Sara Dommer, an assistant professor of marketing at Penn State University, led a study exploring why people cheat on tasks like completing crossword puzzles, Wordle, or counting calories when the rewards are purely intrinsic.
"I found that people do cheat when there are no extrinsic incentives like money or prizes but intrinsic rewards, like feeling better about yourself," Dommer stated.
Her research delves into the concept of diagnostic self-deception. "Diagnostic self-deception is when individuals cheat but deceive themselves by attributing their heightened performance to their innate ability instead of the cheating," she explained. "For this to work, it has to happen via diagnostic self-deception, meaning that I have to convince myself that I am actually not cheating. Doing so allows me to feel smarter, more accomplished or healthier."
Dommer conducted four studies to determine if people will cheat when rewards are solely intrinsic and what drives the feeling of accomplishment despite the cheating. Her findings were published in the Journal of the Association for Consumer Research.
The Calorie Counting Experiment
In the first study, 288 undergraduate students received menu information about three days' worth of meals. They were tasked with entering calorie information into a food-tracking app. The students were divided into two groups: one group received additional calorie counts accompanying the meal descriptions, while the other group did not.
The app listed five potential calorie options for each food entered. For example, the calorie counts for three pancakes with butter ranged from 300 to 560 calories. The group without specific calorie information could have averaged the five options to get a better idea of the true caloric value of each meal.
Instead, participants in the group without specific calorie information tended to enter fewer calories than the group that received specific calorie information. This suggests that people will cheat for intrinsic benefits—in this case, feeling healthier.
Testing Intelligence with an IQ Quiz
In the second study, 195 participants recruited on Amazon Mechanical Turk were divided into a control group and a cheat group. They were asked to complete a 10-question multiple-choice IQ test. Participants in the cheat group were told that correct answers would be highlighted so they could keep track of their progress.
After the exam, participants in the cheat group were asked to enter the number of questions they got correct, while the control group had to estimate their scores. Then the program graded both groups' tests, and participants were asked to predict their score on an additional 10-question IQ test that did not allow for the possibility to cheat.
Dommer found that participants in the cheat group reported higher scores than those in the control group. Based on their self-reporting, participants in the cheat group overestimated their performance on the second IQ test.
"The results suggested that participants in the cheat group engaged in diagnostic self-deception, believing that their performance was due to their intelligence and not cheating, though their scores on the second test suggest otherwise," Dommer said.
Unscrambling Words and Self-Perception
The third study mirrored the second but involved unscrambling letters to find a word. Individuals assigned to the control group had to enter their answers in an answer box, while those in the cheat group were shown the correct answer after three minutes and asked to self-report their scores.
Participants rated on a scale from one to seven how much their intelligence and the task's difficulty contributed to their performance. They also answered the question, "To what extent do you agree that unscrambling words is an accurate test of intelligence?"
Dommer found that those in the cheat group reported successfully unscrambling more words than those in the control group. Compared to the control group, participants in the cheat group were more likely to attribute their performance to their intelligence. They were also more likely to consider the task a legitimate test of intelligence.
"Participants in the cheat group engaged in diagnostic self-deception and attributed their performance to themselves," Dommer noted. She explained, "The thinking goes, 'I'm performing well because I'm smart, not because the task allowed me to cheat.'"
Financial Literacy and Reducing Cheating
In the final study, 231 participants were asked to take a financial literacy test. Participants were divided into control and cheat groups. About half of each group read a statement about how most American adults cannot pass a basic financial literacy test.
Dommer thought that introducing uncertainty about their own financial literacy might make individuals value accuracy over performance and reduce cheating. After taking the test, participants rated themselves on 15 traits related to financial literacy.
She found that introducing the uncertainty statement decreased cheating as individuals sought more accurate measures of their financial literacy. "How do we stop people from engaging in diagnostic self-deception and get a more accurate representation of who they are? One way is to draw their attention to uncertainty around the trait itself. This seems to mitigate the effect," Dommer said.
Challenging Assumptions About Cheating
"Our society tends to think of 'cheating' as a strategic, intentional act," Dommer stated. "This work suggests that at times cheating happens beyond conscious awareness."
"I don't think there's a good cheating or a bad cheating. I just think it's interesting that not all cheating has to be conscious, explicit and intentional," she added. "That said, these illusory self-beliefs can still be harmful, especially when assessing your financial or physical health."
This article was written in collaboration with generative AI company Alchemiq