As an undergraduate, at my school it was practically a requirement to steal silverware from the campus cafeteria. There were students who'd commandeered full sets of china. The desk clerk at my dorm used to say that the only thing we were learning from our college education was "how to steal."
Somehow it didn't seem wrong to us to steal from the cafeteria (though I drew the line at a single setting of silverware). Plus, we'd heard that at other schools, students used the cafeteria trays as sleds after the first winter snow. At least we weren't doing that (though arguably this was only because there are no hills in Chicago).
Clearly this "tradition" of petty theft was something we learned from our classmates -- but what exactly led us to believe that our unethical behavior was "okay"? It could be that since we saw no one getting caught or punished, we decided we wouldn't either. Or perhaps because the behavior was so widespread we never considered that it might be wrong. Or maybe our own sense of morality was modified by what we saw our friends doing.
A team led by Francesca Gino devised a clever study to test some of these explanations for unethical behavior by groups. They recruited 141 undergraduates from Carnegie Mellon University to participate in a study ostensibly about math skills. The students entered a testing room in groups of 8 to 14. They were given a set of 20 problems and told they had just five minutes to complete them -- an impossibly short period given the difficulty of the problems. They were also given an envelope with $10 inside. After the five minutes were over, the students were told to self-grade the test and pay themselves 50 cents for each correct response, depositing the change in a cardboard box for the purpose. All except one group of students put their answers in a paper shredder, so the experimenters would never know if they had cheated. The other group served as a control, and their questions were carefully graded by the experimenter.
Here's the key to the study: In some groups of students, a hired actor took the test along with them, pretending to be finished after just one minute and loudly proclaiming he'd answered all the questions correctly and conspicuously keeping all his money. It would have been obvious to the other students that he cheated. Half the time, he wore a plain white T-shirt, and half the time he wore the shirt of Carnegie Mellon's arch-rival, the University of Pittsburgh. The experimenters didn't track the individual answers to the tests, but they were still able to measure how much cheating occurred by seeing how much money the students kept. Here are the results:
The control group kept an average of $3.50, answering just 7 questions correctly. When cheating was possible and no actor was in the room, the students self-reported a much higher score, keeping $6 each -- obviously many students cheated. But when the actor wore a Pittsburgh shirt, students kept just $4 -- still significantly more than the control group, but significantly less than when they were on their own. Most interesting of all, when the actor wore a plain shirt and the students assumed he was a colleague, they kept significantly more than any other group: an average of $7.50 each, reporting 15 correct responses compared to just 7 in the control group.
Gino's team says this shows that peers have a very strong influence on ethical behavior. When the actor appeared to be a rival rather than a peer, cheating was nearly eliminated. But when he seemed to be a fellow student, cheating increased dramatically. Clearly having an opportunity to cheat and understanding that it was impossible for them to be caught influenced behavior as well: otherwise, there would have been no cheating when no actor was in the room.
In a second experiment, the researchers divided students into three groups: a control group where cheating was impossible, and two shredder groups where cheating was possible. Again, one of the shredder groups had no actor in the room, but in the third group, as the experimenter gave instructions to the group, the actor asked aloud, "So, is it okay to cheat?" The experimenter replied "You can do whatever you want." Here are the results:
When the actor asked if it was okay to cheat, the students retained significantly less money than when there was no actor in the room and cheating was possible. There was still more cheating than in the control group, but simply by bringing attention to the fact that over-reporting answers was, in fact, cheating, the total amount of cheating was reduced.
So it appears that all three of our initial questions about why we cheat play into real-world cheating. We're influenced by our chances of getting caught, by how much attention we're paying to the ethical issues involved, and whether or not people like us are doing it. And we reserve special disdain for our rivals, taking care not to behave in the unethical ways they do. Perhaps if the University of Chicago wants to cut down on theft in their cafeteria, what they really need to do is point out how often those unethical Northwestern students steal silverware.
Gino, F., Ayal, S., & Ariely, D. (2009). Contagion and Differentiation in Unethical Behavior: The Effect of One Bad Apple on the Barrel Psychological Science, 20 (3), 393-398 DOI: 10.1111/j.1467-9280.2009.02306.x
There are huge numbers of fascinating effects like this being described in the Behavioural Economics literature these days.
I'm about to run a similar experiment at my university, although what we're interested in is whether the participants are more likely to cheat the experimenter when he's perceived as coming from a competing college. Feel free to let me know your hunches on how it'll turn out.
Dan Ariely talks about this study in a TED talk. Here's the link.
At my school the justification for stealing from the cafeteria was, "I'm paying enough to go here, I deserve some free forks," etc. It seemed fair at the time, but of course it just drives the cost up more for the next group.
Like @HTF, we had the same justification for "borrowing" stuff from the cafeteria because we planned on leaving it behind when the school year ended. There's also something about the confines of a campus that makes things seem like they don't have consequences. Our group finally got in trouble when we tried to steal one of the couches and put it in our suite. Not a very smart idea.
Fun, but I have to wonder what the participants are really thinking. I suspect a significant portion would smell a rat and not take the study at face value. If the rat-smelling is easier in certain conditions than others, that might be what's driving the results.
Now, that is a really intelligent methodology for research.
I'd be very curious as to what happened if one had three different t-shirts. In addition to the white t-shirt and the t-shirt for the rival school, why not the t-shirt of the school itself? It might be interesting to see if having a more direct explicit connection the school makes more of them perceive the individual as part of their group.
re:#5 -- I agree.
Particularly in the second experiment described, the experimenter's response of "you can do whatever you want" doesn't bring attention to cheating as much as it brings attention to the fact that the experiment is ABOUT cheating.
The described explanation for the difference in results ("over-reporting answers was, in fact, cheating") is the opposite of what was communicated. The experimenter essentially stated that over-reporting is NOT cheating. ("So, we're allowed to steal silverware?" ... "You can do whatever you want.")
Hmm. I don't remember stealing silverware from the dining halls during my years a UChicago undergrad. Perhaps my memory is faulty or my education was incomplete :)
Another possible conclusion is that ethics are much more subjective and contextual than we imagine them to be. What in one situation would be considered unethical, in another would be considered acceptable.
This is hinted at by the final experiment where the students are told they can do whatever they want. Cheating, in this particular context, is no longer seen as unethical.
Two variables that I believe would produce interesting results on the outcomes would be the perceived purpose, or end goal, of the test, and the perceived identity of the person administrating the test.