On the TV show Lie to Me, the lead character confidently declares that on average people lie three times in a ten minute conversation. He is a deception consultant who excels in reading micro expressions on people's faces to determine if they are lying. This character is based on the renowned psychologist Paul Ekman, whose work revolves around the idea that facial emotional expressions are universal and can be analyzed. He is the scientific consultant on the show and in fact deconstructs each episode in his blog, The Truth Behind Lie to Me. But given that it is a dramatic TV show, Lie to Me focuses on people with strong motivation to lie. Would everyday people with no specific motivation still engage in dishonest behavior when given the opportunity? Apparently the answer is yes given the real cost to the economy of low level dishonesty (returning clothes after wearing, taking office supplies, inflated insurance claims etc) which runs into billions of dollars a year. But what explains this behavior? Interesting answers were found by three researchers who ran a series of experiments to investigate this issue.
The basic idea being investigated here is called the Self-Concept. When tempted, people can choose to stay honest and maintain their perception of themselves as honest, or choose to be dishonest for the purpose of financial gain, but in the process compromise their image of themselves. Rather than being a either/or choice, people may choose to strike a balance by engaging in dishonesty that is low enough not to damage the self-concept, but high enough to provide financial gain.
All six of the experiments conducted by the researchers used the same basic setup where people were given some simple problems to solve and paid money based on how many they got right. The differences were in who made the evaluation of what is right. Sometimes an experimenter would check the answers (the control group), while other times the participants would be separately given the answer sheet and would need to figure out the number of answers they got right, tell the experimenter and get paid (the test conditions). In other words, the test conditions were set up in such a way that a person who was inclined to cheat was clearly able to do so. Across all experiments, people in the latter conditions got more answers right than those in the control condition. So it is quite clear that people cheat when given the opportunity.
It is also true across all experiments that low level cheating was happening. Given the simplicity of the problem solving task, it would not be difficult for participants to claim that they got all answers right. But people don't do that. Consistently they claim only a small proportion more than they should. But wait, you say, maybe only a few people cheat and they cheat a lot, while the others are all honest. In other words, the few bad apples theory might be in effect. The researchers checked for that and unfortunately it was not to be. The cheating was broad and shallow, not narrow and deep.
But there were some redeeming situations. When participants with an opportunity to cheat are exposed to religious reminders (The Ten Commandments) or moral reminders (University honor codes) the cheating completely disappeared. This was true even when a particular university did not even have a moral code of conduct! However, when indirect monetary mediums (such as tokens) are used instead of money, the willingness to cheat goes right back up. The researchers suggest this may explain the levels of cheating associated with office supplies, expense reports and such.
How is a person's self-concept being affected during this process? When (unobtrusively) asked to report on their level of honesty there is no difference between the groups, implying that those who are engaging in this level of cheating don't see themselves as cheating. Further, when the monetary incentives for cheating are significantly raised, the dishonesty actually disappears indicating that when people can't rationalize it they don't do it.
Interestingly when other non-participants are asked to predict outcomes, they guess that high level dishonesty would occur and that moral codes would have no impact. In other words, people are not good at predicting how they will act in these situations.
In a study that we conducted, about 30% of respondents said that they help themselves to office supplies. Strangely, this number goes up in companies where free coffee is provided! To see similarly interesting results click here. Feel free to send in questions you have always wanted answered.
The research reported in this write-up was conducted by Nina Mazar, Assistant Professor of Marketing at the University of Toronto, On Amir, Assistant Professor of Marketing at the University of California at San Diego and Dan Ariely, the James B. Duke Professor of Behavioral Economics at Duke University. It was published in the Journal of Marketing Research.