For years, the “opt-out” approach to increase organ donation has been used as a classic—perhaps even tired—example of the power of nudges. The combination of defaults and status quo bias, in contrast with actively opting-in to become a donor, resulted in dramatically higher organ donor rates. Case closed!
Not so fast.
When recent headlines declared that a research paper[1] had thrown this accepted truism into question, a flurry of Internet chatter ensued. Opt-out for organ donation doesn’t work? Defaults are suspect? Against the already fraught context of the field’s replicability crisis, this new finding had the potential to rock basic assumptions of behavioral economics and nudging to its core.
On further inspection, however, the hand-wringing was unwarranted. The “nudge” to increase the number of potential organ donors through an opt-out default had actually worked just fine. But the context of going from organ donor checkbox to viable organs is more complex in reality. It turned out that organs from “opt-out” donors were far less likely to go to waiting recipients when families had the final say, or in situations where organ donation is less culturally accepted. In cases when emotionally distraught families didn’t feel that the presumed consent of a default option truly reflected their loved one’s active choice, it was easy to override: the default choice simply didn’t feel like enough of a decision.
In a September 2018 article Sarah Reid and I introduced a new model intended to define a new landscape for strategic design and behavioral science. In this model, we posited that behavioral and design problem-solving methodologies occupy a shared terrain, with each leading and supporting depending on the nature of the challenge. In situations where the solution centered on behavior change, behavioral science can take the lead while design lends a hand. In other settings, especially when creating the new or developing systemic solutions, strategic design can lead the charge, informed by behavioral insights.
Behavioral science-led, design-informed interventions often center on modifications to “choice architecture,” rethinking the user’s environment to address common behavioral tendencies and help individuals make better choices. This can take the form of adjusting how people perceive inputs and what counts as viable options, or by integrating behavioral guidance into processes to gently nudge us into “good” behavior.
Some use the metaphor of plumbing to illustrate the nature of behavioral solutions: correcting or optimizing infrastructure to help reroute convoluted flows and cut out unnecessary complexity. But installing plumbing in a house with an uneven foundation or optimizing a bathroom that's in the wrong location means only solving part of a problem, and may in fact invite other issues.
The organ donation scenario is a perfect illustration of why behavioral science benefits from design’s more personal contextual lens. Increasing the pool of potential donors is a behavioral challenge well-suited to nudges, but we don’t just need more donors: we need to increase the number of donated organs. In other words, the family’s role in agreeing to donate or withhold organs wasn’t some kind of extra detail, but a critical component of the problem to be solved. Solving for this requires understanding of the situational context to ensure we’re addressing the right problem and increase our chances of implementing a successful solution.
But that’s just a one instance. With a jump into our time machine back to 2008—the year of Nudge’s original release—let’s took at another example where behaviorally-led, design-informed approaches played a part.
The 2008 financial downturn evoked a widespread sense of customer insecurity, which in turn fed a general risk aversion to making major purchases. Car manufacturers were particularly hard-hit: the option of putting down thousands of dollars for a new car that would immediately begin to depreciate was off the table for many consumers.
Knowing that financial incentives can help overcome this hesitation, many car companies offered 0% down or deferred payments to attract potential buyers. While they may not have been explicitly informed by behavioral insights, instinctively these car companies’ offers were grounded in well-known behavioral strategies related to time discounting. No money down meant no immediate loss, and deferred payments relied on our optimism about the future. Yet sales remained flat, or even decreased… no one wanted to buy a car in such uncertain times.
Hyundai was in a worse spot than most. They had no financing arm to negotiate or provide better terms, and thus lacked the economic levers used by other companies. In spite of this deficit they hit upon an alternative: the Hyundai Assurance program.
This program recognized that the uncertain economic conditions were an important underlying cause of consumer reticence to buy, but also that the actual, concrete manifestation of that uncertainty was the vivid fear of losing one’s job. At the time, newspapers were full of stories about major layoffs and more cuts to come; given the availability heuristic and our generally poor perception of statistical probability, it’s hardly surprising that many people assumed they might be next on the chopping block, and were wary of buying a car today only to find themselves unemployed and saddled with car payments tomorrow.
Hyundai’s offer—"certainty in uncertain times"—brilliantly reset the terms of this contextual dilemma. Purchasers who made at least two payments and suffered a significant life disruption, such as involuntary job loss or personal bankruptcy if self-employed, could return the car for up to a year a for full refund and no impact on their credit score. The results were nothing short of miraculous. Hyundai sold 460,000 cars in 2009 while other manufacturers’ sales continued to stagnate, with only 350 people—less than 1 percent—ultimately using the Assurance program.
For the other 99 percent of customers, their lives played out exactly as they might have had the Assurance program never existed. But their perception of the decision was radically different due to Hyundai’s framing of the problem, which re-centered consumers’ perception of what counted as a valid choice by addressing their internal, identity-based choice architecture rather than just an external, environmentally-based choice architecture. Whereas most companies understood loss aversion, Hyundai recognized the importance of understanding the human context in this situation.
This addition of this personal and conextual lens provides a way to get more at the why (and, almost as importantly, why not) behind people’s decisions to engage. While some nudges are aimed squarely at changing behavior in a user’s best interest more transparently—increasing 401k savings is hard to argue with—many behavioral challenges require some degree of user agency or acceptance of what choices are even valid. A greater understanding of the personal and situational context—What’s my story? Who am I? With whom do I feel a sense of kinship?—can help us not only understand people’s motivations with greater clarity, but also grasp to what extent people’s options make sense in the narrative they tell themselves. The power of nudges comes from their ability to side-step human decision-making weaknesses; imbuing them with insight into people’s self-perceptions and what they value can make solutions more effective, and even help inform what is likely to work in the first place.
The implications of identity- and value-based framing show up in other types of decisions as well. Ask your average person whether they’d rather die or get a haircut and you’ll get peculiar looks: the value of one’s life will override the value of one’s hair every time. Yet for patients battling cancer with chemotherapy, this equation is different. Even when they objectively recognize that treatment is necessary to save or prolong their lives, the decision to embark on a chemotherapy regimen does not come lightly. The intensity of this decision-making can be baffling to oncologists, who might see their patient's’ hesitation to lose their locks as vain or strange in contrast with the rational path of starting appropriate treatment.
But is this really so surprising? Hair loss looms large for patients not just as an abstract form of loss aversion, but as a very concrete signal of illness, a marker that one has crossed over from being a person to a patient. It is tangible and vivid in the form of the image that stares back from the bathroom mirror, but also in the ways that others often see and treat people with significant illnesses as less than fully human.
Truly solving for loss aversion in this case requires looking beyond the theory that loss hurts more than gain feels good, or the simple equation of life > hair, to understand what real people value—and thus fear losing—in their actual context: for chemotherapy patients, hair is a proxy for much larger constructs of identity, control, and a sense of self-determination. In the same way that defaults increase the number of organ donors but only indirectly lead to more viable organs available for transplant, behavioral interventions that address medication adherence without also taking a person’s larger context, personal values, and sense of identity into account are powerful, but not sufficient. Seen through this lens, it’s easier to diagnose why some smart-on-paper interventions have gone wrong.
The 1970s introduced us to Scared Straight, a program created to re-direct teens heading toward a life of crime. The scenario went like this: a group of at-risk teens spent the afternoon at a prison, where they came face to face with incarcerated prisoners. Over the course of the visit, captured on film, the prisoners—hardened criminals and legitimately scary dudes—related their stories about where their lives of crime had led them, alternately scolding, threatening, and generally scaring the bejeesus out of these young people.
In some ways this program was ahead of its time, retrospectively applying behavioral insight at a time when the field of behavioral economics was but a glint in Kahneman and Tversky’s eye: vivid stories in place of abstract data to make a compelling case; the threat of loss aversion, in the form of a very concrete loss of freedom in prison; social proof, through direct connections between the youths’ current situations and the former lives of the prisoners. By all indications, the experience was a brutally effective cautionary tale. Interviewed on the way out, the teens were viscerally shaken, fully convinced by the need to course-correct their life trajectories. The footage was turned into an Academy Award–winning film, and variations on the program still exist.
But this wasn’t the success story it initially seemed. Followed over time, the teens involved in the program and others like it actually had worse outcomes than those with more traditional interventions. Why? Hindsight provides some clues. While the short-term effect was certainly powerful, it ignored the broader real-life context that the teens returned to. We know that present-tense experiences have outsize influence on our emotional and rational states, but they fade over time. In this case, as the afternoon dissolved into memory it was replaced by the temptations and social norms of petty crime that the teens were re-surrounded by on a daily basis. However terrifying the experience had been in the moment, it became abstract over time, and knowing you’d lived through it—That’s as bad as it gets? He wasn’t that tough—fed a tough-guy narrative that bolstered the youths’ identities as bad-asses rather than serving as a deterrent.
Let’s end on a success story: solving the stigma of free or subsidized lunches. For middle- and high-schoolers, qualifying for a free lunch can be social kryptonite. Given the choice to receive a free lunch or skip it altogether, the answer is obvious: better to go without food than to suffer the stigma needing it in the first place. Some schools made this choice easier in the wrong way, exacerbating the problem by creating different lines and using alternate payment processes for subsidized lunches that only called attention to the disparity.
From an economic point of view, the choice is still an easy one, but with a different answer: Who turns down free food? Even from a behavioral standpoint, adjustments to choice architecture can only do so much. Smarter placement and framing of options might help move the needle slightly—some New York schools went so far as to invite suited-up professional athletes from local teams to eat the subsidized lunches in an effort to make the option cooler—but the super-power strength of teenage in- and out-groups is far too strong. It’s a behavioral problem, but not one that can be solved through behavioral interventions alone.
Many large urban cities such as Chicago, Boston, and Dallas solved the problem not by adjusting the choice architecture of one half of a two-tiered system but by going after the bigger issue of stigma, leveraging federal programs to offer free school lunch for everyone. Not only does this benefit more kids, but it solved the problem at the root of the issue—the uneven foundation—rather than just fixing the broken plumbing.
In all these examples, the challenge was behavioral at its core, but required an understanding of the complexities of contextual uncertainty, identity and perception, or “construal,” to solve successfully. Where behavioral science contributes a critically important perspective on common cognitive processing errors, fields like design and sociology can help to uncover what feeds and flavors people’s perceptions of choice, contributing a “here and now” sense of user insight that helps to clarify what people value and, thus, what they fear losing most. As these stories indicate, we ignore this at our own peril.
[1] Yiling Lin, Magda Osman, Adam J. L. Harris, Daniel Read. Underlying wishes and nudged choices. Journal of Experimental Psychology: Applied, 2018