat more,” explains Ariely.
But what Ariely and Szalavitz are most interested in is an economic model of cheating, and how this relates to cheating in other spheres (or domains) of life. “We’re studying the ways in which ways people are willing to cheat and not cheat, and how these interact.”
“The standard economic model of cheating is attributed to [Nobel Prize in Economics winner] Gary Becker,” explains Ariely, recalling how Becker found himself choosing between parking illegally and being late for as meeting. “Every time we have an opportunity to cheat we do a cost benefit analysis.”
‘What the hell effect’
Ariely then outlines a standard test for cheating: participants are told to answer a set of questions on a test paper, with a strict time limit. They are then asked to mark themselves, and then shred their papers before reporting their mark to the adjudicators. But what the participants don’t realise is that the shredder is rigged and shreds just the edges of their test paper. Using this method, it’s been shown that most people get four questions right, but report that they got seven.
But, as we’ve already seen, cheating is highly contextual. Ariely has proved in previous experiments that people cheat less if they have been asked to remember 10 commandments before taking the test. They’ve even found that people given fake designer sunglasses to wear for a few hours before a test cheat more than those given real designer sunglasses.
The ultimate inducement to cheat though, is money. “When we test people where there’s monetary incentive to cheat, 63% of people switch at some point.” (ie start cheating) So far, so obvious. But their studies have found something unexpected: while most people start out not cheating, and end up cheating a bit, some people who start cheating go the whole hog. “We call this the ‘what the hell effect’,” chuckles Ariely, “it seems to be driven by the sense that if you’re going to cheat you might as well do it properly”.
But how do we reduce people’s tendency to go all the way, once they have started cheating? Ariely decided to ask a Catholic priest for economic rationale of confession. “Surely the rationale incentive is to cheat on the way to confession?” he enquired. The answer it seems, is the opposite of rational. “We’ve found confession — Catholic or otherwise — is very useful for curtailing the ‘what the hell’ effect,” says Ariely. If people are given a chance to confess their “sins” during a cheating test, this significantly reduces the effect.
So does cheating in one domain make you more likely to cheat in another? In their most recent study, Ariely and Szalavitz surveyed attendees of SXSW as well as a control group with some surprising results. The domains they surveyed were academics (eg. cheating on tests), work (eg. taking credit for someone else’s work), relationships (infidelity), legal cheating (like using someone else’s prescription medicine), games (such as using loaded dice) and technology (like pirating movies).
They found that this tech savvy audience were far more likely to “cheat” in the tech domain than in any other, although Ariely and Szalavitz differ in explanation. They also found that while men generally cheat more than women in each domain, women who do cheat tend to cheat across all the domains. So while men may cheat at, say, golf or cards and nothing else, a woman who cheats is likely to do it in all areas of her life.
‘Sticking to the man’
There also seem to be groupings within these domains. Work cheating and romantic cheating are considered different from everything else and don’t correlate with other cheating. A person may cheat their taxes a little, but that doesn’t mean they will be unfaithful to their spouse, or lie to get a promotion. What’s more, “cheating for money correlates very poorly with all other domains. It seems to be it’s own domain,” says Ariely.
Ariely is most interested in situations where there’s a conflict of interest, since this is a major driver of cheating behaviour. He relates a story in which a doctor he knew offered to tattoo one side of his face with fake stubble so that he would be symmetrical (Ariely was badly burned in an accident and one side of his face is hairless).
When Ariely refused the doctor went on a tirade: “Dan what’s wrong with you? Do you enjoy looking non symmetrical? Do you get sexual favours because women feel sorry for you?” When a puzzled Ariely went to a colleague of the doctor he found out he was about to publish paper on the tattooing and needed one more subject to do so. “This doctor wasn’t a bad person — he just had a huge conflict of interest.
They have also discovered a universal predicator for our tendency to cheat. “The trait correlated most closely to cheating is creativity, because it requires you to both lie and feel good about yourself. Creative people tell better stories. We can create all kinds of stories to justify the morality of our cheating — we’re very good at that.” For instance it’s easier to cheat your insurance company because you can construct the narrative that you’re restoring karmic balance by “sticking it to the man”.
But it seems this narrative is harder to construct when you are obviously harming others. When asked what they would do with the email addresses and Twitter passwords of everyone attending SXSW, less than 10% of those surveyed would even look at the data, while 80% would immediately report it to the company. Contrast this with the idea of getting unfair access to a discount coupon for SXSW, where 65% were willing to use it 55% to share with friends.
Will Ariely’s research ever allow us to eliminate cheating? I think he would laugh at the impossibility of the idea. But what it can teach us is how to design our systems to take this inherent moral flexibility into account. Perhaps a mandatory online confession before we fill in our tax returns might be in order? Sars e-confessions has a nice ring to it after all.