The Unexpected Effects of Nudging
Even the best designed nudges can go awry. What can we learn when they do?
Posted Jul 24, 2019
Vancouver, Canada, discourages the use of plastic bags by charging shoppers five cents per bag, but a local store owner decided to take things one step further. He attempted to shame customers who bought plastic bags by emblazoning them with embarrassing logos, like “Into the Weird Adult Emporium” and “Wart Ointment Wholesale.” Who would want to walk around with a bag implying you just came from a sex shop or have some weird skin thing going on?
Turns out, lots of people. Sales of plastic bags increased as people wanted to acquire the novelty bags. One might hope these bags will be less likely to end up in rivers and landfills, but I have my doubts: How long is that wart joke going to stay funny? Now the store plans to print the snide logos on canvas bags…which is even worse for the environment. In the end, this store may increase its sales but at the expense of our planet.
This is one example of a well-intentioned and reasonably well-considered nudge completely backfiring. People are complex, and when we try to subtly sway them with behavioral science, we must be ready for some unexpected (and sometimes unfortunate) outcomes. This is as true for experts in the field as it is for grocers. Two recent studies that used nudges to support college students show how unexpected behaviors can crop up as a result of even the best-designed interventions.
Discouraging Study Time
In a series of studies at the University of Toronto, researchers wanted to help students in introductory Economics classes improve their grades. The researchers’ central goal was to correct students’ misconceptions about how much effort was required to earn an A and encourage them to increase their study time to achieve their desired grade. They tried several interventions to improve students’ study habits, including goal-setting strategies, strengthening growth mindsets, and real-time coaching and support.
In the end, students studied less than before! What the researchers discovered was that students confronted by the knowledge that they needed to study more to earn an A revised their expectations downward about what grade they wanted. So instead of studying more, they just accepted that they would earn a B or a C given how much they were willing or able to study. Moreover, students who were studying more than necessary to earn an A also cut back on their effort. Study time appeared to have an upper limit—real or perceived—for many students, and their reaction to being nudged was to realign their academic goals to their status quo.
Reducing Student Borrowing
In some cases, nudges achieve the desired effect but not the desired outcome. Many college students are confused by the loan process and some borrow more than they need. For example, 26 percent of students borrow more when their maximum loan amount is increased, suggesting that they’re anchored to that figure rather than carefully considering how much they really need. Over-borrowing, the argument goes, leads to a higher likelihood of default or other future constraints (e.g., delaying homeownership) due to owing more money. Many efforts have been made, therefore, to educate students about college loans and align their borrowing decisions with their needs.
To this end, researchers sent text messages to students at the Community College of Baltimore County (CCBC) designed to:
- Empower students to make their own borrowing decisions and not be anchored by the amounts offered.
- Increase the salience of the future costs of borrowing (e.g., monthly payments).
- Connect students over text to a financial counselor at CCBC.
And it worked! Students borrowed 7 percent less (about $200) during the next term versus the control group, with some of that effect due to students who chose not to borrow at all. However, students who received these texts were 10 percent more likely to fail a course and 6 percent less likely to return the following year. Ultimately, students in the intervention were more likely to default, probably due to worse academic performance preventing them from graduating or from finding higher-paying work. It turns out that the long-term costs of over-borrowing may be offset by the short-term benefits of financial security during college.
What Can We Learn from Unexpected Effects?
These interventions (including the grocery store bags) were well-conceived efforts to alleviate social problems, but all resulted in negative outcomes. Focusing specifically on nudging college students, what can we learn from these unexpected effects?
1. Information is only Step 1. These studies provided students with additional knowledge around studying and borrowing, respectively, mimicking the default way in which we often try to sway behavior: with information. But information can be insufficient (or damaging) without a full understanding of why people act the way they do in the first place. Even in circumstances in which we know people have incomplete or incorrect information, motivational processes may be at work that cause people to interpret that new information in a way that conforms to their current behavior.
2. Meet students where they are. Despite the solid research underlying both studies, the researchers misjudged how students would respond to nudges. Students at the University of Toronto may have had limited capacity to study more and were discouraged by being nudged to exceed their limits. Even students with more time may not perceive it that way, or may not assign the same value that we do to spending additional hours studying for a higher grade. Students at CCBC may not have believed that they needed less money, only that they shouldn’t borrow what they need. These students, therefore, may have spent more time in paid employment or cut back on other expenses (from books to food) in order to make up for that financial shortfall.
3. Experimentation is crucial. It would be easy, but certainly undeserved, to fault these researchers for conducting studies that harmed students, even in some small way. But we cannot shy away from experimentation when it can teach us so much about how nudges operate in real-world contexts. Now these researchers or others can reconsider how best to leverage behavioral science as a means of improving students’ academic performance or making optimal borrowing decisions. Moreover, while nudges are sometimes criticized for producing small effects, they can also uncover key insights while producing only small harms. With this knowledge in hand, broader reforms to help students can take into account these potential problems and avoid proliferating these unexpected effects on a grand scale.
Barr, A. C., Kelli, A. B., & Castleman, B. (2019). The effect of reduced student loan borrowing on academic performance and default: Evidence from a loan counseling experiment. EdWorkingPaper: 19-89. Retrieved from http://edworkingpapers.com/ai19-89
Denning, J. T., & Jones, T. R. (2019). Maxed out? The effect of larger student loan limits on borrowing and education outcomes. IZA Discussion Papers No. 12239. Retrieved from https://www.econstor.eu/bitstream/10419/196737/1/dp12239.pdf
Oreopoulos, P., & Petronijevic, U. (2019). The remarkable unresponsiveness of college students to nudging and what we can learn from it. NBER Working Paper No. 26059. Retrieved from https://www.nber.org/papers/w26059