Women, Men, and Math Problems.

June 10, 2012 BY danariely

In the experiments my colleagues and I have run on cheating, we’ve used a task in which pride about personal performance and ability has no part. The matrix test is merely a search task (wherein participants find the two numbers out of 12 that add up to 10) rather than a skill. It’s not something you’re going to brag to your friends about in all likelihood.

Recent graduate Heidi Nicklaus of Rutgers University was interested in the opposite; she wondered how people’s pride about their perceived and imputed abilities would affect their dishonesty. Specifically, she was interested in gender stereotypes. We’ve all heard the stereotype that men tend to excel at math more than women, and that women can talk and write circles around men with their superior verbal skills. So the question was, if men are more proud of their mathematic ability and women of verbal, it might cause them to cheat more.

In her experiment, Heidi first primed her participants with two comical videos that exaggerate gender stereotypes (see below). Then participants were presented with one of two sets of fake data (presented as legitimate); one supported the math versus verbal aptitude stereotypes, the other countered them. Finally, participants took brief 10-question tests measuring both math and verbal aptitude, and were told they would receive $.50 for each correct answer. Similar to the experiments my colleagues and I have run on cheating, half of participants in each condition could cheat while the other half could not.

The results showed that when people could cheat, they generally did, which is what I’ve always found in cheating experiments. On average, people claimed one extra correct answer than when cheating was not possible (an average of 4 instead of 3 correct answers out of 10 on both math and verbal tests). No news here, so what about the effect of gender stereotypes? Did having them reinforced or, alternatively, countered before taking the test have any influence on cheating?

First, the hypothesis. What Heidi expected to find was that men and women would cheat along stereotypical lines, that is, that men would cheat more on math (to show that they did, indeed, excel in mathematics) and women would cheat more on the verbal portion for the same reason. So it was intriguing when Heidi found that men cheated more on math question than expected, but men and women cheated equally on verbal questions (rather than women cheating more as anticipated).

These findings—that people did not cheat more to keep up with perceived higher achievement by others—are similar to what my colleagues and I have found. In one experiment our results showed, similarly, that people cheated by the same amount regardless of whether they thought their peers solved an average of 4 or 8 out of 20 questions in a given amount of time (reporting an average of 6 correct answers). People cheated as much as they could justify, and apparently others’ performance is not of any great concern in this justification.

Oh, and as for the stereotype that kicked off the experiment: there were no differences in performance on math or verbal questions based on gender. So hopefully this harmful stereotype will fall by the wayside sooner rather than later, since nearly all similar studies yield the same conclusion.

Taxes and Cheating

April 10, 2012 BY danariely

Will Rogers once said that “The income tax has made liars out of more Americans than golf” and I worry that he was correct. During his confirmation hearing to become the Treasury Secretary, it was revealed that Tim Geithner failed to pay Medicare, Social Security, and payroll taxes for several years while he worked for the International Monetary Fund (IMF). When asked by Senator John Kyl (R AZ) during the hearing about the (more than $40,000) “mistake,” which Geithner blamed on the tax software he was using, he replied, “it was very clear that this was an avoidable mistake… You’re right. I had many opportunities to see it.” But he didn’t, apparently, and that was that.

There are many problems here—one of which is the possibility of a double standard that allowed Geithner to get away with this entirely (I am not sure if this is the case or not). I suspect that if he had he been working for a domestic monetary agency, that is, the IRS, he would have faced heavy prosecution, fines, and almost certainly been fired.  Also, as the future head of the Treasury, we might hope that he understands the tax code well enough to do his own taxes. Part of his defense, was, of course, that the code is too complex. Which is true, but in light of this, and his own errors, we might then hope he would be more aggressive about reforming the code, which he has not. The worst part of it, however, is the personal example he provided to the rest of the American taxpayers: do your taxes wrong, omit a few things, and if they catch you all you need is to pay it back — it’s basically okay.

I’m not calling for punishing Geithner (retribution isn’t necessarily helpful, not to mention it’s a little late), but as we draw closer to tax time, it’s worth recalling this incident and how it might affect the American public. In the research my colleagues and I have carried out on dishonesty, we’ve found repeatedly that people become more likely to lie and cheat after witnessing the dishonest behavior of others. In one of our experiments, we tested to see how participants would respond to a blatant act of dishonesty in their midst—would they think they too could cheat and get away with it, or would they perhaps straighten up and fly righter than ever? To find out, we gave participants 5 minutes to solve as many mathematical problems as possible (where they were instructed to find which two numbers out of 12 add up to 10).

In the control, where no cheating was allowed, the average student solved 7 problems, which gave them a pay off of $3.50 out of a maximum of $10 (if they solved all 20 problems). To see how witnessing and act of dishonesty would affect participants, we had one student—a confederate named David—stand up after only a minute and claim he’d solved all 20 matrices. The experimenter merely responded that in that case he could take his earnings and go. So how did the participants respond to this display when asked to self-report the number of matrices they solved? By cheating a whole lot: they claimed an average of 15 correct answers, more than twice the average score when cheating was not allowed.

Seeing someone cheat for their own benefit and then get away with it clearly has an impact on our moral behavior—loosening it to a substantial degree.

So, what does this experiment means for paying taxes? It means that the more we see politicians—the people who make our laws—fudge their taxes (which seems to happen continually), the more likely the rest of us are to adjust our understanding of what is right and wrong about paying our taxes, and do the same.

But there is hope.  When we ran the same experiment with one slight difference, we found that dishonesty decreased dramatically. This time, instead of looking like all the other participants, who were students at Carnegie Mellon University, we had our confederate wear a sweatshirt that located him within a different social group.  This time h was wearing a University of Pittsburgh sweatshirt (Carnegie Mellon’s neighboring and rival university). When the dishonest act was committed by a person from an out-group, we found that cheating decreased dramatically to the lowest level in all the experiments (participants claimed “only” 9 correct problems).

What this means is that if we think of ourselves and our politicians as being part of the same social group, we might follow their footsteps when we hear about another politician or celebrity who hasn’t paid taxes in years.  On the other hand, if we don’t think that we belong to the same social group we might not feel more justified in our own moral indiscretions, and instead be extra careful not to be confused with this other, not so moral, social group.

So the moral of the story is: when you settle in to work on your taxes in the next few weeks, try not to think about the individuals who cheat on their taxes—and if you can’t avoid thinking about them, at least try to separate your own social group from theirs.

Go forth and be financially virtuous.

Dan Ariely is the James B Duke professor of Psychology and behavioral Economics at Duke University and the author of (the soon to be released) The Honest Truth About Dishonesty.

Dishonest Drunks

May 13, 2011 BY danariely

Beer tapsWhen you think about behavioral science research, the image that probably comes to mind is that of laboratories, computers, surveys, electrodes, and maybe even rats — but you may not realize the amount of research conducted in the field. At the Center for Advanced Hindsight we certainly do our share of lab research, but we also like to shake things up and occasionally target the unsuspecting participant in their favorite local setting. For instance, you might find us at a popular eatery, your favorite independent bookstore, a busy shopping center, a science fair, or even driving around in our fancy research mobile.

On one of our latest excursions, we ventured out to Franklin Street, a hot spot for many Chapel Hillians to have a drink (or a few) and a good time. When the night was upon us we set out to answer the question: are you more likely to cheat when you’re drunk?

So we set up two research stations and waited for the bar crawlers to crawl. As the night progressed we surveyed the bar, recruiting bar-goers of varying drunkenness. Participants, many with drink in hand, played a 15-minute computer game that was designed to test their honesty. The game was a simple task where participants chose to pay themselves more or less money based on their choices in the game, and of course some of their decisions turned out to be more honest than others. We visited a wide array of bar scenes from the local band crowd to the underground pool players, the 90’s hip hoppers, the indie rockers, and even the Carrborites — and we found the same thing.

Our data shows a low to moderate correlation between cheating and drunkenness, which may suggest that the more alcohol you consume the more dishonest you become. Were the participants actually more dishonest? One could argue that perhaps that they were less capable of completing the task while intoxicated. Of course, we’ll need to keep looking into the possibility. And you can, too. Next time you are out with some friends, you might want to take a few minutes and conduct an “experiment” on your own.

~Jennifer Fink~

Gray Areas in Accounting

November 8, 2010 BY danariely

Every profession is bound by written and unwritten rules and policies; some of them are set by organizations while others are an integral part of the occupation you choose. For example, doctors have to swear by the Hippocratic Oath, lawyers cannot divulge any privileged attorney-client conversations, and priests cannot reveal what was said to them in the confessional. The same kind of ethical code exists in the profession of accountancy because it is a means of public service. As Robert H. Montgomery put it, Accountants and the accountancy profession exist as a means of public service; the distinction which separates a profession from a mere means of livelihood is that the profession is accountable to standards of the public interest, and beyond the compensation paid by clients.”

However, accountants are plagued by deep ethical dilemmas – there may be times when their employers ask them to twist and tweak the financial position of the company because they’ve had a bad year. The usual spiel given is that they’re definitely going to rake in the profits in the months that follow and that the deficits that have been covered up this year will more than be taken care of in the years to come. So the accountant is left wondering if he/she should be loyal to their professional ethics or show loyalty to the company that has hired them.

To overcome this type of problems, ethics is taught as a subject when you choose to study accountancy, because you are responsible not just to your employer, but also to the general public who believe in your reports and statements and take important decisions based on your word. An important question here if course is how effective are ethics classes, and even if they are successful how long would their influence last (a week? a month? a year? 2 years?).  It is hard to believe that taking one or two classes on ethnics while studying accountancy is going to have a long term effect, and most likely higher standards and more strict definitions of continuing are needed (and maybe also higher frequency of education).

In other cases accountants are torn between reporting misbehaviors that are going on and between minding their own business – should they open up a can of worms and be at the center of a controversy or just take heart in the fact that they were true to the ethics of their profession? This type of cases present the “action inaction bias” where in general people view their own actions as much more important than inactions – which means that accountants are more likely to care about their own actions than about reporting the actions of others.  Yet, accountants must remember that they are accountable for not just their actions, but also their non-actions, if either tend to affect the public adversely.

Overall the study of accountancy, and understanding its challenges is not only important in its own right, but it also provide an important case from which to view problems of conflicts of interest


This guest post is contributed by Omar Adams, he writes on the topic of online accounting degrees . He welcomes your comments at his email id: omaradams47@gmail.com.

Who cheats more?

July 25, 2010 BY danariely

This is a short video from a talk I gave on June 7th at the Booksmith in SF.

The question here, is who cheats more and who cheats less….

Before you watch the video, think about a country that you have family or social links to (not the US) and ask yourself if people in that country cheat more or less than Americans.

A) People in that other country cheat more than Americans
B) Americans cheat more than the people in that other country

Next, try to predict if you think that bankers cheat more than politicians or if politicians cheat more than bankers.

C) Bankers cheat more than politicians
D) Politicians cheat more than bankers

Now if you don’t mind post your 2 answers to Twitter together with the name of the country that you have selected. Please use my username (@danariely), and I will collect the responses from Twitter.

Now, you can watch the video and get the answers

How to commit the perfect crime

June 5, 2010 BY danariely

There is a certain perverse pleasure in contemplating the perfect crime.

You can apply your ingenuity to the hypothetical issues of choosing a target, evading surveillance and law enforcement, dealing with contingencies and covering your tracks afterward. You can prove to yourself what an accomplished criminal mastermind you would be, if you so chose.

The perfect crime usually takes the form of a bank robbery in which the criminals cleverly bypass all security systems using neat gadgets, rappelling wires and knowledge they’ve acquired over several weeks of casing the joint. This seems to be an ideal crime because we can applaud the criminals’ cunning, intelligence and resourcefulness.

But it’s not quite perfect. After all, contingencies by definition depend on chance, and therefore can’t ever be perfectly thought out (and in all good bank-robber movies, the thieves either almost get caught or do).  Even if the chances of being caught are close to zero, do we really want to call this a perfect crime? The authorities are likely to take it very seriously, and respond accordingly with harsh punishment. In this light, the 0.001 percent chance of getting caught might not seem like a lot, but if you take into account the severity of punishment, such crimes suddenly seem much less perfect.

In my mind, the perfect crime is one that not only yields more money, but is one where, if by some small chance you did get caught, no one would care, and the punishment would be negligible.

So, with this new knowledge how would you go about it?

First, the crime would need to be obscure and confusing, making it difficult to detect. Breaking a window and stealing jewelry is too straightforward. Second, the crime should involve many people engaging in the same type of crime so that no one can point a finger at you. This is why looting, though easy to detect, is much more difficult to get a handle on than a single robbery. Third, your crime will need to fall under the shady umbrella of plausible deniability so that if you do get caught, you can always say you didn’t know it was wrong in the first place. With this kind of defense, even if the public cares, the legal system may let you off easy. Moreover, plausible deniability allows you to apologize in the aftermath and ask forgiveness for your “mistake.”

If you really want to go all out, do something you can spin in a positive light, and maybe even create an ideology around it. This way you can then explain how you’re actually on the side of progress. Say, for instance, you’re “providing liquidity” and “lubricating the market” and thereby helping the economy – even if it happens to be by taking people’s money. You can also resort to opaque and promising-sounding language to make your case; you’re “restoring equilibrium,” “eliminating arbitrage” and creating “opportunity” and “efficiency” across the board.

Basically, just bottle snake oil and tell them it will cure, rather than cause, blindness.

Something to avoid, on the other hand, is anything involving an identifiable victim with whom people can sympathize and feel sorry for. Don’t rob one little old lady blind, or any one individual for that matter. It’s part of human nature that we care so much about blue-collar crime, even though the average burglary only costs about $1,300 (according to 2004 FBI crime reports), of which the criminal only nets a few hundred. Crimes like burglaries are the least ideal crime: they’re simple, detectable, perpetrated by a single or just a few people. They create an obvious victim and can’t be cloaked in rhetoric. Instead, what you should aim for is to steal a little bit of money from as many people as possible—little, old or otherwise — it doesn’t matter, as long as you don’t reverse the fortune of any one individual. After all, when lots of individuals suffer just a bit, people won’t mind as much.

So, what is the ideal crime?  Which activity is difficult to detect, involves many people, has plausible deniability, can be supported by an ideology and affects many people just a bit?  Yes, I think you know the answer, and it does involve banks…

Seriously, what we have here is a problem with our priorities. We have tremendous regulations for what is legal and illegal in the domain of possessions and blue-collar crime. But, what about regulations in banking?  It is not that I really think that bankers plan and plot crimes for a living (I don’t), but I do think they are continuously faced with tremendous conflicts of interests, and as a consequence they see reality in a way that fits their own wallets and not their clients.  The recent turmoil in the market is just a symptom of this conflict of interest problem, and unless we remove conflicts of interests from the banking system, we are going to be part of a long stream of perfect crimes.

This blog post first appeared on a website for a new PBS show called Need To Know

Fishing & cheating

July 5, 2009 BY danariely

I found this quote in a wonderful book called 3 men in a boat.  The book was written in 1889 by Jerome K. Jerome, and interestingly it does not seem that much has changed since then.

I knew a young man once, he was a most conscientious fellow and, when he took to fly-fishing, he determined never to exaggerate his hauls by more than twenty-five percent.
“When I have caught forty fish,” said he, “then I will tell people that I have caught fifty, and so on. But I will not lie any more than that, because it is sinful to lie.”

Cashier-Free Honesty Cafes – Will They Work?

June 19, 2009 BY danariely

What if on your next coffee run, you discover that Starbucks has started running on the honor system? All the baristas are gone, and in their place, you find Tupperware filled with coins and bills. Would you pay for your daily soy Latte? Or would you “forget” to shell out the five bucks? Be honest.

This, of course, is only a thought experiment, as I doubt Starbucks will be adopting the honor policy anytime soon. But in another part of the world, it’s a real question that residents are facing on a daily basis. As the New York Times recently reported, the attorney general’s office in Indonesia has been opening thousands of “honesty cafes” as part of its anticorruption campaign.

The idea is that these cashier-free cafes will teach people to be honest and curb the country’s corruption problem (which pervades business, politics, and education) by inducing residents – especially the young – to get into the habit of practicing honesty. As the Times reports, “…the cafes are meant to force people to think constantly about whether they are being honest and, presumably, make them feel guilty if they are not.”

It’s a laudable plan, and a lovely feel-good idea, but will it work? I have my doubts.

First I think that people will also cheat to a certain extent in these honesty cafes (as they do in our experiments). In fact, according to one Indonesian student, they already do: “Some of my friends don’t pay the right amount.”

But that’s not the worst of it. I worry that these cafes won’t just fail to discourage cheating – they will actually lead to more of it. In some of our research, we found that cheating on one occasion makes it easier for people to cheat again on a later task, because it alters their self-concept. (Think of dieting as an analogy: once you break your diet once, it’s that much easier to say, “Oh what the hey, cut me a slice of that chocolate cake; I’ll count calories again tomorrow.”)

With honesty cafés widespread, residents will have more temptations to cheat, more occasion to cheat, and maybe this will make it such that they will find it easier to cheat again in other contexts.

Maybe these cafes are a good idea, maybe it will not have any effect, but I worry that it might make things worse.

My TED talk has been posted

March 17, 2009 BY danariely

These days when there are lots of financial related scandals and cheating this topic might be particularly relevant.

I am not sure it is the best talk I ever gave, but if you have 17 min to spare…

Take me to TED

3 irrational lessons from the Bernie Madoff scandal

March 13, 2009 BY danariely

The first chapter of the Bernie Madoff fiasco has come to a close, with Madoff pleading guilty to 11 charges of fraud yesterday.

Madoff’s massive Ponzi scheme was horrific on many levels. But while we watch the next phase of the scandal, it’s important to ask: What lessons are we going to learn from this? I can see three lessons that relate to my work studying human irrationality — and in particular, some non-useful lessons we might learn.

One lesson that individuals and foundations are likely to take from the Madoff scandal is that in addition to diversifying their portfolio across several investments (stock, bonds, equity, cash), they also need to diversify their investments among several advisors. While the idea of diversifying among advisors has some merit — and it could reduce the exposure risk of another Madoff scandal — it will also make the task of managing portfolios much more difficult and much less efficient. Imagine that you have $1,000,000, split among four advisors. You will need a whole new level of coordination among them so they can have the right amount of cash, bonds, stocks etc., across all of your assets.

And I think that people will begin to over-diversify across investors. Why? Because when we have one large and salient instance in our minds, it can be so powerful that we overemphasize it. This same effect is very apparent in what we call “the identifiable victim effect,” and it is the reason that we overemphasize the risks of a shark attack, and underestimate the risks of riding a bike without a helmet. In general, what we find when there’s one single vivid event is that people overweight it — we focus on it too much. So that’s the first lesson: We’re going to learn from the Madoff scandal, but we are going to overdo it.

Another non-useful lesson that I think we will adopt is to start searching with more vigor for other bad apples. On one hand, it is clearly important to prevent more Madoffs, but at the same time I worry that as a consequence of searching for bad apples, we won’t pay enough attention to other financial behavior that might not be as badly wrong but that can actually have larger financial consequences.

In our research on dishonesty, we found that when we give people the opportunity to cheat, many of them cheat by a little bit, while very few cheat by a lot. In our experiments, we lost about $100 to the few people who cheated a lot — but lost thousands of dollars to the many people who each cheated by a bit. I suspect that this is a good reflection of cheating in the stock market, where the real financial cost of the egregious cheating by Madoff is actually a tiny fraction of all the “small” cheating carried out by “good” bankers.

The risk here is that if we pay too much attention to chasing bad apples, we might pay too little attention to the situations where the small dishonesties of many people can have large consequences (such as paying slightly higher salaries to cronies, making small changes to financial reports, doctoring documents, being slightly dishonest about mortgage terms), and in the process neglect the real economic source of the trouble we are in.

A third bad lesson that I think people will take from this concerns the way we define acceptable levels of cheating. In a study that may parallel Madoff’s egregious dishonesty, we again gave the participants the opportunity to cheat, while solving a puzzle quiz — but this time we hired an actor. This actor, posing as a fellow participant, stood up at the start of the session and declared that he had solved all the puzzles. Now the question is how his behavior would influence the other participants in the room — the ones who were watching him.

What we found is that when the actor wore a plain T-shirt, which made him part of the student group, cheating increased. On the other hand, when the actor wore a T-shirt of the rivaling university, cheating decreased. What this means is that when someone who is part of our own social group cheats, we find it more acceptable to cheat, but when people who are not part of our social group cheat, we want to distance ourselves from these people and cheat less.

Madoff was part of the financial elite — part of an in-group of our financial leaders. Think of all these people who were in his house, who knew him well. So now, when other people in this circle see him cheating, think about the long-term consequences: Would these other people in this financial industry now be more likely to take the immoral path? It doesn’t have to be another Ponzi scheme. It just means that, now that they have been exposed to this extreme level of dishonesty, they might adopt slightly lower moral scruples. Maybe they will start not letting their clients know exactly what they own and what they don’t own, or change a little bit the interest rate that they’re charging them … I don’t think that those in his circle will necessarily become more Madoff-like people, but I do suspect that they will get a substantial relief from their moral shackles. Sadly, that’s his legacy.

So, Chapter One of the Madoff scandal is over, but I worry that the negative downstream consequences of this experience are just starting …