Why Behavioral Reforms are More Likely than Implicit Bias Training to Reduce Racial Conflicts in U.S. Policing

Phillip Atiba Goff, Jillian K. Swencionis, City University of New York, Susan A. Bandes, DePaul University

Michael Brown Jr.’s death at the hands of Ferguson, Missouri, Officer Darren Wilson was a critical point in the national debate about policing, as “Ferguson” became shorthand for growing public consciousness about racial disparities in police use of force. Continuing moral outrage has the potential to inspire essential reforms in policies governing police-civilian encounters. But effective reforms must be based on accurate understandings of human behavior.

One currently favored reform calls for training officers to put aside often unconscious racial biases. “Implicit bias training,” as this policy is called, amounts to a promising intervention but not a panacea, because it often rests on assumptions not supported by the available scientific evidence. We argue that bias training is maximally effective when it is part of broader, regularly evaluated efforts focused on fraught situations known to spark racial tensions in policing.

Myths about Implicit Bias Training

Implicit racial bias – defined as an unconscious mental link between a social group and a feeling or concept – is widely assumed to be a significant factor prompting police to use disproportionate deadly force against Black civilians. And if implicit bias is at fault, the reasoning goes, the solution is to train officers not to harbor these biases. But such rationale is flawed for several reasons. Implicit bias is only one of the psychological and social influences that may lead to racially disproportionate use of force. Too often, such use of force is attributed only to individual, front-line officers, obscuring the importance of situations as well as department culture and policies. Finally, even when implicit biases are partly at fault, the question should be whether implicit bias training, standing alone, can be effective.

  • Conventional wisdom focuses too narrowly on implicit bias. To be sure, implicit bias likely poses a formidable barrier to equitable policing. Pervasive social stereotypes make Black people seem dangerous and prone to criminal behavior – often, as research shows, evoking behaviors such as police use of excessive force that fit those stereotyped fears and expectations. But implicit bias is only one of the mechanisms that can lead to such errors. Another is explicit racial prejudice. And yet another is identity threat, referring to the counterintuitive finding that a concern about being perceived as racist may actually exacerbate racially disparate behavior.
  • More fundamentally, it has not been established that implicit bias training actually reduces either bias or disparate treatment. At best, this kind of training has been shown (in laboratory settings) to cause short-term decreases in trainees’ scores on measures of implicit bias attitudes, but not changes in discriminatory behavior. Becoming aware of one’s biases does not necessarily lead a person to change that behavior. Indeed, there is even some evidence that implicit bias training may backfire, amplifying racial bias. The reason for gaps may be that implicit bias training attempts to change individuals without reforming the culture or chronic situations of the institutions in which they work. Training tries to reform the internal beliefs of individuals, not their behaviors.

Our research supports a different approach. Instead of targeting police officers’ feelings and beliefs, law enforcement agencies should limit the types of situations that facilitate biases.

A Better Path Forward – Change Behaviors and Limit Tense Situations

Our approach acknowledges the extreme difficulty of changing internal beliefs and motivations. The implicit biases people hold are, after all, learned through repeated exposure to racial stereotypes. Training sessions focused on beliefs alone are unlikely to overcome the weight of longstanding mental associations, much less change behaviors.

Fortunately, effective reform does not require altering the beliefs or feelings of law enforcement personnel. It requires changing behaviors. Our approach targets bias-inducing situations. Factors known to increase the risk of racially biased policing include inexperience, time pressure, divided attention, hunger, stress, sleep deprivation, and identity threats. Furthermore, unrestricted discretion is a major culprit – as we explain here using the example of traffic stops.

Many of the Black civilian deaths fueling the current debate occurred during traffic stops for minor offenses such as a burned-out tail light (as in the cases of Walter Scott and Philando Castile) or a missing front license plate (as in the case of Samuel DuBose). Whether or not these stops were racially motivated, the following behaviorally focused steps could avert such tragedies:

  • Studies show that simply limiting police authority to stop motorists substantially reduces biased incidents. In situations where police stop drivers based on reasonable suspicion of criminal activity, Black and White drivers are pulled over at similar rates. In contrast, when officers conduct stops for minor traffic violations, Black drivers are pulled over at rates disproportionate to their population share – and tensions can flare over minor matters.
  • Departmental culture must be consistent with individualized training. Departmental policies communicate values and can back up incentives that encourage fair policing. To communicate egalitarian values, department leadership must actually enforce supportive rules. In the scenario described above, for example, a department can effectively communicate and enforce rules prohibiting traffic stops for minor traffic violations like a burned-out tail light or failure to signal a lane change – and it can make sure that officers are not rewarded for arrests or fines stemming from prohibited stops.
  • An assessment plan to evaluate the effects of the training is crucial. Until more data on outcomes is gathered, belief in the effectiveness of implicit bias training is an article of faith rather than a scientifically-grounded approach. Data must be regularly gathered on the causes of and ways to ameliorate racially disparate policing practices. This will require ongoing partnerships between researchers and law enforcement agencies. At the same time, existing knowledge should be deployed to create effective and workable policies – such as policies that limit or eliminate situational triggers for racially biased behavior. If such situations are minimized, individual biases and fears will have less room to operate.

Read more in Jillian K. Swencionis and Phillip Atiba Goff, “The Psychological Science of Racial Bias and PolicingPsychology, Public Policy, and Law 23, no. 4 (2017): 389-409.

March 2018