Article

Accountability as a Debiasing Strategy: Testing the Effect of Racial Diversity in Employment Committees

I.     Introduction

Despite the advances made in the Civil Rights Era, race is still a salient and politically divisive issue in the United States. There is clear evidence that traditional racism, such as de jure segregation and beliefs in the biological inferiority of African Americans,1 is no longer the primary barrier to equal opportunity and full participation of minorities in the American workplace.2 In the 21st century, racial minorities are limited by increasingly subtle, informal, and institutionally based forms of racism.3 These contemporary forms of bias and discrimination continue to perpetuate disadvantages as employment disparities persist in hiring, compensation, promotion, and other high stakes employment outcomes.4 This “[w]orkplace bias . . . is a reality in organizations large and small,” at all levels of the organization, “in both the private and public sectors.”5

Statistics reveal ongoing barriers to equality in the workplace. For example, in 2017, the median Black male earned 69.3 cents for every dollar the median white male earned.6 Likewise, the median Latino male earned only 70.1 cents for every dollar the median white male earned.7 The wage earnings gap was also experienced among women. For instance, the median white woman made 80.5 cents compared to the earnings of the median white male, and Black and Latina women made 66.6 cents and 61.9 cents, respectively, for every dollar a white male earned.8 At this rate, “Hispanic women will have to wait until 2233 and Black women will wait until 2124 for equal pay.”9Further, while unemployment was about 4.9% in 2016, down from 10% in 2009, African Americans faced a different reality.10 African Americans faced an unemployment rate of 8.4% compared to only 4.3% for whites, showing no improvement over the gap that existed fifteen years earlier.11 Most disturbing, this unemployment gap between African Americans and whites was consistent across all levels of educational attainment.12

While these macro trends of racial inequality are informative, we must also examine the extent to which these differences are attributable to racial bias and discrimination in the workplace.13 The subtle and covert nature of contemporary discrimination poses challenges for social scientists who wish to uncover biased outcomes in concrete terms.14 Experimental methods have helped reveal the true extent of discrimination for equally-matched workers where all factors are held exactly the same except for race.15 For example, in a study by Bertrand and Mullainathan, the researchers sent statistically identical resumes to employers in Boston and Chicago.16 Half of the employers received resumes with African-American-sounding names (Lakisha Washington and Jamal Jones), while the other half received the same resumes with white-sounding names (Greg Baker and Emily Walsh).17 White names triggered a callback rate that was 50% higher than that of equally qualified Black applicants.18 Further, their study indicated that improving the qualifications of applicants benefited white applicants but not African Americans, thus leading to greater discrimination against those with higher skill.19 Overall, they found that the having a white name is a benefit equivalent to eight additional years of employment experience.20

Another large scale experimental study tested the impact of both race and criminal record on employment outcomes.21 In this audit study, matched pairs of individuals, also known as, “testers” applied for real job openings. The study was designed to observe whether employers respond differently to applicants depending on their race and criminal record.22 Unsurprisingly, the study found that a criminal record is a major obstacle to employment and that African Americans are more strongly affected by the impact of a criminal record than their white counterparts.23 However, the experiment also revealed that an African-American job candidate with no criminal background is less likely to be extended a callback than an equally qualified white job candidate with a criminal record.24 The effect of race was substantial, in fact greater than or equal to the effect of having a criminal record on employment outcomes.25

A recent study of law firm partners conducted by Nextions Consulting found that discrimination persists after African Americans enter the workforce. The law firm partners who agreed to participate in the “writing analysis study” received copies of a legal memo.26 The legal memos distributed to all the partners were identical. However, half the partners were instructed that an African-American male associate named Thomas Meyer wrote the memo, and half were told that a Caucasian male associate named Thomas Meyer wrote the memo.27 The reviewers rated the memo thought to be written by a white man an average score of 4.1 out of 5, while they rated the memo thought to be written by an African-American man a score of 3.2 out of 5.28 Furthermore, the partners found an average of 2.9 out of 7 grammar and spelling errors in the memo written by the white writer, while they found 5.8 out of 7 errors in the memo written by the African-American writer.29 This research demonstrates that being hired is just one of many hurdles faced by racial minorities. Once gaining the esteemed employment, race continues to influence outcomes, including the way supervisors evaluate two equivalent employee’s skills and work product, such as their writing.30

While these discriminatory outcomes may be due in part to traditional prejudice and racial animus, contemporary bias is often subtle, unconscious, and institutionally based.31 Whatever the root cause of the bias, the consequences for racial minorities are real. Therefore, the following question remains: What are the best strategies to reduce bias and discriminatory outcomes? This project explores various solutions that will begin to break down racial bias specifically in the employment discrimination context.

In this Article, workplace bias is defined as “differences in career outcomes by gender or race/ethnicity that are not attributable to the differences in skills, qualifications, interests, and preferences that individuals bring to the employment setting.”32 In the context of racial bias, I define “debiasing” as strategies that reduce an individual’s or institution’s implicit or explicit reliance on social stereotypes and societal hierarchies, which too often serve as misguided indicators of a worker’s skills, competence, interests, and value to the organization.33 Debiasing interventions at the individual and institutional level can be effective, inexpensive methods to improve decision making and reduce discriminatory outcomes.34

In addition to providing solutions and strategies, this Article fills a gap in the literature by presenting the first empirical study that examines whether an employment committee’s racial composition influences the decision-making process.35 Experimental results reveal that greater debiasing effects are realized with accountability to a racially diverse committee compared to accountability to a homogeneous committee.36 Members of diverse committees were more likely to value diversity, acknowledge structural discrimination, and favor inclusive promotion decisions.37

This Article is organized into five main parts. Part II presents my conceptual framework for understanding different types of debiasing strategies based on social psychological research. Part III elaborates on the theory of accountability as a debiasing strategy. Part IV presents my empirical study that tests whether the racial composition of the employment committee to which one is accountable influences the extent of debiasing. Part V explores various mechanisms of how racial diversity influences group decision-making processes. Finally, Part VI concludes with implications of this research for antidiscrimination law and organizational policy.

II.     Debiasing Efforts: A Conceptual Framework

The broader question of how we can reduce racial bias has been studied extensively by social psychologists.38 I conceptualize the strategic recommendations resulting from these empirical studies as falling into three major categories: changing minds, changing hearts, and changing structures. The ultimate goal of these strategies is the same—to reduce biased attitudes and decision making—but the core process through which bias is reduced is quite different in each category.39 These strategies are increasingly important because antidiscrimination law has proven ineffective at remedying the types of discrimination we see today. In contemporary organizations, blatant acts of racial animus by identifiable bad actors have largely been replaced by more subtle and covert forms of “second generation discrimination.”40 In reality, many employers are aware that the types of discrimination the law is designed to address are now rare and therefore disregard the law as obsolete.41 Until Title VII catches up to begin remedying these more contemporary forms of bias, we must place greater focus on alternative strategies to reduce bias and inequality.

A.     Changing Minds, Changing Hearts, and Changing Structures

Implicit bias theory has taught us a great deal about how bias can influence judgments and decision making.42 The human brain is biologically wired to categorize information quickly, which can result in automatic biases that are processed beneath our conscious awareness.43 This may cause us to discriminate without even recognizing it. Strategies designed to reduce bias by changing minds aim to shift the cognitive processes within individual decision makers. These strategies regard racial bias as a cognitive defect that occurs through this automatic categorization process that can result in errors.44 These errors may lead to biased outcomes if not checked and remedied through strategies that will alter the way the mind processes information. These debiasing strategies include raising awareness of bias and reducing cognitive load.45 An example of a changing minds strategy in the employment context would be asking a hiring manager to take additional time to review resumes so she does not make snap judgments based on ethnic names that may result in disparate outcomes.46 These strategies are aimed at individuals, and they tend to be temporary in nature.47 Strategies aimed at changing minds are not designed to change one’s affinity for a particular group. Instead, they intend to change the cognitive process in an attempt to reduce the errors that result in stereotypes and biased decision making.48

Strategies that attempt to reduce bias by changing hearts aspire to change core attitudes and emotional responses toward stigmatized groups. These interventions target both implicit bias, as well as explicit bias that individuals consciously endorse.49 The goal is to eliminate bias by creating common human connections.50 These strategies tend to include a moral undertone to influence what an individual feels is the right or fair thing to do.51 These debiasing strategies include contact with minority friends or lovers and working on a diverse team toward a common goal.52 Like changing minds, changing hearts is also aimed at individuals, but it goes beyond cognitive processing in an isolated situation and involves deeper life experiences that alter how a person feels about a racial minority group.53 An example of a changing-hearts strategy in the employment context might be a white male engineer teaming up with an African-American engineer to complete a project to be entered in a competition. Although differences may lead to conflict in the early stages, the bonding that occurs through the teamwork may reduce the initial racial biases held by both individuals.

Efforts that seek to break down racial barriers by changing structures focus on reshaping institutions to effectuate change. While it is great to systematically change minds and hearts one individual at a time, these structural/institutional strategies do not require that. These interventions involve changing rules, organizational policies, practices, and culture to constrain decision makers and limit the ways that bias can affect important outcomes.54 These debiasing strategies include blind screening processes, objective evaluation checklists, affirmative action plans, or committees to review employment decisions.55 An example of a changing structures strategy in the employment context is an employer establishing a compensation review committee and charging committee members with collecting data and reviewing the bonuses recommended by individual managers. Each manager is aware that his decisions will be reviewed by the committee and he will have to explain the reason for any race or gender disparities for similarly situated employees. These individual managers may still hold racial biases, but they are now accountable to a review committee, so they may be more likely to “check” these biases when making decisions.

B.     Changing Structures: Nudging Through Accountability

Unlike changing minds and hearts, which are geared toward changing individual actors, changing structures emphasizes the need to alter institutions to mitigate the effects of these individual racial biases, whether implicit or explicit. Although an integrative effort involving all three components—minds, hearts, and structures—is ideal, I argue that changing structures presents the best opportunity to reduce discrimination both short- and long-term. This strategy is the most practical to implement because research has demonstrated that implicit and explicit biases can be very resistant to change.56

This Article draws from the vast literature on implicit bias, focusing specifically on accountability as a strategy to reduce bias, which falls within the changing structures category. Legal scholars and social psychologists have argued that the impact of racial bias on judgments can be minimized by having a mechanism in place that holds decision makers accountable for their judgments, including the process they use and the criteria they apply in the decision-making process.57 Numerous empirical studies have supported the theory that those “who expect to explain their decisions to another make more accurate and less biased decisions.”58

In this Article, I extend this work by critically examining whether effects of accountability are influenced by race. I draw theoretically from jury studies that reveal whites serving on racially mixed juries, compared to all-white juries, engage in a more thorough decision-making process and exhibit less bias.59 Following research by Sommers (2006) and Sommers et al. (2008),60 I am specifically examining how white participants may be influenced by accountability to racially diverse committees. The beliefs and behavior of whites are of particular interest given that they are at the top of the racial hierarchy in the United States and widely represented in positions of power in employment contexts.61 Building on this existing literature, this Article offers the first empirical examination of how the race of the individual(s) to which one is accountable may influence decision making in the employment context.

Nudging theory explores strategies to yield desired behavior without explicitly telling a person what decision to make or what action to carry out.62 Thaler and Sunstein suggest nudging as a way for public policy to encourage behavior while still leaving actors to freely choose.63 They define a nudge as “any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives.”64 Quotas and race-based hiring are typically illegal and can lead to backlash.65 Thus nudging, through the design of employment committees, may offer an effective debiasing alternative to these strategies that come across as more determinative and heavy handed. In the debiasing context, this theory can help design institutions and decision-making processes in ways that reduce discrimination and promote more inclusive outcomes—without telling managers specifically who or what demographic groups they need to hire. Leaving managers with this level of control will reduce resistance to antidiscrimination efforts and yield more optimal outcomes.

If simply changing the racial composition of a committee can indeed nudge less discriminatory behavior, we can encourage these changes through voluntary organizational policies such as a “Rooney Rule” for hiring committees.66 Alternatively, Title VII can be re-interpreted to hold employers liable under a negligence theory. This would encourage employers to make changes that yield inclusive hires and promotions, such as assembling more diverse committees.67 Ultimately, this would further capitalize on these debiasing effects, and further advance the broader non-discrimination goals of the statute.68 This is a proposal that has been gaining momentum among legal scholars attempting to address implicit biases in the workplace.69 These proposals are discussed in greater depth in the final section.

III.     Accountability as a Debiasing Strategy

Accountability, or the fear of scrutiny, and the motive for cohesion leads individuals to actively anticipate the objections or counter-arguments that might be raised against their positions.70 In doing so, individuals will be more vigilant and more likely to perform the intensive cognitive tasks required for high-quality decision making, such as considering a range of options and evidence, willingness to tolerate inconsistency, and being open to receiving new evidence.71 For example, in experiments where subjects expect to be interviewed about their decision-making processes after they make a judgment, they invest greater cognitive energy, leading to more complex impressions and greater accuracy.72 In an attempt to avoid looking unwise to their audience, participants “survey a wider range of conceivabl[e] relevant cues[,] . . . pay greater attention to the cues they use[,] . . . anticipate counter[-]arguments[,] . . . and monitor[] the cues that are allowed to influence judgment and choice.”73 Prior research on accountability suggests that the type of self-critical and effortful thinking that results in debiasing tends to be activated when those making the decisions discover they are “accountable to an audience (a) whose views are unknown, (b) who is interested in accuracy, (c) who is interested in processes rather than specific outcomes, (d) who is reasonably well informed, and (e) who has a legitimate reason for inquiring into the reasons behind participants’ judgments.”74 While there has been some preliminary discussion about the social characteristics of who exactly the decision maker is accountable to, this prior literature has focused primarily on political ideology, gender, and level of authority of the actor(s) to whom one is reporting, and it has not yet examined the effect of race.75

 A.     Accountability in Employment: Committees as Gatekeepers

Employment committees responsible for hiring, compensation, and promotion frequently serve as gatekeepers to prized opportunities and resources. This type of committee is one possible context in which to create a sense of accountability, or “pressures to justify one’s opinions to others” that can potentially decrease the influence of bias on decision-making.76 While some employers have taken strides to create more racially diverse committees, others remain all-white.77 While prior research suggests that accountability created by using a group-based committee structure will lead to debiasing, it is unclear whether greater debiasing occurs with a racially diverse committee than with a homogeneous committee.

Consider the case of Ms. Patricia Griffis, a black female employee with extensive experience at a local police department. Ms. Griffis applied for a vacancy being left by her supervisor, whom she had worked under for seven years.78 Previously, Ms. Griffis had assumed her supervisor’s responsibilities whenever her supervisor “was absent due to illness or vacation, and twice for several months while [her supervisor] was on maternity leaves.”79 Ms. Griffis’s supervisor evaluated her previous “performance as exceeding expectations in nearly all performance factors” and recognized Ms. Griffis for “training new employees and for being a . . . hard-working employee.”80 When her supervisor vacated her position, Ms. Griffis applied for it.81 An all-white interview board consisting of five members reviewed the applications and conducted interviews.82 A white male applicant was scored the highest by every board member. This candidate received the position despite the fact he had only worked for the City for five months as a probationary employee with training as a records clerk.83

Should the composition of the interview committee be a significant piece of evidence in evaluating the merit of Ms. Griffis’s claim? How much weight should the court give this piece of evidence? Would we expect a more diverse committee to yield a less biased outcome? In general, courts have been reluctant to use the composition of a hiring committee as a factor which suggests discrimination.84 Courts such as the Griffis court usually only consider the composition of the committee when there are other overwhelmingly strong indicators of discrimination.85 This Article builds on implicit bias and accountability theory to empirically test whether jurists should, in fact, highlight the racial composition of hiring committees when considering discrimination claims. I hypothesize that the racial makeup of the committee will shape the decision-making process in ways that make diverse committees less susceptible to bias than homogeneous committees. If supported empirically, then committee demographics should be a more heavily weighted factor when analyzing discrimination claims. In the empirical study described in greater detail below, I find that racial diversity does in fact produce stronger debiasing effects, which supports a growing literature in the behavioral sciences on “nudging” as a theory of behavioral change.

B.     Theory of Why Race Matters

Social psychological theory has helped explain why demographic diversity may influence decision making. This literature has revealed that complex thinking may be restricted on teams that are homogeneous along the lines of race or gender, and other types of diversity.86 Thus, individuals reporting to a homogeneous committee that is of their same race may not be as likely to consider multiple perspectives when making employment decisions. Reporting to a homogeneous committee may cause the individual to conform to the expected perspectives and views of the group, resulting in a phenomenon known as “groupthink”—where group members agree prematurely without critically analyzing all relevant facts and evidence.87

Applying this theory to the employment discrimination context, biased outcomes may occur when a group of all-white or all-male decision makers are in control of resources, power, and mobility in the organization.88 For example, systemic discrimination is often perpetuated by “social closure” through which members of the dominant group consciously or unconsciously exclude outsiders from underrepresented backgrounds.89 This allows homogeneous decision makers serving on high-status committees to “monopolize access to the most desirable jobs” and restrict social networks, while they “develop trust and a sense of mutual obligation . . . based on social similarity.”90

Even well-intentioned, homogeneous decision makers may not be as likely to take multiple perspectives and check their biases. As a result, they may be more likely to make decisions that conform to the status quo.91 For example, when a group of white, male partners are required to collaborate to select who will be promoted to partner, they may select those who are most like them, and fit their stereotyped image of leadership, resulting in “homosocial reproduction.”92 Breaking the homogeneity by assembling committees that are diverse along lines of race and gender can lead to greater complex thinking, reduced racial bias, and greater compliance with antidiscrimination law.93

 C.     From Jury Studies to Employment Committees

While the benefits of diverse committees may make intuitive sense to those who support diversity and inclusion, it is important to examine these effects empirically to know whether racial diversity really does make a difference, and if so, how. The effect of group racial composition has been studied empirically in research exploring jury deliberations and outcomes. For example, Sommers (2006) conducted an experiment where participants deliberated on the trial of a Black defendant as members of a racially homogeneous or heterogeneous mock jury.94 Analysis of the deliberations revealed that white participants on racially mixed juries, compared to those on all-white juries, cited more case facts, made fewer errors, raised more questions about what was missing from the trials, and were more likely to discuss racial issues, such as profiling, during deliberations.95 Even before the discussion began, whites in diverse groups were less punitive toward a Black defendant.96

Sommers et al. (2008) further found that when whites merely anticipate being members of a racially heterogeneous group, they exhibit more thorough information processing.97 In two experiments, white participants “who read about a race-relevant topic exhibited better reading comprehension when they expected a discussion with a diverse group” and were more likely to discuss polarizing social issues than those assigned to an all-white group.98 In these studies, positive effects were observed even when participants were only assigned to a diverse group and anticipated collaborating on a task, but did not actually do so.99

Using data from Florida felony trials between 2000 and 2010, Anwar et al. (2010) examined the impact of a jury pool’s racial composition on trial outcomes.100 This study found that Black defendants are much more likely to be convicted by juries composed of all whites than white defendants under the same circumstances.101 However, it also found that this disparity disappeared when the jury pool contained at least one Black participant.102 They found that when the jury is comprised of all-white jurors, “there is a large (16 percentage point) gap in conviction rates for black versus white defendants.”103 They further found that the gap in conviction rates for black versus white defendants [was] eliminated when there is at least one black member of the jury pool.”104 The authors hypothesized that the presence of Black jurors in the jury pool may have indirectly influenced trial outcomes.105 For example, interracial exposure and interactions among the jury pool pre-trial may have altered the attitudes of the final white jurors who were selected and seated.106 The following study builds on prior research to examine the social psychological processes through which racial diversity may influence group decision making in employment settings.

IV.     An Empirical Test

This Article joins these research findings on juries with legal perspectives on accountability, and offers an empirical test to expand on our understanding of debiasing processes. This split-sample survey experiment seeks to clarify the conditions under which accountability to a committee of peers influences bias and behavior.107 Based on Sommers’ 2006 and 2008 research on jury deliberations, I hypothesize that white participants assigned to a racially diverse committee and charged with making an important employment decision affecting the hiring, compensation, or promotion of other employees will demonstrate superior critical thinking compared to those sitting on an all-white committee. More specifically, I hypothesize that whites accountable to diverse committees will hold more positive views of diversity, will be more likely to identify racial bias, and will be more likely to extend an opportunity to a minority candidate than those accountable to an all-white committee.

A.     Research Questions and Experimental Method

I explore these questions using an experimental method that asks participants to review an employment scenario involving subjective decision making and informal employment practices that are vulnerable to racial bias. In the experiment, white participants were assigned to a committee to evaluate a promotion decision involving “Darnell” and “John.” The study tests whether the information processing, beliefs, and judgments of the participants are influenced by the racial composition of the committee to which he or she is accountable in the decision-making process. The workplace memo reviewed by participants reads:

Dear Member of Max Corp. Committee,

Please carefully review the case and be prepared to share your recommendations with the committee.

Darnell is a fourth year associate at Max Corp. When John, a new associate with previous experience was hired, a senior partner asked Darnell to “show him the ropes” at Max Corp. Darnell, John, and the senior partner would all be working together in the same division. Darnell agreed and felt that this would be a good opportunity to demonstrate his leadership at the company. After a few months, Darnell noticed that John and the partner were getting along very well. The partner praised John’s performance, they frequently went out to lunch, and they were always chatting amongst themselves in the partner’s office. Darnell also noticed that John was receiving more of the assignments with the most prestigious clients.

A year later, John was recommended for promotion, mainly as a result of his performance on a case with a very prestigious client and a fine recommendation from the partner. Although both employees did promising work and had similar evaluations on record, Darnell was not recommended for promotion. Darnell became concerned due to the fact that, of 39 associates who were promoted this year at Max Corp, only 3 were members of a racial minority group.

Darnell has requested that his situation be reviewed.

Building on existing research, I expect a general finding that white participants assigned to a racially diverse committee will make decisions more inclusive of Darnell, the racial minority, and they will be more sensitive to issues of social inequality, diversity, and inclusion than those assigned to homogeneous groups.

1.     Participants

This online survey experiment was distributed through the Institute for Research in the Social Sciences (“IRiSS”) at Stanford University. The participants included Stanford alumni, parents of Stanford students, and community college students. The survey was administered online using Qualtrics survey software. The sample included 182 white participants ranging from 18 to 86 years of age, with an average age of 51. Sixty-eight percent were female. Participants resided in 33 different states. Sixty-seven percent of participants had graduate degrees, 19% held bachelor’s degrees, and 14% were high school graduates. Seventy percent identified as politically liberal, 19% conservative, and 11% identified as neither. This survey sample is broader than a convenience sample of undergraduate or law students for increased generalizability, although not the ideal national probability sample.108

 2.     Procedures and Experimental Design

Participants were randomly assigned to work with either a racially diverse committee or homogeneous all-white committee. This two-group experimental design helps capture the complexity of real-life employment choices and the contextual factors that may play a role in shaping them. To simulate the process of working on a team to make an employment decision, participants were told that they would be assigned to a committee with other research participants to work together to make an employment decision. The participants were provided with the names of the other committee members they would be working with at the outset of the study. The participants assigned to the diverse committee were given the names of one white, one Asian, and one African-American group member, while the participants assigned to the homogeneous committee were provided with three white names. The race of each group member was manipulated by using names commonly held by each racial group.109

Participants were instructed to imagine they were serving on a virtual work committee for a company named Max Corporation. The committee was being asked to work together remotely to make an employment decision. Each participant was asked to review the employment scenario involving Darnell and John. The scenario raises issues of subjective decision making, allocation of opportunities and resources, and a promotion that involves issues of subtle workplace bias. The participants then completed a 15-minute survey which asked them to: (1) answer questions regarding their reactions and suggested decision regarding the scenario; (2) assess whether they are likely to reach an agreement with other committee members on the issue; and (3) respond to survey items regarding diversity values and intergroup attitudes.

3.     Measures

Key dependent variables include: diversity beliefs, measured by several questions regarding whether diversity should be an important goal in organizations; perceived cohesion, measured by the participant’s estimate of the likelihood they will agree with other committee members regarding the promotion decision; addressing inequality, measured by the frequency of “bias words” participants use to describe the promotion scenario; racial attitudes, measured by an adaptation of the Color Blind Racial Attitude Scale (“CoBRAS”);110 and promotion decision, measured by whether the participant recommended the white or minority candidate for promotion. Demographic variables included gender, age, education, and political ideology.

To measure diversity beliefs, the participants were asked several questions regarding whether diversity should be an important goal in organizations. The participants were asked these questions after they responded to questions involving the promotion decision and they were told their responses would not be reviewed by the other committee members. The first question asked generally, “Do you believe it is important to strive for diversity in the workplace?” This general value of diversity was measured on a scale from 0 to 100 with 0 being “Not At All Important” and 100 being “Extremely Important.”

Participants were also asked to answer 12 questions related to why they believe diversity is an important goal. The question started with the stem question: “To what extent do you believe the following factors are important reasons to strive for diversity in the workplace?” A number of specific rationales were listed, such as it “introduces a broader range of viewpoints for workplace decisions”; “creates a more desirable workplace environment”; and “is required to comply with legal requirements.” Responses were recorded on a seven-point scale from 1 being “Not at All Important” to 7 being “Extremely Important.” Respondents were also asked, “Some people believe that diversity is an important goal for organizations, while others believe diversity efforts have gone too far. What is your view?” This response was coded 1 as “Gone too far”, and 0 as an “Important Goal.”

Several studies have indicated that diverse groups experience increased conflict.111 Thus, perceived cohesion was measured by the participant’s estimate of the likelihood they will agree with other committee members regarding the promotion decision. The question asked, “What is the likelihood you will reach an agreement with the other committee members regarding this decision?” This estimate of group cohesion was measured on a scale from 0 to 100 with 0 being “Definitely will Not Agree” and 100 being “Definitely will Agree.” The perceived cohesion scores were compared across conditions using t-test analyses.

I measured addressing inequality, or the extent to which participants acknowledged and addressed potential inequality, by counting the frequency of “bias words” used to describe the promotion scenario. After recommending whether to promote John or Darnell, participants were asked to “Name two reasons John received the promotion” at Max Corp. based on the information provided in the memo. The neutral reasons Max Corp. provided in the memo include “performance” and “partner recommendation.” Reasons that identify subtle forms of bias include words such as “bias,” “racist,” “favoritism,” “socialized” with partner, and “personal” relationship. The frequency of neutral and bias words was calculated and compared across the two experimental conditions.

When measuring racial attitudes, it has been argued that even symbolic racism measures are no longer sensitive to current expressions of attitudes.112 Therefore, the Color Blind Racial Attitudes Scale (CoBRAS) was used to tap into contemporary forms of racial attitude expression.113 In general, a high score on this multi-factor scale indicates that the respondent denies the existence of racism and believes that race does not and should not matter.114 At the end of the study, participants were asked to report whether they agreed or disagreed with a number of statements. Participants were told these responses would not be reported to their fellow committee members. For example, “Social policies, such as affirmative action, discriminate unfairly against white people” and “Racial problems in the U.S. are rare, isolated situations.” Items were measured on a five point Likert scale from “Strongly Disagree” to “Strongly Agree.” Some items were reverse coded as appropriate. Scores on the CoBRA scale were compared across conditions using t-test analyses.115

The promotion decision involved the scenario with John and Darnell at Max Corp.116 Race was primed using a stereotypically African-American name, “Darnell,” and a stereotypically white name, “John.” The survey item read: “Only one person in this division can be promoted. At this point, based on your expertise and opinion, what preliminary recommendation do you wish to submit to the committee?” The response options were: “Definitely Promote John”; “Definitely Promote Darnell”; “Probably Promote John”; or “Probably Promote Darnell.” Participants predominately responded in two categories of the promotion decision measure—“Probably promote John” or “Probably promote Darnell”—so this item was coded into a dichotomous variable with one being “Promote Darnell (Minority Candidate)” and zero being “Promote John (white Candidate).”

B.     Hypotheses

This study was designed to test the following hypotheses:

(1) Participants accountable to the racially diverse committee will be more likely to express positive beliefs about diversity than participants accountable to the all-white committee.

(2) Participants accountable to the racially diverse committee will be more likely to identify and address bias in their responses than participants accountable to the all-white committee.

(3) Participants accountable to the racially diverse committee will be more likely to recommend the minority candidate Darnell for promotion than participants accountable to the all-white committee.

C.     Results

At the first-level of analysis, I tested the hypotheses by comparing participants accountable to all-white committees and participants accountable to heterogeneous committees using two-sample t-tests or proportion tests, depending on the outcome variable in question.117 Mean outcomes by committee racial composition are shown in Table 1. In general, the findings reveal that participants assigned to a racially diverse committee exhibited more positive beliefs toward diversity, were more likely to acknowledge subtle forms of bias, and were more likely to promote a minority candidate than participants assigned to an all-white committee. These findings support the hypothesis that all forms of accountability are not equally as effective at reducing racial bias. Here, mere anticipation of reporting to racially diverse peers yielded less biased beliefs and behaviors, without actual contact, meaningful interaction, or exposure to minority perspectives. 

Table 1. Mean Outcome Variables by Committee Racial Composition

Williams Table 1 

Participants assigned to the racially diverse committee were more likely to express that diversity is an important goal than participants assigned to the homogeneous committee. Seventy-five percent of participants on the diverse committee believed that it is important to strive for diversity while only 68% of participants on the all-white committee held this belief. Participants on the diverse committee were more likely to support many business rationales for diversity that relate to team and company performance such as it “leads to innovation,” “improves ability to serve clients” and “introduces a broader range of viewpoints for decisions” than participants on the all-white committee. Participants on the all-white committee were more likely to endorse the idea that striving for diversity has “gone too far” than participants on the racially diverse committee.

Participants assigned to the racially diverse committee were also more likely to discuss bias when asked about the promotion scenario, than those on an all-white committee. The company memo to participants describing the promotion scenario explained that John received the promotion at Max Corporation due to his performance with an important client and a partner recommendation. Participants were asked in an open-ended format to recall why John, the white candidate, was promoted over Darnell, the minority candidate. Participants in both conditions were equally as likely to describe the neutral reasons the firm provided in the memo: “performance” and “partner recommendation.” However, participants assigned to the diverse committee were more likely to identify and report subtle forms of bias that emerged in the scenario. Twenty-eight percent of participants on the diverse committee used “bias words” to explain why the firm initially promoted John. These words include “bias,” “racist,” “favoritism,” “socialized” with partner, and “personal” relationship. Only 17% of those assigned to the all-white committee identified bias when discussing the firm’s initial promotion decision. Thus, consistent with Tetlock’s (1985) theory, participants accountable to racial minorities “exhibited behaviors that suggested a more active and engaged process of gathering information and of considering the implications of that information.”118

Figure 1. Participants’ Summary of Why Firm Promoted John and not Darnell, by Committee Racial Composition

Screen Shot 2018 06 02 at 5.46.40 PM

Participants assigned to the racially diverse committee were significantly more likely to counter the firm’s decision and recommend the minority candidate for promotion. Thirty-six percent of participants on a diverse committee recommended Darnell for promotion, while only 24% of participants on the white committee recommended Darnell. White women were overwhelmingly more likely to recommend the minority candidate for promotion across both conditions. Overall, 39.2% of females recommended the minority associate for promotion, while only 10.7% of white males recommended the minority associate for promotion.

Figure 2. Proportion of Participants Recommending Darnell for
Promotion, by Gender and Committee Racial Composition

Screen Shot 2018 06 02 at 5.48.11 PM

Participants’ racial attitudes, as measured by the CoBRAS scale, did not vary across conditions. Participants were equally as likely to respond that “Racial problems in the U.S. are rare, isolated situations,” “Racial and ethnic minorities in the U.S. have certain advantages because of the color of their skin,” and “Social policies, such as affirmative action, discriminate unfairly against white people” in both conditions. This finding supports the social psychological theory that intergroup attitudes are now more liberal, with explicit racism less frequently observed.119 Furthermore, debiasing and inclusion strategies may influence behavior through subtle and possibly unconscious processes that are not directly related to explicit racial attitudes.

These findings at the first-level of analysis support the proposition that the effectiveness of accountability depends on whom one is accountable to, including their race. White participants who were accountable to a racially diverse committee exhibited more positive diversity beliefs and were more likely to acknowledge the possible role of bias in the firm’s promotion process than those accountable to an all-white committee. This evidence supports literature on group decision-making that has found diverse groups to engage in greater complex thinking such as willingness to consider multiple perspectives.120 Participants on the diverse committee were also more likely to challenge authority and recommend the minority candidate for promotion against the firm’s initial recommendation.

Encouraged by the exploratory analysis showing the difference between individuals exposed to homogeneous groups and a heterogeneous group, I conducted a logistic regression analysis to investigate whether assignment to a diverse group increases the odds that one would recommend Darnell for promotion, while controlling for gender, racial attitudes, and political ideology.121

Table 2. Likelihood to Promote Darnell (Odds Ratios Reported)

Screen Shot 2018 06 02 at 5.49.52 PM

Model 1 shows that accountability to a diverse group, controlling for no other variables, increases the odds that one will promote Darnell by a factor of 1.84.122 Models 2 and 3 further suggest that this relationship continues even when accounting for an individual’s gender and political ideology. Interestingly, gender has a powerful effect. Males are less likely to promote Darnell by a factor of 0.19. Put another way, women were more than five times as likely as men to promote Darnell. Nonetheless, exposure to the diverse group continues to be statistically significant despite the powerful explanatory value of gender. Likewise, Model 3 shows that liberals, relative to conservatives, are 2.49 times more likely to promote Darnell. But even after controlling for ideology, exposure to a diverse group continued to have an effect on promotion.

V.     Understanding Mechanisms

A.     Research on Benefits of Racial Diversity

The above findings demonstrate benefits of racial diversity, but not due to the typical arguments advanced by social psychologists. The common mechanisms used to explain why racial diversity leads to group benefits include: (1) contribution of diverse perspectives, and (2) self-critical perspective taking.

1.     Contribution of Diverse Perspectives

          Prior research has revealed that a group’s composition can influence its dynamics and performance, but findings on how diversity affects outcomes have been mixed.123 Most research on the benefits of diversity suggests that the different perspectives offered by racial minorities, and other underrepresented groups will improve the decision-making process by broadening the discussion.124 The benefits of diversity frequently cited in this literature include increases in group creativity, information sharing, flexibility, and thoughtfulness.125

The research of social psychologists and economists that suggests racial diversity has potential performance benefits has been presented to the Supreme Court. For example, research by Patricia Gurin et al. (1999) has been cited in briefs and testimony in Grutter v. Bollinger and Fisher v. University of Texas at Austin arguing that contribution of diverse perspectives leads to intellectual engagement, increases academic motivation, enhances critical thinking, and develops social skills on college campuses.126 A related theory is that women and racial minorities need to be represented in organizations, particularly in leadership roles, to speak up for minority interests and “open doors” for other minorities to create a more inclusive environment.127

These theories of group diversity present the traditional “information exchange” explanation for diversity effects, which suggests that group heterogeneity leads groups to bring a wider range of information and perspectives to the table.128 With respect to racial diversity, it is predicted that racial minority group members will bring unique perspectives, making different contributions to a group than whites.129 An example in the employment context is the expectation that racially diverse hiring committees will be more likely to discuss institutional racism, lack of diversity, or subtle barriers to minorities because the committee members are more likely to have personal experience with these issues. This strategy places the burden for initiating the conversations on minority group members, expecting them to “educate” the rest of the group about bias and the need for diversity and inclusion. This can be problematic because it also implies that there exists a uniform “minority experience” to be conveyed.

Evidence from this empirical study suggests a simpler explanation. The findings demonstrate that, on racially diverse committees, bias is reduced and opportunity is broadened even without the racial minorities contributing their “diverse” perspectives or fighting for inclusion. Simply putting whites on a team with racial minorities makes them more likely to acknowledge bias and support diversity without any specific position or diverse perspective asserted by the minority committee members.

2.     Self-Critical Perspective Taking

Numerous studies have supported the theory that perspective taking is an effective strategy to reduce intergroup bias. For example, studies have found that perspective taking, or seeing the other side of the situation, leads to decreased activation and application of negative group stereotypes.130 Other studies have shown that adopting the perspective of one outgroup member leads to more positive evaluations of other outgroup members131 and of the group as a whole.132 When an individual reports to a committee, they believe their decisions are being monitored and they are aware they may be asked to explain their positions. This accountability may make them more likely to engage in perspective taking. Based on the accountability literature, individuals tend to alter their views depending on their audience.133 When it is unclear what views the audience holds, individuals engage in preemptive self-criticism, “think[ing] in more self-critical, integratively complex ways in which they consider multiple perspectives on the issue and try to anticipate the objections that reasonable others might raise.”134 In the case of participants reviewing the promotion scenario at Max Corp., there is some evidence that being assigned to a racially diverse committee leads to perspective taking. White participants on the diverse committee are more likely to identify bias that may be experienced by Darnell, which may mean they are more likely to take the perspective of Darnell.135

Similar to perspective taking, there is also the possibility that serving on a racially diverse committee with members of a minority group makes whites exhibit empathy toward racial minority candidates who may face barriers. Empathy is “the ability to understand and vicariously share the feelings and thoughts of other people.”136 Empathic feelings are core human responses in social and interpersonal life because they enable individuals to be in sync with their social environment and to understand others’ intentions, actions, and behaviors.137 

Although this study was not designed to explore gender differences, it is worthwhile to note that significant gender differences emerged. Empathy was not tested directly, but this gender effect may suggest that white women are more empathetic toward a minority candidate because they perceive a common experience or shared fate. This type of identification with the experiences of racial minorities may cause the women to be more willing to recognize systemic bias in employment situations, and actively combat it. According to this theory, white women may play a critical role in reducing bias in selection processes and increasing the inclusion of underrepresented racial minorities in the modern workplace.

B.     Alternative Mechanisms

This empirical study suggests that white group members on an employment committee may behave and think differently depending on their group’s racial composition. While this Article frames the group process in terms of “accountability,” it is possible that the racial diversity of an employment committee can lead to greater inclusion of a minority candidate and greater awareness of bias through mechanisms other than accountability. Although not tested directly at this stage, alternative explanations include:
(1) contact theory, and (2) social desirability caused by reporting to racial minorities.

1.     Contact Theory

Social psychologists studying how to decrease the effects of racial (and other group-based) biases have offered a “contact theory” of intergroup relations. This research posits several factors that determine whether group contact will generate more positive attitudes toward that group.138 As Jerry Kang and Mahzarin Banaji summarize: “[S]ocial psychologists have distilled the conditions that contribute to a debiasing environment. People must be: (1) exposed to disconfirming data; (2) interact with others of equal status; (3) cooperate; (4) engage in non-superficial contact; and (5) receive clear norms in favor of equality.”139

In the empirical study presented here, the white participants experience a very minimal level of contact with their minority teammates. They simply read their names and expect to collaborate remotely via computer. They do not see their faces, they do not learn any information about them other than their names and where they are from, and they do not get to hear their perspectives on the workplace scenario. The fact that the interaction was so minimal supports a mere exposure theory. Researchers have demonstrated in numerous contexts, that the mere physical presence of others can reduce bias. Deep, prolonged, and meaningful interaction may be beneficial to the team decision-making process, but this research, consistent with Sommers (2006; 2008), demonstrates that extended interaction under perfect conditions is not necessary for some level of debiasing to occur.

2.     Salience and Social Desirability

          Some research has suggested that “membership in a diverse group” makes race salient and “reminds [w]hites of their motivation to avoid prejudice.”140 Accordingly, whites’ contributions and views, including their willingness to acknowledge potential bias, may vary by group composition, as they attempt to avoid prejudice when race is salient.141 One explanation for this is a theory of “social tuning,” which is akin to social desirability bias. Social desirability bias describes “the tendency of [survey respondents] to deny socially undesirable traits or qualities and to admit to socially desirable ones.”142It can take the form of over-reporting what is perceived as “good behavior” or under-reporting undesirable behavior.143

The theory of social tuning asserts “that people generally prefer to have positive interactions with others, and that they bring their own attitudes in line with the presumed views of others in order to have those positive interactions.”144 This is sometimes described as a more general tendency of individuals to conform to a perceived socially shared norm.145 The findings in this study do not provide evidence of a social desirability or social tuning mechanism. Participants’ scores on the CoBRAS, a contemporary measure of racial attitudes, were the same across conditions. This suggests that participants on the diverse committees were not suppressing their true views and adapting to perceived minority views for likeability or to “not seem racist.”

VI.     Discussion and Implications

The foregoing results provide support for the broader notion that changing structures will help reduce the prevalence of racial bias in the employment context. Employers can change the composition of employment committees through a shift in organizational policies, practices, and culture. In turn, this will shape the decision-making process and limit the ways that bias can affect important outcomes. One question that remains is: how can we systematically encourage and incentivize managers and other organizational leaders to assemble racially diverse committees? Two possible solutions are: (1) voluntary organizational strategies aimed at decreasing discrimination and prompting inclusion, and (2) legal interventions that will promote the suggested structural changes and ultimately reduce bias. Neither of these solutions require race conscious employment decisions. Instead, they are forms of nudging that will shape the social context in ways that are more likely to yield outcomes inclusive of marginalized groups, while allowing the decision makers to maintain discretion and control. I discuss each of these two strategies in turn below, along with the broader theory of nudging applied to the employment context.

A.     Rooney Rule for Employment Committees

One example of a voluntary organizational strategy that can promote diverse hiring committees is to adopt a “Rooney Rule” for employment committees. This type of rule will mirror the rule adopted in the National Football League (“NFL”) where each committee making high stakes employment decisions will include at least one racial minority. This type of policy will work for organizations that recognize a problem and/or value diversity and thus take the initiative to reduce bias and discrimination in their workplace. By constructing racially diverse hiring committees, employers have the opportunity to shape the decision-making process in a way that reduces bias and promotes inclusive decision making.

The “Rooney Rule” was adopted by the NFL in 2002 in response to “criticism and threat of legal action over the low percentage of African American head coaches employed in its professional league.”146 The rule was named after the former owner of the Pittsburgh Steelers, who advocated for hiring more African Americans in the leadership ranks of NFL teams.147 The rule was adopted after civil rights lawyers Johnnie Cochran and Cyrus Mehri released a report entitled “Black Coaches in the National Football League: Superior Performance, Inferior Opportunities.”148 The report found that African-American coaches had statistically performed better than their white counterparts but were more likely to be the last hired and first fired.149 In response to the report, the NFL implemented the Rooney Rule requiring teams to interview minority candidates or face a fine.150 In 2003, the NFL gave weight to the rule when it fined the Detroit Lions $200,000 for not interviewing a minority candidate.151

The idea of the Rooney Rule has gained support in other contexts. For example, President Obama urged companies to voluntarily adopt the rule.152 Major tech companies, like Facebook, Pinterest, Microsoft, and Amazon have implemented the rule to promote the hiring of women and racial minorities.153 Likewise, the Pentagon is also considering implementing a similar rule, but there are some detractors from within who are skeptical.154 Furthermore, NFL Commissioner Roger Goddell recently said that he plans to expand the rule to promote more women in the NFL office and executive team offices.155

Although there is some debate about the effectiveness of the Rooney Rule, Madden, and Ruther note that after the implementation of the Rooney Rule, the number of African-American head football coaches went from two to seven in just four years.156 Further, they find no significant hiring disparities between white and black coaches since the Rooney Rule was adopted.157 The authors conclude:

The affirmative efforts to open more head coaching positions in the NFL to African Americans has greatly reduced, and likely eliminated, the racial disadvantage in hire into these positions. Since the Rooney Rule was put into place, there are no racial differences in performance among head coaches in the NFL.158

On the other hand, there is literature which contends that the Rooney Rule has been ineffective. For example, Solow et al. found that after controlling for age, experience, and performance, there is “no evidence that the Rooney Rule has increased the number of minority head coaches.”159 Outside of the academic field, most critiques have noted that although the number of minority head coaches has increased, it is still woefully disproportionate to the percentage of African-American players in the NFL.160 Several critics contend that most teams do not take the requirement seriously and interview sham candidates to simply show compliance.161

While the Rooney Rule is one structural change that shows promise, it may have little effectiveness without other organizational changes. The experimental research presented in this Article demonstrates the importance of who has a seat at the table in interviewing and making these high stakes decisions. Without any diversity among the decision makers, it is less likely that the team will value diversity, acknowledge bias, and ultimately hire a racial minority. In the context of the NFL, for example, most head coach hiring decisions are made by some combination of the team’s owner, presidents, vice presidents, and general managers. In 2016, there were only two minority owners of NFL football teams, a Pakistani-born American businessman and an Asian-American woman.162 There are no people of color who serve as CEOs or Presidents on any NFL team.163 There are only five African-American general managers out of 32 teams.164 Only 10.8% of all team vice-presidents are people of color.165 Given the relative lack of diversity among those making the hiring decisions, it is not surprising that the Rooney Rule has been perceived as limited in its effectiveness. This homogeneity may contribute to the trend of symbolic compliance without any true change in the decision-making process.

To facilitate greater progress, I propose that the current Rooney Rule requiring that a racial minority be interviewed should be coupled with a parallel rule requiring that the hiring committee include at least one or more racial minorities or women. For organizations that seek to promote diversity, this provides one additional strategy in the toolkit that does not require race conscious hiring. The results presented above suggest that simply having racial minorities on the committee will encourage white decision makers to value diversity, acknowledge bias, and make decisions more inclusive of racial minorities.166 This type of accountability process could complement existing voluntary strategies like the NFL Rooney Rule.

B.     How Courts Can Encourage Debiasing Efforts: Negligence Revisited

While voluntary strategies like the Rooney Rule may be a start to begin breaking down bias, many employers will decline to adopt such policies. Additionally, the policies may lack legitimacy when not reinforced by formal antidiscrimination law.167 The enacting of Title VII of the Civil Rights Act in 1964 “outlawed employment discrimination on the basis of race, color, religion, national origin, and gender with the objective of ‘break[ing] down old patterns of racial segregation and hierarchy.’”168 I propose that Title VII be construed to incentivize debiasing among those employers less likely to voluntarily prioritize it. This can be accomplished by employing a negligence theory of discrimination. This type of framework has been gaining traction among legal scholars given the growing understanding of the role implicit and institutional bias may play in the workplace.169 Our current antidiscrimination law fails to address most of the bias that takes place in the contemporary workplace because plaintiffs have difficulty showing intent to discriminate on the basis of a protected classification.170

The negligence framework is premised on two facts. First, social science research has continued to show the widespread nature of implicit bias in employment contexts171 and employers today are aware of the impact of bias in the workplace.172 Second, Title VII has failed to account for the more subtle forms of bias that plague the contemporary workplace.173 Many discrimination cases face obstacles because a plaintiff has to show intentional discrimination, much as one has to show intent in intentional torts.174

A negligence framework for employment discrimination reconciles these two facts by arguing that bias will not cease unless companies take proactive measures to mitigate the harms caused by implicit, explicit, and institutional bias. Under this approach, companies would have a duty to mitigate bias and the related concepts of group think and social closure by taking steps to diversify their hiring, promotion, and compensation committees.175 If companies do not mitigate such bias, then this approach would hold them liable for negligent discrimination.176

Although modifying Title VII may initially come across as overly ambitious, this proposal would simply adapt disparate treatment law to be more closely aligned with judicial interpretation of Title VII in the harassment context. Judges would re-interpret Title VII to hold employers liable for discrimination on a negligence standard for failing to take steps to counter potential discrimination that remains rampant in our society.177 I argue that this relatively modest proposal is attainable as federal courts have created new theories of liability under Title VII in the past, without actually amending the statute (e.g., sexual harassment liability in Meritor/Faragher-Ellerth; disparate impact liability in Griggs; liability for failure to accommodate differences). This re-interpretation has been proposed by other scholars in the past.178 Persisting employment inequalities and Title VII’s failure to address contemporary bias is a compelling reason to revisit this proposal.179

Courts already hold employers liable for negligence in the context of harassment by co-workers. The Supreme Court in Vance v. Ball State University explained that an employer is liable if the employer is “negligent in failing to prevent harassment from taking place.”180 Accordingly, in a coworker harassment case, if the employer has reason to be aware of the harassing conduct and fails to take appropriate actions to remedy it, liability will attach.181 Given the widespread nature of bias, particularly in leadership positions, underrepresented professions and industries, and organizations where individuals have complained of discrimination, employers should know that bias may potentially lead to adverse employment actions. Thus, I argue that employers should be expected to take steps to de-bias the workplace and if not, will be held liable under a negligence standard.

Employers are required to maintain a working environment free of harassment and intimidation on the basis of race, sex, and other protected characteristics.182 Title VII provides that: “Employers are required to:
(1) establish and maintain a workplace free of harassment; (2) take steps to eliminate harassment when it occurs; and (3) take steps to redress or remedy harassment when it occurs.”183 Racial harassment decisions do not focus on intent to discriminate by either the harasser or the employer, but rather on a failure to do right.184 Similarly, in the disparate treatment context, employers would be expected to take steps to counter implicit, explicit, and institutional forms of bias present in the workplace. The employer could follow guidance set forth in the debiasing framework I presented in Part II of this Article. Some examples of steps managers and other leaders may take to reduce bias include: education, training, and dialogue; data collection to track potential inequities; goal setting to address under-representation and/or pay gaps; greater accountability in the form of a Chief Diversity Officer; and, of course, taking steps to diversify candidate pools and employment committees.185 In cases where employers fail to take these steps and discrimination occurs, the employer would be held liable under a negligence framework, even in the absence of any concrete showing of intent to discriminate or racial animus.

Under a negligence framework, individuals would not have to prove that discriminatory intent animated an adverse employment decision.186 Rather, plaintiffs could allege that the company did not take the duty of care necessary to prevent adverse outcomes which negatively affected a marginalized group.187 Such a legal regime would result, for example, in a different ruling in the monumental sex discrimination case Wal-Mart Stores, Inc. v. Dukes.188 In Wal-Mart, Betty Dukes and her co-plaintiffs claimed that Wal-Mart’s decentralized system for promoting individuals resulted in unfair discrimination towards women.189 The Supreme Court denied the certification of the women as a class because the plaintiffs were unable to show that Wal-Mart had a discriminatory animus behind its policies.190 But under a negligence framework, courts would not have to determine whether there was malicious discriminatory intent.191 Instead, the plaintiffs in Wal-Mart would simply have to show that Wal-Mart failed to fulfill its duty of preventing discrimination by adopting debiasing strategies.192 Likewise, if employers know that they may be held liable for the bias produced and reinforced by using homogeneous committees, they are more likely to make proactive changes that fulfill the nondiscrimination goals of Title VII.193

 C.     Limitations of Reforms

Although a Rooney Rule for employment committees and a negligence framework are promising strategies to encourage debiasing in employment decisions, there are a few potential limitations that should be considered. First, if an employer diversifies its hiring committees in response to a voluntary organizational policy or a new legal framework, it may attempt to use this to shield itself from liability. For example, an employer can simply add a racial minority with less rank/status to the committee and use this fact to show their efforts to de-bias as evidence that they did not have a discriminatory intent. They may simply argue “we had an African American on the committee, thus we are a nondiscriminatory employer and we fulfilled our duty to debias.” This can be argued even in cases where the step was just symbolic to signal compliance, and the committee still discriminated in the end. This will be handled similar to the harassment context, which requires employers to take steps to prevent harassment. While appointing a racial minority to an employment committee is one step toward fulfilling this duty, it does not fill the duty in itself. The employer may still be held liable if overall, it has failed to meet the standard of care to minimize implicit bias and discrimination occurs.

Next, there may be a concern that these strategies may “over-correct” for bias and employers may choose minorities who are not best for the position. Given that racial minorities often need to be better to just be considered,194 and once considered, they tend to be judged with greater scrutiny,195 this should not be our primary concern. If we ever reach the day where racial minorities are no longer underrepresented in the workplace and social science evidence does not reveal widespread discrimination, then we may want to be more attentive to striking the right balance of equity.

Lastly, there is the possibility that encouraging more diverse hiring committees will over-burden racial minorities with committee work, particularly when they are underrepresented in their fields. For example, if there are only token numbers of racial minorities, they may be asked to serve on every hiring, promotion, compensation committee, etc. to fulfill the requirements of the new policy. Ideally, this would highlight issues of under-representation and provide further motivation to expand the diversity in the workplace. Further, while the work may be time consuming, the benefits to racial minorities serving on committees outweigh the costs. Committee members often gain exposure to other decision makers in the organization, which can lead to valuable relationships, and offers an opportunity to demonstrate their leadership.196 Racial minorities are too often excluded from these roles, so their inclusion further promotes the broader integration goals of Title VII.

D.     Nudging Nondiscrimination

Encouraging firms to incorporate minorities on employment committees, whether through voluntary action or changes in legal rules, is a debiasing strategy that can be seen as a form of “nudging.”197 Nudging is a concept in behavioral economics that was first introduced by Thaler and Sunstein.198 They argued that traditional economics had oversimplified life with its conception of homo economicus—a fully rational person who is constantly calculating what is in its best interest.199 Recognizing that such a view failed to recognize true human behavior, Thaler and Sunstein suggested nudging as a way for public policy to encourage behavior while still leaving actors to freely choose.200 To be considered “a mere nudge, the intervention must be easy and cheap to avoid. Nudges are not mandates.”201 Thaler and Sunstein argue that “[w]orkplaces, corporate boards, universities, religious organizations, clubs, and even families might be able to use, and to benefit from, small exercises in libertarian paternalism.”202

Nudging has been used to subtly influence behavior in a variety of areas such as improving the rate of returns on taxes,203 the use of energy in homes,204 increasing the number of students getting tetanus shots,205 and enrolling people in 401(k) programs.206 Many of the changing structures strategies in Part II are consistent with nudging theory.207 In terms of employment discrimination, one example of a nudge would be using blind screenings, a technique which increased the representation of women in orchestras during the 1970s and 1980s.208 Another nudging technique that employers can use to reduce the gender gap is requiring candidates to be evaluated for a position jointly rather than one at a time because evaluating comparatively tends to facilitate less stereotyped thinking on the part of the decision-maker.209 Changing the way committees are assembled can be seen as a form of nudging.

VII.     Conclusion

By making decision makers accountable to diverse committees, the structure of the decision-making process will be altered in a way that encourages more hiring and promotion of racial minorities, while not over-deterministically demanding it. The evidence presented in this Article suggests that whites on diverse committees are more likely to value diversity, acknowledge bias, and make decisions inclusive of racial minorities. Thus, incorporating accountability—specifically to diverse employment committees—may improve minority outcomes and help achieve the goals of Title VII, while still maintaining the committee’s autonomy for making decisions.


  1. [1]. Black and African American are used interchangeably throughout this Article.

  2. [2]. Susan Sturm, Second Generation Employment Discrimination: A Structural Approach, 101 Colum. L. Rev. 458, 459–60 (2001) (describing “[c]ognitive bias, structures of decision making, and patterns of interaction” as the replacements of traditional or deliberate racism); see Eduardo Bonilla-Silva, The Structure of Racism in Color-Blind, “Post-Racial” America, 59 Am. Behav. Scientist 1358, 1361–63 (2015) (describing “new racism” as the covert replacement of traditional racism which permeates society resulting in minorities being systematically disadvantaged).

  3. [3]. Samuel R. Bagenstos, The Structural Turn and the Limits of Antidiscrimination Law,
    94 Calif. L. Rev. 1, 2 (2006) (discussing how workplace structures and not overt policies or attitudes about race cause inequality); Tristin K. Green, Discrimination in Workplace Dynamics: Toward a Structural Account of Disparate Treatment Theory, 38 Harv. C.R.-C.L. L. Rev. 91, 91 (2003); Sturm, supra note 2, at 468–69; see Anthony G. Greenwald & Linda Hamilton Krieger, Implicit Bias: Scientific Foundations, 94 Calif. L. Rev. 945, 950–52 (2006) (defining implicit bias as an unconscious preference for or aversion to specific groups of people and describing how it can cause a person to act contrary to avowed beliefs); Jerry Kang & Kristin Lane, Seeing Through Colorblindness: Implicit Bias and the Law, 58 UCLA L. Rev. 465, 519–20 (2010) (discussing that we still live in a racially discriminatory society because of implicit bias); Linda Hamilton Krieger
    & Susan T. Fiske, Behavioral Realism in Employment Discrimination Law: Implicit Bias and Disparate Treatment, 94 Calif. L. Rev. 997, 1027–61 (2006) (describing four tenets of social psychology and using them to refute the way in which an individual must prove Title VII discrimination because of implicit bias).

  4. [4]. See Marianne Bertrand & Sendhil Mullainathan, Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination, 94 Am. Econ. Rev. 991, 992 (2004) (finding job applicants with white-sounding names were 50% more likely to receive callbacks for interviews than applicants with African American-sounding names); John T. Jost et al., The Existence of Implicit Bias Is Beyond Reasonable Doubt: A Refutation of Ideological and Methodological Objections and Executive Summary of Ten Studies that No Manager Should Ignore, 29 Res. Organizational Behav. 39, 47–48 tbls.1 & 2 (2009) (finding that people display implicit biases regarding “racial and ethnic outgroups,” sex, citizenship, and social status, and these implicit associations predict social and organizationally significant behaviors, such as the medical choices, voting preferences, and employment); Devah Pager, The Mark of a Criminal Record, 108 Am. J. Soc. 937, 959–61 (2003) [hereinafter Pager, Mark of a Criminal Record] (finding that “a criminal record presents a major barrier to employment” and African Americans are “more strongly affected by the impact of a criminal record” than their White counterparts); Devah Pager & Hana Shepherd, The Sociology of Discrimination: Racial Discrimination in Employment, Housing, Credit, and Consumer Markets, 34 Ann. Rev. Soc. 181, 200 (2008) (finding that despite progress since the early 1960s, “discrimination does continue to affect the allocation of contemporary opportunities; and ... remains an important factor in shaping contemporary patterns of social and economic inequality”); Barbara F. Reskin, Including Mechanisms in Our Models of Ascriptive Inequality, 68 Am. Soc. Rev. 1, 14–16 (2003) (Different levels of “mechanisms influence levels of ascriptive inequality” and “[a]lthough researchers try to explain observed inequality, theories about actors’ motives guide the search for explanation, and it is all but impossible to know actors’ motives. . . . If we are serious about explaining variation in equality, our theories and analytic models must include indicators of causal mechanisms.”); Arin N. Reeves, Nextions, Written in Black & White: Exploring Confirmation Bias in Racialized Perceptions of Writing Skills (2014), http://www.
    nextions.com/wp-content/uploads/2017/05/written-in-black-and-white-yellow-paper-series.pdf (finding “confirmation bias unconsciously causes supervising lawyers to more negatively evaluate legal writing by an African American lawyer” than by a Caucasian lawyer).

  5. [5]. William T. Bielby, Minimizing Workplace Gender and Racial Bias, 29 Contemp. Soc. 120, 120 (2000).

  6. [6]. U.S. Dep’t of Labor, Bureau of Labor Statistics, Usual Weekly Earnings of Wage and Salary Workers: Fourth Quarter 2017 tbl.2 (2018), https://www.bls.gov/news.release/
    pdf/wkyeng.pdf; see also Mary C. Daly et al., Fed. Reserve Bank of S.F., Disappointing Facts About the Black-White Wage Gap 2 (2017), https://www.frbsf.org/economic-research/files/
    el2017-26.pdf (“In 1979, the average black man in America earned about 80% of the average white man ($15 versus $19 per hour).”).

  7. [7]. U.S. Dep’t of Labor, supra note 6.

  8. [8]. Id.

  9. [9]. Pay Equity & Discrimination, Inst. for Women’s Pol’y Res., https://iwpr.org/issue/
    employment-education-economic-change/pay-equity-discrimination (last visited Mar. 10, 2018).

  10. [10]. U.S. Dep’t of Labor, Bureau of Labor Statistics, Labor Force Characteristics by Race and Ethnicity, 2016, at 1 (2017) [hereinafter Labor Force Characteristics], https://
    www.bls.gov/opub/reports/race-and-ethnicity/2016/pdf/home.pdf; U.S. Dep’t of Labor, Bureau of Labor Statistics, The Recession of 2007–2009, at 2 (2012), https://www.bls.gov/
    spotlight/2012/recession/pdf/recession_bls_spotlight.pdf.

  11. [11]. U.S. Dep’t of Labor, Labor Force Characteristics, supra note 10, at 1.

  12. [12]. Valerie Wilson, African Americans Are Paid Less than Whites at Every Education Level, Econ. Pol’y Inst. (Oct. 4, 2016), http://www.epi.org/publication/african-americans-are-paid-less-than-whites-at-every-education-level.

  13. [13]. Analyzing statistical trends discussed above allows researchers to identify disparities between groups and track their movement over time. One limitation of this approach is that it can be difficult to account for the range of factors that may contribute to the inequities. These type of statistical trends always leave open the possibility that the racial disparities we believe are due to discrimination may in fact be caused by some other unmeasured factors.

  14. [14]. Pager & Shepherd, supra note 4, at 200.

  15. [15]. Experimental audit studies examining hiring decisions have revealed strong evidence of racial discrimination. A review of these studies finds a preference for white candidates over racial minorities ranging from 50% to 500%. See Devah Pager, The Use of Field Experiments for Studies of Employment Discrimination: Contributions, Critiques, and Directions for the Future, 609 Annals Am. Acad. Pol. & Soc. Sci. 104, 114 (2007).

  16. [16]. Bertrand & Mullainathan, supra note 4, at 996.

  17. [17]. Id. at 992.

  18. [18]. Id.

  19. [19]. Id.

  20. [20]. Id.

  21. [21]. Pager, Mark of a Criminal Record, supra note 4, at 955–62.

  22. [22]. Id. at 946–48.

  23. [23]. Id. at 955–60.

  24. [24]. Id. at 958.

  25. [25]. Id. at 957–59.

  26. [26]. Reeves, supra note 4.

  27. [27]. Id.

  28. [28]. Id.

  29. [29]. Id.

  30. [30]. See id.

  31. [31]. See Bagenstos, supra note 3, at 2; Green, supra note 3, at 91; Greenwald & Krieger, supra note 3, at 950–52; Kang & Lane, supra note 3, at 467–68; Susan Sturm, Lawyers and the Practice of Workplace Equity, 2002 Wis. L. Rev. 277, 281.

  32. [32]. Bielby, supra note 5, at 121.

  33. [33]. Id.

  34. [34]. See generally Kang & Lane, supra note 3 (discussing implicit bias within the law); Alexandra Kalev et al., Best Practices or Best Guesses? Assessing the Efficacy of Corporate Affirmative Action and Diversity Policies, 71 Am. Soc. Rev. 589 (2006) (discussing the efficacy of corporate affirmative action and the impact of diversity policies); Susan Sturm, The Architecture of Inclusion: Advancing Workplace Equity in Higher Education,29 Harv. J.L. & Gender 247 (2006) (discussing a framework for developing inclusive institutions).

  35. [35]. In this context, I define employment committee as a group of people, typically members of the organization, entrusted with the charge of reporting on and/or making employment decisions such as hiring, promotions, and salary decisions.

  36. [36]. See infra Part IV.C.

  37. [37]. See infra Part IV.C.

  38. [38]. See generally Nilanjana Dasgupta, Mechanisms Underlying the Malleability of Implicit Prejudice and Stereotypes: The Role of Automaticity and Cognitive Control, in Handbook of Prejudice, Stereotyping, and Discrimination 267 (Todd D. Nelson ed., 2009) (discussing contemporary social psychology theories on implicit bias and ways to control unconscious prejudice); Calvin K. Lai et al., Reducing Implicit Prejudice, 7 Soc. & Personality Psychol. Compass 315 (2013) (providing a critical analysis of current implicit prejudice reduction methods and mechanisms that influence the expression of implicit prejudice); Lisa Legault et al., Ironic Effects of Antiprejudice Messages: How Motivational Interventions Can Reduce (but Also Increase) Prejudice, 22 Psychol. Sci. 1472 (2011) (discussing the effect of prejudice-reduction policies and the resultant decrease, and the unintentional increase in prejudice); Margo J. Monteith et al., Putting the Brakes on Prejudice: On the Development and Operation of Cues for Control, 83 J. Personality & Soc. Psychol. 1029 (2002) (describing a study concerning the development and operation of cues for control “to understand how control can be exerted over (automatic) prejudiced responses”).

  39. [39]. A specific strategy can strive to reduce bias through one of these processes, or a combination of the processes. As this Article will demonstrate, diversity training is an example of a strategy that may involve a combination of changing minds, changing hearts, and changing structures.

  40. [40]. See Sturm, supra note 2, at 461.

  41. [41]. Susan Sturm, Rethinking Race, Gender, and the Law in the Twenty-First Century Workplace,
    12 Performance Improvement Q. 20, 37 (1999).

  42. [42]. See generally Greenwald & Krieger, supra note 3 (highlighting the scientific research on implicit bias).

  43. [43]. Rachel D. Godsil, Answering the Diversity Mandate: Overcoming Implicit Bias and Racial Anxiety, N.J. Law., Feb. 2014, at 25, 26–27; Kristin A. Lane et al., Implicit Social Cognition and Law, 3 Ann. Rev. L. & Soc. Sci. 427, 428–31 (2007); Jeffrey J. Rachlinski & Cynthia R. Farina, Cognitive Psychology and Optimal Government Design, 87 Cornell L. Rev. 549, 555–56 (2002); see Greenwald & Krieger, supra note 3, at 950–52, 959–62.

  44. [44]. See Linda Hamilton Krieger, The Content of Our Categories: A Cognitive Bias Approach to Discrimination and Equal Employment Opportunity, 47 Stan. L. Rev. 1161, 1216–17 (1995) (describing that in social cognition theory, “biases in social judgment operate automatically and must be controlled ... through subsequent ‘mental correction’”).

  45. [45]. See Irene V. Blair & Mahzarin R. Banaji, Automatic and Controlled Processes in Stereotype Priming, 70 J. Personality & Soc. Psychol. 1142, 1155–59 (1996) (finding that stereotypes may be automatically activated and can only be controlled and eliminated through awareness; this is especially so when the individual is under low cognitive constraints); Richard F. Martell, Sex Bias at Work: The Effects of Attentional and Memory Demands on Performance Ratings of Men and Women,
    21 J. Applied Soc. Psychol. 1939, 1941–42 (1991) (finding bias can be limited by allowing more time to consider decisions, that is by decreasing cognitive load—the amount and complexity of information to be processed in any given time frame); Devin G. Pope et al., Awareness Reduces Racial Bias 6–7 (Nat’l Bureau of Econ. Research, Working Paper No. 19765, 2013), http://
    www.nber.org/papers/w19765.pdf (finding that exposure to media attention raising awareness of racial bias, even of subtle forms, results in reduced levels of bias).

  46. [46]. One example of raising awareness is educating an individual on bias. If the manager is trained on the effects of bias, she has access to additional information that may make her think about the hiring process differently and consider potential bias. Reducing cognitive load is lowering the amount of information the manager needs to process in a short time. Asking the manager to review 20 resumes in one hour rather than 100 resumes in one hour reduces the cognitive load and encourages slower, more thoughtful deliberations rather than snap judgments that may rely on stereotypes.

  47. [47]. See generally Blair & Banaji, supra note 45 (discussing the automatic nature of biases and suggesting ways for perceivers to reduce the effects of their bias); Dasgupta, supra note 38 (discussing contemporary social psychology theories on implicit bias and ways to control unconscious prejudice).

  48. [48]. See supra note 47 and accompanying text.

  49. [49]. See supra note 47 and accompanying text.

  50. [50]. See supra note 47 and accompanying text.

  51. [51]. See supra note 47 and accompanying text.

  52. [52]. Christopher L. Aberson et al., Implicit Bias and Contact: The Role of Interethnic Friendships, 144 J. Soc. Psychol. 335, 340 (2004) (finding that implicit bias was reduced when individuals had contact “with close African American friends”); Nilanjana Dasgupta & Shaki Asgari, Seeing Is Believing: Exposure to Counterstereotypic Women Leaders and Its Effect on the Malleability of Automatic Gender Stereotyping, 40 J. Experimental Soc. Psychol. 642, 654 (2004) (finding the longer the period of exposure to counterstereotypes, the greater the decrease in automatic stereotypic beliefs); Nilanjana Dasgupta & Luis M. Rivera, When Social Context Matters: The Influence of Long-Term Contact and Short-Term Exposure to Admired Outgroup Members on Implicit Attitudes and Behavioral Intentions, 26 Soc. Cognition 112, 119–21 (2008) (finding that short-term exposure to admired outgroup members lowered levels of implicit prejudice and reduced discriminatory voting intentions); Patricia G. Devine et al., The Regulation of Explicit and Implicit Race Bias: The Role of Motivations to Respond Without Prejudice, 82 J. Personality & Soc. Psychol. 835, 835, 845–46 (2002) (finding that “explicit race bias [is] moderated by internal motivation to respond without prejudice,” while “implicit race bias [is] moderated by the interaction of internal and external motivation to respond without prejudice”); Muzafer Sherif, Superordinate Goals in the Reduction of Intergroup Conflict, 63 Am. J. Soc. 349, 349, 355–56 (1958) (finding that “introduction of goals ... compellingly shared by” group members and requiring “collaborative efforts of all” parties was “effective in reducing tension between groups” with “unfavorable attitudes and derogatory stereotypes of” one another). But see Jerry W. Robinson, Jr. & James D. Preston, Equal-Status Contact and Modification of Racial Prejudice: A Reexamination of the Contact Hypothesis, 54 Soc. Forces 911, 917–20 (1976) (concluding that some situations involving interracial contact are perceived as equal status by whites, but are not perceived the same way by Blacks and do not always significantly reduce prejudice for both groups).

  53. [53]. Aberson et al., supra note 52, at 337; Sherif, supra note 52, at 355–56.

  54. [54]. See Reskin, supra note 4, at 10–12 (discussing the effect of Title VII on employers’ implementation of policies, practices, and structures to curb ascriptive inequality); Sturm, supra note 2, at 489–520 (describing examples of successful instances of implementing policies and procedures to produce gender-neutral employment decisions).

  55. [55]. See Hal R. Arkes & Victoria A. Shaffer, Should We Use Decision Aids or Gut Feelings?, in Heuristics and the Law 411, 419–21 (Gerd Gigerenzer & Christoph Engel eds., 2006) (finding bias in decision making is decreased when decision aids are used at key decision points; these techniques provide an objective framework to structure and evaluate the thought process); Adam Benforado & Jon Hanson, The Great Attributional Divide: How Divergent Views of Human Behavior Are Shaping Legal Policy, 57 Emory L.J. 311, 336 (2008) (finding that if people think they are being monitored or may have to explain their decisions, they are more motivated to act in an unbiased way); Monica Biernat & Melvin Manis, Shifting Standards and Stereotype-Based Judgments, 66 J. Personality & Soc. Psychol. 5, 19 (1994) (finding that concrete, objective indicators and outcomes reduce standard stereotypes); Claudia Goldin & Cecilia Rouse, Orchestrating Impartiality: The Impact of “Blind” Auditions on Female Musicians, 90 Am. Econ. Rev. 715, 726–28 (2000) (observing that the hiring process for orchestral musicians was made less biased by the use of a blind audition); Kalev et al., supra note 34, at 610–12 (finding that reporting progress on an annual affirmative action plan reduces bias and inequality and that having a leader in the organization specifically responsible for diversity and equity issues increases accountability and reduces inequality); Jennifer S. Lerner & Philip E. Tetlock, Accounting for the Effects of Accountability, 125 Psychol. Bull. 255, 270 (1999) (finding that accountability to others with unknown views “is most likely to attenuate bias to the extent that a given bias results from ... lack of self-critical attention to the judgment process and ... failure to use relevant cues”); Sturm, supra note 2, at 479–89 (arguing that a successful way to combat more subtle and complex forms of workplace inequity is by linking efforts of relevant stakeholders such as lawyers, courts, employees, workplaces, and related organizations in a regulatory approach that encourages employers to proactively implement efforts to improve).

  56. [56]. Alafair S. Burke, Neutralizing Cognitive Bias: An Invitation to Prosecutors, 2 N.Y.U. J.L.
    & Liberty 512, 523 (2007) (“Unfortunately, the empirical evidence also suggests that cognitive bias is stubborn, and that education is an unlikely panacea.”).

  57. [57]. See, e.g., Tristin K. Green & Alexandra Kalev, Discrimination-Reducing Measures at the Relational Level, 59 Hastings L.J. 1435, 1444 (2008); Jerry Kang et al., Implicit Bias in the Courtroom, 59 UCLA L. Rev. 1124, 1169–72 (2012); see also Bielby, supra note 5, at 124;
    Lerner & Tetlock, supra note 55, at 256 (describing the phenomenon of people altering behavior to conform to the expectations of the audience); Nancy Levit, Megacases, Diversity, and the Elusive Goal of Workplace Reform, 49 B.C. L. Rev. 367, 427–28 (2008) (concluding that “accountability and responsibility for organizational change” are the most effective methods of promoting changes in the workplace with respect to equality and diversity); Philip E. Tetlock, Accountability and Complexity of Thought, 45 J. Personality & Soc. Psychol. 74, 81 (1983) [hereinafter Tetlock, Complexity of Thought] (finding that accountability only facilitates complex information processing “when [people] do not have the lazy option of expressing [the] views ... of the person to whom they feel accountable to”); Philip E. Tetlock, The Impact of Accountability on Judgment and Choice: Toward a Social Contingency Model, 25 Advances Experimental Soc. Psychol. 331, 341 (1992) (finding that when decision makers know the views of those to whom they are accountable, they tend to adopt those views).

  58. [58]. Mary D. Brtek & Stephan J. Motowidlo, Effects of Procedure and Outcome Accountability on Interview Validity, 87 J. Applied Psychol. 185, 189 (2002); Thomas E. Ford et al., The Role of Accountability in Suppressing Managers’ Preinterview Bias Against African-American Sales Job Applicants, 24 J. Pers. Selling & Sales Mgmt. 113, 114 (2004); Neal P. Mero & Stephan J. Motowidlo, Effects of Rater Accountability on the Accuracy and the Favorability of Performance Ratings, 80 J. Applied Psychol. 517, 523 (1995); Philip E. Tetlock & Jae Il Kim, Accountability and Judgment Processes in a Personality Prediction Task, 52 J. Personality & Soc. Psychol. 700, 706 (1987); cf. Tetlock, Complexity of Thought, supra note 57, at 80 (describing findings from the study by stating that people tended to think about issues in more complex, multidimensional terms to anticipate arguments raised in opposition).

  59. [59]. See Samuel R. Sommers, On Racial Diversity and Group Decision Making: Identifying Multiple Effects of Racial Composition on Jury Deliberations, 90 J. Personality & Soc. Psychol. 597, 604 (2006) [hereinafter Sommers, Racial Composition on Jury Deliberations]; Samuel R. Sommers et al., Cognitive Effects of Racial Diversity: White Individuals’ Information Processing in Heterogeneous Groups, 44 J. Experimental Soc. Psychol. 1129, 1134 (2008) [hereinafter Sommers et al., Cognitive Effects].

  60. [60]. See supra note 59 and accompanying text.

  61. [61]. Sommers, Racial Composition on Jury Deliberations, supra note 59, at 597; Sommers et al., Cognitive Effects, supra note 59, at 1129.

  62. [62]. See Ryan Calo, Code, Nudge, or Notice?, 99 Iowa L. Rev. 773, 775 (2014); see also On Amir & Orly Lobel, Stumble, Predict, Nudge: How Behavioral Economics Informs Law and Policy, 108 Colum. L. Rev. 2098, 2107 (2008) (discussing how nudging can improve people’s well-being rather than more coercive means).

  63. [63]. Richard H. Thaler & Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness 6–8 (2009).

  64. [64]. Id. at 6.

  65. [65]. See Ricci v. DeStefano, 557 U.S. 557, 563 (2009); Johnson v. Transp. Agency, 480 U.S. 616, 626–27 (1987).

  66. [66]. See infra Part VI.A. The “Rooney Rule” was adopted by the National Football League (“NFL”) where each time a head coach is hired at least one racial minority candidate must be interviewed. See infra notes 144–47 and accompanying text. This can be adapted to require that any committee making high stakes employment decisions will include at least one racial minority.

  67. [67]. See, e.g., Leora F. Eisenstadt & Jeffrey R. Boles, Intent and Liability in Employment Discrimination, 53 Am. Bus. L.J. 607, 670 (2016).

  68. [68]. Id. at 628, 670.

  69. [69]. See David Benjamin Oppenheimer, Negligent Discrimination,141 U. Pa. L. Rev.899,
    919–20, 922–25 (1993) (advocating for the negligent tort standard to be used in disparate treatment cases). Several circuits have established a negligence standard for harassment claims, leading scholars to push for the same standard for implicit bias claims. See Faragher v. City of Boca Raton, 524 U.S. 775, 799 (1998) (noting that some lower courts had “uniformly judg[ed] employer liability for co-worker harassment under a negligence standard”); Freeman v. Dal-Tile Corp., 750 F.3d 413, 422–23 (4th Cir. 2014) (“Similar to the reasoning we set forth for employer liability for co-worker harassment, ‘an employer cannot avoid Title VII liability for [third-party] harassment by adopting a “see no evil, hear no evil” strategy.’[sic] Therefore, an employer is liable under Title VII for third parties creating a hostile work environment if the employer knew or should have known of the harassment and failed ‘to take prompt remedial action reasonably calculated to end the harassment.’” (alteration in original) (citation omitted)); Dunn v. Wash. Cty. Hosp., 429 F.3d 689, 691 (7th Cir. 2005) (“[T]he plaintiff bears the burden of showing that the employer knew of the problem (usually though not always this requires the employee to show that a complaint was made) and that the employer did not act reasonably to equalize working conditions once it had knowledge.”); Galdamez v. Potter, 415 F.3d 1015, 1022 (9th Cir. 2005) (“An employer may be held liable for the actionable third-party harassment of its employees where it ratifies or condones the conduct by failing to investigate and remedy it after learning of it.”); Watson v. Blue Circle, Inc., 324 F.3d 1252, 1259 (11th Cir. 2003) (“When, as in this case, the alleged harassment is committed by co-workers or customers, a Title VII plaintiff must show that the employer either knew (actual notice) or should have known (constructive notice) of the harassment and failed to take immediate and appropriate corrective action.”); Turnbull v. Topeka State Hosp., 255 F.3d 1238, 1244 (10th Cir. 2001) (holding that “an employer may be responsible for sexual harassment based upon the acts of nonemployees” under a “negligence analysis”).

  70. [70]. Ronald Chen & Jon Hanson, Categorically Biased: The Influence of Knowledge Structures on Law and Legal Theory, 77 S. Cal. L. Rev. 1103, 1185 (2004) (explaining that accountability causes individuals “to muster the additional cognitive resources required for data-driven processes, utilizing those processes rather than the schematic processes that are more prone to error.” (citing Jacques-Philippe Leyens et al., Stereotypes and Social Cognition 135–37 (1994))); see also Leyens et al., supra, at 138.

  71. [71]. Leyens et al., supra note 70, at 136.

  72. [72]. See id. at 136–37; see Tetlock & Kim, supra note 58, at 700.

  73. [73]. Lerner & Tetlock, supra note 55, at 263.

  74. [74]. Id. at 259.

  75. [75]. See Christine M. Beckman & Damon J. Phillips, Interorganizational Determinants of Promotion: Client Leadership and the Attainment of Women Attorneys, 70 Am. Soc. Rev. 678, 696–98 (2005) (finding that law firms accountable to corporate clients with women in CEO or legal counsel positions had higher growth rates for women partners than those reporting to clients without women in these leadership positions); Jennifer J. Dose & Richard J. Klimoski, Doing the Right Thing in the Workplace: Responsibility in the Face of Accountability, 8 Emp. Responsibilities
    & Rts. J. 35, 49 (1995) (finding that the effectiveness of accountability strategies depend primarily on the relationship between the principal, “(party to whom one is accountable)” and the agent (employee)); Tetlock, Complexity of Thought, supra note 57, at 80–82 (describing responses to perceived political ideology of the audience).

  76. [76]. Tetlock, Complexity of Thought, supra note 57, at 74 (“[A]ccountability motivates cognitive work only when [people] do not have the lazy option of expressing views that they are confident will gain the approval of the person to whom they feel accountable ... .”); see Lerner & Tetlock, supra note 55, at 263; Levit, supra note 57, at 372 (“[T]he implementation of specific practices that make people accountable for change is more effective in increasing diversity than educating employees about stereotypes and biases.”).

  77. [77]. See Hiring Committees 2017–2018, PrawfsBlawg (July 21, 2017), http://prawfsblawg.
    blogs.com/prawfsblawg/2017/07/hiring-committees-2017-2018.html.

  78. [78]. Griffis v. City of Norman, No. 99-6420, 2000 WL 1531898, at *1 (10th Cir. Oct. 17, 2000).

  79. [79]. Id.

  80. [80]. Id.

  81. [81]. Id.

  82. [82]. Id.

  83. [83]. Id.

  84. [84]. In Turner v. Public Service Co. of Colorado, part of the plaintiff’s discrimination claim was that she was not hired because the interviewing was all-male. Turner v. Pub. Serv. Co. of Colo., 563 F.3d 1136, 1145 (10th Cir. 2009). The court rejected this argument, noting that “[n]or does the fact that Turner’s interview panel consisted of four men raise any concerns. Turner proffered no evidence that any of the interviewers held discriminatory attitudes or participated in past discrimination.” Id. at 1146. Similarly, in Kunej v. Labor Commission, the court found that a male plaintiff’s claim that he was discriminated against because of the all-female composition of the hiring committee “fail[ed] to explain how this should have affected the Commission’s findings regarding pretext or disparate impact.” Kunej v. Labor Comm’n, 306 P.3d 855, 860 n.2 (Utah Ct. App. 2013). Finally, in Bennet v. Roberts, an African-American teacher alleged that African Americans were underrepresented in the school district she applied to and that she was not hired because of her race. Bennet v. Roberts, No. 96 C 6917, 2001 WL 290188, at *1 (N.D. Ill. Mar. 16, 2001). In that case, the plaintiff focused predominantly on the all-white committee that interviewed her as evidence of systematic discrimination. Id. at *8. That court dismissed her claim because the plaintiff was unable to show a system-wide practice of using all-white committees. Id.

  85. [85]. In Hemmings v. Tidyman’s Inc., a female plaintiff who previously was given outstanding reviews and moved up the corporate ladder at a supermarket chain was interviewed by an all-male hiring committee, who ultimately chose the only male candidate for the position. Hemmings v. Tidyman’s Inc., 285 F.3d 1174, 1179 (9th Cir. 2002). The jury awarded Hemmings nearly
    $2 million in lost wages and punitive damages. Id. at 1182, 1191 n.20. Tellingly, however, there were substantial other factors which created strong inferences of discrimination besides the composition of the committee. See id. 1179, 1188–90. One member, for example, of the all-male hiring committee candidly admitted to the plaintiff that the board did not hire her “because the board ‘did not want to work with an emotional woman.’” Id. at 1179. Other factors included company-wide statistical discrimination in pay towards women, clear signs of retaliation on the part of the company against the plaintiff, and a co-plaintiff who was routinely paid inferior wages despite doing the same job that previous supervisors held. Id. at 1188–90. Thus, the composition of the committee was just one factor among a myriad of other factors which helped build the case for discrimination.

  86. [86]. See Frances J. Milliken et al., Diversity and Creativity in Work Groups: A Dynamic Perspective on the Affective and Cognitive Processes that Link Diversity and Performance, in Group Creativity: Innovation Through Collaboration 36–37 (Paul B. Paulus & Bernard A. Nijstad eds., 2003) (finding that readily detectible diversity may have a lower level of initial group identification, but this can be mitigated by careful attention to the management of a group’s interaction processes); Karen A. Jehn et al., Why Differences Make a Difference: A Field Study of Diversity, Conflict, and Performance in Workgroups, 44 Admin. Sci. Q. 741, 744 (1999) (finding that a diversity in knowledge bases positively influenced group performance, but the positive influence was mediated by conflicts about what tasks should be completed and how to do it); Elizabeth Mannix & Margaret A. Neale, What Differences Make a Difference?: The Promise and Reality of Diverse Teams in Organizations, 6 Psychol. Sci. Pub. Int. 31, 32 (2005) (concluding that realizing the economic benefits of workplace diversity requires “leadership that values a variety of opinions” and an “organizational culture [that] value[s] openness and stimulate[s] personal development”); Frances J. Milliken & Luis L. Martins, Searching for Common Threads: Understanding the Multiple Effects of Diversity in Organizational Groups, 21 Acad. Mgmt. Rev. 402, 406 (1996) (finding that while observable forms of diversity have been found to negatively affect group identification, there is some indication that racial diversity may positively influence some group-level cognitive outcomes including the number of alternatives considered and the overall quality of ideas in a decision-making task). See generally Scott E. Page, The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies (2007) (finding that groups of people with varying perspectives perform better than experts with like-minded perspectives); Sommers et al., Cognitive Effects, supra note 59 (finding that while diverse groups exchange a wider range of information than homogenous groups, the effects of diversity do not occur solely though information exchange).

  87. [87]. See Irving L. Janis, Groupthink: Psychological Studies of Policy Decisions and Fiascoes 3–7 (1983); Tony Simons et al., Making Use of Difference: Diversity, Debate, and Decision Comprehensiveness in Top Management Teams, 42 Acad. Mgmt. J. 662, 664 (1999) (arguing that debate is needed to benefit from diversity and to counteract the “groupthink” phenomenon); cf. Milliken & Martins, supra note 86, at 412 (finding that when there are outsiders on boards of directors there is a higher likelihood of “environmental analysis and more creative solutions to organizational problems” that might disrupt the status quo).

  88. [88]. Examples of employment discrimination cases where this type of social closure has been alleged include Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 345 (2011); Price Waterhouse v. Hopkins, 490 U.S. 228, 235–36 (1989); Jones v. Nat’l Council of Young Men’s Christian Ass’ns of the U.S., 48 F. Supp. 3d 1054, 1092 (N.D. Ill. 2014); Pippen v. State, 854 N.W.2d 1, 8 (Iowa 2014).

  89. [89]. William T. Bielby, Accentuate the Positive: Are Good Intentions an Effective Way to Minimize Systemic Workplace Bias?, 95 Va. L. Rev. Online 117, 123 (2010); see also Kevin Stainback, Social Contacts and Race/Ethnic Job Matching, 87 Soc. Forces 857, 857 (2008) (“[A]mong individuals using contacts in job searches, same-race contacts dramatically increase race/ethnic matching for all status groups. On the other hand, using a crossrace contact, while a rare event, strongly discourages this process.”); Donald Tomaskovic-Devey, The Gender and Race Composition of Jobs and the Male/Female, White/Black Pay Gaps, 72 Soc. Forces 45, 46, 59–62 (1993) (finding that the race pay gap is closely tied to job closure (38%) and the racial composition of the job).

  90. [90]. Bielby, supra note 89; Tomaskovic-Devey, supra note 89, at 64; see Stainback, supra note 89, at 860.

  91. [91]. Aberson et al., supra note 52, at 344 (finding that the idea of perspective taking, where a person takes the viewpoint of a member of the outgroup, is also useful in changing implicit bias); see Andrew R. Todd et al., Perspective Taking Combats Automatic Expressions of Racial Bias, 100 J. Personality & Soc. Psychol. 1027, 1038 (2011) (finding a positive impact of perspective taking on automatic expression of racial bias).

  92. [92]. Rosabeth Moss Kanter, Men and Women of the Corporation 48, 62–63 (1977) (defining “homosocial reproduction” as a management selection process in which “managers reproduce themselves” as they guard the position for applicants that are similar to them); Virginia Ellen Schein, The Relationship Between Sex Role Stereotypes and Requisite Management Characteristics, 57 J. Applied Psychol. 95, 99 (1973) (“[S]uccessful middle managers are perceived to possess those characteristics, attitudes and temperaments more commonly ascribed to men in general than to women in general.”); Raina Brands, ‘Think Manager, Think Man’ Stops Us Seeing Woman as Leaders, Guardian (July 15, 2015, 2:17 PM), https://www.theguardian.com/women-in-leadership/2015/
    jul/15/think-manager-think-man-women-leaders-biase-workplace (describing a “think manager, think male” phenomenon that “occurs because the traits we typically associate with leaders—forceful, dominant, strong, competent or even heroic—are stereotypically associated with men”).

  93. [93]. See Aberson et al., supra note 52, at 345 (establishing that interethnic contact situations may be an important step to reducing prejudice); see also Page, supra note 86, at 144–74 (discussing that experts with similar opinions can be outperformed by people with broader perspectives); Todd et al., supra note 91, at 1039 (discussing particular studies that suggest perspective is an effective strategy for countering automatic expressions of racial bias and facilitating more favorable interracial interaction).

  94. [94]. Sommers, Racial Composition on Jury Deliberations, supra note 59, at 600.

  95. [95]. Id. at 605–06.

  96. [96]. Id. at 603.

  97. [97]. Sommers et al., Cognitive Effects, supra note 59, at 1134.

  98. [98]. Id. at 1132.

  99. [99]. Id. at 1134.

  100. [100]. See Shamena Anwar et al., The Impact of Jury Race in Criminal Trials 2 (Nat’l Bureau of Econ. Research, Working Paper No. 16366, 2010), http://www.nber.org/papers/w16366.pdf.

  101. [101]. Id.

  102. [102]. Id. at 2–3.

  103. [103]. Id. at 14.

  104. [104]. Id. at 4 (“[T]he black-white conviction gap declines by an average of 16 percentage points in all trials in which there is at least one black member of the jury pool ... .”).

  105. [105]. Id. at 17 n.39.

  106. [106]. Id.

  107. [107]. I rely on experimental methods to systematically examine whether a committee’s racial composition is a causal mechanism that influences the participants’ beliefs and employment decisions. Experiments are designed to create theoretically relevant aspects of natural environments under controlled conditions. In real organizations, the environment is very complex and there are infinite confounding factors that interfere with our ability to examine a particular part of a theory. By testing the effects of accountability to a diverse committee in a controlled experiment, we can know with greater certainty that the manipulated variable, not other extraneous variables, causes the effect.

  108. [108]. These demographics may differ from the general population of hiring committees, providing a conservative test of the hypotheses. For example, women and liberals are significantly over-represented in the sample. These groups tend to be more likely than men and conservatives to appreciate the benefits of diversity and to be conscious of potential bias. See Ellen E. Kossek
    & Susan C. Zonia, Assessing Diversity Climate: A Field Study of Reactions to Employer Efforts to Promote Diversity, 14 J. Organizational Behav. 61, 61–81 (1993); Geoffrey Maruyama & José F. Moreno, University Faculty Views About the Value of Diversity on Campus and in the Classroom, inDoes Diversity Make a Difference? Three Research Studies on Diversity in College Classrooms 9, 9–23 (2000), https://www.aaup.org/NR/rdonlyres/97003B7B-055F-4318-B14A-5336321FB742/0/DIV
    REP.PDF; Moving Mind-Sets on Gender Diversity: McKinsey Global Survey Results, McKinsey & Co. (Jan. 2014), https://www.mckinsey.com/business-functions/organization/our-insights/moving-mind-sets-on-gender-diversity-mckinsey-global-survey-results; Political Typology Reveals Deep Fissures on the Right and Left: Appendix 1: Typology Group Profiles, Pew Research Ctr. (Oct. 24, 2017), http://www.people-press.org/2017/10/24/appendix-1-typology-group-profiles-2. Thus, the results may actually under-represent the magnitude of bias.

  109. [109]. Diverse committee names: Darius Washington, Bethany Nelson, and Peter Chang. Homogeneous committee names: Connor Edwards, Bethany Nelson, and Jeff Goldberg. Names of each origin were gathered from common websites that provided stereotypical names based on race. See Most Common Last Names for Whites in the U.S., Mongabay.com, https://names.
    mongabay.com/data/white.html (last visited Mar. 10, 2018); Top 20 ‘Whitest’ and Blackest’ Names, ABC News (Sept. 21, 2006), http://abcnews.go.com/2020/top-20-whitest-blackest-names/story?
    id=2470131.

  110. [110]. See Helen A. Neville et al., Construction and Initial Validation of the Color-Blind Racial Attitudes Scale (CoBRAS), 47 J. Counseling Psychol. 59, 67 (2000) (“CoBRAS is reliable and has initial construct, concurrent, discriminant, and criterion-related validity.”).

  111. [111]. Carsten K.W. De Dreu & Laurie R. Weingart, Task Versus Relationship Conflict, Team Performance, and Team Member Satisfaction: A Meta-Analysis, 88 J. Applied Psychol. 741, 747 (2003).

  112. [112]. Eduardo Bonilla-Silva, Racism Without Racists: Color-Blind Racism and the Persistence of Racial Inequality in America 4–7 (5th ed. 2018).

  113. [113]. See Neville et al., supra note 110, at 59.

  114. [114]. See id. at 63.

  115. [115]. See id. at 66.

  116. [116]. Adapted from Green, supra note 3, at 108–11.

  117. [117]. In statistics, t-tests are a type of hypothesis test that allows you to compare means between two independent groups. It is applied to compare whether the average difference between the two groups indicates a “real” difference or if it is due instead to random chance.

  118. [118]. Mero & Motowidlo, supra note 58, at 523; see Philip E. Tetlock, Accountability: A Social Check on the Fundamental Attribution Error, 48 Soc. Psychol. Q. 227, 231 (1985).

  119. [119]. See Bonilla-Silva, supra note 112, at 29–30.

  120. [120]. See Page, supra note 86, at 144–74 (discussing the ways in which diversity impacts a group’s ability to perform over individual ability).

  121. [121]. A logistic regression is form of statistical analysis that allows a researcher to see the influence of different factors on a dichotomous (or binary) outcome (or dependent) variable. It is a variation of linear regression analysis. Like linear regression, it is predictive in nature, but unlike linear regression, it is specially tailored to simple dichotomous outcome variables. In this analysis, the dichotomous outcome variable is whether Darnell was recommended for promotion (yes-no).

  122. [122]. An odds ratio is the odds that an outcome will occur given an exposure to a variable, compared to the odds of the outcome occurring in the absence of that exposure. Thus, an odds ratio of 1.84 means that when a participant was exposed to a diverse group, the odds that the participant would promote Darnell were 1.84 times greater than a participant who was not exposed to a diverse group.

  123. [123]. Mannix & Neale, supra note 86, at 43 (explaining that studies “have shown that surface-level social-category differences ... more likely ... have negative effects on the ability of groups to function effectively” than underlying differences).

  124. [124]. See Page, supra note 86, at 144–74; Patricia Gurin, Expert Report of Patricia Gurin, 5 Mich. J. Race & L. 363, 422 (1999) (reviewing studies that show “students, indeed, acquire a very broad range of skills, motivations, values, and cognitive capacities from diverse peers when provided with the appropriate opportunities to do so”); Patricia Gurin et al., Diversity and Higher Education: Theory and Impact on Educational Outcomes,72 Harv. Educ. Rev. 330, 334 (2002). For the appendices to Gurin’s report, see Expert Report of Patricia Gurin, Regents U. Mich., http://diversity.umich.edu/
    admissions/legal/expert/gurintoc.html (last updated Sept. 5, 2012).

  125. [125]. L. Richard Hoffman & Norman R.F. Maier, Quality and Acceptance of Problem Solutions by Members of Homogeneous and Heterogeneous Groups, 62 J. Abnormal & Soc. Psychol. 401, 407 (1961); Rebecca K. Lee, Implementing Grutter’s Diversity Rationale: Diversity and Empathy in Leadership, 19 Duke J. Gender L. & Pol’y 133, 143 (2011) (“Organizations ... should adopt the core diversity model, which aims to promote the sharing of information by drawing upon the experiences and ideas of diverse members ... .”); Charlan Jeanne Nemeth, Dissent as Driving Cognition, Attitudes, and Judgments, 13 Soc. Cognition 273, 287 (1995); Katherine W. Phillips et al., Diverse Groups and Information Sharing: The Effects of Congruent Ties, 40 J. Experimental Soc. Psychol. 497, 508 (2004); Harry C. Triandis et al., Member Heterogeneity and Dyadic Creativity, 18 Hum. Rel. 33, 46 (1965); see Stephen M. Rich, What Diversity Contributes to Equal Opportunity, 89 S. Cal. L. Rev. 1011, 1086–88 (2016).

  126. [126]. See supra note 124 and accompanying text.

  127. [127]. See Boris Groysberg & Katherine Connolly, Great Leaders Who Make the Mix Work,
    91 Harv. Bus. Rev. 68, 71, 76 (2013).

  128. [128]. See, e.g., Valerie P. Hans & Neil Vidmar, Jury Selection, in The Psychology of the Courtroom 39, 42 (Norbert L. Kerr & Robert M. Bray eds., 1982); Hoffman & Maier, supra note 125, at 407; Jehn et al., supra note 86, at 758.

  129. [129]. See Peters v. Kiff, 407 U.S. 493, 503–04 (1972).

  130. [130]. See, e.g., Adam D. Galinsky & Gordon B. Moskowitz, Perspective-Taking: Decreasing Stereotype Expression, Stereotype Accessibility, and In-Group Favoritism, 78 J. Personality & Soc. Psychol. 708, 721 (2000).

  131. [131]. See, e.g., Margaret Shih et al., Perspective Taking: Reducing Prejudice Towards General Outgroups and Specific Individuals, 12 Group Processes & Intergroup Rel. 565, 573 (2009).

  132. [132]. C. Daniel Batson et al., Empathy and Attitudes: Can Feeling for a Member of a Stigmatized Group Improve Feelings Toward the Group?, 72 J. Personality & Soc. Psychol. 105, 117 (1997); John F. Dovidio et al., Perspective and Prejudice: Antecedents and Mediating Mechanisms,
    30 Personality & Soc. Psychol. Bull. 1537, 1543–44 (2004); Adam D. Galinsky & Gillian Ku, The Effects of Perspective-Taking on Prejudice: The Moderating Role of Self-Evaluation, 30 Personality
    & Soc. Psychol. Bull. 594, 600 (2004); Theresa K. Vescio et al., Perspective Taking and Prejudice Reduction: The Mediational Role of Empathy Arousal and Situational Attributions, 33 Eur. J. Soc. Psychol. 455, 468 (2003); Jacquie D. Vorauer & Stacey J. Sasaki, Helpful Only in the Abstract?: Ironic Effects of Empathy in Intergroup Interaction, 20 Psychol. Sci. 191, 196 (2009).

  133. [133]. Lerner & Tetlock, supra note 55, at 263.

  134. [134]. Id. at 257; see also Philip E. Tetlock et al., Social and Cognitive Strategies for Coping with Accountability: Conformity, Complexity, and Bolstering, 57 J. Personality & Soc. Psychol. 632,
    638–40 (1989) (subjects who feel accountable to others are more likely to research, present evidence, and make decisions to please their audience, while subjects who do not feel accountable tend to engage in more self-justification and less self-criticism).

  135. [135]. See supra Part IV.C.

  136. [136]. See Frederique de Vignemont & Tania Singer, The Empathic Brain: How, When and Why?, 10 Trends Cognitive Sci. 435, 435(2006).

  137. [137]. Id.

  138. [138]. See Gordon W. Allport, The Nature of Prejudice 281 (1954); John F. Dovidio et al., Reducing Contemporary Prejudice: Combating Explicit and Implicit Bias at the Individual and Intergroup Level, in Reducing Prejudice and Discrimination 137, 147–48 (Stuart Oskamp ed., 2000). For reviews and updates of the contact hypothesis, see On the Nature of Prejudice: Fifty Years After Allport 8–9 (John F. Dovidio et al. eds., 2005); John F. Dovidio et al., Intergroup Bias: Status, Differentiation, and a Common In-Group Identity, 75 J. Personality & Soc. Psychol. 109, 109–10 (1998); Thomas F. Pettigrew & Linda R. Tropp, A Meta-Analytic Test of Intergroup Contact Theory, 90 J. Personality & Soc. Psychol. 751, 751–52 (2006).

  139. [139]. Jerry Kang & Mahzarin R. Banaji, Fair Measures: A Behavioral Realist Revision of “Affirmative Action, 94 Calif. L. Rev. 1063, 1101 (2006).

  140. [140]. Sommers, Racial Composition on Jury Deliberations, supra note 59, at 601 (citing Valerie P. Hans & Neil Vidmar, Jury Selection, in The Psychology of the Courtroom 39 (Norbert L. Kerr & Robert M. Bray eds. 1982).

  141. [141]. Samuel R. Sommers & Phoebe C. Ellsworth, Race in the Courtroom: Perceptions of Guilt and Dispositional Attributions, 26 Personality & Soc. Psychol. Bull. 1367, 1376 (2000); Samuel R. Sommers & Phoebe C. Ellsworth, “Race Salience” in Juror Decision-Making: Misconceptions, Clarifications, and Unanswered Questions, 27 Behav. Sci. & L. 599, 600–01 (2009); Samuel R. Sommers & Phoebe C. Ellsworth, White Juror Bias: An Investigation of Prejudice Against Black Defendants in the American Courtroom, 7 Psychol., Pub. Pol’y, & L. 201, 223 (2001).

  142. [142]. Derek L. Phillips & Kevin J. Clancy, Some Effects of “Social Desirability” in Survey Studies,
    77 Am. J. Soc. 921, 923 (1972); see Roger Tourangeau & Ting Yan, Sensitive Questions in Surveys, 133 Psychol. Bull. 859, 859–60 (2007).

  143. [143]. Phillips & Clancy, supra note 142, at 923; Tourangeau & Yan, supra note 142, at 863.

  144. [144]. Katharine T. Bartlett, Making Good on Good Intentions: The Critical Role of Motivation in Reducing Implicit Workplace Discrimination, 95 Va. L. Rev. 1893, 1950 (2009); see Brian S. Lowery et al., Social Influence Effects on Automatic Racial Prejudice, 81 J. Personality & Soc. Psychol. 842, 843 (2001); Stacey Sinclair et al., Social Tuning of the Self: Consequences for the Self-Evaluations of Stereotype Targets, 89 J. Personality & Soc. Psychol. 160, 160–61 (2005).

  145. [145]. See Luigi Castelli & Silvia Tomelleri, Contextual Effects on Prejudiced Attitudes: When the Presence of Others Leads to More Egalitarian Responses, 44 J. Experimental Soc. Psychol. 679, 684 (2008); see also Christian S. Crandall et al., Social Norms and the Expression and Suppression of Prejudice: The Struggle for Internalization, 82 J. Personality & Soc. Psychol. 359, 359 (2002) (finding that individuals “closely adhere to social norms when expressing prejudice, evaluating scenarios of discrimination, and reacting to hostile jokes”); Gretchen B. Sechrist & Charles Stangor, Perceived Consensus Influences Intergroup Behavior and Stereotype Accessibility, 80 J. Personality & Soc. Psychol. 645, 651 (2001) (reporting experimental findings that when individuals learn “one’s beliefs are shared (or not shared) with others,” it influences their attitudes, behavior, and “the strength of the attitude–behavior relationship”).

  146. [146]. Brian W. Collins, Note, Tackling Unconscious Bias in Hiring Practices: The Plight of the Rooney Rule, 82N.Y.U. L. Rev.870, 871 (2007) (footnote omitted).

  147. [147]. Patrick K. Thornton, The Legacy of Johnnie Cochran, Jr.: The National Football League’s Rooney Rule, 33 T. Marshall L. Rev.77, 87 (2007).

  148. [148]. Johnnie L. Cochran, Jr. & Cyrus Mehri, Black Coaches in the National Football League: Superior Performance, Inferior Opportunities (2002), http://media.wix.com/
    ugd/520423_24cb6412ed2758c7204b7864022ebb5d.pdf; see Thorton, supra note 147, at 77–78, 82.

  149. [149]. Cochran & Mehri, supra note 148, at ii, 8–12.

  150. [150]. Thorton, supra note 147, at 77–78.

  151. [151]. William C. Rhoden, Sports of The Times; Lions Need a Better Idea in the Front Office, N.Y. Times (July 31, 2003), http://www.nytimes.com/2003/07/31/sports/sports-of-the-times-lions-need-a-better-idea-in-the-front-office.html.

  152. [152]. See Christina Passariello, Tech Firms Borrow Football Play to Increase Hiring of Women, Wall St. J. (Sept. 27, 2016, 4:06 AM), https://www.wsj.com/articles/tech-firms-borrow-football-play-to-increase-hiring-of-women-1474963562; Valentina Zarya, Why is the “Rooney Rule” Suddenly Tech’s Answer to Hiring More Women?, Fortune (Aug. 10, 2015), http://fortune.com/2015/08/10/
    rooney-rule-diversity-in-tech.

  153. [153]. Passariello, supra note 152.

  154. [154]. Tom Vanden Brook, Pentagon Proposal on ‘Rooney Rule’ for Minority Officers Raising Internal Concerns, USA Today (Apr. 13, 2016, 7:13 PM), http://www.usatoday.com/story/news/politics/
    2016/04/13/pentagon-proposal-rooney-rule-minority-officer-internal-concerns/82953928.

  155. [155]. Kevin Patra, Roger Goodell: NFL Creating a Rooney Rule for Women, NFL (Feb. 6, 2016, 6:01 PM), http://www.nfl.com/news/story/0ap3000000632320/article/roger-goodell-nfl-creating-a-rooney-rule-for-women.

  156. [156]. Janice Fanning Madden & Matthew Ruther, Has the NFL’s Rooney Rule Efforts “Leveled the Field” for African American Head Coach Candidates?, 12 J. Sports Econ. 127, 128 (2011).

  157. [157]. See id. at 140.

  158. [158]. Id.

  159. [159]. Benjamin L. Solow et al., Moving on Up: The Rooney Rule and Minority Hiring in the NFL,
    18 Labour Econ.332, 333 (2011).

  160. [160]. Approximately 69% of players in the NFL currently are African-American, but the NFL has had at best 25% of all head coaches be “coaches of color.” See Richard Lapchick et al., The Inst. for Diversity & Ethics in Sport, The 2016 Racial and Gender Report Card: National Football League 5, 28 tbl.4 (2016), http://nebula.wsimg.com/1abf21ec51fd8dafbecfc2e0319
    a6091?AccessKeyId=DAC3A56D8FB782449D2A&disposition=0&alloworigin=1.

  161. [161]. Jason Reid, Rethinking the NFL’s Rooney Rule for More Diversity at the Top, FiveThirtyEight (May 20, 2016, 1:22 PM), http://fivethirtyeight.com/features/rethinking-the-nfls-rooney-rule-for-more-diversity-at-the-top; Mike Sando, Rooney Rule in Reverse: Minority Coaching Hires Have Stalled, ESPN (July 19, 2016), http://www.espn.com/nfl/story/_/id/17101097/staggering-numbers-show-nfl-minority-coaching-failure-rooney-rule-tony-dungy.

  162. [162]. Lapchick et al., supra note 160, at 7.

  163. [163]. Id. at 10, 31 tbl.7.

  164. [164]. Id.

  165. [165]. Id. at 11.

  166. [166]. See supra Part IV.C.

  167. [167]. Jamillah Bowman Williams, Breaking Down Bias: Legal Mandates vs. Corporate Interests,
    92 Wash. L. Rev. 1473, 1481 (2017).

  168. [168]. Id. (quoting Johnson v. Transp. Agency, 480 U.S. 616, 628 (1987)); see Civil Rights Act of 1964, Pub. L. No. 88-352, § 701–16, 78 Stat. 241, 253–66 (codified as amended at 42 U.S.C. § 2000e (2012)); Johnson, 480 U.S. at 628 (quoting United Steelworkers v. Weber, 443 U.S. 193, 208 (1979)).

  169. [169]. See, e.g., Oppenheimer, supra note 69, at 922–25.

  170. [170]. Handbook of Employment Discrimination Research: Rights and Realities
    xiii–xxiv, 33–34 (Laura Beth Nielsen & Robert L. Nelson eds., 2005) (finding that while “there has been very significant growth in the number of complaints filed with the EEOC and in federal courts (filings tripled from 8,000 in 1989 to almost 24,000 in 1998), success rates for plaintiffs (estimated at less than 20% for federal cases with opinions) [are low] as the courts have moved in the direction of requiring direct proof of discriminatory intent, [making] [a]ffirmative action ... in employment nearly impossible to practice, [and making] ... harassment under Title VII easier to defend against [for employers]).

  171. [171]. See, e.g., Bertrand & Mullainathan, supra note 4, at 992 (discussing that in employment callbacks there is a 50% gap as “[a]pplicants with White names need to send about 10 resumes to get one callback whereas applicants with African-American names need to send about 15 resumes”); Pager, Mark of a Criminal Record, supra note 4, at 955–60 (“[T]he effect of a criminal record appears more pronounced for blacks than it is for whites... . [T]he employment barriers of minority status and criminal record are compounded, intensifying the stigma toward this group.”). See generally Reeves, supra note 4 (finding confirmation bias in partners evaluating legal memos for errors as “evaluators unconsciously found more of the errors in the ‘African American’ Thomas Meyer’s memo” compared to the same memo from a Caucasian lawyer).

  172. [172]. Stephanie Bornstein, Reckless Discrimination, 105 Calif. L. Rev. 1055, 1057–58 (2017).

  173. [173]. See Green, supra note 3, at 91 (describing discrimination since Title VII “as a perpetual tug on opportunity and advancement” away from the underrepresented); Greenwald & Krieger, supra note 3, at 950–52 (defining implicit bias as an unconscious preference for or aversion to specific groups of people and describing how it can cause a person to act contrary to avowed beliefs); Kang & Lane, supra note 3, at 473–89 (discussing that we still—post-Title VII of the Civil Rights Act of 1964—live in a racially discriminatory society because of implicit bias); Krieger
    & Fiske, supra note 3, at 1027–61 (describing four tenets of social psychology and using them to refute the way in which an individual must prove Title VII discrimination because of implicit bias); Sturm, supra note 2, at 520–22 (describing the need of employers to implement policies and practices to ensure equal treatment because the law did not go so far as to dictate specific employment decisions that would address the more subtle forms of bias); Sturm, supra note 31, at 281 (describing the need for complex solutions to second generation workplace inequity in a post-Title VII world).

  174. [174]. Oppenheimer, supra note 69, at 922–25 (“[T]he wrong prohibited by Title VII is an intentional wrong requiring proof in some form of a conscious discriminatory motive. That wrong, at its core, is an intentional tort—an intentional wrongful interference with the rights of another person.”).

  175. [175]. See Bagenstos, supra note 3, at 18 (arguing that the law “‘would hold employers directly liable under Title VII for organizational choices, institutional practices, and workplace dynamics that enable the operation of discriminatory bias on the basis of protected characteristics[,]’ [and] [i]f a plaintiff established ‘that the employer’s institutional structures or practices unreasonably enabled the operation of discriminatory bias in the workplace,’ [the plaintiff] would be entitled to an injunction requiring the employer to change its organizational structures to reasonably minimize bias” (footnotes omitted) (quoting Green, supra note 3, at 145, 147)); Oppenheimer, supra note 69, at 967 (describing the standard of care for employers protecting employees from harassment as a reasonableness standard, which requires them to exercise great care); see also Martha Chamallas, Structuralist and Cultural Domination Theories Meet Title VII: Some Contemporary Influences, 92 Mich. L. Rev. 2370, 2398 (1994) (“I imagine that under a structuralist approach the plaintiff’s prima facie case would consist of a showing of dramatic underrepresentation of the plaintiff’s group, satisfactory performance by the plaintiff on objective measures, and evidence of a subjective, largely standardless selection process. In such a case, an employer would be held liable unless it could show that it had taken adequate measures to guard against stereotyping.”).

  176. [176]. For an interesting alternative to negligent discrimination, see generally Bornstein, supra note 172 (arguing that interventions for combatting workplace bias are in many cases so well-known and cheap to implement that companies should be held to the lower threshold of “reckless discrimination” because they know techniques available to shield implicit biases from decision making).

  177. [177]. For a review of inequities, see supra Part I.

  178. [178]. See, e.g., Oppenheimer, supra note 69.

  179. [179]. See Handbook of Employment Discrimination Research, supra note 170, at 6; Sturm, supra note 2, at 460–61.

  180. [180]. Vance v. Ball State Univ., 570 U.S. 421, 449 (2013).

  181. [181]. See Burlington Indus., Inc. v. Ellerth, 524 U.S. 742, 759 (1998).

  182. [182]. See EEOC Dec. No. YSF 9-108 (1969), Empl. Prac. Guide (CCH) ¶ 6030.

  183. [183]. Oppenheimer, supra note 69, at 946–47 (“Employers were expected to prevent racial harassment from infecting the workplace. They were expected to be careful, vigilant, and protective of minority employees in confronting harassment when it occurred. Liability would be imposed when the employer failed to comply with the standard of care established by the statutory prohibition on discrimination, either by failing to establish a harassment-free workplace, or by failing to respond appropriately when harassment occurred.”).

  184. [184]. See id.

  185. [185]. See Kalev et al., supra note 34, at 593–94; see also Frank Dobbin & Alexandra Kalev, Why Diversity Programs Fail, Harv. Bus. Rev., July–Aug. 2016, at 52, 53–55.

  186. [186]. See Oppenheimer, supra note 69, at 967–72.

  187. [187]. Professor Stone argues that there has been a change in the workplace from a traditional hierarchical structure to a new “boundaryless workplace.” See Katherine V.W. Stone, The New Psychological Contract: Implications of the Changing Workplace for Labor and Employment Law, 48 UCLA L. Rev. 519, 599–614 (2001). She warns that this new workplace structure could perpetuate discrimination in the form of unequal training for minorities and women, exclusion by male cliques in the workplace, and invisible authority structures. Id. at 605–08. Stone notes that applying a negligence standard will do little to address these forms of discrimination in today’s workplaces, and calls for companies to bring in outside dispute resolution teams to resolve discrimination among co-workers. Id. at 611–14.

  188. [188]. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011).

  189. [189]. Id. at 344–45.

  190. [190]. See id. at 353–56.

  191. [191]. Richard Thompson Ford, Bias in the Air: Rethinking Employment Discrimination Law, 66 Stan. L. Rev.1381, 1387 (2014) (describing the way in which the Wal-Mart employees from the above case would not have to provide discrete acts of discrimination, but rather a failure to take due care to prevent discrimination).

  192. [192]. Id.

  193. [193]. Oppenheimer, supra note 69, at 971.

  194. [194]. See Bertrand & Mullainathan supra note 4, at 1002–03; Sonia K. Kang et al., Whitened Résumés: Race and Self-Presentation in the Labor Market, 61Admin. Sci. Q. 469, 491–94 (2016) (finding that for applicants who did not reveal their race, “whitened résumés led to more callbacks than unwhitened résumés”).

  195. [195]. Reeves, supra note 4; Gillian B. White, Black Workers Really Do Need to Be Twice as Good, Atlantic (Oct. 7, 2015), https://www.theatlantic.com/business/archive/2015/10/why-black-workers-really-do-need-to-be-twice-as-good/409276.

  196. [196]. Deborah Thompson Eisenberg, The Restorative Workplace: An Organizational Learning Approach to Discrimination, 50 U. Rich. L. Rev. 487, 493 (2016) (Companies decrease their susceptibility to implicit bias through restorative practices. She describes restorative practices as “provid[ing] a continuum of proactive dialogic processes to promote stronger relationships or ‘social capital.’ Restorative practices are founded on the basic proposition that ‘[h]uman beings change their behavior based upon the bonds’ that they form. Those bonds can be developed through regular opportunities for interaction and dialogue, grounded in principles of respect, reciprocity, and accountability.” (second alteration in original) (footnote omitted) (quoting M. Diane McCormick, Susquehanna Township School District Adopts Anti-Bullying Program, PennLive (May 27, 2012, 6:13 PM), http://www.pennlive.com/midstate/index.ssf/2012/05/susque
    hanna_township_school_di_18.html)).

  197. [197]. Thaler & Sunstein, supra note 63, at 6–8.

  198. [198]. Id.

  199. [199]. Id. at 6.

  200. [200]. Id. at 6–8.

  201. [201]. Id. at 6.

  202. [202]. Id. at 255.

  203. [203]. Id. at 67.

  204. [204]. Id. at 69.

  205. [205]. Id. at 72.

  206. [206]. Id. at 111.

  207. [207]. See supra Part II.B.

  208. [208]. Bornstein, supra note 172, at 1057.

  209. [209]. Iris Bohnet,What Works: Gender Equality by Design 126–28 (2016); Iris Bohnet et al., When Performance Trumps Gender Bias: Joint Versus Separate Evaluation, 62Mgmt. Sci. 1225, 1225-27 (2016); Iris Bohnet, Gender Equality: A Nudge in the Right Direction, Fin. Times (Oct. 13, 2010), https://www.ft.com/content/59d7d2f6-d6a7-11df-98a9-00144feabdc0.

*

Associate Professor, Georgetown University Law Center. J.D., Stanford Law School; Ph.D. (Sociology), Stanford University.



Thank you to Richard Bales, Jeffrey Hirsch, Theresa Beiner, Veronica Root, Anthony Baldwin, and Wendy Greene for your feedback on early drafts of this Article. I would also like to thank Richard Banks, Kenworthey Bilz, and Darrell Miller for their valuable comments at the Culp Colloquium held at Duke Law School. Thank you to my colleagues Sheryll Cashin, Brian Galle, Paul Butler, and David Super for their insights. Lastly, thank you to my research assistants from Georgetown University Law Center, including Stephen Benz, Brooke Martin, and Lindsay Lincoln whose hard work was invaluable.