Become a Patron


Excerpted From:  L. Song Richardson, Cognitive Bias, Police Character, and the Fourth Amendment, 44 Arizona State Law Journal 268 (Spring 2012)(210 Footnotes) (Full Document)



The Fourth Amendment's proactive policing doctrine fails to attend to the cognitive biases that can affect officer judgments of suspicion. Well-established social psychological literature reveals that non-conscious and efficient cognitive processes can bias decision-making in predictable ways. For instance, fundamental attribution error can cause individuals to ignore the situational determinants of ambiguous behaviors and instead, to attribute these actions to an individual's disposition. This error occurs more frequently when the target of scrutiny is nonwhite because of the effects of non-conscious racial bias. Hence, as a result of psychological biases, officers are more likely to attribute the ambiguous behaviors of nonwhites to criminality and the identical behaviors of whites to external factors.

Ignoring the influence of cognitive biases on police decision-making creates perverse results at odds with the Fourth Amendment's normative goals. Rather than protecting individuals against arbitrary police intrusion, the doctrine unintentionally creates incentives for policing that encourages it. Thus, jurisprudential ignorance of cognitive biases has costs.

* * *

The Supreme Court's proactive policing doctrine was meant to prevent arbitrary encroachments upon individual liberty while simultaneously promoting effective law enforcement. However, it fails to adequately achieve either goal because it is based upon a behaviorally inaccurate conception of how officers make judgments of criminality. Subparts A and B explain why the doctrine fails to reduce arbitrary and ineffective policing. Subpart C offers some suggestions for reform.

A. Jurisprudential Ignorance

The seminal proactive policing case is Terry v. Ohio. In Terry, the Court held that officers were permitted to conduct stops and frisks of suspicious individuals, even in the absence of probable cause, as long as their suspicions of criminality were reasonable. Understanding that the decision might lead to unjustified and arbitrary encroachments upon liberty, the Court imposed two safeguards--articulation and particularity. These safeguards require officers to articulate the specific and particular facts that led them to conduct the Terry stop. A court then reviews these facts to determine if they provide a reasonable basis for the officer's suspicions.

Articulation and particularity are important because they permit judicial review of the stop and frisk as well as of the officer's credibility. Furthermore, they allow courts to determine as a normative matter whether certain criteria should never form the basis of a reasonable suspicion.

However, these safeguards have proven ineffectual because they do not account for the influence of psychological biases on police judgments. An example can help illustrate this.

Imagine that an officer on the beat observes two individuals engaged in a heated discussion on the street that eventually leads to an ambiguous physical contact. In one scenario, the two individuals are black. In the second scenario, they are white. Assume that both sets of individuals are engaged in criminal activity.

Cognitive biases may cause the officer in scenario one to interpret the ambiguous contact he observes as violent and aggressive. This will lead him to approach the individuals and, because of his interpretation of the contact as violent, conduct a frisk for his own safety. As a result of the frisk, the officer will find evidence of criminal activity and make an arrest. When the officer later has to articulate the particular facts that led him to conduct the Terry stop, he will have no difficulty doing so. After all, observing two individuals arguing in the street and then observing one shove the other certainly could lead an officer to reasonably conclude that he had just witnessed a criminal assault. Meanwhile, the officer in scenario two will likely interpret the contact between the white individuals as noncriminal horseplay. As a result, these individuals will not be stopped and frisked.

Thus, the effects of psychological processes on officer judgments can result in nonwhites being subjected to police action more often than their white counterparts, even when engaged in identical behaviors. The doctrinal safeguards of articulation and particularity, as currently operationalized, do very little to ameliorate these effects. As the above example demonstrates, it is easy for an officer to articulate the particular facts that led him to feel suspicious. However, articulation of these facts does not prevent an officer from acting based upon implicit biases that make him more suspicious of nonwhites than whites. An officer may feel genuinely suspicious, without realizing that those feelings were affected by non-conscious biases and that identical behaviors of a white individual may not have attracted his attention. Articulation and particularity, therefore, do not provide protection against arbitrary policing; they do not prevent officers from stopping and frisking nonwhites in circumstances where whites would not be. Additionally, courts are not equipped to determine whether the officer's feelings of suspicion, based upon the articulated facts, were the result of non-conscious biases.

B. Exacerbating Psychological Biases

The doctrine not only fails to prevent arbitrary policing, but also fosters cognitive biases by strengthening the association between nonwhites and crime. One way the doctrine achieves this is by making clear that race is relevant to the question of whether a reasonable suspicion of criminality exists. Allowing officers to consider race in the reasonable suspicion analysis communicates to officers that it is appropriate to associate race with criminality; after all, why else does the court permit them to consider a target's race, even when they do not have any specific information that a suspect of a particular race was engaged in criminal activity?

The same is true regarding proxies for race such as the moniker high crime neighborhoods. Labeling a neighborhood as high in crime is:

almost always determinative in legitimating the police conduct of stopping an individual. The conclusion in legal opinions, among scholars, and on the street is the same: a high-crime area designation almost always shifts the analytical balance toward a finding of reasonable suspicion.

Despite its importance to the reasonable suspicion analysis, however, courts rarely require any empirical proof to justify the officer's conclusion concerning the nature of the neighborhood. Rather, courts typically rely solely upon officer testimony and officers often describe entire neighborhoods as high crime areas. However, empirical methods exist for determining whether a neighborhood is high in crime. Those methods demonstrate that high crime areas are usually limited to certain blocks within a larger neighborhood.

The neighborhoods typically described as high in crime are disproportionately urban, nonwhite, and poor. This fact, coupled with the reality of residential segregation, means that entire neighborhoods of racial minorities are labeled as high crime. By failing to require officers to support their claims with empirical evidence, courts allow officers to view nonwhite neighborhoods as hotbeds of criminal activity.

Consideration of location or race is not always inappropriate. If the police are looking for a specific suspect who is known to live in a certain location or have a certain racial appearance, then it may be appropriate for officers to rely on these criteria. However, courts do not limit consideration of race or racial proxies to these situations.

The problem with decisions allowing blanket consideration of race and location is that they exacerbate the influence of psychological biases on police judgments of criminality by strengthening the association between race and crime. This, in turn, makes officers more likely to interpret the ambiguous actions of nonwhites with suspicion and more likely to act with aggression upon confronting them, all the while being unaware that psychological biases influenced their judgments and behaviors. Finally, self-fulfilling prophecy effects may lead to negative interactions that confirm and strengthen the negative stereotypes, even when the individual confronted turns out to be uninvolved in criminal activity.

Doctrinal fostering of cognitive biases is particularly problematic because the nature of policing already exacerbates them. For one, officers engaged in proactive policing are, by definition, searching for evidence of criminal activity. Research led by Stanford psychologist Jennifer Eberhardt demonstrates that when thinking about crime, civilians and officers alike non-consciously think about blacks. This activates negative racial stereotypes that can affect the interpretation of ambiguous behaviors. Her research concludes that:

[n] ot only are Blacks thought of as criminal, but also crime is thought of as Black. . . . [S] imply thinking of crime can lead perceivers to conjure up images of Black Americans that ready these perceivers to register and selectively attend to Black people who may be present in the actual physical environment.

This ready association between blacks and crime already increases the likelihood that officers will activate implicit biases while investigating crime. Enabling and encouraging these negative associations, as current doctrine does, will only exacerbate their influence on police decision-making.

Furthermore, officers are often deployed to urban, majority-black neighborhoods--the neighborhoods most often referred to as high crime neighborhoods--to engage in proactive policing. Even without the doctrine's encouragement, behaviors in these neighborhoods already are more likely to be viewed with suspicion than identical behaviors in similarly situated white neighborhoods. Finally, officers engaged in proactive policing often make quick decisions based upon limited information. Rapid decision-making increases reliance on stereotypes and heuristics. For all these reasons, the doctrine's negative effects on cognitive biases are multiplied.

While the doctrine encourages police interactions with nonwhites in a manner that exacerbates cognitive bias, doing so does not increase law enforcement productivity. Reconsider the scenarios concerning two individuals involved in a heated discussion resulting in physical contact. This time, assume that both sets of individuals are just horsing around. The white individuals in scenario two will not be stopped and frisked, but the black individuals in scenario one will be. As ambiguous behaviors continue to be interpreted with more suspicion for nonwhites than for whites, the hit rates (the number of stops and frisks that lead to evidence of criminal activity) will be lower for blacks than for their white counterparts. The hit rate data that exists often demonstrates this outcome.

The case of Illinois v Wardlow provides a compelling example of how the Court's failure to account for cognitive biases can nurture them while failing to increase productivity. In Wardlow, police were patrolling a neighborhood believed to be a hotbed of narcotics trafficking. Officers riding in the last car of a four-car caravan noticed Mr. Wardlow standing next to a building. Mr. Wardlow fled after looking in their direction. The officers caught him and upon conducting a frisk, found him in possession of a handgun.

The sole basis for the officers' suspicions of criminality was Mr. Wardlow's flight from the police in an area believed to be high in crime. Mr. Wardlow challenged this basis for the stop and frisk, arguing that flight from officers in a high crime area alone was insufficient to create a reasonable suspicion. The Court rejected this argument. It held that while mere presence in a high crime location is insufficient to support a particularized and reasonable suspicion of criminality, it is sufficient when coupled with flight upon noticing a police officer.

The Court's holding comports with the common sense intuitions of many that people do not flee from the police unless they have something to hide. In fact, the Court specifically condoned officers' use of commonsense judgments and inferences about human behavior in making reasonable suspicion judgments, despite the fact that these assumptions about human behavior are often flawed.

This turns out to be the case with respect to the relationship between criminality and flight in areas purportedly high in crime. In their important article arguing for a more transparent criminal procedure doctrine, Professors Meares and Harcourt examined a study, available at the time Wardlow was decided, which sheds light on the relationship between flight and criminality. The study, conducted by the New York State's Attorney General's Office, scrutinized the New York Police Department's stop and frisk practices over a one year period.

The study sought to determine how predictive certain criteria were of criminal activity. It found that when officers conducted stops and frisks based upon criteria that clearly [met] the constitutional standard of reasonable suspicion, the ratio of stops to arrests was around 7.3 to 1. Conversely, when stops were based upon constitutionally unjustified criteria, the stop to arrest ratio was 29 to 1. Comparing these ratios to stops conducted because the suspect fled led to results at odds with the Court's commonsense intuitions.

Stops made because the suspect fled had a stop to arrest ratio of 26 to 1, very close to the ratio associated with stops based upon criteria that fail to satisfy the reasonable suspicion test. In high crime areas, the stop to arrest ratio was still low, 20 to 1. When flight was motivated by the presence of a police officer, the ratio was 15.8 to 1. Remarkably, when the flight was motivated by the presence of a police officer in a high crime area, the very facts involved in the Wardlow decision, the ratio of stops to arrests was 45 to 1. Professors Meares and Harcourt argued that this astounding ratio:

is suggestive that in high-crime urban communities where the population is disproportionately minority, flight from an identifiable police officer is a very poor indicator that crime is afoot. The data thus reveal an irony. The [Wardlow] Court relied on commonsense to come to a conclusion that two indicators of criminality (flight from police and presence in a high crime neighborhood) more reliably justify police action than just one indicator (flight from a police officer generally). This commonsense conclusion may well hold up outside of high crime urban areas. In New York City, however, the data suggest otherwise.

These results are likely surprising to many since the association between criminality on the one hand, and poor, urban, nonwhite communities on the other, is so strong. Hence, it is easy to overlook the situational factors that can explain flight from the police in these areas. Yet, as noted by Justice Stevens in his dissent in Wardlow,

Even assuming we know that a person runs because he sees the police, the inference to be drawn may still vary from case to case. Flight to escape police detection, we have said, may have an entirely innocent motivation: It is a matter of common knowledge that men who are entirely innocent do sometimes fly from the scene of a crime through fear of being apprehended as the guilty parties, or from an unwillingness to appear as witnesses. . . . [A] reasonable person may conclude that an officer's sudden appearance indicates nearby criminal activity. And where there is criminal activity there is also a substantial element of danger . . . . These considerations can lead to an innocent and understandable desire to quit the vicinity with all speed.

Fundamental attribution error coupled with implicit racial bias can explain why situational influences are often ignored when individuals flee from the police in neighborhoods purportedly high in crime. Police on the lookout for criminal activity are steeped in non-conscious stereotypes that affect their decision-making. When they observe someone fleeing for no apparent reason, their immediate reaction is to chase the individual rather than to consider whether any other behaviors of the individual support the inference of criminality. Quick decision-making increases the use of stereotypes and heuristics. Furthermore, behaviors are more likely to be interpreted as suspicious when the target is nonwhite because of implicit racial bias. Hence, when cognitive biases are considered, it is not surprising that the police are often incorrect when they stop individuals who flee from the police in urban, nonwhite neighborhoods.

The problem with the Court's holding in Wardlow, then, is that it fosters existing cognitive biases. As a result of the Court's decision, police are entitled to assume that flight in response to the police in high crime neighborhoods is evidence of criminality. Rather than encouraging officers to consider the situational influences on a target's behavior, especially in urban, nonwhite neighborhoods, the court encourages officers to focus on dispositional explanations instead. By coupling suspicion with flight in these neighborhoods, the decision not only communicates to officers that their suspicions of criminality are justifiable, but also entitles them to act upon their suspicions. The decision has the effect of solidifying cognitive biases because the more relationships are practiced, the more likely they are to become automatic. The more often officers stop blacks, believing (even incorrectly) that they are acting suspiciously, the stronger the connection between blacks and criminality, both consciously and non-consciously.

Another problem with decisions like Wardlow that make race and racial proxies legally relevant is that they will likely influence police trainings. After Wardlow, for instance, police likely are trained that flight in neighborhoods they believe are high in crime will always justify a stop and frisk, without the necessity of engaging in further information gathering. Thus, rather than creating a doctrinal rule that encourages officers to pause, reflect and gather more information before conducting a stop and frisk, and thereby potentially correcting for the influence of cognitive bias on their decision-making, the decision encourages the quick, hunch-like thinking that tends to rely upon cognitive shortcuts or biases.

In sum, there is reason to believe that cognitive bias causes officers to disproportionately conduct stops and frisks of nonwhites. Cognitive biases can also explain why officers are more accurate when they make stops of whites; officers will require more evidence of criminal activity before they will interpret the ambiguous behaviors of whites as criminal. The current Terry doctrine does not protect nonwhites from being approached, stopped, and frisked more often than whites not because they are necessarily acting more suspiciously, but because psychological processes may increase the likelihood that officers will attribute their ambiguous behaviors to a criminal disposition rather than to situational influences. In fact, the doctrine is structured in ways that appear to nurture the cognitive biases that burden nonwhites.

C. Doctrinal Reform

This section imagines a Terry doctrine that accounts for the effects of psychological processes on police judgments of criminality. The goal is to frame the doctrine in a manner that not only encourages officers to exercise more care before acting on their suspicions but also promotes structural changes in policing institutions. The changes will help create a Fourth Amendment jurisprudence that is more consistent with the norm of reducing arbitrary policing.

1. Considering Hit Rates

Psychological biases are at their zenith when people make quick judgments based upon limited information. Hence, promoting high epistemic motivation will help to reduce their effects. One way to achieve this is by requiring officers to present evidence of their hit rates for Terry stops during motions to suppress. When officers understand that their accuracy will matter to courts in determining the reasonableness of a Terry stop, they will likely be motivated to exercise more care before conducting one. Importantly, hit rates will not be the only criterion courts consider. Rather, the officer's hit rate is meant to provide courts with more information to consider in the Terry equation.

* * *

2. More Meaningful Articulation

Imposing a more meaningful articulation requirement will also increase epistemic motivation. Currently, officers are only required to state the particularized facts that led them to suspect criminality. However, as the thought experiments described above demonstrate, merely stating the facts that led the officer to feel suspicious does not prevent arbitrary police intrusions based upon racial hunches.

Courts can increase the efficacy of the articulation requirement by requiring officers to state not only how their experience and training relate to their judgments of suspicion on a particular occasion, but also how they know that the factors they articulate correlate with criminality. For instance, if fleeing from the police in a high crime area is given as evidence of criminality, officers should present data demonstrating the probative value of this criterion. By requiring officers to collect this data, they will be forced to pay more attention to the relationship between the criterion they utilize and its relationship to criminality. Thus, a more robust articulation requirement will bolster epistemic motivation.

* * *

3. Limiting Consideration Race and Racial Proxies

Courts should also avoid making categorical judgments concerning the relevance of race and racial proxies. Otherwise, the association between race (or racial proxies) and criminality receives the court's imprimatur. This lends credence to the association between race and crime that strengthens the association, thereby increasing cognitive biases.

Furthermore, making race relevant increases the likelihood that police departments will train their officers to use race as one of the factors that makes behaviors reasonably suspicious. If anything, courts should hold that these criteria, in the absence of specific information that make them relevant, will never serve to establish reasonable suspicion. This will force officers to gather more facts to support their judgments of criminality since they will no longer be able to rely upon the target's race to bolster a weak reasonable suspicion case. In turn, gathering more facts will help reduce psychological biases' effects.

* * *

IV. Police Character

Likely, the most effective way to reduce the effects of psychological biases on policing is through institutional reform. Examining the institutional environment in which officers operate is critically important, including exploring how existing department incentives and culture influence officer motivation and information processing. I have explored these issues in depth elsewhere and will not rehearse those observations here.

Rather, in this part, I will focus on officer disposition. Officers' dispositions can affect not only how much influence cognitive biases have on their behavior but also their motivation to change. Some officers are simply more likely than others to engage in careful decision-making and to aspire to reduce the effects of cognitive bias on their behavior. * * *

In sum, focusing on officer disposition when making hiring decisions may result in a pool of officers who are more likely to be motivated to reduce the effects of cognitive biases on their behavior. This is important because one's peers can influence racial attitudes. These officers can help change the culture of police departments, increasing not only the likelihood that institutional changes will be made, but also their effectiveness. These officers will be more likely to be motivated to reduce the effects of cognitive bias on their behaviors by participating in trainings, by exercising more care in making judgments of criminality, and by supporting the institutional changes that are critical to mitigating cognitive bias. All this can lead to policing that is more protective of individual privacy while simultaneously increasing officer effectiveness and accuracy.

Associate Professor, American University, Washington College of Law. J.D., The Yale Law School; B.A., Harvard College.