The Yale Law Journal


The Implicit Racial Bias in Sentencing: The Next Frontier

31 Jan 2017


A prominent life scientist recently declared that the Higgs boson particle, the Internet, and implicit bias are the three most important discoveries of the past half-century.1 In President Obama’s commencement address at Howard University last year, Obama stated: “And we knew . . . that even the good cops with the best of intentions—including, by the way, African-American police officers—might have unconscious biases, as we all do.”2 Why has implicit racial bias worked its way into a presidential address? More importantly, after focusing so long on explicit biases, what do we need to know and do about the pervasive problem of implicit racial bias in the courtroom?3

As I and many others, including Professor L. Song Richardson, argue, implicit racial bias is now the most pervasive problem affecting the criminal justice system. In her review of Nicole Gonzalez Van Cleve’sbook Crook County: Racism and Injustice in America’s Largest Criminal Court, Professor Richardson eloquently discusses the explicit racial bias that Van Cleve vividly portrays happening in Cook County, Illinois courtrooms.4 As Professor Richardson, in her insightful critique, reveals, the pernicious and invisible-to-the-naked-eye effects of implicit bias in the shadows of the courtrooms and courthouses in Cook County—which Gonzalez Van Cleve does not address—present the more vexing problems.5 While the structural racial bias resulting from the “systemic triage” that Professor Richardson explores is an urgent problem, so too is the next frontier—the emerging discovery of implicit racial bias arising out of the relationship between skin tone, Afrocentric features, and sentencing.6

Social scientists, academics, lawyers, judges, and court administrators have recently demonstrated a heightened interest in implicit bias. While the recognition and limited study of implicit racial bias in the courtroom is not much older than the discovery of the Higgs boson, at least one criminal defense lawyer recognized it in a motion for new trial nearly ninety years ago. Lena Olive Smith—the first black female member of the Minnesota bar, a renowned civil rights lawyer of her time,7 and one of my personal heroes—called attention to the racial dynamics of a 1928 state court prosecution in which a black man was tried for raping a white woman before an all-white jury. In a motion for a new trial, Smith wrote:

The court fully realizes I am sure, that the very fact that the defendant was a colored boy and the prosecutrix a white woman, and the entire panel composed of white men—there was a delicate situation to begin with, and counsel for the State took advantage of this delicate situation . . . . [P]erhaps [the jurors] were, with a few exceptions, conscientious in their expressions [of no race prejudice]; yet it is common knowledge a feeling can be so dormant and subjected to one’s sub-consciousness, that one is wholly ignorant of its existence. But if the proper stimulus is applied, it comes to the front, and more often than not one is deceived in believing that it is justice speaking to him; when in fact it is prejudice, blinding him to all justice and fairness.8

The definition Smith offers of implicit bias (albeit not labeled as such) matches any contemporary cognitive scholars’ definition.9

While scholars have started to look at a host of implicit bias issues in the courtroom, most have done so with mock juries or studies that more closely resemble social science labs than real courtrooms.10 This Essay focuses on a single aspect of implicit racial bias in the courtroom: judicial sentencing.

Three critical points articulated by Professor Richardson are important to reiterate. First, implicit racial bias and other implicit biases exist even, and sometimes particularly, in egalitarian individuals. In fact, such individuals are less likely to be aware of these implicit biases, because they lack explicit biases.11 I am a prime example. Given my personal relationships and professional background as a former civil rights attorney, I did not consider myself racially biased. You can imagine how shocked I was, after taking my first Black/White Implicit Association Test (IAT) more than a decade ago, to discover I had strong anti-black implicit biases.12 Second, the effects of implicit biases in the courtroom are invisible to the naked eye.13 Finally, Professor Richardson is correct that, in extremely busy courts like Cook County, Illinois, where courtroom participants are overwhelmed with more cases than proper resources, such conditions create a rich environment for systemic implicit racial biases to thrive and infect every aspect of courtroom criminal proceedings.14 Professor Richardson astutely observes that multi-tasking courtroom professionals in Cook County and other overwhelmed criminal courts face time pressures that prompt them to make quick discretionary decisions—”the classic situations in which implicit biases are likely to influence decisions and judgments.”15 However, the corollary is not true. No cognitive social scientist or implicit bias scholar has suggested that implicit biases arise only when there are severe time pressures. Thus, it would be unwise to assume implicit bias in courtrooms exists in Pittsburgh, but not Pocatello, and in Chicago, but not Chico.

In this Essay, I will first briefly discuss the IAT and its objective role as the most recognized, studied, and accepted test in revealing implicit racial bias. Next, I turn to cognitive blind spots, especially judicial blind spots, that lead to implicit racial bias in sentencing. Cutting edge empirical studies of inmate populations in several states strongly suggest implicit racial bias contributes to increases in the length of sentences based on offenders’ darker skin tone and more pronounced Afrocentric features.

i. implicit racial bias in judicial sentencing

A. Implicit Association Test

The IAT,16 in its most common format, is a computerized online test that takes about ten minutes (including a “brief questionnaire of explicit attitudes, stereotypes, and related judgments about the topic, a short demographics survey, and administration of an IAT”) and is the primary and most popular tool for determining whether and what type of implicit biases individuals have.17 Its wide use is due to its “ease of administration, adaptability to a variety of topics, large effect sizes, and good reliability . . . .”18 A description of a typical administration of the IAT is in the following footnote.19

A plethora of studies have found that people harbor biased implicit associations against stereotyped group members like Blacks.20 Brian Nosek (one of the founders of the IAT, along with Anthony G. Greenwald and Mahzarin R. Banaji) reviewed data from more than 2.5 million completed IATs and self-reports across seventeen topics (e.g., race, skin-tone, religion, weight, etc.).21 Regarding the Black v. White IAT, a dark skin v. light skin (skin-tone) IAT, and a Black children v. White children IAT, Nosek found that White participants demonstrate a strong implicit pro-White preference.22 This was also true for Native Americans, Hispanics, Asians, and multi-racial individuals, demonstrating that “the result is more than an own-group preference effect.”23 Only Blacks did not, on average, express a pro-White implicit bias.24 These studies are important because implicit racial preferences often predict behavior and decision-making.25

B. Judges and Scomatas

The tendency to favor in-groups and disfavor out-groups is likely one of the most prevalent findings in social science.26 That implicit racial bias may affect sentencing should come as no surprise. There “is rich and overlapping literature” documenting implicit racial bias by white Americans favoring whites over blacks,27 and commentators “almost universally agree” that racial disparities are pervasive in the U.S. criminal justice system.28 Scholars in criminal law have used implicit racial bias analysis to explain virtually every aspect of racial discrepancies, from police procedures like stop-and-frisk to arrest rates, prosecutorial charging decisions, and plea bargaining and sentencing.29

Judges, like all vertebrates, have visual blind spots or scotomas (from the Greek word for darkness).30 We also have cognitive blind spot bias—that is, the ability to see bias in others, but not in ourselves.31 In one study, Professor Jeffery Rachlinski and co-authors found that 97% of state court administrative law judges attending an educational conference rated their ability “to avoid racial prejudice in decisionmaking” in the top half of other judges at the conference.32 Of course, that is mathematically impossible. The authors worried that “this result means that judges are overconfident about their ability to avoid the influence of race . . . .”33 In my recent national empirical study, I found that 92% of senior federal district judges, 87% of non-senior federal district judges, 72% of U.S. magistrate judges, 77% of federal bankruptcy judges, and 96% of federal probation and pre-trial services officers ranked themselves in the top 25% of respective colleagues in their ability to make decisions free from racial bias.34 Again, mathematically impossible.

Justice Anthony Kennedy recently penned an excellent definition of the cognitive blind spot bias in judges, without naming it, when he wrote about the unconstitutional failure of a state supreme court justice to recuse himself in a criminal case: “Bias is easy to attribute to others and difficult to discern in oneself.”35 Because of this very strong cognitive blind spot bias, judges and other courtroom actors are unlikely to question whether their decisions and actions are influenced by either explicit or implicit racial considerations. Without this self-examination, including taking IAT tests, judges, prosecutors, defense lawyers, probation officers and other court personnel are highly unlikely to accept any personal responsibility for their own complicity in sustaining a racialized system. As Professor Richardson observed, “courtroom actors need not be consciously biased in order for race to have pernicious and disturbing consequences on behaviors and judgments.”36 Thus, an objective of this Essay and the training of judges and court staff about the IAT and implicit biases is to bring awareness not only to blind spot cognitive biases, but also to how these biases potentially allow implicit racial and other biases to flourish. Unquestionably, there is a growing awareness of the effects of implicit bias in the legal system. Yet, at a training last year for 500 trial court judges in Florida, fewer than ten responded that they had previously taken an IAT test. That is quite typical of other implicit bias trainings I have done for other courts. Experience in these trainings also indicate judges are unaware of their blind spots. If judges and court personnel are both unaware of their blind spots and implicit biases, this can easily deceive them into “believing,” as Lena Olive Smith observed nearly ninety years ago, that “it is justice speaking to [them]; when in fact it is prejudice, blinding [them] to all justice and fairness.”37

C. Implicit Racial Bias in Sentencing

Only two studies have given actual trial judges IATs—remarkably, also in conjunction with judicial sentencing decisions.38 The only two studies on implicit bias in judges established, unnervingly, that judges (state judges in the first study and both state and federal in the second) have equal or greater implicit racial biases than members of the general public.39 The results of these two studies are both complex and highly nuanced—the subject of their own law review articles—and are only generally summarized here.

In the first study, when state judges were explicitly told about the juvenile offenders’ race in a fight scenario where the juvenile was charged with battery, White judges convicted just as often with White and Black offenders.40 However, Black judges were statistically significantly more likely to convict the offender when he was identified as White rather than Black.41 The focus of the study, though, was on the relationship between IAT scores and length of punishment. Judges (Black and White) who expressed a strong pro-White or pro-Black IAT preference did not sentence White and Black offenders differently.42 In a different part of the study, judges were asked to assign one of seven different dispositions in two juvenile cases.43 The race of the juvenile was not explicit, but race was subliminally primed by words on a computer screen. Judges primed with African-American words44 who had IAT White-positive/Black-negative scores treated the juvenile more harshly. Judges primed with African-American words who had IAT White-negative/Black-positive scores treated the juvenile less harshly.45

Turning to the second study, the authors, of which I am one, also found that both state and federal trial judges, like the state judges in the first study, had implicit biases equal to or greater than members of the general public.46 We studied implicit biases against so-called privileged minorities, Asians and Jews. Because the vast majority of implicit bias studies focus on implicit bias regarding Blacks, we chose Jews and Asians, who are perceived by many to be “model minorities or “success stories.”47 While social science research on these two groups show positive stereotypes related to, for example, education and business acumen, there are also strong negative stereotypes, such as slyness, lack of trustworthiness, and financial fraudulence.48

Here is a very brief summary of just a few of the thirteen findings from our study:

· All judge cohorts studied—federal district judges, U.S. magistrate judges, and state trial judges from eight states selected at random—possessed similarly strong implicit biases against Asians and Jews;

· Federal and state judges displayed strong to moderate implicit bias against Asians (relative to Caucasians) on the stereotype IAT, such that Asians were associated with negative moral stereotypes (e.g., greedy, dishonest, scheming) and Caucasians were associated with positive moral stereotypes (e.g., trustworthy, honest, generous);

· Federal and state judges displayed strong to moderate implicit bias against Jews (relative to Christians) on the stereotype IAT, such that Jews were associated with negative moral stereotypes (e.g., greedy, dishonest, scheming) and Christians were associated with positive moral stereotypes (e.g., trustworthy, honest, generous);

· Federal district judges gave (marginally) longer sentences to Jewish defendants than Christian defendants. There were no significant differences in how these judges sentenced White as compared to Asian defendants;

· Magistrate judges’ sentences did not vary significantly based on the defendant’s group membership;

· State judges, contrary to prediction, sentenced White defendants to significantly longer sentences than Asian defendants;

· For federal judges, the political party of the appointing president did not predict different IAT scores;

· And judges’ self-reported agreement with Asian stereotypes was correlated with their agreement with Jewish stereotypes.49

At bottom, these two studies—the only ones to administer IATs to actual judges—reached two similar but important conclusions. While judges have equal to or greater implicit biases against Blacks, Asians, and Jews than White members of the general public, for the most part, judges are able to control biases when deciding on the length of sentences. In the Rachlinski study, sentencing did not reflect racial bias, except when race was subliminally primed.50 Of course, in the real world, the race of a defendant being sentenced is explicit, not subliminally primed. The Rachlinski study also found that on the question of “conviction” rather than the length of a sentence, Black judges, who demonstrated a pro-Black bias on their IATs, were less likely to convict a Black defendant relative to a White defendant.51 In the second study, we found that state court judges gave longer sentences to the White defendant than the Asian defendant even though their IAT scores demonstrated a strong anti-Asian bias.52 Also, we found that federal judges gave longer sentences (a finding of marginal significance) to Jewish defendants than Christian defendants and that this was correlated with anti-Jewish and pro-Christian IAT scores.53

While these two studies are important, they have significant limitations. The presentation of hypothetical sentencing scenarios in a research context may not accurately reflect how real-world sentencing decisions are made—especially in the overburdened and under-resourced courts like Cook County that Professors Van Cleve and Richardson discuss. Indeed, Professor Richardson theorizes that “systemic triage”—the pressurized decision-making by courtroom participants in overburdened courts—is ripe for implicit bias to racialize justice.54 This is true regardless of the absence of judges’ and other courtroom participants’ conscious or explicit racial bias.55

As Rachlinski and his colleagues concluded, implicit biases among judges are widespread and can influence judgments, but when judges are aware of potential biases, they seem to have the cognitive skills to avoid their influence.56 The authors’ data did not permit them to determine “whether a desire to control bias or avoid the appearance of bias motivates judges in their courtrooms the way it” did in their study.57 Certainly awareness of a judge’s implicit biases is a first step. Professor Richardson notes that “awareness of implicit bias” and “doubting one’s objectivity” are promising interventions.58 However, it remains unclear whether awareness alone, even coupled with doubting one’s objectivity will allow judges to minimize their implicit biases in decision-making. One recent study observed that judges are now well sensitized to Black-White racial bias and are seemingly “able to avoid it for that reason.”59 This conclusion was based on a discussion at the beginning of the Pizzi article that concluded many studies examining the length of sentences do not find disparity based solely on race.60

Before any definitive conclusions can be drawn about the effects of judicial racial implicit biases on sentencing, more and larger empirical studies must be performed. These studies should include Hispanic defendants, because at least at the federal level, there are currently more incarcerated Hispanic (35.2%) than Black (34.4%) or White (27.0%) offenders.61 More importantly, because any individual or specific judges’ implicit biases could affect the length of their sentences, it is imperative to raise judicial awareness of the potential impact that implicit biases may have on sentencing and judicial decision-making. This needs to be accomplished through judicial training.

But even if judges are sensitized to avoid sentencing disparity based solely on race, judges are not well sensitized to the new frontier of implicit racial bias: bias against those with darker skin tone and greater Afrocentric features. I now turn to this emerging issue, where most sentencing judges are totally unaware of the potential for implicit racial bias based on skin tone and Afrocentric features.62

ii. the next frontier: skin-tone, afrocentric features, and the length of sentencing

Professor Richardson recognized that skin color triggers automatic and unconscious implicit racial biases.63 Skin tone bias has deep roots in America. The widespread assumption about the inferiority of those with dark skin partly justified both the colonial and antebellum eras of American history.64 Gunnar Myrdal noted the passage from Charles S. Johnson: “The evil of and ugliness of blackness have long been contrasted in popular thinking with the goodness and purity of whiteness.”65 Social scientists have established the link between darker skin and lower educational achievement, family income, and socioeconomic status.66 Dark-skinned blacks are also more likely to grow up in segregated neighborhoods and less likely to marry and be elected to political office.67

Emerging empirical research strongly suggests that the newest frontier of implicit racial bias in sentencing is the relationship between darker skin tones, stronger Afrocentric features, and longer sentences.68 This cutting-edge research is based upon sentencing data and images of Afrocentric facial features and skin tone of actual offenders from Florida, Minnesota, Oregon, and North Carolina. The studies used sophisticated regression analyses to determine that the variables of darker skin tone and greater Afrocentric facial features were the cause of longer sentences. Thus, it is not race alone, but Afrocentric features like darker skin tone, wider noses, coarser hair, darker eyes, and fuller lips that influence the length of a criminal sentence, because defendants with these characteristics are perceived as more likely displaying a Black stereotype of aggressiveness, criminality, dangerousness, and recidivist law-breaking.69 After posting a short summary of this article on a criminal law blog, Judge Richard Kopf, a federal district judge in Lincoln, Nebraska, replied: “I confess that I have known this—the Afrocentric feature effect—in my heart for a long time. It is very difficult for me to overcome—it as almost like it is hard-wired.”70

Indeed, strong Afrocentric features can lead to dire consequences, including death. In one well-known study, Professor Jennifer Eberhardt and colleagues established that, where a white victim was involved, a Black defendant with strong Afrocentric features was twice as likely to be given the death penalty by a Philadelphia jury than a Black defendant with weak Afrocentric features.71

Studies have shown, with consensus, that it is easy to classify both Blacks and Whites on a scale of strength of Afrocentric features.72 Thus, it should not be surprising that judges in sentencing are fully cognizant of the strength of Afrocentric features in people they sentence, albeit subconsciously and therefore implicitly. Nor should it be surprising that, when the more well-known overt category of race (Black-White) does not seem to have a strong influence on the length of a sentence because of judges’ abilities to mitigate race, skin tone and Afrocentric features do.73 This suggests that much greater awareness of the potential effects of darker skin tone and greater Afrocentric features on the length of sentences is needed.

I strongly recommend that each state and federal court entity explore the urgent need to train not only sentencing judges but other professionals that have major input into the sentencing process, like probation officers, who write pre-sentence reports, and defense and prosecution lawyers, who also recommend the length of sentences. It is especially important to train probation officers because of their high blind spot bias to racial discrimination74 and their singular role in preparing pre-sentence reports and frequently recommending a sentence. This training should be conducted by a member of the growing body of social scientists, law professors (often with a deep understanding of and/or an advanced degree in cognitive psychology), and other scholars in the implicit bias field.75 The training should include each judge taking a number of IATs; a detailed explanation of how the IAT works, including how it is scored; social science studies on IAT scores and decision-making; and the emerging neuroscience of implicit bias, especially the fMRI studies on race, face identification, and amygdala (the brain’s fear center) activation. Additionally, implicit bias training should be incorporated into new judges’ training. And, because of the dramatic increase in the volume and breadth of research on implicit bias, it should also be part of any continuing education of judges.76 One of my suggestions in my training is to eliminate the photograph of the offender on the front page of the pre-sentence report. The photograph is a classic psychological prime that can easily trigger implicit bias in the judges’ evaluation of the rest of the pre-sentence report.

Progressive courts should consider collaborating with academic social scientists to study the effects of this new frontier in implicit bias on sentencing in their own courts. In my training of judges on implicit bias, only a tiny percentage—far less than 1%—have been aware of racial implicit bias and IAT scores as they relate to sentencing. Even fewer knew anything about the relationship between darker skin tone and more Afrocentric facial features subconsciously impacting the length of sentences. That is why implicit bias and its relationship to darker skin tone and more Afrocentric features is truly the “New Frontier.” I am optimistic that once sentencing judges become aware of how these subconscious implicit biases work, awareness will help them combat the pernicious effects. Further empirical research is necessary to determine whether increased judicial awareness of the potential impact of offenders’ darker skin tones and greater Afrocentric features minimizes the length of sentences.


Professor Gonzalez Van Cleve’s ethnography, Crook County: Racism and Injustice in America’s Largest Criminal Court, reveals massive overt racism in the criminal courts of Cook County. Professor Richardson’s critique of the book establishes that, in addition to overt racism, there are more serious problems of implicit racial bias in the criminal justice system that Professor Gonzalez Van Cleve did not address. I fully concur and offer my own critique that, at least for sentencing discrimination, the new frontier in empirical research strongly suggests that skin tone and Afrocentric facial features are important and crucial variables underlying implicit racial bias. As cognitive social scientists and members of the academy explore this new frontier, and as judges are educated and informed about the effects of skin tone and Afrocentric features in judicial decision-making, it is my hope that these important issues will be lifted from the shadows of American courtrooms and judges’ subconsciousness.

Mark W. Bennett is in his twenty-third year as a district judge in the U.S. District Court for the Northern District of Iowa. He has written and lectured extensively about implicit bias in the legal system, having helped train more than 1,500 state and federal judges from Alaska to Florida and many more lawyers. He has conducted empirical research on state and federal judges about how implicit bias affects their decision-making. He is also the first judge in the nation to instruct jurors on implicit bias.

Preferred Citation: Mark W. Bennett, The Implicit Racial Bias in Sentencing: The Next Frontier, 126 Yale L.J. F. 381 (2017),