Breaking the Silence: Can Behavioral Science Confront Structural Racism?

A few months after George Floyd’s murder catalyzed a national reckoning, in the middle of a pandemic disproportionately devastating Black and brown communities, and the day after Karon Hylton-Brown, a Black man, was killed during a police chase for not wearing a helmet while using a scooter, we facilitated what would, in another time, have been a straightforward behavioral science workshop. This particular event focused on encouraging people in U.S. cities to choose nonmotorized transport like scooters and bikes in order to reduce carbon emissions. But 2020 had a way of setting fire to what had previously been “normal,” or “comfortable.” and this experience was no different.

As seasoned researchers who have worked at the federal, state, and local levels of government using applied behavioral research, we’ve participated in and facilitated this type of conversation many times before. In those conversations, racism and racist structures are sometimes mentioned, but the focus is on identifying individual-level behavioral bottlenecks and barriers. Ultimately, this leads to individual-level behavioral interventions and leaves institutionalized racism unaccounted for and unchallenged.

But this was a different moment, and it called for a different kind of convening. Rather than opening the session, as we usually would, with an introduction to behavioral insights broadly and its application to “small, overlooked barriers” in the transportation space, we began by discussing the links between racism and transportation.

Our opening speaker, transit and health equity researcher Charles Brown, pointed to the abundant evidence of racist patterns in transportation including: police profiling of Black and brown people (especially men), race-based ticketing while walking, biking, and driving, and historical redlining policies that limit mobility, access, and well-being for BIPOC populations. We have led and participated in similar behavioral design workshops in domains such as poverty, health, and education, but the racist structures underlying these issues had never been centered in this way. We felt a wave of relief and exhaled a breath that we didn’t realize we had been holding.

As a field, we have failed to account for structural racism in our designs and interventions because the challenge feels too big for our tools and methods.

Our relief was short-lived. When we tried to pivot from Brown’s presentation, asking the group which policies and programmatic changes could help address issues like police profiling, they were silent. That silence was a call to action.

As a field, we have failed to account for structural racism in our designs and interventions because the challenge feels too big for our tools and methods. While our hesitancy is understandable—applied behavioral science alone won’t dismantle structural racism—we’ve confused our limitations with powerlessness.

Each of us—especially social scientists who study and seek to influence societal behaviors—has a role to play in the fight toward racial justice. For example, seriously considering racism in our research processes can expose and quantify the magnitude of racist practices, allowing us to minimize intervention side effects. This practice can even begin to challenge, if not dismantle, race-based barriers in our intervention designs. When we fail to play our role—even if our contribution seems small and insufficient—we become part of the problem. If behavioral interventions are simply laid over the roots of America’s racist policies and practices, our approaches and nudges will unintentionally reinforce existing unjust systems.

It’s time to investigate our own barriers and biases and lay out a pathway toward antiracist behavioral design. The field and tools of behavioral science can play a role in exposing racist structures and illuminating the challenges that lay ahead, but only by first examining what we can do better.

A first order problem is that as a field we rarely name, let alone challenge, underlying racist structures. Instead, we design around the racist elephant in the room. For example, a text message reminding people that scooters are accessible in their city might help encourage some groups to pick one up on their next trip, but it fails to name, challenge, or bring public awareness to the underlying issues of police profiling, redlining, and housing segregation that impede Black people’s freedom of movement.

Each of us—especially social scientists who study and seek to influence societal behaviors—has a role to play in the fight toward racial justice.

One of the field’s foundational insights is that human behavior is often triggered by situational features. However, we have failed to acknowledge that systemic racism—and its counterpart, white privilege—are ever-present, though usually unacknowledged, situational features of American life.

Our failure to acknowledge this reality shows up at the very beginning of research, warping the way we define a problem and how we go about solving it. For example, we start many of our projects by creating behavioral maps, which identify the forces in people’s environment that move them toward or away from certain behaviors. These maps help us target where an intervention might help encourage, or discourage, certain behaviors. But structural barriers like redlining and employment discrimination, while they may be acknowledged, rarely become the places where scientists ultimately focus their intervention design efforts. BIPOC communities’ behavioral maps feature different, heavier barriers that are not easily nudged to the side. For example, a text message might remove one downstream barrier by making the process of finding a bike or scooter easier, but it doesn’t acknowledge or challenge thornier upstream boulders like residential segregation or police profiling.

Beyond behavioral mapping, a foundational challenge is that we often fail to collect quantitative and qualitative data disaggregated by race and ethnicity. This prevents us from identifying how our behavioral interventions play out in different communities. This kind of incomplete data blinds us to a comprehensive understanding of the barriers preventing people from making decisions and taking actions. For instance, in the transportation space, most available research indicates that the lack of protected bike lanes is a major barrier to increased biking. But a sliver of research conducted specifically with people who identified as Black, Latino, or mixed-race suggests “arrested mobility” issues—like fear of police profiling and police violence that impact BIPOC communities, especially men—may also play a significant role. The fear of police violence and profiling is driven by redlining policies that segregated Black neighborhoods and then policed that segregation, often with murderous results. Yet this fear does not appear on a standard survey exploring barriers to biking, because this concern is virtually nonexistent for whites.

Even when we do collect disaggregated data, we fail to collect nuanced data. For example, disaggregation by race hides differential treatment by skin color that behavioral interventions mostly do not capture or even acknowledge. In America, social outcomes vary by skin tone. Darker skinned individuals (whether African American, Latino, or mixed), on average, make less moneylive in more segregated areas, and are more likely to be arrested than those with similar racial backgrounds but with lighter skin. We know skin tone is a barrier; yet we fail to consider or collect relevant data.

Another lens on this problem is the relative lack of racial and ethnic diversity in our field. Research indicates that “racial and ethnic” researchers are more likely to collect more nuanced data on race (for instance, using multiple race-related items rather than a binary variable). That is, diverse researchers collect diverse data. And yet, as a field, we have failed to fully include diverse partners in our work. In fact, the behavioral and social science workforce is less representative of racial and ethnic minorities than are biomedical sciences or engineering. In many subfields, BIPOC researchers are less likely to receive funding. In 2003, National Science Foundation research found that Black and Latino people made up only four percent and three percent, respectively, of the behavioral and social sciences research workforce. Ten years later, in 2013, representation had increased by exactly one percentage point. We can’t measure what we don’t ask about, and we can’t ask about what we don’t know to ask. Diverse voices, researchers, and perspectives help to ensure we are asking the right questions.

The field and tools of behavioral science can play a role in exposing racist structures and illuminating the challenges that lay ahead, but only by first examining what we can do better.

This is the time to act. The current moment of cultural reckoning has cracked open a space to imagine and work toward inclusive, antiracist behavioral design. And it has revealed the stakes: the promise of applied behavioral science to improve lives will remain unrealized in America until, as a field, we acknowledge and address our racist structures. While behavioral scientists alone may not be able to dismantle America’s racist superstructure, there are steps we can take immediately.

First, we can begin to expose and quantify the magnitude of racist practices by redesigning our research methods to ensure we are collecting more reliable and nuanced data on race and ethnicity, rather than continuing to accept the limitations of administrative data. For example, our hypothesized study on text messages to increase the use of bikes and scooters would not challenge underlying transportation barriers, but it could expose and estimate the size of the problem by intentionally collecting a diverse sample and documenting the differential use of bikes/scooters and barriers faced by BIPOC communities. Collecting this information may mean placing a higher value on rigorous qualitative data and insights (especially in preliminary research phases). And it means incorporating participatory formative research processes that engage diverse communities in designing the questions and objectives of a study.

In addition, we must diversify the field by recruiting BIPOC students early and intentionally, as efforts like the PhD Project have already started to do. All of this must be done with thoughtful engagement to ensure communities are included meaningfully and to avoid the tokenization of BIPOC scholars and partners.

Though these recommendations are not comprehensive, they represent a way forward—which begins by looking at our past. Pioneers of social psychology developed principles that enable us to consider and address the social contexts that create and maintain structural racism. As the beloved Lee Ross, whom the field recently lost, and his coauthor Richard Nisbett state in their classic work from 1991, The Person and The Situation, “There is a clear need for social scientists to identify ethnic differences, to explain them, … and to find ways to diminish their capacity to fuel conflict.” It’s time to go back to these roots and engage psychological research in the service of the broader public good. To paraphrase the poet Amanda Gorman, our field isn’t broken, it’s simply unfinished.