Helen Hershkoff∞

This essay addresses but does not purport to resolve some of the difficult issues raised by the social movement #MeToo and its relation to judicial decision making. A useful starting point for discussion is the decision of the Chief Justice of the United States to establish a Working Group to examine the sufficiency of safeguards then in place to ensure “an exemplary workplace for every judge and every court employee.”1 This decision came shortly after a federal appeals judge had resigned from office rather than defend against allegations that he had sexually harassed his judicial clerks.2 Following investigation and recommendations,3 the Judicial Conference of the United States, as well as the Administrative Office of the U.S. Courts and some circuit courts, provided new guidance and took other steps to address sexual misconduct by judges—and others—in the workplace. The process for reporting such misconduct will now be more accessible to complainants; moreover, information about workplace complaints and decisions will be separately categorized in the judiciary’s statistical reports, and retrievable through an electronic database. In addition, the Working Group clarified that complaints that are finally resolved by remedial action (other than by private censure or reprimand) are posted on a public website. Presumably those interested, especially litigants and potential claimants, will monitor this information, and law schools will take it into account when recommending that students apply for judicial clerkships. Enhanced training conducted by the Federal Judicial Center and others also will be provided to judges to reduce the incidence of sexual misconduct.4

These developments are important, salutary, and critical to preserving public trust in the judicial system. Significantly, the Working Group underscored that its report did “not conclude its work,” and that it intended “to monitor ongoing initiatives.”5 Moreover, comments submitted to the Working Group suggested that the termination of a disciplinary action “should not prevent the Judiciary from continuing an institutional review to determine if there are systemic problems within a court … that require correction.”6 In that spirit, this essay shifts attention from complaints against judges of alleged sexual misconduct in the workplace, to complaints of alleged gender-related workplace violations—for example, claims of pay equity, employment discrimination, and sexual harassment—that judges who have been accused of sexual misconduct may be tasked to decide. Should litigants know that a presiding judge is the subject of a pending sexual misconduct complaint or was the subject of a complaint found to be unsubstantiated? In other institutional contexts, it would be highly unusual to disclose information about disciplinary complaints before the matters are investigated and if they are not confirmed. However, courts are not ordinary institutions. Is there a danger that being the subject of a sexual misconduct claim, if later found to be unsubstantiated, might negatively impact the judge’s ability to credit claims of civil rights violations in an unrelated lawsuit, despite the judge’s best intentions to be impartial and to have an open mind?7

Inter-disciplinary research about decision making suggests that “[u]nconscious bias is everywhere,” as Professor Iris Bohnet writes.8 She urges that institutions adopt a conscious strategy of information disclosure on the view that transparency will make them fairer and more accountable. Arguably, raising questions about information disclosure that focus only on gender and judicial decision making (and, then, only with regard to judges who are or have been the subject of a dismissed sexual misconduct complaint) addresses too small a piece of the problem (or potential problem), with too little attention paid to gender bias generally and to other implicit biases, such as those relating to race and socioeconomic class, that likewise may require attention.9 On the other hand, some might argue that enhanced disclosure is not needed in any of these areas because norms of judicial self-governance, such as accountability, collegiality, and respect for the rule of law, work to immunize judges from forms of bias that infect decision making outside of the courthouse, and obviate the need for de-biasing strategies within the courthouse.10 Indeed, putting the words judicial bias—let alone implicit bias—in the same sentence as judging for many would be a non-starter, to be dismissed as an oxymoron that runs counter to the national image of Blind Justice,11 as well as the conventional trope that a Supreme Court Justice does the job best by calling balls and strikes without regard to personal qualities.12 Quite apart from the events that triggered the Working Group, commentators have raised questions about the possible negative impact of implicit bias on judicial decisions in civil rights cases involving gender discrimination in the workplace.13 On the other hand, some commentators have warned against invoking implicit bias as a criticism of any specific judicial decision;14 more generally, the Chief Justice in a different context than gender relations has cautioned against basing recusal motions on an “amorphous ‘probability of bias.’”15

Justice Ruth Bader Ginsburg wrote more than two decades ago that paying “close attention to the existence of unconscious prejudice can prompt and encourage those who work in the courts to listen to women’s voices”; further, she explained, it “heightens appreciation that progress … requires a concerted effort to change habitual modes of thinking and action.”16 The very narrowness of the inquiry presented in this essay may make it a useful window into larger questions about unintended barriers to our nation’s ideal of equal access to justice. Part I surveys the judiciary’s current information-disclosure regime about judicial sexual misconduct; Part II identifies particular cognitive biases that may generate or exacerbate implicit bias and negatively affect judicial decision making; and Part III considers the sufficiency of norms of judicial self-governance to counteract any unintended spillover effects of gender misconduct complaints in unrelated cases involving gender-related claims. The issues are complex, and reasonable minds certainly will differ on how best to secure judicial impartiality—both in appearance, and in actuality.17 This essay, which some might regard as provocative, is intended to be exploratory, not prescriptive.

In general, information about judicial misconduct comes from administrative complaints or recusal motions. Federal employees and litigants alleging that a judge has engaged in sexual misconduct may file administrative complaints under the Judicial Conduct and Disability Act and the Rules for Judicial-Conduct and Judicial-Disability Proceedings, 18 or through the judiciary’s internal framework of Employment Dispute Resolution Plans.19 The complaints trigger single-case investigation and individual discipline; to my knowledge, they are not analyzed with an eye toward identifying problems that might suggest a need for systemic or institutional reform. With limited exceptions, the complaint process is completely confidential. Orders will be made public through the clerk’s office of the relevant regional circuit and on that court’s website only after final action and a remedy has been imposed (other than a private censure or reprimand); the name of the subject judge but not the complainant will be disclosed.20 The system is not designed to reveal substantive information about pending complaints or those that are found to be unsubstantiated. Until the Working Group’s report, the judiciary did not separately identify information about sexual misconduct complaints in publicly available statistics, and this information was not accessible through electronic databases.21 In addition to the process for administrative complaints, litigants may file motions for recusal (and the judge may self-recuse) under 28 U.S.C. § 455—“Disqualification of justice, judge, or magistrate,” and 28 U.S.C. §144—“Bias or prejudice of a judge.” Information from motions for recusal can be gathered and analyzed from the case files themselves or from databases that are not designed specifically to analyze gender-related bias as a basis for the motion.22

The availability bias—the heuristic that treats examples that quickly come to mind as representative—suggests a modicum of caution before designing reforms based on the latest headline about a particularly bad actor.23 To date, the number of sexual misconduct complaints against judges has been impressively low.24 Nevertheless, filing rates could be the result of the cumbersome complaint process that the Working Group has sought to improve.25 The hope, of course, is that the Working Group’s recommendations of training and education will preemptively reduce explicit incidents of sexual misconduct in the future.26 Moreover, by publishing statistics about complaints, increased transparency could create institutional incentives to curtail unacceptable behavior.27

Whether to undertake further disclosure of complaints depends, in part, on whether it would be useful as a deterrent or as a marker of improvement still to be achieved. To answer this question, some comparative considerations might be useful. For example, it might be beneficial to explore the kinds of judicial biases that currently are treated as disqualifying when recusal motions are granted, when judges undertake self-recusal, or when an appeals court reassigns a matter because the district court judge has shown an inability to handle the case.28 Likewise, one might want to distinguish cases in which judges have not been expected to self-recuse, for example, because of their status as a member of a particular racial group or engagement in private conduct as a partner in a same-sex relationship.29 These comparisons could provide insight into how courts conventionally assess a judge’s point of view, experience, and conduct when challenged as factors that are said to affect impartiality or its appearance. In theory, professional norms afford sufficient protection against forms of generic bias that affect all judges and legal situations.30 The American legal system has long put faith in norms as a constraining influence on public decision making; that approach, however, has come under serious stress during the Trump Administration.31 In any event, even if broader disclosure of pending complaints or of complaints that are not substantiated is determined to be neither appropriate nor useful, questions nevertheless might remain about the impact of the complaint process on subject judges in unrelated lawsuits involving gender issues.

The Model Code of Judicial Conduct Rule 2.3 states that “[a] judge shall perform duties of judicial office … without bias,” and that “[a] judge shall not, in the performance of judicial duties, by words or conduct manifest bias ….”32 The word “bias” is freighted with negative undertones—in particular, irrational and even hostile treatment of a person or an idea. It suggests a close-minded attitude, a disregard of facts, and a reflexive rejection of reasoned argument. All of these qualities are antithetical to the role of the judge in democratic society.33 But judges are human,34 and the psychological research about human decision making underscores the ubiquity of bias,35 understood as “a pattern of judgment that systematically departs from the prescription of a normative rule.”36 Judges, like others, use heuristics to make decisions—mental shortcuts that in some situations may produce error and illusion.37 And judges, like others, are subject to emotion that in some situations may produce less than optimal outcomes in court decisions.38

Cognitive bias refers to the everyday shortcuts that human beings use to make decisions.39 The existence of cognitive bias is analytically separate from implicit gender bias,40 stereotypes about women that may negatively impact or distort decisions involving equal treatment in the workplace or other private and public spaces.41 In some situations, however, the ordinary cognitive biases that characterize decision making could lead to implicit gender bias, notwithstanding the best of judicial intentions.42 Generally, implicit bias, as psychologists understand that term, is not a ground for recusal or disqualification.43 Whether implicit bias nevertheless should be a matter of concern to the federal courts is a separate question—one that was not addressed by the Working Group, but may be worth considering. The narrow inquiry here is whether a sexual misconduct complaint filed against a judge, later found to be unsubstantiated, has negative spillover effects when the judge presides over an unrelated claim of gender equality. If such effects do exist, one might want to know whether their scope and nature differ in significant ways from those of unsubstantiated complaints based on different forms of conduct. One also would be interested to know the tilt or direction of the bias; conceivably, wrongful accusations could make a judge more sympathetic—and not less sympathetic—to gender claims. 44

In the context of judicial misconduct complaints, a number of conventional cognitive biases could generate the unintended result of negative implicit gender bias among those judges who have been subject to unsubstantiated sexual harassment claims. For example, the availability heuristic is a mental shortcut in which information that is more easily recalled is considered more common or relevant. One study showed that the biased recall of more violent crimes, which are more frequently publicized, leads to the assumption that harsher sentences should be implemented.45 In addition, persons wrongfully accused of illicit conduct likely are changed by that experience, and, feeling wronged, may bring a perspective of skepticism when faced with comparable charges against others.46 Relatedly, the heuristic known as confirmation bias raises the question: whether judges who have gone through an investigative and disciplinary process may tend to accept evidence that conforms to their experience, while discounting contrary evidence.47 A judge’s experience could produce the illusion that complainants are to be treated as unreliable, or that a stricter standard of evidence ought to be imposed on evidence that validates the story of an accuser in a case that is now bending before the judge.48 Unsurprisingly, false or unsubstantiated accusations against laypersons have been linked with emotion bias (including anger and skepticism).49 Depending on when a disciplinary action is filed within the span of a judge’s career, the judge could be subject to a path-dependent or primary-effect bias in which an early event generates responses that reinforce later decisions.50 Finally, judges who have gone through the complaint process may believe that the experience impairs the impartiality of others who have faced claims of sexual harassment, but that they are immune, a situation referred to as overconfidence or “Bias Blind Spot.”51

The Model Code of Judicial Conduct states “[a] judge shall uphold and promote the independence, integrity, and impartiality of the judiciary.”52 A distinct set of negative effects may arise when a judge is unaware of a bias that unconsciously, but negatively, affects impartiality. The litigants may be denied a fair process for their dispute; the public may suffer an erosion of trust in an important democratic institution; and judges may experience a decline in collegiality and reputation.53 An institution’s self-awareness of the existence of bias is an important step to determining whether improvement is needed. As the President-elect of the American Bar Association warned in August 2018: “[T]he challenges in both the civil and criminal justice arenas … are greatly exacerbated—if not partly caused—by the hesitancy of many in the legal profession and the business community to acknowledge and seek correction of the shortcomings of the justice system,” including those that are “rooted in a lack of cultural and language awareness, racism or sexism, bias, [or] power inequities.”54

The Working Group’s assignment was not to investigate whether sexual misconduct complaints against judges have spillover effects on court decision making; nor was its focus whether sustained workplace complaints against a specific judge carry an unacceptably high risk to impartial adjudication that presumptively should disqualify the judge from hearing cases within particular legal categories. However, analogous questions have been raised in other contexts affecting the judiciary. For example, a great deal of law is designed to control jury decision making, and the federal judiciary recognizes that a potential juror’s experiences and attitudes may signal biases that could impair the appearance of or actual impartiality of lay decision making. For this reason, courts have held valid preemptory challenges to remove a juror who has been falsely accused of a crime.55 By contrast, trust in the judge’s ability to exercise sound judgment is the foundation of the legal system and its promise of fairness. To be sure, it may be that norms of judicial self-governance are sufficiently robust to overcome impermissible biases. However, unfounded judicial misconduct claims may pose novel challenges to the sufficiency of judicial self-governance norms. Indeed, paradoxically, in this context, judicial norms may be an obstacle to the self-reflection associated with de-biasing strategies.

Self-governance norms operate in different ways depending on court hierarchy. Consider the norm of collegiality. Collegiality typically refers to the process of group decision making among a multi-member court. A definition of collegiality—and its stated rationale—by an influential federal appellate judge is that of “a process that helps to create the conditions for principled agreement, by allowing all points of view to be aired and considered.”56 Collegiality, similar to other group decision making, “forces individuals to discuss their reasoning in a way that facilitates de-biasing.”57 Yet, some commentators express concern that the norm of collegiality produces conformity, self-censorship, and deference to consensus as a means of horse-trading.58 In the context of #MeToo, the norm potentially could incentivize members of the bench to “look away” if the attitude or conduct, although inappropriate, is considered tolerable under the circumstances.59 Indeed, the homogeneity of members of a multi-panel court could tend toward “group-think,” and encourage an echo effect in which unacceptable behaviors are recast as appropriate to the role.60 What may make a difference to the gender de-biasing of a collegial court is the presence of a woman on the bench,61 but this is a long-term reform not likely to be available in the near future.  The norm of collegiality operates differently in first-instance courts. District court judges preside over and decide cases alone, and so the sharing of ideas operates outside the space of the decisional courtroom. It is generally assumed that the norm of collegiality has less efficacy when individuals work alone, but whether this is the case for judges is a question to consider.62

Another norm of judicial self-governance is accountability, which is related to respect for the rule of law and the appropriate roles of the three branches of government.63 District court judges are subject to a strong norm of accountability in the sense that their opinions are subject to review by a higher court that ensures decisions are within precedential boundaries.64 An important part of accountability is the writing of decisions and the giving of reasons, both of which allow the district court to engage in a process of self-evaluation and enable the appeals court to provide meaningful oversight. 65 The robustness of the accountability norm may be questioned, given trends in appellate decision making, including: the increase in resolutions without opinions with precedential value;66 the role of judicial clerks in appellate decision writing;67 and the significance of discretion in many procedural rulings.68 The problems have been fully ventilated by former Judge Nancy Gertner, who has explained the ways in which caseload pressures and case management practices encourage district court judges to place excessive reliance on heuristics that tend to reinforce a bias against civil rights plaintiffs in decisions that often evade meaningful review.69 Most telling, she noted that her colleagues saw no problem in the decisional results.70

The judiciary’s internal examination of these possibilities could be helpful in identifying whether a problem exists. If it does, different solutions could reasonably be considered. In particular, the Working Group has recommended enhanced training programs that will help judges be alert to bias, and presumably be better equipped to regulate inappropriate emotion and to develop de-biasing strategies as a preventive approach.71 Excellent programs already exist and provide a foundation for further institutional reform.72 Moreover, although consideration of these issues implicates important confidentiality interests,73 the investigation of trends can be undertaken without disclosing the name of the subject judge or of the complainant.

#MeToo has raised important questions for American society about the prevalence of sexual harassment in the workplace. The Chief Justice exercised leadership in calling for the judiciary to address employment problems identified within the courthouse. But the primary role of the courthouse is to provide justice. The literature on behavioral psychology invites further consideration regarding whether ordinary cognitive biases may in some settings produce forms of implicit gender bias that mar the impartiality of a judge’s decisional process. The narrow (and so-far infrequent) situation of unsubstantiated sexual misconduct claims against a judge opens a window onto this broader question. To be sure, some might argue that even surfacing the question puts the courts on a slippery slope that not only threatens the privacy interests of the complainant and subject judge, but also jeopardizes judicial independence by inviting a host of baseless claims. Both concerns have merit, but surely the judiciary has institutional guardrails sufficient to prevent these undesirable ends. Through tough self-examination, the Working Group has moved the federal courts closer to the goal of “an exemplary workplace for every judge and every court employee.”74 Further self-examination might now be in order to move closer to the ideal of equal justice for all.