Subtitles and transcripts of podcasts

Subtitles and transcripts are available for some podcasts linked to this blog, for example for all videos on Under the video, select Subtitles available in, and to the right of the video, select Open interactive transcript.
For Russian readers / Для русскоязычных читателей: Для некоторых подкастов, описанных в блоге, например, для подкастов на сайте, есть субтитры и записи текстов. Под видео выберите Subtitles available in Russian, а справа от видео выберите Open interactive transcript.

Monday, October 29, 2012

Dan Ariely: The Honest Truth About Dishonesty: How We Lie to Everyone - Especially Ourselves

In The Honest Truth About Dishonesty: How We Lie to Everyone - Especially Ourselves, Professor of Psychology and Behavioral Economics at Duke University Dan Ariely describes the recent studies on psychology and economics of cheating. The experiments are so cool, and the book is so well-written, that I literally could not put it down. The research offers practical methods to decrease cheating. The new theory proposed by Ariely contradicts the currently prevalent rational model of crime, based on cost-benefit analysis, which is used to set our current policy.

One of the many different methods to measure dishonesty involves an experiment where volunteer subjects (typically, college students) were asked to quickly solve as many math problems as possible. The subjects were then paid 1 dollar (the amount varied) for each of the problems they solved, or rather claimed to have solved. In the "no cheating" condition, the subjects submitted the sheet with the solved problems to the experimenter, who checked the results and paid the subjects. In the "cheating possible" condition, the subjects were asked to shred the sheet with the problems after they were done, and then tell the experimenter how many they had solved (without him checking). The subjects did not know that the shredder was rigged and shredded only the sides of the sheet, allowing the researchers to check later how many problems the subjects actually had solved. Below are some of the surprising findings.
Very few people cheated by a lot. Most people cheated by a small amount. In the "no cheating" condition, they solved on average 4 out of 20 problems, and in the "cheating possible" condition, they claimed to have "solved" 6 problems. Thus, they cheated by claiming an extra 2 problems. Most of the money "lost" due to cheating was due to many people cheating by a small amount, rather than a few "bad apples". Out of 30,000 people tested, there were 12 big cheaters who, combined, stole $150. But there were also 18,000 little cheaters who, combined, stole $36,000!
Dan Ariely proposes a new "fudge factor" theory of dishonesty: "On one hand, we want to view ourselves as honest, honorable people. On the other hand, we want to benefit from cheating and get as much money as possible. Thanks to [cognitive flexibility], as long as we cheat by only a little bit, we can benefit from cheating and still view ourselves as marvelous human beings."
According to Ariely, here is the summary of forces that shape dishonesty:

Increase dishonesty:
Ability to rationalize. If the subjects are paid, instead of real $1 bills, in tokens, which they can change into $1 bills almost immediately, they increase cheating by 2-fold. According to the theory, they rationalize cheating because it's not money that they are stealing, but tokens. Think about implications of this for our increasingly cashless economy, or compensation based on stock options vs. cash. In another experiment, people who were annoyed by the experimenter, cheated almost twice as much as those who were not (to annoy the subjects, the experimenter picked up his cell phone in the middle of explaining the task to them). Here, the subjects might have rationalized cheating as justice. Ability to rationalize is the most general factor increasing dishonesty, and partly overlaps with other more specific factors below, such as creativity, one immoral act, etc.
Conflict of interest in action (Lewis Farrell,

Conflicts of interest. In an experiment designed to resemble a relationship between financial advisers and their clients, the "advisers" cheated the clients by inflating the "values" of items by 25%. Amazingly, adding disclosure only increased their cheating by another 25%, while the clients did not discount this increased level of cheating nearly enough. In other words, disclosure under these conditions was actually harmful to the "clients", who were being cheated.

Creativity. Subjects who engaged in creative activities (such as writing, painting, etc) more frequently, viewed themselves as more creative, or chose a greater number of creative adjectives (such as insightful, inventive, etc) in a word selection task, also cheated more often. Interestingly, intelligence was not correlated with dishonesty.
One immoral act. Subjects were asked to wear sunglasses before the experiment. The group that was told they were wearing fake Chloé sunglasses cheated twice as often as the group that was told they were wearing authentic sunglasses. (Actually, the sunglasses were authentic in both groups.)
Being depleted. If the subjects' willpower (self-control) was depleted, for example by getting them tired before the experiment, they cheated 3-fold more. Compare this with other large and systemic effects of depletion in the earlier post on willpower.
Others benefiting from our dishonesty. When the payoff from cheating was shared 50-50 with another subject, people cheated 60% more than they do alone. The cheating increased even more when they cheated altruistically, that is, when they knew that 100% of the payoff would go to another subject.
Watching other behaving dishonestly. Culture that gives examples of dishonesty. The subjects were college students solving math problems as described above. When another "subject" (actually, a paid actor who was wearing the sweatshirt of the same university) stood up after one minute and said that he had "solved" all 20 problems (an obvious lie), the rest of the subjects increased their cheating by 60%. Interestingly, if the cheating "subject" belonged to a rival group (wearing a rival university sweatshirt), the cheating decreased by 60%, relative to the regular "cheating possible" condition!
Enron CEO Jeff Skilling (above) and other executives promoted a culture of dishonesty at Enron (

No effect:
Amount of money to be gained. The amount of cheating did not change between the groups of subjects who were offered either 0.5, 1, 2, 5 or 10 dollars per each of the math problems solved.
Probability of being caught. The amount of cheating did not change between the groups of subjects who were asked to either (a) shred the entire sheet, (b) shred half the sheet, or (c) shred the entire sheet and pay themselves from a jar with cash.
Concerns about standing out. The amount of cheating did not change between the groups of subjects who were told in advance that an average subject in the past solved 4 (which was true) or 8 problems.
Decrease dishonesty:
Pledges and moral reminders in advance. Subjects who were asked before the experiment to recall the Ten Commandments did not cheat at all. This was independent on the level of religious belief, and on the number of the Commandments that the subject could recall. In fact, none of the 450 subjects, all UCLA students, could recall all 10. Amazingly, the results held true for atheists as well. Swearing on a Bible (for atheists!) also resulted in no cheating. MIT and Yale students who were asked to sign the pledge "I understand that this experiment falls under the guidelines of the MIT/Yale honor code" before the experiment, also showed no cheating. This is despite the fact that neither university has an honor code!
Signatures in advance. Subjects who were asked to sign the form with the math problems at the top (before solving the problems) cheated 4-fold less than those who were asked to sign at the bottom (after solving the problems). Note that we are typically asked to sign most forms, such as tax forms and insurance claims, at the bottom - precisely where they have less effect.
Supervision. Being closely supervised (all actions watched over) by another unfamiliar subject resulted in no cheating. However, if those who watch and those who are being watched were asked to talk to each other and get to know each other before the supervised task, the cheating was much higher.


RSA Animate - The Truth About Dishonesty (animation set to excerpts of the RSA talk by Dan Ariely, 11 min)
RSA - Free beer: The truth about dishonesty (full talk at the RSA, audio, 1 hr) 
The Leonard Lopate Show: Dan Ariely tells the Truth about Dishonesty (audio, 35 min)
Dan Ariely - The Honest Truth about Dishonesty | Point of Inquiry (audio, 35 min)

Dan Ariely also wrote other books, for example this great book about the amazing biases of our everyday thinking: Predictably Irrational,Revised and Expanded Edition: The Hidden Forces That Shape Our Decisions.

No comments:

Post a Comment