Incentives to cheat

See all posts

A fascinating joint study out of two business schools suggests that contingent compensation (read: performance pay) encourages cheating and sometimes even gives rise to asset misappropriation (read: stealing), important words of caution when considering performance pay for teachers.

Researchers divided the participants for the study into three tracks and then gave them a series of questions to solve. Each track provided a different form of compensation to the test takers: a flat-salary track which paid $50 per person regardless of how well they did on the test; a bonus track which paid according to the number of questions solved correctly, with a maximum payout of $50; and lastly, a penalty track, which gave participants $50 in advance, but took money away for any wrong answers.

In all three tracks the participants scored their own tests and then reported their scores to the experimenters. Little did the participants know that the experimenters would not take their word for the results, but instead would scavenge the trash bins, fish out the scratch work, and systemically verify the honesty of the reporting.

Not surprisingly, those on the bonus and penalty tracks were more likely to misrepresent their results than those who got $50 no matter how they did.

What was more surprising is that participants on the penalty track were more likely to cheat than those on the bonus track. The authors of the study explain that this finding comports neatly with the venerable results of Kahneman & Tversky's Prospect Theory (1979), which suggest that people are typically loss-averse: a person is pained more by a dollar lost than pleased by a dollar gained. In this case, subjects trying to protect money they perceived as theirs were more ready to cheat than those attempting to collect money they did not yet have.

But couldn't the penalty gang have stopped at cheating? Did they have to resort to petty thievery? It turns out that the penalty track participants were most likely to steal the attractive pens handed out to them at the start of the test. The authors conjecture that placement on the penalty track may have triggered a different norm of reciprocity in the subjects' attitude towards the experimenters, engendering antagonism and resulting in the aforementioned asset misappropriation.

These findings -- if accepted and corroborated -- seem to suggest that a system of performance pay -- especially, but not only, a punitive one -- carries with it risks of misconduct. Naturally, when apples are cut unequally, some people will reach out for a slice they don't deserve; others will begrudge not getting the slice they think is theirs.