Implicit Bias and Equal Protection: A Paradigm Shift


Yvonne Elosiebo



In a society that touts grand egalitarian principles, one must reckon with the reality of racial, financial, carceral, educational, and health disparities. Certain groups consistently differentially perform in almost every metric, despite the contention that they are accorded the same opportunities. This article explores how unconscious bias has played a central and unacknowledged role in creating disparate outcomes among racial groups. While many people claim to be racially unbiased or even colorblind, various studies that rely on the principles undergirding social cognition and the Implicit Association Test show that those same individuals unconsciously express preferences for and attribute positive characteristics to individuals in their social and racial cohorts. They simultaneously express negative behaviors toward and attribute negative characteristics to those outside of their social and racial groups, most notably black individuals. Supreme Court precedent has exacerbated the disparate outcomes in the carceral and legal system by failing to incorporate the findings of implicit bias research into its decision-making and continuing to require that plaintiffs show discriminatory intent in order to establish a constitutional violation of the Equal Protection Clause of the Fourteenth Amendment. This article suggests that the Supreme Court adopt a new standard for Equal Protection violations—discriminatory negligence. Recognizing that actors are mostly unaware of their biases, a discriminatory negligence theory creates a duty when the actor is externally confronted with evidence of her bias, and a prima facie case is established when the actor fails to remedy the effects of that bias. The article then explores the proposed standard in practice.

I. Introduction

Inequality in the United States is pervasive. In 2015, Blacks and Hispanics made up 32% of the total United States population, but 56% of people incarcerated.1 In 2014, black people alone made up 2.3 million of the 6.8 million total incarcerated persons.2 African Americans3 “are imprisoned for drug offenses at 13 times the rate of their white counterparts.”4 In Pennsylvania and around the country, “at any given poverty level, school districts that have a higher proportion of white students get substantially more funding than districts that have more minority students.”5 Whites “have a longer healthy life expectancy than [B]lacks.”6 And the unemployment rate for African Americans is consistently around twice the national average and more than double the unemployment rate for whites.7 Despite the increasingly stark disparities, the United States continues to proclaim a national identity of racial equality, egalitarianism, and democratic ideals.

Why do these racial inequalities still exist in 2018? While a number of historical and political actions have contributed to the status quo, courts also play a direct role in these unequal outcomes.8 In 1976, the United States Supreme Court set the standard for proving unconstitutional racial discrimination in Washington v. Davis.9 Justice White, writing for the Court, required plaintiffs to show discriminatory intent—or that an actor had the specific purpose to discriminate—against a particular racial group in order to prevail on such a claim.10 Discriminatory intent is difficult to prove. Some have argued that by making the remedy for being impermissibly discriminated against effectively unreachable, the Supreme Court eliminated the remedy altogether.11 Recent studies on implicit bias show that unconscious racial bias, rather than overt racial bias, is a better indicator of people’s tendencies to discriminate on the basis of race.12 Implicit bias is a phenomenon by which people outwardly profess to hold race-neutral views of others, but often unconsciously treat people of different backgrounds less favorably than others.13 In other words, many people do not have an intent to discriminate and rarely discriminate purposefully, but instead act more favorably toward certain people with whom they identify more and, unconsciously, less favorably toward people with whom they identify less because of unconscious biases formulated through human development and social interactions. The notion to which the Supreme Court adheres—that only an overt intention to discriminate on the basis of race qualifies as impermissible bias—is outdated.

This article proposes to revise the standard for proving unconstitutional racial discrimination to better align with current research on implicit racial bias. Rather than requiring plaintiffs to show discriminatory intent to prove racial bias, courts should only require a showing of discriminatory negligence, based in part on implicit bias research. Negligence is one of several mental states widely recognized in both tort law and criminal law.14 Negligence, as opposed to purpose or intent, knowledge, and recklessness, is the only mental state that finds culpability even where the actor is not (but should be) aware of the risk that her actions create and perpetuate.15 As such, discriminatory negligence is a more accurate standard for discrimination claims, where the perpetrator should have been aware that her actions are impacted by racial bias and have a high risk of producing a racially disparate impact.16 Government bodies or public officials particularly should be aware of such a risk because they have a public duty to treat every individual, and to create laws, equitably, ensuring a “fair” distribution of benefits and burdens across all groups. Given this new and improved understanding of how bias operates, influences actions, and leads to discrimination,17 the Supreme Court should adopt this new standard of review and overrule Washington v. Davis.

The first section examines the historical and modern discourse surrounding racial discrimination in the United States and the role that the Supreme Court may be inadvertently playing to exacerbate racial discrimination. The section examines Washington v. Davis, which served as the impetus for modern jurisprudence articulating the standard for unconstitutional racial discrimination, and a number of cases in which the Davis standard is applied. The second section delves into the mechanics of implicit bias, how it affects every person’s actions and associations, and its limitations and challenges. The final section discusses the proposed standard of discriminatory negligence and how to prevail on a claim of unconstitutional racial discrimination. The section also examines how the new standard might play out in the real world, using Washington v. Davis as a case study. The article closes with a discussion of why other solutions, such as eliminating the mens rea element altogether, were not pursued, as well as the potential impact of the changed standard.

II. Argument

A. Notions of Equality

The United States prides itself on being the land of the free, a land of equal opportunity for all, the Democracy of democracies. The founding fathers stated “that all men are created equal, that they are endowed by their Creator with certain unalienable Rights.”18 Abraham Lincoln stated in the Gettysburg Address that the United States was “conceived in liberty and dedicated to the proposition that ‘all men are created equal.’”19 The Fourteenth Amendment to the Constitution mandates that no state shall “deprive any person of life, liberty, or property without due process of law; nor deny any person within its jurisdiction the equal protection of the laws.”20 Indeed, its principal author, John Bingham, wrote that it “protect[s] by national law . . . the inborn rights of every person” within the jurisdiction of the United States, particularly when those rights are violated by States.21 The Fifteenth Amendment states that “[t]he right of citizens of the United States to vote shall not be denied or abridged by the United States or by any State on account of race, color, or previous condition of servitude.”22 Though this amendment speaks specifically about voting, it represents the prevailing egalitarian views at the time that continue into the present. The Universal Declaration of Human Rights, to which the United States became a signatory in 1948, states that “[a]ll human beings are born free and equal in dignity and rights”23 and that “[a]ll are equal before the law and are entitled without any discrimination to equal protection of the law.”24 The United States also became a signatory to the International Convention on the Elimination of All Forms of Racial Discrimination in 1966.25 Yet despite these statutory, treaty-based, and formal pronouncements committing the United States to racial equality, the data reveal a disconnect between policy and practice.

The United States Supreme Court also holds these egalitarian ideals as guiding standards. In Brown v. Board of Education, Chief Justice Warren stated that the Court first interpreted the Fourteenth Amendment as “proscribing all state-imposed discriminations against the Negro race.”26 The Court in Regents of the University of California v. Bakke stated that Congress enacted Title VI to “halt federal funding of entities that violate a prohibition of racial discrimination similar to that of the Constitution.”27 The Supreme Court reiterated this ideal in Fisher v. University of Texas at Austin by stating, “Distinctions between citizens solely because of their ancestry are by their very nature odious to a free people, . . . contrary to our traditions and hence constitutionally suspect.”28 Because Americans think of principles such as freedom and equality as universal and accorded to all men without regard to race, they conclude that eventually, in an essentially good and fair country, these high ideals will prevail over the more entrenched values that divide people.29 Not only have Americans touted these notions of equality for many years past, they also vehemently denounce actions that overtly challenge these ideals.30

1. The Reality of Racial Bias in the United States

Despite these lofty ideals of equality for all, national statistics tell a very different story. A 2010 study by the Agency for Healthcare Research and Quality found that Blacks, American Indians, and Alaska Natives receive worse health care than whites in 40% of its measures.31 Auto-loan borrowers of color tend to pay more than similarly-situated white auto-loan borrowers because of discriminatory borrowing programs.32 In Pennsylvania, a study found that funding gaps in education existed solely based on the racial composition of the school.33 A school district’s funding level was reduced solely by the increased presence of minority students.34 The U.S. Department of Housing and Urban Development found that “[r]eal estate agents and rental housing providers recommend and show fewer available homes and apartments to minority families, thereby increasing their costs and restricting their housing options.”35 Nationwide, people of color are more likely to be arrested and prosecuted and are more likely to receive harsher sentences upon conviction.36 The United States’ failure to achieve equal treatment and equitable outcomes in health, lending, education, housing, and criminal justice perverts the very ideals upon which the country was founded.

Racial profiling by law enforcement and the correlating criminalization of people of color have recently gained an unprecedented amount of attention.37 Studies have shown that whites and minorities largely commit drug crimes at the same rates and that whites commit the majority of violent and property crimes; yet people of color are arrested and convicted in disparate numbers.38 Both Democratic and Republican administrations admit that “racial profiling is unconstitutional, socially corrupting, and counter-productive,”39 yet the practice continues and has become “an affront to the promise of racial equality.”40 The international community has defined racial profiling as “the practice of police and other law enforcement officers relying, to any degree, on race, color, descent or national or ethnic origin as the basis for subjecting persons to investigatory activities or for determining whether an individual is engaged in criminal activity.”41 Yet local law enforcement agencies all over the United States use race in policing.42 Unfortunately, it would be nearly impossible for a class of victims of such profiling to prevail on a claim of unconstitutional racial discrimination under the current standard. The victims would have to prove that the policymaker had the specific purpose of targeting their race with the discriminatory action. Problematic is the fact that the policymaker almost certainly also had the legitimate goal of promoting public safety. If public safety is the primary goal of a policy or program and racial discrimination is an unfortunate byproduct, this arrangement is perfectly acceptable under current Supreme Court jurisprudence.

In New York City, for example, when it comes to the highly controversial practice of Stop-and-Frisk, studies have shown that “police are stopping hundreds of thousands of law abiding New Yorkers every year.”43 Indeed, a precinct commander in the New York Police Department, described as “on the path to making chief one day,” was recorded instructing a junior officer to use race as a factor when determining whether there was reasonable suspicion to subject individuals to the Stop-and-Frisk protocol.44 The vast majority of those stopped are black and Latino.45 In 2003, police stopped New Yorkers 160,851 times.46 Eighty-seven percent of those stopped were innocent.47 In the aggregate, of the roughly 140,000 innocent individuals who were stopped, approximately 85% were black or Hispanic while only 12% were white.48 Over the years, the racial disparity in these numbers has not improved. In 2013, of the 191,851 stops that the police made, 88% were totally innocent, and still approximately 85% of all the stopped persons were black or Hispanic whereas only 11% were white.49 In 2015, while the overall number of stops dropped dramatically to 22,565, the racial breakdown has remained roughly the same at 81% of all the stops being of black or Hispanic individuals and 11% being of white individuals.50 The innocent victims, who are “byproducts” of the policy, have virtually no recourse under current law. Racial profiling and, particularly, the practice of Stop-and-Frisk produce disparities in stops, arrests, and subsequent jail time that fail to live up to the Court’s pronouncement that distinctions based on race “are by their very nature odious to a free people whose institutions are founded upon the doctrine of equality.”51

New York is far from the only culpable state. The U.S. Department of Justice (DOJ) recently published a detailed analysis of law enforcement practice in Ferguson County, Missouri, following the controversial killing of an unarmed black teenager by an armed white police officer.52 The DOJ found that the Ferguson Police Department (FPD) engaged in a

pattern of unconstitutional policing, [which] has also shaped its municipal court, leading to procedures that raise due process concerns and inflict unnecessary harm on members of the Ferguson community. . . . Ferguson’s police and municipal court practices have sown deep mistrust between parts of the community and the police department, undermining law enforcement legitimacy among African Americans in particular. 53

. . . Ferguson’s approach to law enforcement both reflects and reinforces racial bias, including stereotyping . . . overwhelmingly impact[ing] African Americans. Data . . . shows that African Americans account for 85% of vehicle stops, 90% of citations, and 93% of arrests made by FPD officers, despite comprising only 67% of Ferguson’s population. African Americans are more than twice as likely as white drivers to be searched during vehicle stops even after controlling for non-race based variables . . . , but are found in possession of contraband 26% less often than white drivers, suggesting officers are impermissibly considering race as a factor when determining whether to search. African Americans are more likely to be cited and arrested following a stop . . . and are more likely to receive multiple citations during a single incident.54

The report suggests that the FPD brings certain offenses almost exclusively against African Americans.55 The report also showed that African Americans feel the brunt of discriminatory policing particularly when individual discretion is involved. For instance, the disparate impact of FPD’s enforcement practices on African Americans is 48% greater when speeding citations are issued on the basis not of radar and laser, but of the officer’s own visual assessment, suggesting that the officer’s personal racial bias against black people affects the size of the fine, to African Americans’ detriment.56 The perpetrating officers did not need to harbor a conscious racial bias for the effect to be meaningful. Approximately “90% of documented force used by FPD officers was used against African Americans.”57 In cases of excessive force, every incident where a police dog bit a person involved a victim who was black.58 Ninety-six percent of arrestees for merely an outstanding municipal warrant were black.59 In Chicago, the DOJ found that

overall, complaints filed by white individuals were two-and-a-half times more likely to be sustained than complaints filed by black individuals . . . . [F]or each allegation contained in a complaint, a white complainant is three-and-a-half times more likely to have the allegation sustained—and the officer held accountable for his or her misconduct—than a black complainant . . . .60

A 2016 DOJ report on the Baltimore Police Department (BPD) found that “BPD engages in a pattern and practice of discriminatory policing against African Americans. Statistical evidence shows that the Department intrudes disproportionately upon the lives of African Americans at every stage of its enforcement activities.”61

Courts, too, proved to be complicit in this racialized criminal justice system: Blacks in Ferguson are 68% less likely than others to have their cases dismissed by judges.62 The fact remains: though African Americans nationally make up only 13% of the population,63 are less likely than whites to be found with drugs during a stop,64 and are equally as likely to commit crimes,65 they bear the brunt of the criminal justice system’s punishments, from Stop-and-Frisk, to arrest, to the number and seriousness of charges brought, to conviction, to sentencing. The disparate treatment of Blacks by police departments around the country foils Lincoln’s idea that all men are created equal.

2. The Supreme Court May Be Exacerbating Racial Bias

Though the Supreme Court has announced broad notions of equality and pledges to uphold the pronouncements of the United States Constitution, many of its decisions seem to do the exact opposite. The Supreme Court holds those claiming racial discrimination by a government entity to a very high standard.66 The current discriminatory purpose standard was first articulated in the seminal case of Washington v. Davis.67 In that case, two African-American men who had applied to but were rejected for positions in the District of Columbia Police Department filed a class action lawsuit against the department alleging that it used racially discriminatory hiring procedures that systematically excluded black candidates.68 These hiring procedures included a written test that African Americans disproportionately failed.69 The plaintiffs convinced the United States Court of Appeals for the District of Columbia that the test that excluded African Americans, Test 21, did not satisfy the crucial element of having a direct relationship to performance on the policemen’s job.70 This relationship requirement, created by Title VII, served the partial purpose of minimizing the effect of arbitrary exams meant to screen out unwanted yet qualified candidates.71

The Supreme Court rejected the more rigorous Title VII standard and ultimately held that, though disproportionate racial impact is not irrelevant and though an invidious discriminatory purpose may sometimes be inferred from the totality of the relevant facts, including the fact that a law bears more heavily on one race than another, a law that is neutral on its face and serves ends that the government otherwise has the power to pursue is not invalid under the Equal Protection Clause simply because it had a greater effect on one race than another.72 The Court placed the primary burden on the plaintiffs by requiring them to prove that the defendant acted with an invidious racially discriminatory purpose, in addition to showing a disparate impact, to pass the first hurdle73 in establishing a claim of unconstitutional racial discrimination.74 Washington v. Davis set the stage for what would become an almost impossible task—showing that practices that disproportionately hurt a protected class of people were created for the purpose of hurting that class. By setting the burden of proof so high, the Court did the opposite of what the founding fathers meant for the Constitution to achieve and made it harder for plaintiffs to seek redress for claims of unconstitutional discrimination.

3. The Aftermath of Washington v. Davis

While Washington v. Davis failed to set an effective standard that would protect minorities from discrimination, it clarified a previously murky understanding of the constitutional standard for proving racial discrimination. Why did the Court deem disparate impact insufficient to infer invidious discriminatory purpose? The Court clearly distinguishes between statutory violations and constitutional violations: the standard under Title VII,75 a statutory provision, for showing unlawful discrimination on the basis of race is not the constitutional standard.76 Whereas Title VII requires a showing of only racially disparate impact, the Equal Protection Clause of the Fourteenth Amendment requires a showing of racially discriminatory purpose. A statutory violation is not equivalent to a constitutional violation: “[a] purpose to discriminate must be present.”77

The Court declined to find invidious discriminatory purpose in Davis for three reasons. First, it did not see how a law that was racially neutral on its face, even if it had a disparate impact, could be unlawfully discriminatory.78 It communicated that a law that by its words treats every person equally regardless of race must at least be given that interpretation. Yet it is well known that a facially neutral law can still be used to accomplish invidious discriminatory ends.79 The Court focused on the race-neutral qualifications that the examination tested, but it should have focused on the test itself, which bore no demonstrated relationship to job performance.80

Second, the Court stated that the affirmative recruitment efforts of the police department warranted the conclusion that the test did not have the purpose of unlawful discrimination against Blacks.81

[T]he test is neutral on its face and rationally may be said to serve a purpose the Government is constitutionally empowered to pursue. . . . [T]he affirmative efforts of the Metropolitan Police Department to recruit black officers, the changing racial composition of the recruit classes and of the force in general, and the relationship of the test to the training program negated any inference that the Department discriminated on the basis of race.82

The Court was persuaded that because the police department made efforts in other areas to diversify its police force, the test in question therefore must not have been racially motivated. Yet this justification misses the issue. The plaintiffs did not challenge general police recruitment practices—“[t]he validity of Test 21 was the sole issue before the [district] court.”83 The Court purported to answer the question of whether Test 21 was unlawfully racially motivated by pointing to other practices of the department that were purportedly racially inclusive. This is not a valid analysis of Test 21 itself. As Justice Stevens said in his concurrence, “neither [the good-faith efforts to recruit black police officers] nor the subjective good faith of the District administration, would save Test 21 if it were otherwise invalid.”84 Affirmative efforts to recruit do not constitute proof that there is no discrimination in the actual hiring of recruits, nor that the test used is not unconstitutionally discriminatory.

Finally, the Court stated that not requiring plaintiffs to show invidious discriminatory purpose would be impracticable and essentially a usurpation of the legislative role. It would potentially “invalidate a whole range of tax, welfare, public service, regulatory, and licensing statutes that may be more burdensome to the poor and to the average [B]lack than to the more affluent white. . . . [E]xtension of the rule . . . should await legislative prescription.”85 However, Marbury v. Madison long ago clarified that it is “emphatically the province and duty of the judicial department to say what the law is,” to weigh proposed law against the Constitution, and, having established a standard to review such claims, to determine which laws, on aggregate, burden certain groups unconstitutionally more than others.86 The Court had in fact already done what it claimed it was uncomfortable doing. When the Court set the intent standard, Congress had not previously established a mental state requirement for a constitutional claim of racial discrimination.87

When Congress eliminated the mental state requirement altogether under Title VII, this did not invalidate a litany of laws governing employment practices. In fact, under the impact regime, plaintiffs have still found it challenging to win claims in court.88

The constitutional standard as articulated by Justice White required that the plaintiffs show “invidious discrimination.”89 When the Court stated that “the constitutional standard for adjudicating claims of invidious racial discrimination is [not] identical to the standards applicable under Title VII,” it distinguished invidious racial discrimination, perhaps the defining characteristic of the constitutional violation, from other racial discrimination.90 Indeed, the invidious nature of the discrimination must be shown in order to reach constitutional status. Webster’s New World College Dictionary defines invidious as “such as to excite ill will, odium, or envy; giving offense” and “giving offense by discriminating unfairly.”91 Laws with disproportionate effects along racial lines that are otherwise neutral and serve legitimate ends are not unconstitutional.92 However, if those laws are shown to be invidious, then they may be unconstitutional.93 Laws that benefit one race more than another need not be invalidated; to reach constitutional muster, a law must be shown to burden another race. The distinction lies in the burdening. For example, a law that benefits Native Americans does not necessarily affect Hispanics, Blacks, or whites at all.94 Laws that benefit one race at the expense of another race, however, establish invidiousness and may reach constitutional muster. Laws that burden one race without affecting another race likewise manifest the invidious element that the Court so adamantly demands, and they should therefore reach constitutional muster.

Justice White further parsed invidious racial discrimination from other discrimination in Davis when he distinguished between racial disproportion and invidious racial discrimination, essentially using racial discrimination and racial disproportion as synonyms. If a law discriminates against one group more than another, i.e., if a law disproportionately affects one group more than another, then that law is not per se unconstitutional without more. While the Due Process Clause of the Fourteenth Amendment contains an equal protection component prohibiting the United States from invidiously discriminating among individuals and groups, a law or other official act is not unconstitutional solely because it has a racially disproportionate effect without invidiousness.95

Justice White explained this concept using Strauder v. West Virginia as an example.96 He stated that the fact that a particular jury does not statistically reflect the racial composition of the community does not in itself make out an unconstitutional invidious discrimination.97 In explaining the difference between racial discrimination and invidious racial discrimination, he singled out the invidious element—the injury: “[a] purpose to discriminate must be present which may be proven by systemic exclusion of eligible jurymen of the proscribed race or by unequal application of the law to such an extent as to show intentional discrimination.”98 He continued that “the State [may] not deliberately and systematically deny to members of [the defendant’s] race the right to participate as jurors.”99 In other words, it is not racially discriminatory to allow only two black jurymen in a county that is 80% Black onto a jury, but complete exclusion—allowing zero black jurymen onto the jury—is racially discriminatory. Distilling the standard from this pronouncement, Justice White essentially found two things problematic: (1) negative treatment such as exclusion or denial and (2) an intent to do this, as illustrated by deliberate and systematic exclusion.100

Justice White used Wright v. Rockefeller to further illustrate his point. The challengers in that case did not prevail on a claim of racial gerrymandering because they “had not shown that the statute ‘was the product of a state contrivance to segregate on the basis of race.’”101 One interpretation of Justice White’s words is that racial discrimination, or a law that has a disproportionate impact on one race, is not itself injurious. It is injurious, however, for that law to exclude, deny, and segregate. Moreover, the injury must be deliberate or systematic. It is this injury that limits the scope and impact of this new standard on current legislation.

Thus, the Court’s concern that an entire “range of tax, welfare, public service, regulatory, and licensing statutes”102 will be invalidated upon adopting a lower standard will likely not manifest. If those statutes primarily benefit an insular group without explicitly burdening another group, then they are constitutional. However, if a statute continually burdens one insular minority, then both under the legal framework and as a matter of public policy that law has an invidious element and should reach constitutional muster.

a. The Washington v. Davis Standard Applied in Subsequent Civil Rights Cases

Washington v. Davis substantially curtailed the ability for victims to sue and win against government and private entities for employing practices that have racially skewed and exclusory effects. Plaintiffs have found it extremely difficult to successfully bring a claim of unconstitutional racial discrimination under the Equal Protection Clause of the Fourteenth Amendment.103 This is particularly true in the context of policing, prosecution, and sentencing practices.104

       McCleskey v. Kemp is a well-known case that demonstrates this difficulty.105 In that case, McCleskey was convicted of murder and two counts of armed robbery and sentenced to death.106 He appealed for habeas corpus relief on the ground that the Georgia death sentencing process was unconstitutional.107 He relied on the Baldus Study, which

show[ed] a disparity in the imposition of the death sentence in Georgia based on the race of the murder victim and, to a lesser extent, the race of the defendant. . . . The raw numbers . . . indicate[d] that defendants charged with killing white persons received the death penalty in 11% of the cases, but defendants charged with killing [black persons] received the death penalty in only 1% of the cases.108

The study also “found that the death penalty was assessed in 22% of the cases involving black defendants and white victims” and only “3% of the cases involving white defendants and black victims.”109 Additionally, “prosecutors sought the death penalty in 70% of the cases involving black defendants and white victims” and only “19% of the cases involving white defendants and black victims.”110 Finally, one of the models concluded that even after controlling for thirty-nine potentially explanatory non-racial variables, defendants charged with killing white victims were still 4.3 times more likely to receive the death sentence than defendants charged with killing black victims.111

The Court applied the Davis standard for proving an equal protection violation—a showing of purposeful discrimination and, as a corollary, that that discrimination had a discriminatory effect on the plaintiff.112 Despite the powerful evidence presented in the Baldus Study, the Court found that the appellant had not offered any evidence specific to his own case that would support an inference that racial considerations played a part in the case’s outcome. The Court held that in order for McCleskey to prevail, he would have to make the impossible showing that the Georgia legislature maintained the death penalty statute because of an anticipated racially discriminatory effect—in other words, that the statute had a discriminatory purpose.113 Washington v. Davis rendered ineffective any redress in McCleskey v. Kemp.

In defining discriminatory purpose, the Court held that the term “implies more than intent as volition or intent as awareness of consequences. It implies that the decision-maker . . . selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon an identifiable group.”114 Requiring the plaintiffs to show that the State of Georgia imposed the death penalty with a purpose to invidiously racially discriminate may have been appropriate when overt acts of discrimination were commonplace.115 However, they are uncommon today. Thus, despite the overwhelming proof in McCleskey that the processes in place that led to death penalties had a vastly disparate impact on black defendants as opposed to white defendants, the Court did not find that the legislature’s act or failure to act by not correcting these racial disparities constituted purposeful discrimination against Blacks. The Court stated that legislatures have wide discretion when creating criminal laws and penalties, and as long as there are “legitimate reasons for the Georgia Legislature to adopt and maintain capital punishment, [the Court] will not infer a discriminatory purpose on the part of the State of Georgia.”116 Though McCleskey had demonstrated that the State’s process of imposing the death penalty on Blacks was racially imbalanced, after controlling for every other explanation, the Court still declined to find an equal protection violation because of the intent requirement established in Davis. The consequences have been dire: roughly 42% of all death-row inmates across the nation are black, nearly three times the proportion in the general population.117

       United States v. Armstrong is another case in which the Court applied the Davis standard to deny relief to black defendants who alleged racially skewed prosecutorial practices.118 The defendants had been charged with selling crack cocaine and using a firearm in connection with drug trafficking.119 They asserted that the prosecutor unconstitutionally selected them for prosecution because they were black, and they filed a motion for discovery.120 Their allegation was based on a study conducted by the Federal Office of the Public Defender that indicated that every crack case for which the prosecutor brought federal charges in 1991 involved a black defendant.121 None was brought against a white person.122 Though the Supreme Court ultimately decided the case on other grounds, it used the Davis standard to require the defendants to show both disparate impact and discriminatory intent.123 The Court ruled that the defendants’ study did not demonstrate an equal protection violation because it did not include “some evidence tending to show the existence of the essential elements of” a selective-prosecution claim,124 which includes a showing that the government’s discriminatory selection of him for prosecution was invidious and based on an impermissible consideration such as race.125 Despite a showing of disparate treatment in all of the 1991 cases, Armstrong failed under the Davis standard. In Armstrong’s wake, a study by the Brennan Center has shown that “the [United States Sentencing Commission] detected notable differences in prosecutorial decisions to seek sentence enhancements for certain federal offenses involving a firearm depending on the race of the defendant.”126

In Village of Arlington Heights v. Metropolitan Housing Development Corporation, a nonprofit organization that built housing developments for low- and moderate-income families brought a lawsuit against the Village of Arlington Heights.127 The organization had requested a rezoning approval for a tract of land located in a primarily white area, on which it was to build a low-income housing development.128 After a number of public meetings in which current residents of the area expressed vocal opposition to the housing plan while referring to “the social issue”—the desirability or undesirability of introducing low- and moderate-income racially integrated housing at the requested location in Arlington Heights—the plan commission denied the rezoning application.129 The court of appeals reversed the district court’s rejection of the plaintiffs’ claim and found that, examining the case in light of its historical context and ultimate effect, the zoning refusal would have a disproportionate impact on Blacks.130

The Supreme Court reversed the court of appeals on the grounds that the nonprofit organization and other plaintiffs “simply failed to carry their burden of proving that discriminatory purpose was a motivating factor in the Village’s decision.”131 The Court based its decision on the sequence of events leading up to the decision,132 the history of zoning in the area, the fact that the usual procedures were used, and the fact that the subject of the testimony at the hearings was almost exclusively on the zoning aspects.133 The Court adopted the district court’s assessment that the evidence, which included that some opponents to the project were motivated by opposition to minority groups, did not warrant the conclusion that such improper opposition motivated the defendants. Acknowledging that the decision “does arguably bear more heavily on racial minorities,”134 the Court stated that “disproportionate impact is not irrelevant, but it is not the sole touchstone of an invidious racial discrimination.”135 Though the evidence arguably met the Davis standard, the Court did not find it sufficient and again flouted the lofty pronouncements of the Declaration of Independence that all men are created equal. Contemporary results of this ruling show that low-income housing projects in the United States’ biggest metropolitan areas that use federal tax credits—the nation’s biggest source of funding for affordable housing—are disproportionately built in majority nonwhite communities.136

Finally, in City of Mobile Alabama v. Bolden, the Supreme Court again used the Davis standard to deny redress to petitioners claiming invidious racial discrimination.137 The plaintiffs brought a class action lawsuit against the Mobile City Commission and its three commissioners alleging that the at-large system of municipal elections violated their constitutional rights because it diluted the voting strength of black residents in Mobile.138 The district court and court of appeals found for the black plaintiffs based on several pieces of evidence: though Blacks made up 35.4% of the population,139 no black person had ever been elected to the commission, leading these lower courts to conclude that the processes leading to nomination and election were not open equally to black residents; the persons elected to the commission discriminated against black persons in municipal employment and in dispensing public services; Alabama has a substantial history of racial discrimination; the city officials had not been as responsive to the interests of black persons as to those of whites;140 and the mechanics of the at-large electoral system were themselves proof that the votes of black persons were being invidiously canceled out.141

The Supreme Court reversed, making two findings. First, it held that the plaintiffs’ right to vote had not been denied or abridged because, unlike the facts of Smith v. Allwright and Terry v. Adams where the plaintiffs were precluded from voting, the plaintiffs here could exercise their right to vote.142 The Constitution does not guarantee the right to win a seat or to proportional representation—only a right to vote.143 Although gerrymandering is one of the most well-known ways in which lawmakers redraw political districts on the basis of race,144 absent an overt invidious racial purpose under Davis, a state may constitutionally redraw its political boundaries in any manner it chooses.145 In this way, Davis continues to deny plaintiffs equal protection within our legal system.

Second, the Supreme Court reversed the court of appeals and district court decisions that said that the at-large voting system was unconstitutional because of its diluting effects. Even though it recognized that legislative apportionments could violate the Fourteenth Amendment if their purpose were invidiously to minimize or cancel out the voting potential of racial or ethnic minorities, the Court stated that a plaintiff must prove that the disputed plan was “conceived or operated [as a] purposeful devic[e] to further racial . . . discrimination.”146 It cited Zimmer v. McKeithen,147 a pre-Davis case, but then rejected the criteria that the case set out to prove discriminatory motive, while admitting that there were “indicia” of some evidence of discriminatory purpose.148 Throwing out each of the elements that the lower courts found in plaintiffs’ favor, the Court articulated that the district court’s findings of fact showed that black persons register and vote in Mobile “without hindrance”; that the fact that black candidates have been defeated is not a constitutional deprivation; that evidence of current discrimination in other areas by officials is tenuously relevant to the constitutional invalidity of the electoral system; that historical discrimination cannot forever condemn governmental action that is not itself unlawful in the present; and that any electoral system will tend to naturally disadvantage a voting minority.149 On these bases, the Court held that plaintiffs had not proved that the at-large electoral scheme represented purposeful discrimination against black voters. The Court took pains to invalidate each of the findings of racial discrimination that the lower courts found, again using Davis to deny plaintiffs their just desserts. The result? Districts in which the majority of residents are people of color are “detrimental to the minorities. . . . ‘If you have too high a percent [of] African Americans in a House district, it does dilute the overall representation of African American interests,’” and it limits black influence in surrounding districts.150

The cases discussed in this section examine the application of the discriminatory intent standard established in Davis. They represent the current state of the law. It is nearly impossible to prove the underlying motive of an individual or municipality, and plaintiffs now rarely win under the constitutional requirement of showing discriminatory impact and invidious discriminatory purpose. As one observer aptly put, “the Court will not invalidate a facially neutral statute unless the law is shockingly oppressive, anachronistic, and inexplicable on other than racial grounds.”151 Quite to the contrary of what the founding fathers intended, “[s]ubtler mechanisms of discrimination, under this standard, must inevitably escape constitutional scrutiny. Absent a legislative solution, the result may well be continued second-class citizenship without a remedy.”152

B. The Mechanics of Implicit Bias

Implicit bias is a growing field of study that has recently gained recognition for its application in various fields, including law. The basic premise of implicit bias is that though people outwardly profess to hold to anti-discriminatory notions, they simultaneously and often unconsciously treat people with whom they least identify, or to whom they attribute negative characteristics, less favorably than people who carry preferable attributes.159

Many well-known researchers have extensively studied the phenomenon of implicit bias. One of those researchers, Jerry Kang, has connected implicit bias with human interactions. His work makes salient the manner in which implicit bias can affect the conduct of police officers, prosecutors, and judges. Studies of implicit bias draw heavily from the field of social cognition.160 In order to understand how implicit bias works, it is necessary to understand how the human brain and human intuition work.

Humans are constantly bombarded with information—so much that it seems practically impossible to process all of it and fashion actions or reactions based on that information within mere seconds.161 But we do. Humans receive environmental stimuli, process them, and then encode them into short- and long-term memories.162 Working with such a vast amount of stimuli, humans by necessity use schemas, or “cognitive structure[s] that represent[] knowledge about a concept or type of stimulus, including its attributes and the relationships among those attributes,”163 to constantly simplify and process the information.164 Schema-based thinking happens “nearly instantaneously,” and operates on both inanimate objects and on human beings.165 We classify new acquaintances into a number of social categories, including gender, age, race, disability, and role.166 This is the first way in which our automatic information processing affects our interactions with people and the world.

More specifically, we function according to racial schemas. Kang identifies three components of racial schemas that influence our interactions with others. We as perceivers categorize individuals into racial categories according to society’s racial mapping.167 While every society may have a different set of mapping rules, once a person is mapped, implicit and explicit racial meanings are triggered, which then influence our interactions.168

Racial meanings include cognitive and affective components.169 “The cognitive component includes thoughts or beliefs about the category, such as generalizations about their intelligence or criminality.”170 The affective component includes emotions, feelings, and evaluations that range from positive to negative, or good to bad.171 Stereotypes are cognitive beliefs about groups, whereas prejudice is positive or negative feelings about those groups.172 Once schemas are activated, racial meanings can generate both thoughts and feelings that singularly and jointly influence a person’s reactions.173

Determining which schemas influence human interactions depends on variables such as what gets activated first, what catches attention, and what can be easily retrieved from memory, for instance by priming.174 However, “the scientific consensus is that racial schemas are not of minor significance[,] . . . are ‘chronically accessible,’ and can be triggered by . . . mere appearance.”175 Visual and physical stimuli play particularly important roles.176 Visual information, in fact, is so powerful that even subliminal, or subconscious, stimuli can still activate the racial schemas.177 This is because racial schemas “operate automatically—without conscious intention and outside of our awareness.”178 Always accessible, once racial schemas are activated their racial meanings influence interactions.179 These racial meanings can influence us to treat people differently based on race.180

1. Explicit Bias Measures Are Dissociated from Measures of Implicit Bias

Social scientists have long tried to measure people’s biases using explicit measurements, but these methods have not been a reliable predictor of people’s behavior.181 One way to measure explicit bias is through direct questioning. Though racial stereotyping and prejudice has substantially declined over the last fifty years, studies show that self-reported evaluations of bias are not trustworthy.182 Self-reporting individuals may feel uncomfortable expressing their anxiety, ambivalence, or dislike for a specific racial group.183 This is particularly true in a society where social rhetoric favors declarations of racial equality. Indeed, “overwhelming evidence [shows] that implicit bias measures are dissociated from explicit bias measures.”184 Dissociation is “a discrepancy between our explicit and implicit meanings.”185 For example, I may report on a survey that I do not have any biases against my Hispanic friends. After all, my husband is Hispanic, I might say. Yet at the same time, my Implicit Association Test (IAT) score, a method of measuring subconscious associations discussed more fully below,186 may reveal that I associate Hispanic with illegality faster than I associate Hispanic with a master’s degree. Despite my outward expression of race neutrality, my IAT score tells a different story—that I may hold negative attitudes toward the very group that I claim to treat equally. It is noteworthy that self-reported explicit biases and people’s IAT scores revealed no correlation, but unconscious implicit biases and IAT scores, as measured through our latencies,187 did.188 Implicit measures provide more accurate metrics for our unconscious behavior than explicit measures.

Using large datasets from the web, researchers Brian Nosek, Mahzarin Banaji, and Anthony Greenwald showed that whites exhibited only some explicit preference for themselves over Blacks, but exhibited a much greater implicit preference for themselves over Blacks.189 The disparity between the explicit and implicit measure is the dissociation. The studies on implicit bias reveal that implicit mental processes may draw on racial meanings that, upon conscious consideration, we would expressly disavow.190 Evidence shows that implicit bias against a social category predicts disparate behavior toward individuals mapped to that category. This occurs “notwithstanding contrary explicit commitments in favor of racial equality.”191

2. Understanding Implicit Bias Through Racial Schemas: The Crash Test and the Speed Test

Recognizing the unreliability of explicit measures of bias, scientists have instead turned to implicit measures of bias. Social cognitionists John Bargh and Mark Chen developed a computer crash experiment to demonstrate that the mere image of a black face, even if subliminal, can activate a black racial schema.192 In the test, research subjects are asked to count whether an even or odd number of circles appears on a computer screen.193 The computer was designed to crash after the 130th iteration, and the subjects instructed to start over.194 Hidden cameras caught the participants’ frustration and hostility. Unbeknownst to the participants, half of them were subliminally flashed a young black male face prior to each iteration and the other half were shown a young white male face. Confirming the researchers’ suspicions, those who were shown the black man’s face reacted with greater hostility to the computer crash.195 This experiment reveals that we do not have to “see” or consciously register the black male face for it to influence our behavior.

Subsequent research studies have shown a connection between subliminal priming, similar to that in the computer crash experiment, and later tasks such as evaluations, interpretations, and speed tasks. The IAT, initially developed by Samuel Gaertner and John McLaughlin, uses sequential priming procedures to measure the automacity of the racial meanings in racial schemas.196 Put simply, the IAT measures how quickly we associate two concepts by measuring millisecond differences in reaction times: latencies.197 The test first primes participants with a particular stimulus, such as a word or face, which activates particular racial schemas.198 These schemas in turn activate racial meanings which may alter performance on some subsequent task.199 The study measures speed: if the primer and task are schema-consistent, the response should be faster; if they are inconsistent, the response should be slower, thereby measuring how closely two concepts are associated.200 The test first chooses two categories to compare201—for instance, black and white—and then selects two sets of stimuli using words or images that correspond to racial meanings or stereotypes associated with those categories.202 “For example, words such as ‘violent’ and ‘lazy’ are chosen for Blacks, and ‘smart’ and ‘kind’ for [w]hites.”203 After being primed with a black or white face, participants must hit a key on the left or right side of the keyboard as fast as possible.204 “In half the runs, the black face and black-associated word are assigned to the same side of the keyboard,” making them schema-consistent.205 “In the other half, they are assigned opposite sides” of the keyboard, making them schema-inconsistent arrangements.206 The same was done for the white face and white-associated stimuli.207

The prediction was that “[t]asks in the schema-consistent arrangement [w]ould be easier,” and thus faster.208 “[T]he time differential between the two arrangements [] provides a measure of implicit bias.”209

The studies revealed that participants were faster at recognizing a positive word, such as “smart,” if they had just been primed with the word “white” instead of “Black.”210 Race-neutral reconciliatory explanations, “such as overall speed of participants’ reactions, right- or left-handedness, and familiarity with test stimuli [were] . . . shown not to undermine the IAT’s validity.”211

The “first wave” of implicit bias research showed that socially dominant groups have implicit biases against subordinate groups, i.e., white over non-white.212 Almost one hundred studies documented people’s tendency to automatically associate positive characteristics with their in-groups more easily than with out-groups and to associate negative characteristics with out-groups more easily than with their in-groups.213 This is the case across races.214

3. Implicit Bias Is Connected with Brain Processes

The millisecond time differentials measured in the IAT are connected with actual brain processes. Jerry Kang has studied functional magnetic resonance imaging (fMRI).215 fMRIs can measure the neural activity that is activated in the amygdala, a subcortical structure in the human brain that is associated with “emotional learning as measured by fear conditioning,” “memory[] and evaluation,” and the “expression of learned emotional responses that have been acquired without direct aversive experience.”216 The amygdalae of white participants activate to a much greater extent when they are subliminally primed with images of black people than with white faces.217 Amygdala activity is “significantly correlated with participants’ IAT scores.”218 This demonstrates that “observable behavior maps . . . to some neural activity. . . . [T]he IAT is measuring something real and significantly connected to emotion-laden racial mechanics.”219 Importantly, there is no correlation between explicit measures of bias (openly admitting one’s biases) and IAT scores or amygdala activity.220

4. Implicit Bias Affects Our Actions

Not only do we have implicit biases, but these biases affect our behavior in relevant ways. One famous study shows that résumés bearing “African-American-sounding” names, such as Lakisha, received far fewer callbacks than identical résumés bearing “[w]hite-sounding names,” such as Emily.221 This influence is possible because people act through racial schemas, translate the schemas into racial meanings, and then act. The experiment sent résumés to over 1300 help-wanted ads.222 The “white-named” résumés received 50% more callbacks.223

A study called Shooter Bias demonstrates how implicit racial bias influences our perceptions of, and reactions to, criminality.224 Researchers Joshua Correll, Bernadette Park, Charles Judd, and Bernd Witenbrink had participants play a video game that placed photos of white and black individuals holding either a gun or an innocuous object such as a wallet, soda can, or cell phone into diverse backgrounds.225 Participants were instructed to shoot the target if the person had a gun and not to shoot if the person in the photograph was not carrying a gun.226 Extreme time pressure was designed to force errors as the participants quickly decided whether to shoot or not to shoot.227 Consistent with racial stereotypes and prejudice, the participants were more likely to mistake black targets as armed (and shoot) when in fact they were not armed.228 Conversely, they were more likely to mistake white targets as not armed (and not shoot) when in fact they were armed.229

City administrations must regularly make policy about which persons to arrest, which neighborhoods to target, which crimes to prosecute, which defendants to set bail for, what plea bargains to offer, and which defendants to find guilty. The discussion above demonstrates the ways in which biases operating on an unconscious level have the potential to influence each of these policy decisions and the actors carrying them out. After all, whether a police officer associates with and “sees” a wallet or a gun could mean life or death, as was the case with Amadou Diallo230 and others.

In light of racial schemas and the implications of the IAT, the standard for reviewing claims of unconstitutional race discrimination against individual police officers, municipalities, police commissioners, police departments, other policymakers, and private entities, should change. To decline to judicially respond to the very real phenomenon of implicit bias is to ignore the best scientific evidence we have about how our behavior is influenced by complex superpositions of mental processes that range from the controlled, calculated, and reasoned to the automatic, unintended, and unnoticed.

5. Limitations to Implicit Bias Research

Implicit bias research has received some criticism, most prominently from Gregory Mitchell and Philip Tetlock.231 Though implicit bias research has been widely accepted among researchers to establish the IAT’s ability to validly predict discriminatory behavior,232 Mitchell challenges the ability of IAT tests to predict behavior in real-world situations—their predictive validity.233 Mitchell’s conclusion that all IAT research is provisional at best rather than predictive234 ignores the large number of studies indicating otherwise. One study performed by Jens Agerström and Dan-Olof Rooth demonstrates that the IAT is capable of predicting real-world discrimination against obese job applicants.235

That study found that hiring managers who possessed a stronger automatic association between obesity and low performance were less likely to invite obese applicants to an interview—in some instances, eight percentage points less likely.236 The automatic obesity stereotypes predicted discriminatory hiring far beyond explicitly touted hiring preferences and obesity stereotypes.237 While researchers have acknowledged the limitations of implicit bias research particularly through IAT studies, they nevertheless tout its predictive power and construct validity.

Agerström acknowledges the limitations of implicit bias research—that some of the previous IAT data is weak and inconclusive, that some data have excluded outliers that would otherwise affect outcomes, that mismatches have been discovered where a person holding an anti-Black bias acts positively toward a black subject,238 and that most of the IAT research has been performed in controlled laboratory environments and not in real-world settings.239 Nevertheless, according to Agerström, the study’s results “strongly suggest” that the IAT can be a valuable tool for assessing obesity-based bias occurring in actual hiring situations.240 Given that the IAT was designed to assess discriminatory biases, the finding that the IAT successfully predicted hiring discrimination, even when controlling for explicit stereotypes, provides support in favor of the test’s construct validity: “[it] suggests that the IAT captures automatic bias that to at least some extent has a person-centered evaluative component and does not merely reflect environmental associations.”241 In other words, not only was Agerström able to reproduce IAT results in real-world as opposed to laboratory settings, but even despite the lack of data establishing full-stop causality in race discrimination, his results suggest that IAT effectively captures and conveys that real-world actions are to some extent based on automatic implicit associations. In real-life situations where a substantial portion of the population can be affected by these biases, this conclusion cannot be ignored.

Rooth also performed a study examining the potential effect of implicit biases on actual behavior that led him to conclusions similar to those of Agerström.242 His study examined real hiring managers’ automatic associations and discriminatory behavior in offering callback interviews.243 The study showed that the probability of inviting individuals with résumés bearing Arab-Muslim-sounding names, as opposed to Swedish-sounding names, to an interview decreased by five percentage points when the recruiter had a stronger negative implicit association toward Arab-Muslim men by one standard deviation.244 Moreover, the negative correlation between IAT score and probability of inviting a person with an Arab-Muslim-sounding name to an interview was consistent and strong. No such stable and statistically significant correlations were found for explicit measures and interview invitation probabilities.245 The results indicate that some recruiters implicitly, but do not explicitly, discriminate.

Notwithstanding the previous discussion, Kang and others have directly answered the objections of Mitchell and Tetlock. In response to the three main objections, Kang and Kristin Lane first reject Mitchell’s criticism that IAT research is “junk science” and that scientists fudge data.246 In making this objection, Mitchell problematically applies a double standard to IAT research, seeking perfect results from IAT studies—an essentially unachievable goal since statistics trade on probabilities—but not demanding the same scientific rigor from the studies that support his position.247

He challenges every validity, methodology, definition, and dataset that undergirds IAT research.248 He does not do the same for his self-coined concept of “stereotype threat,” for instance, which is the idea that the IAT effect might be explained by the “fear of being labeled a bigot.”249 Mitchell simultaneously relies on a single study to support this idea, takes for granted that it even has predictive validity, and does not examine any limitations of the concept or the study.250 In essence, Kang and Lane expose that Mitchell “manufactur[es] scientific doubt,” just as the tobacco industry did with the Environmental Protection Agency when the latter linked tobacco and second-hand smoke to cancer and death.251 They acknowledge that on-the-merits skepticism and criticism are important to any scientific and evidence-based theory.252 However, Mitchell is engaging in “backlash scholarship” by using straw men to deter institutions from responding to new scientific discoveries, while never requiring perfect proof of the assumptions that underlie the status quo.253

Kang and Lane also respond to the “hardwired resignation” objection, which posits that implicit biases are hardwired into humans and there is nothing we can do about them.254 They note that (i) there is great variability between individuals’ degrees of implicit biases such that all humans are not homogenously biased and incapable of changing their predetermined actions and (ii) “the causal link between biases and behavior can be disrupted through procedural and structural reforms.”255

Finally, Mitchell provides a “rational justification objection,” arguing that implicit biases reflect reality, making it rational to act upon on them.256 This theory problematically assumes that “[w]e carry accurate probabilities in our heads” and that “we act on the basis of those probabilities rationally,”257 neither of which is true. The IAT provides substantial evidence that implicit biases are indeed dissociated from explicit biases.258 Most people’s racial experiences come from vicarious, rather than direct, experiences, and mainstream media is filled with inaccurate, interest-driven, and biased portrayals of racial others.259 Recent findings show that television exacerbates implicit biases against racial minorities.260 People see “‘illusory correlations[]’ [w]henever two salient events are noticed together”—for example, seeing a member of a racial minority near a crime scene.261 People also have preferential recording and recall of stereotype-consistent data.262 People have strong motivational desires to exaggerate the good traits of in-groups and the bad traits of out-groups.263

People suffer from ultimate attribution error, which causes them to view negative attributes of out-groups as permanent, stable, and dispositional while viewing the same negative attributes of in-groups as malleable, contingent, and the result of bad luck.264 Indeed, fifty years of scientific inquiry confirms that people are in fact vastly inaccurate calculators of probabilities, and further, a large amount of evidence shows that we do not actually process that information rationally.265 For instance, participants in a study who had to choose a partner for a trivia game reported that weight did not make a difference in the partner, but they in fact sacrificed between 10.53 and 12.31 IQ points to work with the thinner partner.266 While people may be naturally inclined to think of themselves as rational beings who make decisions based on objective evidence, they must nevertheless acknowledge implicit associations and bias. The legal system must also acknowledge them.

C. The Supreme Court Should Respond to Implicit Bias by Creating a New Standard

As the discussion on implicit bias and dissociation illustrates, in a society whose normative attitude touts racial equality, it is politically incorrect, publicly condemned, and extremely uncomfortable to admit that one is purposefully racially biased. Self-reports indicate that people hold race-neutral views and that they treat people of different races equally.267 Yet as measurements such as IAT show, no matter how close one may outwardly feel to a person of a particular race, unconscious racial biases remain active: “that we are not even aware of, much less intending, such race-contingent behavior does not magically erase the harm.”268 If the harms of implicit bias manifest in our daily interactions, nothing prevents them from manifesting in a legislator’s office, in police interactions, in the prosecutor’s office, or in the courtroom. Simply willing for these biases to go away does not objectively or permanently strip them of influence. In fact, one study shows that people who were instructed to avoid using race were less able to avoid the influence of race.269

The Court should heed these discoveries: faced with the evidence, it seems increasingly outdated to require discriminatory intent in an age when people openly deny and genuinely believe themselves to be rid of racial bias, yet unconsciously make racially biased decisions. This disconnect between outward professions and inner influences, rendering nearly impossible the ability to prove discriminatory purpose in court, is why Washington v. Davis should be overruled. The Supreme Court should create a standard that is more in line with scientific findings regarding our current manifestations of racial discrimination, thereby prompting a more genuine commitment to the egalitarian principles of the founding fathers.

1. Taking a Step Back: A Survey of Mental States in Criminal and Tort Law

Whereas proving unconstitutional racial discrimination requires a showing of intentional discrimination, lawmakers use a variety of other mental states to evaluate culpability in other contexts.270 Tort law and criminal law both use a spectrum of mental states in their culpability determinations.271 Criminal law embraces a comprehensive and complex hierarchy of mental states.272 Under the Model Penal Code (MPC), criminal law recognizes four types of mental state, or mens rea: intent or purpose, knowledge, recklessness, and negligence.273 The mental state requirement could refer to “the actor’s conduct, the attendant circumstances, and the result that he brings about (or seeks to bring about) by his conduct.”274 Under the MPC, a person acts purposely or intentionally when it is her conscious object to cause a certain result, or she is aware, believes, or hopes that a circumstance exists.275 A person acts knowingly when she is practically certain that her conduct will cause a certain result, or if she is aware of a high probability that a circumstance exists.276 A person acts recklessly when she “consciously disregards a substantial and unjustifiable risk that [harm] will result from [her] conduct.”277 She acts negligently when she is unaware, but should be aware, “of a substantial and unjustifiable risk that [harm] exists or will result from [her] conduct.278 Under negligence, otherwise identical to recklessness, the actor should have been aware of the risk of harm to another.279 Outside of the MPC, a different and “traditional . . . view sees recklessness as culpable indifference to risk.”280

Tort law also contemplates mental states.281 Tort law comprises three broad categories: intentional torts, negligence, and strict liability.282 While under the Second Restatement of Torts intent encompasses both purpose and knowledge, the Third Restatement separates them.283 Intent or purpose is defined as an actor purposefully causing harm by acting with the desire to bring about that harm.284 Knowledge is when an actor knowingly causes harm by engaging in conduct believing that that harm is substantially certain to result.285 Next in the hierarchy is recklessness. “Tort recklessness, like criminal recklessness, requires a greater deviation from the standard of care of the reasonable person than negligence requires.”286 The Second Restatement distinguishes recklessness as a ground for tort liability and recklessness as a ground for punitive damages. The former is defined as the reckless disregard for the safety of another—when the actor does an act or intentionally fails to do an act that is her duty to do, knowing of facts which would lead a reasonable person to realize that her conduct creates an unreasonable risk of physical harm to another.287 The latter is defined as “outrageous” conduct that shows reckless indifference to the rights of others.288 It views the latter as a narrower concept: “[p]unitive damages are not awarded for mere inadvertence.”289 Finally, negligence is defined in the Second Restatement of Torts as conduct that falls below the established standard “for the protection of others against unreasonable risk of harm.”290

Under constitutional equal protection doctrine, government actors are subject to strict scrutiny when they purposefully discriminate against a suspect class.291 If government actors knowingly harm a suspect class as a byproduct of a government action of significant interest or importance, they act constitutionally. However, the actor should be liable if she shows insufficient concern or respect for the interests of minorities. Such indifference is arguably “more serious than acting in the face of knowledge” that minorities will be harmed.292

The Supreme Court has acknowledged that indifference can be a form of discrimination.293 In reviewing a claim under the Rehabilitation Act of 1973, the Court noted that Congress recognized discrimination against individuals with disabilities “to be most often the product, not of invidious animus, but rather of thoughtlessness and indifference—of benign neglect.”294

Eric Schnapper, a professor at the University of Washington School of Law, has argued that courts should not simply prohibit discriminatory ends, but should also require the least discriminatory means.295 He argues that many times decisions that result in a racial impact are not based on “an affirmative desire to harm [B]lacks, but on a greater willingness to see a given burden borne by [B]lacks than by whites.”296

The fact that tort law and criminal law have successfully woven a broad spectrum of mental states into their legal rubrics would support the position that constitutional law can also incorporate a broader spectrum of mental states into its framework. Constitutional law cannot remain anachronistic in its adherence to racial animus anchored in intentional harm. It may arguably be more effective at achieving its goals of equal treatment before the law if it incorporates notions of recklessness and negligence. Not only do such mental states more accurately reflect the functioning of the human brain, but applying them will also ease the burden on plaintiffs to demonstrate the mental state of a defendant, making the constitutional protection against racial discrimination more accessible and more in tune with the egalitarian standard of the Constitution.

2. Discriminatory Negligence as an Alternative Framework

For this reason, the standard for showing racial discrimination should be changed to disparate impact plus discriminatory negligence. Discriminatory negligence, borrowing from the notion often used in the tort and criminal law contexts to illustrate mental state, means that an actor should have been aware that the acts that she committed created an unwarranted risk of harm, and those actions fell below the established standard for the protection of others against unreasonable risk of harm.297 Unconscious acts that fall below the standard of equal treatment align perfectly with modern notions of implicit bias and make negligence a more accurate standard for proving unconstitutional racial discrimination in a court of law. Thus, while it is almost impossible to prove discriminatory intent, since the accused will simply deny that she intended to exclude anyone on the basis of race, under a discriminatory negligence standard a plaintiff would still be able to show racial discrimination by showing that the actions of the defendant created an unwarranted risk of racial exclusion or disparate impact. Further, even though a defendant may not have been aware that her actions created that risk, she should have been aware. The standard for proving negligence would, similarly to tort law, rely on the reasonable person standard based on facts in the world at the time. That is, the standard would inquire whether a reasonable person would have known that a certain action would have caused a certain racially disparate result. If so, then the defendant should be found liable of engaging in unconstitutional racial discrimination.

3. Discriminatory Negligence in Practice

Analyzing the facts of Washington v. Davis using this new standard shows how it would apply in a modern-day courtroom. The main question in Davis was whether Test 21 unlawfully discriminated on the basis of race to exclude Blacks from promotion in the police department. The district court noted that respondents did not actually claim any intentional discrimination or purposeful discriminatory acts.298 However, the Supreme Court requires discriminatory purpose as proof of the invidiousness of the discriminatory effect to establish the constitutional claim.299 By looking at all of the facts, including that Test 21 bore no relationship to job performance, that it screened out an alarmingly high number of black applicants, and that the police department nevertheless used the test for its recruitment program, respondents claimed that Test 21 was unconstitutional because it had a highly discriminatory impact on black applicants. Under a negligent discrimination claim with the same set of facts, the plaintiffs would need to show that although the Metropolitan Police Department may have been unaware that neutral Test 21 tended to disproportionately exclude black applicants, it should have known that discrimination was a risk, given the number of black applicants who failed the test. The three reasons that the Court gave for rejecting the plaintiffs’ claims would fail under the new standard.

First, the Court stated that it had difficulty understanding how a racially neutral qualification for employment is nevertheless racially discriminatory simply because a greater proportion of Blacks fail to qualify than members of other racial groups. Under the negligence test, the plaintiffs would have to establish through empirical evidence that the police department should have been aware that exclusion was systematic. One might do this through evidence of prior complaints to the police department from rejected applicants or published studies outlining the disparate representation in policing. This burden of proof may require a showing similar to that required by theories of failure to supervise or failure to train under Monell liability when bringing a claim against a municipality.300 Under Monell v. Department of Social Services, both of these theories require a showing of prior notice, which may be established through any number of means, including the submission of complaints, police reports, or studies on the challenged behavior.301

The second reason on which the Court relied—that the Metropolitan Police Department had made affirmative efforts to recruit black officers—would not necessarily come out differently under the discriminatory negligence standard. The police department could argue that its affirmative efforts were a response to the filtering effect that the test was producing against black applicants and that it was not responsible for being unaware of any further threat of discriminatory impact. The police could also argue that the changing general racial composition of the recruit classes and force belies discriminatory negligence and the requisite unawareness of a risk of discriminatory or disparate effect—since the demographics of the police department were changing, it did not have a heightened responsibility to be aware of unfair racial outcomes. The plaintiffs could argue that the department’s affirmative recruitment efforts before the test had no bearing on whether the test effectively nullified those efforts and that the department should have nevertheless been aware of the discriminatory filtering, and thereby invidious, effect of the test, despite its pre-testing efforts.

            The Court also stated that invalidating an otherwise neutral statute that in practice benefits and burdens one race more than another would be far-reaching and would raise serious questions about, and perhaps invalidate, a whole range of tax, welfare, public service, regulatory, and licensing statutes that may be more burdensome to the poor and the average black person than to the more affluent white.302

This fear corresponds more closely with the problems of adopting a no-mens-rea regime, rather than a new negligence regime. The new regime’s filtering effect would positively impact current laws overall. Therefore, under the new standard, the goal remains the same as under the old standard—to eliminate invidious racial discrimination.

The plaintiffs’ ultimate argument was that Test 21 had a disproportionate impact on Blacks that amounted to unconstitutional racial discrimination and that, although Title VII did not apply to their claim at the time, the Equal Employment Opportunity Commission’s interpretation of Title VII’s job-relatedness requirement should be given deference by the Court as the standard to adopt for constitutional purposes, which Test 21 clearly violated.303 Plaintiffs could prove discriminatory negligence by providing empirical evidence that the department should have known that the test tended to exclude Blacks—such as complaints by rejected test-takers previously submitted to the recruiting department or relevant policymaker; news stories published in media outlets exposing the disparaging phenomenon; statements by a policymaker within the department that she is aware of the phenomenon; or perhaps even proof of general knowledge among police department personnel, such as witness statements that the test was having this result, of which the police department should have been aware. Plaintiffs would then have to show that despite this, the police department took no steps to verify and correct the invidious discriminatory impact that the test was having. The burden would then shift to the police department to show either that it was unreasonable for the department to have known of the tendency or that it did in fact take affirmative steps to verify and correct the invidious impact of the test.

Ultimately, the Supreme Court answered the wrong question in Davis. The Court analyzed job-relatedness by the test’s relationship to the training program rather than the test’s relationship to “work behavior” provable by “empirical data demonstrating that the test is predictive of or significantly correlated with important elements of work behavior which comprise or are relevant to the job[s] . . . for which candidates are being evaluated.”304 The question is whether the test meets the correct standard, and the standard is in dispute. Under a discriminatory negligence standard, plaintiffs would have needed to demonstrate that the test’s invalidity had an invidious discriminatory impact.

4. A Mens Rea Element Should Be Maintained for Constitutional Claims

Some might recommend eliminating the mens rea requirement altogether and making the standard strict liability, requiring only a showing of disparate impact, identical to the Title VII standard. This position is unsupportable for two main reasons. First, mental culpability is a foundational element of American notions of justice.305 Indeed, it is fundamental to the United States justice system. For example, a person with a mental disability is not eligible for the death penalty because the Supreme Court has recognized that “society views [such] offenders as categorically less culpable than the average criminal.”306 A punishment is tailored to an offender’s “moral guilt.”307 Putting someone to death for crimes that she “had no intention of committing or causing does not measurably contribute to the retributive end of” the criminal law.308 Similarly, in civil law, though mens rea is not always required to establish liability, an intentional breach may increase the damages owed and the scope of liability.309

Second, courts have previously declined to hold entities to the Title VII standard for a constitutional claim. One of the Supreme Court’s concerns in Washington v. Davis was that eliminating the mens rea requirement would lead to the invalidation of numerous other laws and statutes. Every law or statute favors a constituency. However, those laws that do not negatively affect other groups should be maintained while those that disfavor certain groups should be eliminated. It would be untenable to strike down every one of these pieces of legislation simply because it was proved that it had a disparate impact favoring one group.310 The Court should be guided by the extent to which one group has borne more of the burdens or benefits of society’s laws in the aggregate. When a proposed law tips the balance, it should be found unconstitutional. It is the government’s job to seek the welfare of the marginalized, the poor, and other interest groups, which requires legislation that favors certain groups. In other words, the distinction between invidious racial discrimination and mere racial discrimination that Justice White outlines may be an appropriate one within the American historical context. This article proffers a compromise that is grounded in science and that also recognizes that some actions may ultimately comprise unconstitutional racial discrimination against certain groups.

5. Potential Impact of a New Standard

The potential impact of a policy change would not be onerous. It is not certain that a lowered mens rea standard would create a litany of litigation in courts around the country, as some may fear. Plaintiffs under this proposed standard would still have to demonstrate negligence, which is not a simple task.311 It is, however, predictable that plaintiffs may find the requirement of demonstrating that unconstitutional racial bias influenced the defendant’s actions more attainable. Courts should want to remedy instances of racial bias, however small, whether explicit or implicit. Racial bias hurts society—it causes victims to question the validity of the courts and the system of impartial justice that this country’s founders worked hard to realize. With the new standard in place, plaintiffs would internalize a sense of justice that they often feel they have been denied. In a society where the rule of thumb is racial equality, and yet tendencies toward racial bias affect every decision that we make, the judiciary should join the right side of history and embrace a standard that protects ethnic minorities from unaware majorities.

III. Conclusion

This article has sought to give a historical overview of the ideals that the United States openly espouses, including a brief review of the legal standard that the Supreme Court created in order to demonstrate unconstitutional racial discrimination and some of the major Supreme Court cases that were influenced and decided by this standard. It explored the complexities of implicit bias and how it undergirds changing the legal standard. Implicit bias affects our interpersonal interactions without our knowledge and influences how we actively see and interpret the world and the people within it, even when we affirmatively proclaim that it does not. To resolve this paradox, the Supreme Court should change the standard for showing racial discrimination from discriminatory intent to discriminatory negligence. This lower standard ensures a more responsive judicial system by “account[ing] for the most accurate model of human thought, decisionmaking, and action provided by the sciences.”312 Making this simple yet transformative change would assist the Court to adjudicate from the perspective of restoring justice to society’s victims rather than providing mere piecemeal (and arguably unsuccessful) deterrence to society’s perpetrators. Such a change would catapult the current judicial system into an age of jurisprudential decision-making girded in science and data and closer to the equality and justice that the founding fathers envisioned.

Suggested Reading

Panel I: Defund Means Defund Andrea Ritchie (she/her) is a Black lesbian immigrant whose writing, litigation, and advocacy has focused on policing of women and LGBT people of color for the past two decades. She is currently a researcher with