July 16, 2010
Are You Paying Employees to Cheat?
Do Incentives, Bonuses Encourage Cheating?
It's not surprising that some university students in recent experiments actually cheated to gain a little additional money. After all, numerous studies and surveys over the years demonstrate that large percentages of students - from high school to college level - cheat to gain better grades.
What's new about the just-published experiments at the University of Guelph, in southwestern Ontario, Canada, is the finding that cheating is greatest when the reward is offered in a target-focused productivity activity.
And co-author of the study, Fei Song, believes the findings relating to cheating by student participants could be similar to what might be happening with incentives and bonuses in the workplace among employees doing their jobs. Song, a business professor at Ryerson's Ted Rogers School of Management, puts it this way:
"We sampled from a population of university students. To what extent is it legitimate to generalize from the behavior of such a sample to a broader population of job-seekers? We argue that sampling from a student population in this context is appropriate because most of the students were actually working at part-time jobs, seeking jobs at the time of the study, or would soon enter the job market."
How the experiments were set up
The researchers from Ryerson University, Toronto, and the University of Guelph (U of G) asked more than 200 students at the U of G to create as many words as they could from anagrams, in one minute efforts.
Students worked in three groups using three different pay-for-performance activities as a financial incentive:
• A target-based activity (with students only getting compensation if they achieved a pre-set goal). As Song explains, "a person would be paid a lump-sum bonus if his or her performance level is at or above a pre-set, absolute benchmark." (In the experiment, the student earned $3.60 for each round in which he or she created nine or more words.)
• A linear piece-rate activity (with students getting a bonus for each correct word they created.) "A person would be paid $X for each piece he or she produces. In other words, there is a linear relationship between one's pay and one's performance," Song explains. (In the experiment, the student earned $0.40 for each word he or she created.)
• A tournament-based bonus system (with student's pay tied to the performance rate of other students). "A person would be paid $X for achieving a performance level above a certain - such as 85th - percentile, relative to his or her peers in the same budgetary period," Song explains. "In other words, a person would be paid a lump-sum bonus if his or her performance level is at or above a relative, floating benchmark." (In the experiment, the student earned $3.60 for each of the rounds in which his or her performance was at or above the 85th percentile, relative to other participants in the same session.)
The experiments involved more than just creating words, though. After each round, each participant would then check the work of another participant. Workbooks were exchanged at random. After the work was checked by the rater, each workbook was returned to the participant to whom it belonged.
Participants then were given permission to correct any mistakes made by the rater. It's at this step that students could cheat. The participants were asked to record the number of correct words they had created on the performance record sheet.
Unbeknownst to participants, the researchers were able to match participants' workbooks with their self-reported performance records (without being able to identify the individuals). The cheating participants were those who earned their bonuses by self-reporting more correct words than identified by the raters.
Final cheating results: The study reports the results this way: "The total number of overclaimed words was more than two and a half times as high in the target-based treatment than in any of the other three treatments." Further, "Warren Buffet once remarked, 'Managers that always promise to "make the numbers" will at some point be tempted to make up the numbers' (Buffett, 2003). Our experiment provides strong support for this insight."
Conclusions: The researchers conclude from these experiments that "financially salient numerical targets produce more cheating than do other compensation systems." A linear piece-rate system produces significantly less cheating than one based on a prespecified target. A tournament compensation system based on relative performance also results in significantly less cheating than a target-based system. Finally, "cheating is more likely under a target-based scheme the closer a participant is to the target."
Commenting on the results of the experiments, Song adds: "We feel that the power of incentives sometimes can be too powerful in the sense that people will be tempted to do unethical things."
The study raises two important questions, with Song's answers:
Q.1. In the study's experiments the students were cheating to gain a few extra dollars. What if they were striving to gain points that they could use to redeem merchandise from a catalog, for example. Would the level of cheating in such a situation be comparable to the level of cheating to gain a few extra dollars?
•Answer: "Incentives can come in many forms. Money is just one kind of incentive. As long as the form of the incentive is something that people value and strive for - which can be non-monetary such as status and prestige - we believe that it would have the same effect as money on people."
Q.2. What do you suggest employers and managers do to reduce employees' temptations to cheat in their efforts to achieve performance goals?
•Answer: "I would not necessarily rule out using a target-based bonus system. I believe it is very much like a two-edged sword. It is effective in inducing effort and thus increasing productivity, while at the same time it can also tempt certain people to cheat and make up their numbers. I think the key is to couple the target-based bonus system - for that matter, any pay-for-performance system - with a comprehensive internal monitoring and audit system, and/or ethics training."
Is Cheating, Fraud Widespread?
The University of Guelph experiments raise an important question: If university students will cheat just to gain a few additional cents or dollars in anagram games, how widespread might the occurrence of cheating and fraud be in actual workplaces?
For an answer, we asked Michael G. Kessler, founder and president of Kessler International, a firm that conducts thousands of investigations and audits of companies around the globe. In 1999, Kessler International released a study showing that 79 percent of employees steal from their employer.
Kessler isn't confident that the extent of cheating by university students in playing anagrams is a direct reflection of how extensive cheating might be in actual work activities. Says Kessler, "I would like to... determine if it [the study] takes into account the risks an employee would face if caught versus a student. Age differences might also play a role. A real world situation might provide completely different results."
Kessler suggests that it's not just employees working to achieve production incentives who may be tempted to cheat. "In this economy not only employees working based upon commission but salaried employees must be monitored."
The company's 1999 study found: 21 percent of employees will never steal from the employer... 13 percent will steal from the employer... and 66 percent will steal if they see others do so without consequences. The study also found that 92 percent of employees would lie on their job applications if asked [had] "they ever stole from a previous employer."
Would a similar study today get different results? Kessler believes so. As he points out, "the level of cheating and stealing in the world has increased as the morals and ethics of the workplace decrease."