Gender and Consensus on the Courts
During group deliberation, women tend to be the harbingers of compromise, vying for consensus on even the most contentious topics. However, women cannot succeed in this endeavor if gender bias prevents their voices from being heard. In federal courts, collaboration is essential to reaching the simple majority needed to resolve a case. While the courts already function under norms of collegiality during deliberations, the practice of writing a separate opinion to acknowledge disagreement with a majority opinion remains relatively common. In this Article, I test two competing theories as to the effects of women judges on consensus. On one hand, I posit that women’s tendency toward democratized decision-making should lead to greater consensus. If women who are judges are more likely to compromise with their peers, it seems likely that case deliberation will be more consensual when women are part of the process. Alternatively, the possibility of gender bias might lead to greater levels of dissensus when a panel includes both men and women. If men are skeptical of positions taken by women and/or women perceive their voices as not being heard by men, dissensus may abound when women join panels. Using both empirical and qualitative evidence, I find support for both theories. Specifically, employing a dataset of more than 11,000 federal appellate opinions and 27,000 judge-case pairings, I find that the presence of women judges generally leads to greater dissensus on the courts. However, the data simultaneously indicate that all-woman panels reach the highest levels of consensus. A systematic analysis of the recorded oral histories of 29 federal appellate judges confirms that, while women may be more likely to pursue consensus building, they will sometimes fail, possibly due to perceptions of bias against them. However, this gendered dissensus may also simply be a byproduct of diverse deliberation that leads to more well-reasoned decision-making, a potentially desirable practice for the court.
In recent years, the gender composition of the federal courts has changed drastically. Of the 228 federal judges appointed by former President Biden, 63% are women, far eclipsing the appointment rates of any other President.[1] This has led to an increase in the total percentage of women on the federal bench by more than 6% in less than four years.[2] This recent shift in the gender composition of the court begs questions about its effect on the behavior of women judges. This Article analyzes such behavior, seeking to understand whether and how women on the federal bench impact group deliberation and consensus on the court.
Specifically, I theorize that women judges will have a higher tendency toward collaboration as compared to peer men. I predict that this will lead to differences in panel consensus depending on the gender of the judge authoring a majority opinion and the gender composition of that panel.
I test my hypotheses on the federal appellate courts—the bodies through which the majority of legal policy is set—since the Supreme Court hears just 1% of all cases appealed to it each year.[3] The American appellate courts are, at their most basic, deliberative institutions. Panels of judges must come to some sort of agreement to process a given case. This is a normatively good thing; judges are more likely to attain the “correct” decision when they work together than they would individually.[4] Further, political theory has generally advocated for more deliberation in democratic decision-making.[5] While scholars have dedicated much thought to deliberation on the courts,[6] few have attempted to discern the effects of gender on judicial deliberation.[7]
Additionally, while women on the court have been studied in their own right, their impact on deliberation is rarely a central inquiry. In fact, the vast majority of research in this area focuses solely on one dimension of judging: case outcomes.[8] If and when scholars endeavor to analyze gender effects at any earlier stages in the judicial process, they usually focus on interruption behavior at oral argument on the Supreme Court.[9] This leaves a clear gap in the literature. An early survey of women on the bench found that women judges have different experiences, attitudes, and perspectives as compared to men.[10] Thus, it is reasonable to assume that gender may also influence individual judicial behavior. By studying the judicial process rather than the raw outcomes, I propose that we can gain a deeper understanding of the reasons behind these outcomes.
In the context of group decision-making more generally, women have been found to enhance the deliberating environment and heighten consensus.[11] However, women may sometimes be hindered from effectively participating in a deliberation due to the behavior of men.[12] While the courts already function under norms of collegiality, it is unclear whether women judges will enhance or impede consensus, given what we know from the research on other groups. I endeavor to disentangle this relationship. Through my analysis, I make contributions to the bodies of literature on gender and politics, judicial politics, and deliberative democracy more generally.
Part I of this Article provides an overview of the relevant literature on collegial courts and the impact of gender on deliberation and consensus. I begin by describing why the federal appellate courts function under norms of collegiality and the circumstances under which judges might defy such norms. Then, I review the contributions women make to group decision-making. Both as leaders and in the background of group discussions, women have the potential to enhance unanimity. However, this may depend on the gender composition of a group, and there are competing theories about how women will behave when they are heavily outnumbered by men in a given group. I thus provide an overview of the literature that suggests—possibly due to subconscious gender bias—that including women in a group of judges may lead to greater conflict on the court. Finally, I integrate these theories on general deliberation with relevant studies on women in the courts.
In Part II, I lay out my key expectations. Because there is some tension in the literature, I present competing theories. Under my first set of hypotheses, women will enhance consensus on the courts: compromising to ensure the majority opinions they author are unanimous and preserving collegiality norms by choosing not to author concurrences and dissents. Conversely, I propose that women might exacerbate dissensus on the court. Because men on the court may harbor some subconscious bias against women, they may be more likely to author separate opinions when a woman is in charge of the majority opinion, leading to lower levels of unanimity. Further, if women feel they are not being heard by a majority author, they may actually be more likely to author separate opinions themselves. Either way, these findings will likely be conditional upon the gender composition of a panel. While each proposed theory provides different implications for a single “token” woman on a panel, both would posit that all-woman panels should lead to the greatest levels of consensus. These expectations are derived from a holistic consideration of the literature on gender, deliberation, and the courts.
Part III describes my empirical data collection methods and key variables. The data include nearly 9,200 majority opinions and nearly 1,900 non-majority opinions in fourteen issue areas over the past 24 years. I test my hypotheses on these data in Part IV. Ultimately, I find more evidence for my second set of hypotheses. That is, women seem less likely than men to bring about consensus unless they are surrounded by other women. Specifically, judges are seemingly more likely to author separate opinions, and these separate opinions tend to be harsher (e.g., in the form of a dissent rather than a concurrence) when a majority author is a woman compared to when a majority author is a man. Furthermore, when they are not charged with authoring the majority, women tend to be more likely to opt into writing a separate opinion than men. The exception is all-woman panels, which exhibit higher levels of unanimity than both gender-heterogeneous and all-male panels.
In Part V, I draw upon qualitative evidence to clarify the possible mechanism behind my findings. Examining a number of oral histories of women and men on federal appellate courts elucidates three themes related to gendered dissensus. First, women seem to approach judging more collaboratively than do men, even if the empirical evidence indicates that this method often fails. Second, women describe heightened gender consciousness and experience with bias on the basis of sex that might reduce their ability to successfully deliberate with men. Third, while many women discussed their experiences with gender bias, nearly as many women recounted moments of allyship by men in the legal profession. This might indicate that the implicit bias experienced during judicial deliberation is often subtle rather than overtly hostile.
This Article concludes with a discussion of possible implications in Part VI. Namely, there may be some concern that women judges are expending more time and resources on both majority and separate opinions when there is dissensus. However, perhaps this additional work leads to more robust judicial decisions. In other contexts, diversifying a decision-making apparatus has been correlated with enhanced reasoning.[13] If conflict on the court leads to more well-considered opinions, perhaps we should strive for greater dissensus. As long as this does not fracture judicial opinions to the extent that public perceptions of judicial legitimacy are threatened, collegial conflict may be desirable.
While substantial work in the realm of judicial politics has endeavored to discern the effect of women judges on case outcomes, there is a relative dearth of research analyzing women’s impact on the process of judicial decision-making. This is somewhat surprising given the large corpus on both how deliberation functions in the courts and how women affect group deliberation more generally. I begin with an overview of the federal appellate courts, summarizing the findings on how judges work together to come to a decision. I especially focus on how custom encourages collaboration on the courts. Next, I provide a synopsis of the literature on gender and group decision-making. I consider both how women differ from men as team leaders and how they function in the background of deliberations. I also analyze how the behavior of women varies as the gender composition of a group changes. Finally, I outline research on gender in the judiciary as it relates to this project.
The federal appellate courts are a fundamentally deliberative institution. That is, judges sitting on a panel must make a group decision as to the outcome in a given case.[14] Leading up to this decision, panels have plenty of opportunities for interaction. For one, judges can communicate their preferences to each other through questions during oral argument.[15] Privately, after oral argument, judges generally attend a case conference to cast tentative votes on the holding and discuss individual reasoning.[16] It is at this initial conference that the presiding judge—either the chief or seniormost judge—assigns out majority authorship.[17] The assigned author will then circulate draft opinions for comments by other judges on the panel.[18] Judges who feel that their comments have not been taken into consideration by the majority author may consider writing a discretionary concurrence—or even a partial dissent—to have their voices heard.[19] Eventually, the opinion(s) are sent out into the world, often as published in the corresponding Federal Reporter.
Most scholars agree that the majority author has the most influence over the content of the ultimate opinion, even though disposing of the case is not an individual endeavor.[20] Thus, the choice as to which judge will author the majority is a critical one. This decision is not necessarily random. One study on opinion assignment found that appellate judges with certain subject-matter expertise were more likely to be assigned to author relevant appeals.[21] In state courts of last resort, women may be more likely to be assigned authorship, especially if a case deals with a gendered issue.[22] This latter point echoes a study on the Courts of Appeals, which finds that women and more liberal judges are more likely to be assigned authorship of sexual harassment cases.[23] However, another study of the federal appellate courts finds that opinion assignment is more so a function of institutional factors, such as the workload of judges on a given panel.[24]
Even if a given author has disproportionate influence on the content of a majority opinion, this does not mean that no other judge will be involved in the process or that the voices of the other judges on the panel do not matter. Because opinions require majority support, the federal appellate courts are generally described as operating under norms of collegiality and consensus. That is, judges are expected to compromise on their individual policy preferences in a case to accord with their fellows on a panel.[25] It tracks that judges would abide by such customs. The judiciary technically has no enforcement mechanism for its decisions and thus relies on notions of institutional legitimacy to ensure any sort of policy impact.[26] Signals of divisiveness on the court through the issuance of separate opinions have the potential to threaten public perceptions of legitimacy.[27] However, that has not prevented the publication of such opinions. In fact, concurrences and dissents are a relatively common practice on the courts today, though that isn’t to say that unanimous opinions are a notion of the past.[28]
The persistence of unanimity is indicative of the collaboration among panel judges in the content of a majority opinion. Studies have found that majority authors craft their opinions to capture the median member of the court, or at the very least the median member of the majority coalition.[29] This tendency need not be strategic on the part of the majority author. Some scholars propose that judges may be susceptible to conformity bias that leads to a deferral to the group preference.[30] This could impact the psychology of either the majority author or others on the bench. Or, perhaps judges are genuinely open to a new point of view expressed by their peers and thus experience authentic shifts in judgement.[31] Regardless of motivation, empirical evidence has found that judges do in fact adapt their partialities as the composition of the court varies, indicating that consensus is important to the court.[32] Furthermore, past cooperation with a judge in a majority coalition has been found to discourage future dissent from that judge when she goes on to author a majority.[33]
Nonetheless, separate opinions do remain pervasive on the court. Many scholars have undertaken efforts to understand when and why judges decide to author separate opinions. After all, this is a wholly discretionary decision that requires substantial time and effort.[34] In addition to imposing costs on the individual author and possibly threatening public perceptions of the court, authorship of separate opinions imposes a tangible cost on majority authors, who must allocate more resources to respond to the views of their dissenters in an opinion. One study on the Courts of Appeals found that the presence of a dissent was correlated with a 20% increase in the length of the corresponding majority opinion.[35]
Perhaps there are case-level factors that could explain this correlation. For example, empirical evidence has shown that separate opinions are more likely to be authored as a case becomes more complex and that concurrences (but not dissents) are more likely to be authored as a case becomes more politically and/or legally salient.[36] There may also be judge-level factors that impact consensus on a court. On the Supreme Court, judges tend to be more likely to author separate opinions the more ideologically distant they are from the majority author.[37] At the level of state courts of last resort, one study found that individual judge characteristics influence consensus on a case. Specifically, as a court becomes more homogenous on social, legal, and political traits, separate opinions are less likely to be authored.[38] Judges may also author separate opinions to simply guard their reputation in some way.[39] For example, Justice Neil Gorsuch tends to champion Tribal sovereignty.[40] Thus, it would have been surprising if Justice Gorsuch had failed to author a dissent when the Court recently found that the United States did not have a duty to secure water for the Navajo Reservation under its Treaty with the Nation.[41]
Of course, even if these factors contribute to the authorship of a separate opinion, such authorship is predicated on something simpler: two judges disagreeing in some way. Dissents tend to express heightened disagreement with the majority, often arguing that the case should have come down in the opposite direction or been otherwise dismissed for procedural reasons. One set of scholars has found that majorities and dissents on a case cite to completely different legal precedent.[42] This suggests that when judges disagree over an outcome, it is not due to a disagreement over the interpretation of a specific test, but rather over which cases and tests should be applied in the legal analysis. In the case of a dissent in part, the judge may wholly agree with one aspect of the majority while differing on another aspect.
Concurrences express lesser degrees of dissensus. A concurring judge agrees with the holding of a case but may disagree with the reasoning used to attain that holding (usually deemed a “concurrence in judgment”). Or, the judge may agree with the reasoning but still have something additional to say, perhaps about the caselaw that the reasoning is predicated upon. In one proposed typology, a concurrence might be: (1) groundlaying; (2) signaling; (3) preserving; (4) defending; or (5) weakening.[43] Groundlaying concurrences are those that disagree with the test used in a case and propose an alternate rule for such cases.[44] Signaling concurrences acknowledge a tangential legal issue or precedent that wasn’t properly raised in the instant case but ought to be revisited in the near future.[45] Preserving opinions have little value to future opinions and are authored simply for a judge to preserve their current thoughts.[46] Defending concurrences are those that champion a certain jurisprudence more broadly.[47] Finally, a weakening opinion narrows the scope or authority of the majority.[48] These latter types of concurrences are most likely to be designated as concurrences “in part.”
There is still substantial research to be done on the authorship of separate opinions. For example, it isn’t wholly clear what the broader value is of concurrences and dissents. While some scholars have posited that separate opinions might impact future adjudications and legislative activity, others have argued that these speculated effects have never been empirically validated.[49] Nonetheless, it is worth recalling here that separate opinions are authored in only a minority of cases. In the 2023–2024 term, even the Supreme Court—widely reported to be highly polarized—issued unanimous rulings in nearly 75% of its cases.[50] Unanimity is likely higher on the Courts of Appeals, where the small number of judges (namely, three) on a typical panel heavily encourages conformity.[51]
B. Gender and Consensual Leadership
Norms of consensus may be even stronger when a woman is at the helm of a group deliberation. A vast literature exists in the business realm typifying management styles and attributing some such styles as more “feminine” or “masculine.”[52] For example, women may be perceived as “connective” leaders—i.e., those focused on their interpersonal relationships within the workplace.[53] Gender role theory casts women as empathetic and community-focused, and theories of management expect these traits to persist in women leaders.[54] Since connective leaders tend to focus on collective rather than individual success, they prioritize fostering communication and trust within their subordinates and consider the interactions among networks within their organizations.[55] Such leaders also tend to engage in social exchanges with their colleagues to bolster working relationships.[56] In addition to focusing on relationships, connective leaders tend to prioritize collaboration.[57] The primary way that women and men managers have been found to differ is that women tend to be more democratic in their leadership styles.[58] That is, women are more likely to reach out for input from others, listen to the views of subordinates, and seek consensus in decision-making.[59]
Traditional notions of leaders cast them as agentic—i.e., aggressive, ambitious, independent, and competitive—rather than connective.[60] But communal notions of leadership may be more desirable in modern management. Because the behavior of followers is integral to the success of leaders, contemporary theories of leadership call for integrating subordinates into certain management decisions.[61] Connective leadership has been found to correlate with higher levels of organizational effectiveness as compared to other management styles.[62] This may be because collaborative decision-making often leads to more well-considered outcomes.[63] Furthermore, workers may experience higher levels of job satisfaction under connective leaders.[64] Likewise, laypersons tend to substantially prefer working with managers who are more relational.[65] Feminine styles of leadership may also be most preferable in moments of crisis, where collaboration, communication, and trust are most imperative.[66]
Given that workers prefer connective leaders and that women in management roles tend to be more connective, we expect employees to prefer reporting to women superiors. But research finds that women in leadership face a double bind.[67] That is, if women act within their gender role by leading relationally and collaboratively, they face criticism for their soft leadership style.[68] However, if women use agentic leadership strategies—in conformity with typical expectations for leaders—they are criticized for acting outside of their gender role.[69] This “think manager, think male” ideology may persist today. For instance, a recent study found that, although workers substantially prefer participatory leadership, they still overwhelmingly favor men as managers.[70] However, this may simply be because men still tend to dominate leadership positions.[71]
C. Gender and Group Deliberation
Women clearly impact group decisions through their leadership styles, but this isn’t the only way that gender affects group decisions. Specifically, the gender composition of a group may impact both considered outcomes and the deliberative environment more generally. Further, the presence of a woman in a group ensures that women’s needs and interests are considered during the deliberation.[72] Women have been found to have a preference for more altruistic and universalistic solutions as compared to men, who tend to be more self-interested and competitive.[73] Importantly, women can influence men to be more universalistic, with majority-women groups more likely to prefer equitable solutions over solutions that would maximize self-gain.[74] Conversely, male-dominated teams come to decisions that are not only less equitable but also less accurate.[75] Not only are groups with more women likely to come to more accurate solutions, but such groups also tend to display higher levels of collective intelligence post-deliberation.[76]
This may be because women are more socially sensitive and allow for more equality in conversational turn-taking.[77] Women in group settings are more likely to express agreement with other speakers, ask for others’ opinions, and pause to give the floor to others as compared to men.[78] Greater proportions of women in a group have also been correlated with fewer interruptions and a greater ability to stay on-topic.[79] Interestingly, there are even gendered differences in the way group members use humor, with women more likely to use jokes to build cohesion.[80] Perhaps because of these effects on the deliberation space, groups with more women have been found to come to decisions that are more likely to be unanimous.[81] Or, such decisions might be more likely to be unanimous because women tend to facilitate compromise during joint decisions.[82]
However, women may have difficulty positively impacting deliberative environments if they are unable to be heard.[83] Traditionally, social role theory casts women as listeners rather than speakers.[84] It may be because of this that men tend to dominate turn-taking in group discussions.[85] Research shows that, although women express a greater willingness to deliberate, they are often undermined by the negative deliberation behavior—i.e., interruptions—of men.[86] Even when a group is filled with mostly women, men dominate conversation, though changing the decision rule in a group may encourage women to speak more.[87] Thus, when gender-homogenized groups of women do not have to contend with the domineering behavior of men, turn-taking in discussions is much more equitable.[88]
Conversely, as already established, groups with only a few “token” women may suffer from the domineering behavior of men in a way that further limits the participation of these women.[89] Tokenization can have broader consequences outside of group deliberation. When a woman is tokenized in a workspace, she is more visible and thus subject to certain performance pressures.[90] This creates a psychological incentive for her to either blend into the background as much as possible or embrace the spotlight and focus on overachieving.[91] Of course, the token women who choose to over-achieve are likely already exceptional at their work.[92] For example, women who are elected to legislative office tend to be of a higher political quality as compared to their male peers.[93] It thus makes sense that token women in state legislatures tend to be at least as successful, if not more successful than men at passing bills.[94] So, while women tend to fare better in deliberation when they comprise a majority in the group, it may be possible for women to succeed while tokenized, contingent on individual-level factors.
Of course, the success of women in masculinized fields, especially in the context of court consensus, will also depend on the behavior of men. To that extent, it’s worth noting that there may be a systemic gender bias that leads to dissensus when a woman joins a panel of judges. While much of the judicial process is somewhat of a black box, survey data shows that women on the court perceive bias against them from male colleagues.[95] This perception has been somewhat validated at certain stages of the judicial process. In the venue of state courts, women have also been found to be disproportionately assigned writing assignments that are “less important.”[96] Gender bias is also on display at the court’s most public occasion for judging—oral argument. Specifically, empirical evidence demonstrates that male Supreme Court Justices (along with male attorneys) interrupt women Justices at a higher rate than they do other men.[97] This latter finding is significant to the instant study since Justices are more likely to vote against other Justices that they interrupt at high rates.[98]
More generally, women judges may confront obstacles simply because they are members of an outgroup in the masculinized occupation of judging.[99] Psychological evidence suggests that women in masculinized careers may be seen as “unfit” workers by peer males.[100] Women in such careers who conform to more masculinized or agentic work styles may be punished by peer males for breaking gender norms, creating a so-called double bind.[101] Recent initiatives by former presidents Obama and Biden have greatly increased the percentage of women serving on federal courts.[102] However, this may have only exacerbated hostility toward women judges. When women are seen as benefiting from diversity initiatives, men may view them as less competent.[103] Additionally, men may be resistant to the integration of more women into masculinized careers, possibly because they feel this threatens their status in the workplace hierarchy.[104]
Outside of panel dynamics, the behavior of women judges may also be influenced by a need to prove themselves to non-court actors. Importantly, women on the court appear to experience bias from political actors and the public. For example, male senators are more likely to “grill” women nominees on their judicial philosophies and areas of perceived expertise during confirmation hearings.[105] Senators also interrupt women nominees at higher rates than they do men.[106] Similarly, male attorneys in front of the Supreme Court are more likely to interrupt female Justices than male ones during oral argument.[107] This bias among lawyers is also reflected in the multiple studies demonstrating that ABA ratings of judges tend to be biased against women.[108] Furthermore, surveys have found that a large plurality of Americans believe women judges are biased in favor of women.[109]
The public likely need not be so concerned about such bias. While research on gender and judging is still burgeoning, several scholars have already endeavored to discern whether there are gendered effects on case outcomes.[110] Generally, the findings point to no.[111] This finding may persist because both men and women on the federal courts have gone through similar legal socialization processes—i.e., education at the same top 14 law schools—and face identical constraints on the bench that ought to lead them to the same legal conclusions on a case.[112] This tracks with legal theories of judicial decision-making, which hold law to be the primary determinant of case outcomes and thus would predict that judicial decisions ought not differ based on factors such as a judge’s gender.[113] If case outcomes are dependent solely on the law, any judge ought to come to the same conclusion in a given case.[114] And empirical evidence has shown that judges are at least somewhat constrained by the law when reaching the disposition of a case.[115] For example, judges have been found to defer to agency actions that they might otherwise ideologically oppose.[116]
However, even if women and men on the bench tend to come to the same legal conclusions, that does not mean they function identically during the decision-making process that culminates in these decisions. Motivated by the finding that women leaders tend to be more collaborative, one researcher tested and found evidence that women district court judges foster settlement more often and more frequently.[117] On state high courts, women judges have been found to exert more independence by authoring separate opinions even when the majority author is a co-ideologue.[118] As majority authors, women on state high courts have been found to foster greater unanimity in certain cases.[119] On the Courts of Appeals, women appear to author opinions that are longer and include both more citations and “deeper” citations as compared to opinions authored by men.[120]
Furthermore, scholars have found that—in the rare cases where women’s views differ from those of men—a woman may be able to persuade her peers to take her view of the case. Specifically, for cases involving sexual harassment suits, the presence of a woman on a panel increases the chance that the case will be disposed of in favor of the plaintiff.[121] Additionally, court outcomes may change as more women attain judgeships. One study finds that, as women reach a critical mass on a court, they tend to be more likely to rule in favor of criminal defendants.[122] Another finds that increasing the proportion of women on a bench is correlated with a decreased sex disparity in criminal sentencing.[123] Taken together, this research indicates that women do seem to have an impact on deliberation in the courts.
I test competing theories as to the impact of women on panel consensus. Because women have a tendency toward democratized leadership and collaborative work styles, one possibility is that women will facilitate consensual decisions on the court, ensuring heightened unanimity for panels including them. Conversely, implicit gender bias among men on the court might lead to greater dissensus when a woman is present on a panel. However, either way, panels with all women ought to lead to the highest levels of unanimity. Additionally, consensus on the courts may be contingent on other factors, such as the gender composition of a panel and the issue area of the case. As judges deliberate over a case’s outcome, both the gender of the majority author and the number of women on that panel may matter. Thus, the literature on gender, deliberation, and courts all coalesce to engender a series of both panel- and judge-level hypotheses.
A. Women Judges as the Harbingers of Compromise
My first set of hypotheses is predicated on the assumption that women will have a positive effect on group deliberations. Initially,
I consider the impact of majority-author gender on consensus. In the context of judicial opinions, majority authors field several iterations of draft opinions to their fellow panel members for comments. The literature on leadership tells us that women tend to be more democratic in their decision-making.[124] If women who author majorities hold democratized views of leadership, they should be more likely to incorporate comments from their colleagues into a finalized opinion as compared to men. This prediction is bolstered by the general literature indicating that women tend to be more compromising than men.[125] With their comments taken into consideration, panel members will likely be deterred from authoring separate opinions to express their views—or at least mitigate the strength of their dissensus in any separate opinion authored. For example, a judge planning to author a full concurrence upon circulation of a draft majority might ultimately decide to enter a concurrence in part if the text of the majority is subsequently modified to placate that judge. This prediction can be succinctly stated thus:
H1A: Panels will come to greater consensus on a case when the majority is authored by a woman.
However, consensus may also depend on the other judges on a panel.[126] Thus, I consider the behavior of women on a panel who are not tasked with authoring a majority opinion.[127] Again, I rely on the literature that finds women tend toward greater collaboration and compromise.[128] If this behavior extends to women judges, then they should be less likely to author separate opinions, even if they feel that their views have not been taken into account by the author of the majority opinion. Furthermore, collegial norms that disincentivize separate opinion authorship will likely be heightened for women, especially when tokenized. Thus:
H1B: Women will be less likely than men to write separate opinions.
Taken together, this first set of predictions hypothesizes that both woman-majority-authorship and the presence of women on a panel will lead to greater consensus in case outcomes.
Conversely, it could be the case that women judges may contribute to dissensus on the courts. Again, I start with the case of a majority opinion authored by a woman. Recall that there is evidence that men on the court tend to disproportionately interrupt peer women,[129] and that interruption behavior leads to conflict that is reflected in the divisions of case outcomes.[130] This alone is enough to predict that majority opinions authored by women might be less unanimous than those authored by men. However, other factors might exacerbate this effect. Namely, judging is still a heavily masculinized career. If men on the court feel that women are not a correct fit for their position—especially the leadership position of authoring a majority—they may be less likely to value that woman’s opinion. Additionally, because men are accustomed to playing the societal role of speaker rather than listener, it may feel natural for them to author a concurrence—thereby speaking up—to more appropriately represent their point of view even if they agree with the majority opinion. More insidiously, perhaps implicit gender bias influences men on the court to harbor some distrust of the work product of women judges, leading such men to violate norms of collegiality to author separate opinions when a woman is tasked with majority authorship. In any case, I predict that:
H2A: Panels will come to lesser consensus on a case when the majority is authored by a woman.
Further, even if women judges may be reluctant to violate collegial norms to author discretionary opinions, they may choose to do so if it is the only way for their voices to be heard. Although opinion writing is generally a collaborative task, the fact remains that the majority author has the final say in the ultimate text of that opinion. Thus, it is wholly within the discretion of that author to integrate (or not integrate) the comments of their peers on the bench. If implicit gender bias functions such that men authoring a majority discount the comments of peer women, women will need to publish a separate opinion if they want their thoughts on the case to be published. Thus:
H2B: Women will be more likely than men to write separate opinions.
In this case, both woman-authorship and the mere presence of women on a court will seemingly lead to greater panel dissensus.
Finally, I anticipate that the highest levels of consensus will be observed when a panel comprises all women. This follows from either theory, regardless of whether women enhance or impede consensus on the courts. If women are more democratizing as majority authors and more compromising as panel members, there should be very little reason for an all-woman panel to author separate opinions. Recall that groups with a majority of women tend to reach more unanimous decisions in non-judicial contexts.[131] On the other hand, if men experience gender bias that leads to dissensus on mixed-gender panels, this dissensus should be assuaged by the elimination of any men on the panel. Even if women judges are not more collaborative than men under this latter theory, panel women ought to be less likely to author separate opinions since they are less likely to feel voiceless qua women when a peer woman authors the majority.[132] Thus, in either case:
H3: All-woman panels will exhibit higher levels of consensus compared to gender-heterogeneous panels.
The dataset includes information from 11,022 federal appellate opinions spanning various issue areas from 2000 to 2024. Data was gathered using Westlaw to search all published cases in each specific issue area over the specified time frame.[133] I begin in 2000 as this is the year in which women comprised 20% of all judges in the federal appellate courts. The data include 9,167 majority/plurality opinions and 1,855 separate opinions. Of the separate opinions, about 35%are concurrences, about 41% are dissents, and about 25% are concurrences in part and dissents in part. These separate opinions occur in 1,721 cases, including 114 cases with three total opinions and ten cases with four total opinions (i.e., where the majority author writes a separate opinion along with both other members of the panel). There are also nine cases in the dataset where a judge flags their disagreement with the majority via a footnote but does not author a separate opinion.
To determine the gender of each judge in the dataset, I used a series of regular expressions[134] to capture the names of all judges on a panel for a given case as well as the majority opinion author. For those cases with separate opinions authored, I manually checked each case for the author and the type of the separate opinion. Once I had judge names, I matched these to unique judicial identifier codes from the Federal Judicial Center (“FJC”). The FJC publishes an up-to-date dataset of demographic information for all judges. Thus, once the unique identifier was coded into the dataset, I merged in demographic data—including gender—for each judge on the panel. About 24 percent of opinions in the dataset are authored by women, and about 24 percent of panel judges in the dataset are women.
Using the FJC data, I also calculate the number of women on each panel. Because my scope is federal appellate courts (and I exclude en banc decisions), nearly all of the cases in the dataset were decided by a panel
of three judges.[135] The majority of panels include either no women (about 42 percent) or one woman (about 44 percent); about 12 percent of panels include two women and about 1 percent of panels are composed of
all women.
I use a few different measures of consensus to test my panel-level hypotheses (H1A, H2A, and H3): (1) unanimity; (2) number of separate opinions; and (3) indexed level of consensus. At the individual level,
I employ binary variables that track the authorship of separate opinions (both cumulatively and by type). All variables were hand-coded. The binary unanimity variable takes a value of “1” if the majority opinion is the only opinion in the case and a “0” otherwise. Over 80% of cases in the dataset are unanimous.
For the second dependent variable, cases may have just one opinion or as many as four opinions. However, I reverse the coding for this variable to ensure interpretation is consistent with other dependent variables. Thus, “number of separate opinions” ranges from 0–3 where a “3” represents that all three judges have agreed (and thus only a single majority opinion was published) and a “0” represents that no judge could agree, creating four separate opinions on the matter. The average case in the dataset produces a value of about 2.8 for this variable, in alignment with the fact that most of the opinions in the dataset are unanimous.
The variable for indexed consensus is on a ten-point index ranging from 0–1 calculated based on both the number of separate opinions and the qualitative types of separate opinions in a case. The typology of a separate opinion gives a relative proxy of the strength of disagreement with the majority. So, for example, a case with a single, regular concurrence represents only minor disagreement among judges, whereas a case with one dissent and one concurrence in judgment represents quite a lot of disagreement. On this index, “0” represents cases with the greatest dissensus (with two separate opinions written that partially concur and partially dissent); a “1” represents cases with the highest level of consensus—i.e., unanimity. Table 1 presents all possible designations on the index with the number of cases that fit into each designation.[136]
Table 1: Consensus Index and Number of Designated Cases
| Index Level | Types of Cases | # Cases |
| 1 | unanimous majority with no separate opinions | 7436 (~81%) |
| 0.9 | majority + concurrence in part majority + concurrence in footnote | 21 (~0.2%) |
| 0.8 | majority + concurrence | 338 (~4%) |
| 0.7 | majority + concurrence in judgment majority + concurrence dubitante majority + dissent in footnote | 155 (~2%) |
| 0.6 | majority + dissent in part majority + concurrence in part and dissent in part (“partial”) | 479 (~5%) |
| 0.5 | majority + dissent majority + concurrence in part + concurrence in part | 613 (~7%) |
| 0.4 | majority + concurrence + concurrence majority + concurrence in judgement + concurrence | 13 (~0.1%) |
| 0.3 | majority + concurrence in part + partial majority + concurrence in part + dissent in part | 3 (~0.03%) |
| 0.2 | majority + concurrence in judgment + partial majority + concurrence + partial majority + concurrence + dissent in part majority + concurrence in part + dissent | 29 (~0.3%) |
| 0.1 | majority + concurrence + dissent majority + concurrence in judgment + dissent majority + partial + partial majority + dissent in part + dissent in part majority + concurrence (any) + concurrence (any) + concurrence (any) | 71 (~0.8%) |
| 0 | majority + partial + dissent majority + concurrence (any) + concurrence (any) + dissent majority + concurrence (any) + dissent in part + dissent in part | 9 (~0.1%) |
The judge-level data include 27,501 total judge-case pairings. Many judges are repeat players in this dataset, with 972 unique judges. Theoretically, every time a judge is on a panel for a case, they have the opportunity to write an opinion. They may be designated to write the majority opinion or may take it upon themselves to author a separate opinion based on how they feel about the ultimate text of the majority. To test the gendered likelihood of writing separate opinions (H1B and H2B), I create a variable simply called “writes” that takes a value of “1” if the judge writes any sort of opinion on the case and a “0” otherwise.
I also create binary variables that track whether a judge authors a concurrence, a dissent, or a combination concurrence/dissent (“partial”).[137] I exclude majority authors from the judge-level analyses, since they are highly unlikely to write a separate opinion in addition to their assigned majority.[138] With the exclusion of majority authors, the data includes 18,334 judge-case pairs.
[1]. John Gramlich, How Biden Compares with Other Recent Presidents in Appointing Federal Judges, Pew Research Center (Jan. 9, 2025), [https://perma.cc/JXQ8-8D4D].
[2]. Profile of the Legal Profession: Judges, American Bar Association(Nov. 18, 2024), [https://perma.cc/3WNK-RWEP] (reporting that 33% of federal judges were women as of August 1, 2024 compared to 27% at the end of 2020).
[3]. How the Supreme Court Decides Which Cases to Hear, The Pew Charitable Trusts (Nov. 24, 2020), [https://perma.cc/EZQ4-6MW8].
[4]. Pauline T. Kim, Deliberation and Strategy on the United States Courts of Appeals: An Empirical Exploration of Panel Effects, 157 U. Pa. L. Rev. 1319, 1321 (2009).
[5]. See Amy Gutmann & Dennis F. Thompson, Why Deliberative Democracy? 19-20 (Princeton Univ. Press 2004).
[6]. See generally Donald B. Hutt, Deliberation and Courts: The Role of the Judiciary in a Deliberative System, 64 Theoria 77 (2017).
[7]. But see Susan B. Haire, Laura P. Moyer & Shawn Treier, Diversity, Deliberation, and Judicial Opinion Writing, 1 J. L. & Cts. 303, 307–08 (2013) (considering the impact of gender and racial diversity on panels of federal appellate courts).
[8]. See, e.g., Christina L. Boyd, Lee Epstein & Andrew D. Martin, Untangling the Causal Effects of Sex on Judging, 54 Am. J. Pol. Sci. 389, 406 (2010).
[9]. Adam Feldman & Rebecca D. Gill, Power Dynamics in Supreme Court Oral Arguments:
The Relationship between Gender and Justice-to-Justice Interruptions, 40 Just. Sys. J. 173, 187 (2019); Tonja Jacobi & Dylan Schweers, Justice, Interrupted: The Effect of Gender, Ideology and Seniority at Supreme Court Oral Arguments, 103 Va. L. Rev. 1379, 1482 (2017).
[10]. Elaine Martin, Men and Women on the Bench: Vive la Difference?, 73 Judicature 204, 208 (1990).
[11]. Hristina Nikolova & Cait Lamberton, Men and the Middle: Gender Differences in Dyadic Compromise Effects, 43 J. Consum. Res. 355, 367 (2016).
[12]. Afsoun Afsahi, Gender Difference in Willingness and Capacity for Deliberation, 28 Soc. Pol. 1046, 1059–60 (2021).
[13]. See Scott Page, The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies 335 (Princeton Univ. Press 2007).
[14]. Kim, supra note 5, at 1320.
[15]. Timothy R. Johnson, Ryan C. Black & Justin Wedeking, Pardon the Interruption: An Empirical Analysis of Supreme Court Justices’ Behavior during Oral Arguments, 55 Loy. L. Rev. 331, 335-36 (2009).
[16]. Donald McG. Rose, Oral Argument in the Sixth Circuit, 19 U. Tol. L. Rev. 229, 229 (1988) (discussing this practice as it relates to the Sixth Circuit).
[17]. Timothy R. Johnson, Conference, Opinion Writing, and Oral Arguments, Oral Arguments and Decision Making on the United States Supreme Court 89 (Robert J. Spitzer ed., State Univ. of N.Y. Press 2004).
[18]. See, e.g., Richard J. Lazarus, The (Non)Finality of Supreme Court Opinions, 128 Harv. L. Rev. 540, 556 (2014) (describing the “harsh criticism” draft opinions often receive during circulation at the Supreme Court). In some state courts, draft opinions may be circulated prior to oral argument. See, e.g., Mark Hummels, Distributing Draft Decisions Before Oral Argument on Appeal: Should the Court Tip Its Tentative Hand – The Case for Dissemination, 46 Ariz. L. Rev. 317, 329–30 (2004); Daniel J. Bussel, Opinions First—Argument Afterwards, 61 UCLA L. Rev. 1194, 1196 (2014). Even in federal courts, where this is typically not the case, scholars have argued that oral argument has no deliberative value, since judges have preferences on court outcomes that are not malleable. Jeffrey A. Segal & Harold J. Spaeth, The Supreme Court and the Attitudinal Model Revisited 280 (2002); David W. Rohde & Harold J. Spaeth, Supreme Court Decision Making 153 (1976). Other scholars have pointed to examples where judges switched their votes after oral argument, indicating that their minds were open
to reasoned change during this process. See, e.g., Timothy R. Johnson & Thomas K. Pryor, Oral Arguments, in Routledge Handbook of Judicial Behavior 221,
226–27 (Robert M. Howard & Kirk A. Randazzo eds., 2017); Eve M. Ringsmuth, Amanda C. Bryan & Timothy R. Johnson, Voting Fluidity and Oral Argument on the
U.S. Supreme Court, 66 Pol. Res. Q. 429, 430 (2013).
[19]. Lazarus, supra note 19, at 556.
[20]. Chris W. Bonneau, Thomas H. Hammond, Forrest Maltzman & Paul J. Wahlback, Agenda Control, the Median Justice, and the Majority Opinion on the U.S. Supreme Court, 51 Am. J. Pol. Sci. 891, 902 (2007); Jeffrey R. Lax & Charles M. Cameron, Bargaining and Opinion Assignment on the U.S. Supreme Court, 23 J. L., Econ. & Org. 276, 296 (2007).
[21]. Jonathan Remy Nash, Expertise and Opinion Assignment on the Courts of Appeals: A Preliminary Investigation, 66 Fla. L. Rev. 1599, 1657 (2015) (finding that judges who served on the Sentencing Commission were more likely to be assigned sentencing appeals and former bankruptcy judges were more likely to be assigned bankruptcy cases).
[22]. Erin B. Kaheny, John Szmer & Robert K. Christensen, Status Characteristics and their Intersectionality: Majority Opinion Assignment in State Supreme Courts, 8 Pol., Groups, & Identities 894, 911 (2020). Note that state court judges are different from federal judges in several ways—namely, they tend to be elected rather than appointed. Thus, it should not be taken as given that any findings about state judges throughout this Article automatically apply to federal judges.
[23]. Sean Farhang, Jonathan P. Kastellec & Gregory J. Wawro, The Politics of Opinion Assignment and Authorship on the U.S. Court of Appeals: Evidence from Sexual Harassment Cases, 44 J. Legal Stud. S59, S81 (2015).
[24]. Jennifer Barnes Bowie, Donald R. Songer & John Szmer., The View from the Bench and Chambers: Examining Judicial Process and Decision Making on the U.S. Courts of Appeals 82 (Gregg Ivers & Kevin T. McGuire eds., 2014). Finding that while this effect is substantively small, it is still greater than the null effects found for strategic and ideological considerations on opinion assignment. Further, in interviews, several judges stated that institutional concerns were at the forefront of their decisions on opinion assignment. Id. at 69–71.
[25]. E.g., Paul J. Wahlbeck, James F. Spriggs II & Forrest Maltzman, The Politics of Dissents and Concurrences on the U.S. Supreme Court, 27 Am. Pol. Q. 488, 491 (1999). Some have argued that the court should operate under norms of full consensus—contending that a requirement of unanimity would ensure opinions to be both fairer and better.
See David Orentlicher, Judicial Consensus: Why the Supreme Court Should Decide Its Cases Unanimously, 54 Conn. L. Rev. 303, 344 (2022).
[26]. Dino P. Christenson & David M. Glick, Chief Justice Roberts’s Health Care Decision Disrobed: The Microfoundations of the Supreme Court’s Legitimacy, 59 Am. J. Pol. Sci. 403, 403 (2015).
[27]. See Cass R. Sunstein, Unanimity and Disagreement on the Supreme Court, 100 Cornell L. Rev. 769, 804–05 (2015). An early version of the Code of Judicial Conduct (the
“Judicial Canon of Ethics”) expressly discouraged judges from authoring separate opinions, though the 1974 replacement contained no such language. See also Wahlbeck, Spriggs & Maltzman, supra note 26, at 489.
[28]. Sunstein, supra note 28, at 784 (finding that 62% of cases in the Supreme Court’s 2013 term were unanimous). Some scholars argue that unanimity in the courts fluctuates over time depending upon the strength of consensual norms in a given period.
See Gregory A. Caldeira & Christopher J. W. Zorn, Of Time and Consensual Norms in the Supreme Court, 42 Am. J. Pol. Sci. 874, 900 (1998).
[29]. See Cliff Carrubba, Barry Friedman, Andrew D. Martin & Georg Vanberg, Who Controls the Content of Supreme Court Opinions?, 56 Am. J. Pol. Sci. 400, 408–09 (2012) (theorizing and finding that the median member of the winning coalition is what matters, in contrast to previous research positing that the median of the bench shapes majority content).
[30]. E.g., David Schkade, Cass R. Sunstein & Reid Hastie, What Happened on Deliberation Day?, 95 Calif. L. Rev. 915, 925 (2007).
[31]. See, e.g.,Harry T. Edwards, The Effects of Collegiality on Judicial Decision Making, 151 U. Pa. L. Rev. 1639, 1645 (2003).
[32]. Scott R. Meinke & Kevin M. Scott, Collegial Influence and Judicial Voting Change: The Effect of Membership Change on U.S. Supreme Court Justices, 41 L. & Soc’y Rev. 909, 931 (2007).
[33]. Wahlbeck, Spriggs & Maltzman, supra note 26, at 502.
[34]. Caldeira & Zorn, supra note 29, at 877. These costs likely cause dissent rates to decrease as caseloads increase in a given circuit. See Lee Epstein, William M. Landes & Richard A. Posner, Why (And When) Judges Dissent: A Theoretical and Empirical Analysis, 3 J. Legal Analysis 101, 131 (2010).
[35]. Epstein, Landes & Posner, supra note 35, at 104.
[36]. Wahlbeck, Spriggs & Maltzman, supra note 26, at 502-03.
[38]. See generally John W. Patterson & Gregory J. Rathjen, Background Diversity & State Supreme Court Dissent Behavior, 8 Polity 610, 619 (1976).
[39]. Epstein, Landes & Posner, supra note 35, at 102.
[40]. Michael J. Sanders, Original Understandings of Tribal Sovereignty: Native Americans in Justice Gorsuch’s Jurisprudence, Am. Indian. L. Rev. (forthcoming 2026).
[41]. Arizona v. Navajo Nation, 599 U.S. 555, 570 (2023).
[42]. Anthony Niblett & Albert H. Yoon, Judicial Disharmony: A Study of Dissent, 42 Int’l Rev. L. & Econ. 60, 70 (2015).
[43]. Charles C. Turner & Lori Beth Way, Classifying Supreme Court Concurrences: The Case of Justice Clarence Thomas, 31 J. Pol. Sci. 139, 148 (2003).
[49]. Compare Sunstein, supra note 28, at 802-03 with Steven A. Peterson, Dissent in American Courts, 43 J. Pol. 412, 432 (1981).
[50]. Fatima Yazdi, Amid Controversy, Supreme Court Shows Surprising Unanimity in Most Rulings—So Far, Capital News Service (May 16, 2024), [https://perma.cc/65Y3-XLEN]. Appendix A reports time trends in unanimity from 2000-2024 within my dataset.
[51]. Burton M. Atkins, Judicial Behavior and Tendencies Towards Conformity in a Three Member Small Group: A Case Study of Dissent Behavior on the U.S. Court of Appeals, 54 Soc. Sci. Q. 41, 43 (1973).
[52]. See, e.g., Alice H. Eagly & Steven J. Karau, Role Congruity Theory of Prejudice Toward Female Leaders, 109 Psychol. Rev. 573, 576 (2002).
[53]. Jean Lipman-Blumen, Connective Leadership: Female Leadership Styles in the 21st-Century Workplace, 35 Soc. Persp. 183, 200 (1992). Various works on this topic conceive of this form of leadership as “relational,” “participatory,” “authentic,” “democratic,” and “transformative,” among other terms. I employ “connective” as it seems to most accurately capture both the relational and participatory aspects of this leadership style.
[54]. Alice H. Eagly & Mary C. Johannesen-Schmidt, The Leadership Styles of Women and Men, 57 J. Soc. Issues 781, 783 (2001).
[55]. Mary Uhl-Bien, Relational Leadership Theory: Exploring the Social Processes of Leadership and Organizing, in 27 Leadership, Gender, & Org. 75, 87-92 (Patricia H. Werhane & Mollie Painter-Morland eds., Springer 2011).
[57]. Lipman-Blumen, supra note 54, at 184.
[58]. Alice H. Eagly & Blair T. Johnson, Gender and Leadership Style: A Meta-Analysis, 108 Psychol. Bull. 233, 242 (1990). See also Katie L. Badura, Emily Grijalva, Daniel A. Newman, Thomas Taiyi Yan & Gahyun Jeon, Gender and Leadership Emergence: A Meta-Analysis and Explanatory Model, 71 Personnel Psychol. 335, 348 (2018) (finding that a gender gap in leadership style persists nearly 30 years after the Eagly & Johnson study, though it has decreased over time).
[59]. Adrianna Kezar & Marissiko M. Wheaton, The Value of Connective Leadership: Benefiting from Women’s Approach to Leadership While Contending with Traditional Views, 21 About Campus 19, 21 (2017).
[60]. Eagly & Johannesen-Schmidt, supra note 55, at 783.
[61]. Bruce J. Avolio, Promoting More Integrative Strategies for Leadership Theory-Building, 62 Am. Psychol. 25, 26 (2007).
[62]. Dean Elmuti, Heather Jia & Henry H. Davis, Challenges Women Face in Leadership Positions and Organizational Effectiveness: An Investigation, 8 J. Leadership Educ. 167, 180 (2009).
[63]. Michael W. Kramer & David A. Crespy, Communicating Collaborative Leadership, 22 Leadership Q. 1024,1035 (2011).
[64]. Soonhee Kim, Participative Management and Job Satisfaction: Lessons for Management Leadership, 62 Pub. Admin. Rev. 231, 237 (2002).
[65]. Kenneth S. Rhee & Tracey H. Sigler, Untangling the Relationship Between Gender and Leadership, 30 Gender in Mgmt.: An Int’l J. 109, 123 (2015). In addition to preferring to work for participatory managers, participants also rated such leaders as more effective.
[66]. See, e.g., Abbie Griffith Oliver, Michael D. Pfarrer & François Neville, Grand Challenges and Female Leaders: An Exploration of Relational Leadership During the COVID-19 Pandemic, 63 Bus. & Soc’y 954, 960 (2024).
[67]. Eagly & Karau, supra note 53, at 576.
[69]. Id. at 577. See also Jun-Yeob Kim, Ning Hsu, Daniel A. Newman, P.D. Harms & Dustin Wood, Leadership Perceptions, Gender, and Dominant Personality: The Role of Normality Evaluations, 87 J. Res. Personality 1, 6 (2020) (finding that women leaders are especially penalized when they display dominant personalities). But see Anyi Ma, Shelby Rosette & Christy Zhou Koval, Reconciling Female Agentic Advantage and Disadvantage with the CADDIS Measure of Agency, 107 J. Applied Psychol. 2115, 2115 (2022) (finding that women leaders are punished for possessing some, but not all agentic traits; women leaders are actually encouraged to possess the agentic traits of competence, diligence, and independence). Id. at 2133.
[70]. Rhee & Sigler, supra note 66, at 123.
[71]. Elmuti, Jia & Davis, supra note 63, at 168. This may be because there are still many barriers to advancement for women—such as discrimination, stereotyping, and the demands of family life. Id. at 179.
[72]. Christopher F. Karpowitz & Tali Mendelberg, The Silent Sex: Gender, Deliberation & Institutions 178-79 (2014).
[73]. Carole Kennedy, Gender Differences in Committee Decision-Making: Process and Outputs in an Experimental Setting, 25 Women & Pol. 27, 34-35 (2003). See also Linda Kamas, Anne Preston & Sandy Baum, Altruism in Individual and Joint-Giving Decisions: What’s Gender Got to Do with It?, 14 Feminist Econ. 23, 41-42 (2008) (finding that, in addition to women’s tendency for individual altruism in dictatorship games, they can influence men to be more altruistic in joint-giving games).
[74]. Kennedy, supra note 74, at 37-38.
[75]. Jeffrey A. LePine, John R. Hollenbeck, Daniel R. Ilgen, Jason A. Colquitt & Aleksander Ellis, Gender Composition, Situational Strength, and Team Decision-Making Accuracy: A Criterion Decomposition Approach, 88 Organizational Behav. & Hum. Decision Processes 445, 466 (2002) (finding that the proportion of men on a team was exponentially correlated with inaccuracy in a war gaming task). Furthermore, all-male groups were much more overaggressive in their decisions than were groups of all women. Id. at 467.This is likely because men tend to be more risk-taking and competitive and thus more likely to pick extreme responses in a given situation. Id. at 451. See also Graham D. Fenwick & Derrick J. Neal, Effect of Gender Composition on Group Performance, 8 Gender, Work, & Org. 205, 216 (2001) (finding that groups with more women score higher on certain group performance metrics).
[76]. Anita Williams Woolley, Christopher F. Chabris, Alex Pentland, Nada Hashmi & Thomas W. Malone, Evidence for a Collective Intelligence Factor in the Performance of Human Groups, 330 Sci. 686, 688 (2010). “Collective intelligence” is defined as the ability of a group to “perform a wide variety of tasks.” Id. at 687.
[78]. Lynn Smith-Lovin & Dawn T. Robinson, Gender and Conversational Dynamics, in Gender Interaction & Ineq. 122, 123-24 (Cecelia L. Ridgeway ed., 1st ed. 1992).
[79]. Pamela Ban, Justin Grimmer, Jaclyn Kaslovsky & Emily West, How Does the Rising Number of Women in the U.S. Congress Change Deliberation? Evidence from House Committee Hearings, 17 Q.J. Pol. Sci. 355, 382 (2022).
[80]. Dawn T. Robinson & Lynn Smith-Lovin, Getting a Laugh: Gender, Status, and Humor in Task Discussions, 80 Soc. Forces 123, 143 (2001).
[81]. Kennedy, supra note 74, at 36-37.
[82]. Nikolova & Lamberton, supra note 12, at 367. Interestingly, compromise in group decision-making might also be affected by the gender composition of the audience for a given deliberation. That is, men tend to be less compromising when they are observed by other men; women, conversely, are more accommodating when monitored by other women. Gary Charness & Aldo Rustichini, Gender Differences in Cooperation with Group Membership, 72 Games & Econ. Behav. 77, 83 (2011).
[83]. Christopher F. Karpowitz, Tali Mendelberg & Lee Shaker, Gender Inequality in Deliberative Participation, 106 Am. Pol. Sci. Rev. 533, 534 (2012).
[84]. Elizabeth S. Parks & Kristen Barta, Are You My Mother? Perpetuating Gender Inequality Through Listening Expectations and Relational Roles, 8 J. Res. Gender Stud. 28, 32-33 (2018).
[85]. See, e.g., Denisa Kostovicova & Tom Paskhalis, Gender, Justice and Deliberation: Why Women Don’t Influence Peacemaking, 65 Int’l Stud. Q. 263, 272 (2021) (finding that men seriously dominate the deliberative environment in one set of peacemaking negotiations);
Nancy S. Marder, Gender Dynamics and Jury Deliberations, 96 Yale L.J. 593, 596 (1987) (reviewing findings that men participate in jury deliberation at higher rates than women).
[86]. Afsahi, supra note 13, at 1059-60. See also Tali Mendelberg, Christopher F. Karpowitz & J. Baxter Oliphant, Gender Inequality in Deliberation: Unpacking the Black Box of Interaction, 12 Persp. Pol. 18, 29 (2014).
[87]. Karpowitz & Mendelberg, supra note 73, at 138-39 (finding that women talk more under majority rule as more women join the group, but this effect does not persist under unanimous rule).
[88]. See Karpowitz, Mendelberg & Shaker, supra note 84, at 540.
[89]. Karpowitz, Mendelberg, & Shaker, supra note 84, at 536 (finding that token men participate in group discussion at a greater rate than token women).
[90]. Rosabeth Moss Kanter, Men and Women of the Corporation: New Edition 210-12 (1993).
[91]. Id. at 220. Kanter’s initial research focuses on women in one anonymous corporation, but similar findings have been replicated in a variety of contexts. See, e.g., Janice D. Yoder, Rethinking Tokenism, 5 Gender & Soc’y 178, 179 (1991).
[92]. Kanter, supra note 91, at 219.
[93]. See generally Sarah A. Fulton, Running Backwards and in High Heels: The Gendered Quality Gap and Incumbent Electoral Success, 65 Pol. Res. Q. 303, 308 (2012) (using political activists to rate the “political quality” of legislators).
[94]. See Kathleen A. Bratton, Critical Mass Theory Revisited: The Behavior and Success of Token Women in State Legislatures, 1 Pol. & Gender 97, 121 (2005).
[95]. Martin, supra note 11, at 207.
[96]. Robert K. Christensen, John Szmer & Justin M. Stritch, Race and Gender Bias in Three Administrative Contexts: Impact on Work Assignments in State Supreme Courts, 22 J. Pub. Admin. Res. & Theory 629, 641 (2012).
[97]. See Jacobi & Schweers, supra note 10, at 1482. Men on the bench also appear to display gender biases against female attorneys and female clerks. Dana Patton & Joseph L. Smith, Lawyer, Interrupted: Gender Bias in Oral Arguments at the US Supreme Court, 5 J. L. & Cts. 337, 345 (2017) (finding that judges interrupt female lawyers sooner, more frequently, and with longer comments as compared to male attorneys); see also Shane A. Gleason, Jennifer J. Jones & Jessica Rae McBean, The Role of Gender Norms in Judicial Decision-Making at the U.S. Supreme Court: The Case of Male and Female Justices, 47 Am. Pol. Res. 494, 514 (2019) (finding that male Justices are more likely to vote for a case when the emotional content of a brief matches the gender norms of the attorney writing the brief); see also Alex Badas, Bailey K. Sanders & Katelyn E. Stauffer, The Role of Judge Gender and Ideology in Hiring Female Law Clerks, J. L. & Cts. 1, 20 (2024) (finding that male judges and conservative judges are less likely to hire female clerks); see generally Connie Lee, Gender Bias in the Courtroom: Combating Implicit Bias Against Women Trial Attorneys and Litigators, 22 Cardozo J. L. & Gender 229 (2016) (reviewing studies on gender bias in the courts and proposing tactics to combat this bias). Interestingly, men on the courts may also be biased when it comes to female litigants, though in this case benevolent sexism may work in favor of the individual in some contexts. See Jeffrey J. Rachlinski & Andrew J. Wistrich, Benevolent Sexism in Judges, 58 San Diego L. Rev 101, 133 (2021) (finding judges favor mothers over fathers in family court and dole out more lenient sentences to female defendants); see also Max Schanzenbach, Racial and Sex Disparities in Prison Sentences: The Effect of District-Level Judicial Demographics, 34
J. Legal Stud. 57, 90 (2005) (finding that benevolent sex disparities in criminal sentencing decrease as more women join the court).
[98]. Tonja Jacobi & Kyle Rozema, Judicial Conflicts and Voting Agreement: Evidence from Interruptions at Oral Argument Interruptions, 59 B. C. L. Rev. 2260, 2311 (2018).
[99]. American Bar Association, supra note 3.
[100]. Madeline E. Heilman, Gender Stereotypes and Workplace Bias, 32 Res. Org. Behav.
113, 116-17 (2012).
[101]. See Eagly & Karau, supra note 53, at 576.
[102]. See Matthew E. Baker, Christina L. Boyd, Aidan N. González & Karson A. Pennington, Race and Gender in Judicial Selection in the United States, in Research Handbook on Judicial Politics 143, 147 (Michael P. Fix & Matthew D. Montgomery eds., 2024).
[103]. Suzette Caleo & Madeline E. Heilman, What Could Go Wrong? Some Unintended Consequences of Gender Bias Interventions, 7 Archives Sci. Psychol. 71, 73-74 (2019).
[104]. See, e.g., Amy Jones, Rhiannon N. Turner & Ioana M. Latu, Resistance Towards Increasing Gender Diversity in Masculine Domains: The Role of Intergroup Threat, 25 Group Processes & Intergroup Rel. NP24, NP42 (2022) (finding this effect in samples of STEM students and professionals).
[105]. See Paul M. Collins, Jr., Lori Ringhand & Christina L. Boyd, Supreme Bias: Gender and Race in U. S. Supreme Court Confirmation Hearings, 117 (2023) (finding that male senators are more likely than their female colleagues to aggressively question female nominees, especially regarding legal philosophy and qualifications).
[106]. Id. at 145-46 (documenting that female nominees are interrupted more frequently than male nominees in confirmation hearings).
[107]. Jacobi & Schweers, supra note 10, at 1482.
[108]. See, e.g., Maya Sen, How Judicial Qualification Ratings May Disadvantage Minority and Female Candidates, 2 J. L. & Cts. 33, 59 (2014) (finding that women and minority judges receive lower ratings than male and white judges with similar professional, educational, and political characteristics); Susan Navarro Smelcer, Amy Steigerwalt & Richard L. Vining, Jr., Bias and the Bar: Evaluating the ABA Ratings of Federal Judicial Nominees, 65 Pol. Res. Q. 827, 833 (2012) (finding that men receive a “Well Qualified” rating from the ABA at a rate 12.3%higher than that of women); see also Rebecca D. Gill, Implicit Bias in Judicial Performance Evaluations: We Must Do Better Than This, 35 Just. Sys. J. 271, 319 (2014) (finding that Judicial Performance Evaluations fielded to attorneys tend to be biased against women and minority judges).
[109]. Yoshikuni Ono & Michael A. Zilis, Do Americans Perceive Diverse Judges as Inherently Biased?, 11 Pol., Groups, & Identities 335, 338 (2023) (finding that Americans view women and Hispanic judges as biased toward their in-groups).
[110]. See, e.g., Christina L. Boyd, She’ll Settle It?, 1 J. L. & Cts. 193, 211 (2013) (finding women judges are more likely to foster settlements); Todd Collins & Laura Moyer, Gender, Race, and Intersectionality on the Federal Appellate Bench, 61 Pol. Res. Q. 219, 225 (2008) (finding minority women on the Court of Appeals are more likely to support the claims of criminal defendants); Donald R. Songer & Kelly A. Crews-Meyer, Does Judge Gender Matter? Decision Making in State Supreme Courts, 81 Soc. Sci. Q. 750, 759 (2000) (finding that women judges tend to vote more liberally in certain kinds of cases and that Democratic women on state high courts tend to prefer the more liberal position in some cases that are not necessarily gendered); Jennifer L. Peresie, Female Judges Matter: Gender and Collegial Decisionmaking in the Federal Appellate Courts, 114 Yale L.J.1759, 1779-81 (2005) (discussing why there may be gendered differences in certain types of case outcomes).
[111]. See, e.g., Boyd, Epstein & Martin, supra note 9, at 406 (finding no gendered difference in outcomes except in cases dealing with sex discrimination); Peresie, supra note 111, at 1776 (finding women may be more likely to rule for plaintiffs in sex harassment and gender discrimination cases).
[112]. Darrell Steffensmeier & Chris Hebert, Women and Men Policymakers: Does the Judge’s Gender Affect the Sentencing of Criminal Defendants?, 77 Soc. Forces 1163, 1168 (1999). Though minimal differences in judging may still exist. Id. at 1186 (finding “many similarities yet some differences in the sentencing decisions of women and men judges” that sit on Pennsylvania courts).
[113]. Legal theories posit that judges base their decisions primarily on the law. See, e.g., Cesare Beccaria, On Crimes and Punishments 14 (Aaron Thomas, ed., Aaron Thomas & Jeremy Parzen, trans., 2008) (1764). Scholars of the legal model are sometimes known as “legalists” or “formalists.” Id.
[114]. Richard A. Posner, How Judges Think 175 (2008) (discussing the Dworkinian view that there is “one right answer to every legal question”); Ronald Dworkin, Law’s Empireviii-ix (1968) (arguing that even “hard” cases usually have a correct answer).
[115]. See, e.g., Frank B. Cross, Decisionmaking in the U.S. Circuit Courts of Appeals, 91 Calif. L. Rev. 1457, 1514 (2003); Rachel K. Hinkle, Legal Constraints in the US Courts of Appeals, 77 J. Pol. 721, 731 (2015) (considering judicial treatment of binding precedent in Fourth Amendment cases).
[116]. Kent Barnett, Christina L. Boyd & Christopher J. Walker, Administrative Law’s Political Dynamics, 5 Vand. L. Rev. 1463, 1499 (2018).
[117]. Boyd, supra note 111, at 211.
[118]. Stephen J. Choi, Mitu Gulati, Mirya Holman & Eric A. Posner, Judging Women, 8
J. Empirical Legal Stud. 504, 525 (2011).
[119]. Meghan E. Leonard & Joseph V. Ross, Gender Diversity, Women’s Leadership, and Consensus in State Supreme Courts, 41 J. Women Polit. & Pol’y 278, 294 (2020) (finding that woman-authored opinions on employment discrimination and search and seizures tend to be more unanimous than corollary opinions authored by men). But see
Mikel Norris, Beyond Consensus: Gender, Chief Justices, and Leadership on State Supreme Courts, 43 J. Women Polit. & Pol’y 134, 146 (2022) (surveying chief justices in state courts and finding that women do not believe their primary responsibility is to form consensus on the court; instead, they are more likely to place importance on interactions with state legislatures).
[120]. See Laura P. Moyer, John J. Szmer, Susan B. Haire & Robert K. Christensen, “All Eyes are on You”: Gender, Race, and Opinion Writing on the U.S. Courts of Appeals, 55 L. & Soc’y Rev. 452, 464 (2021) (hereinafter Moyer et al.) (operationalizing “deep cites” as those that discuss a case in depth). Id. at 462. See also Haire, Moyer & Treier, supra note 8, at 320 (finding that panels with a majority of women/minority judges produce opinions that are longer and include more points of law, as measured by Westlaw keycites); Kaleigh A. Ruiz, Gender Differentials in the Content of Judicial Opinions, 57 Ariz. St. L. J. 1825, 1861-62 (2025) (women judges write differently about issues as compared to men). See also Michael A. Livermore, Nina Varsava, Keith Carlson & Daniel N. Rockmore, Gendered Judicial Opinions, 1J. L. & Empirical Analysis 1, 12 (2024).
[121]. Boyd, Epstein & Martin, supra note 9, at 402-06; Peresie, supra note 111, at 1778; Sean Farhang & Gregory Wawro, Institutional Dynamics on the U.S. Court of Appeals: Minority Representation under Panel Decision Making, 20 J. L., Econ., & Org. 299, 320 (2004).
[122]. Paul M. Collins, Kenneth L. Manning & Robert A. Carp, Gender, Critical Mass, and Judicial Decision Making, 32 L. & Pol’y 260, 271-73 (2010).
[123]. Schanzenbach, supra note 98, at 74.
[124]. Eagly & Johnson, supra note 59, at 242.
[125]. Nikolova & Lamberton, supra note 12, at 367.
[126]. In fact, the discretionary behavior of separate opinion authorship likely provides some of the most substantial evidence for gendered behavior on the courts given that majority authorship is assigned out by a presiding judge. While one could argue that this assignment process is plausibly random, there are likely many factors impacting a presiding judge’s decision to assign out a majority. For example, if a presiding judge assumes women can better facilitate compromise, they might systematically assign women judges to author more contentious cases, confounding the results predicted by my first theory.
[127]. While a majority author could always choose to author a separate opinion in addition to the majority, this is a highly unlikely occurrence.
[128]. Nikolova & Lamberton, supra note 12, at 367.
[129]. E.g., in the case of interrupting women attorneys at heightened rates during oral argument. Jacobi & Schweers, supra note 10, at 1482.
[130]. See Jacobi & Rozema, supra note 99, at 2311.
[131]. Kennedy, supra note 74, at 38.
[132]. Additionally, we might think that preferences on the outcome of a case converge as a group becomes gender-homogenous, though we should be cautious about this assumption. See, e.g., Alexander Bolton & Sharece Thrower, Studying Bureaucracy in a Diverse Democracy, 50 Congress & the Presidency 135, 142 (2023) (noting that
“all groups have some level of heterogeneity consequently ignored”).
[133]. Those fourteen issue areas are: (1) abortion; (2) disability advocacy; (3) affirmative action; (4) campaign finance; (5) capital punishment; (6) Commerce Clause; (7) Contracts Clause; (8) piercing the corporate veil; (9) criminal appeals; (10) environmental protection;
(11) gender discrimination; (12) race discrimination; (13) sexual harassment; and
(14) takings. These issue areas were initially selected by Sunstein, Schkade & Ellman to test the attitudinal model of judicial decision-making—i.e., whether the ideology of a judge explains her decision in a particular case. See Cass R. Sunstein, David Schkade &
Lisa Ellman, Ideological Voting on Federal Courts of Appeals: A Preliminary Investigation, 90 Va. L. Rev. 301, 311-13 (2004). Later work on whether gender factored into the attitudinal model utilized the Sunstein, Schkade & Ellman database but eliminated cases on criminal appeals. See Boyd, Epstein & Martin, supra note 9, at 397. I replicated the collection method employed by these scholars, with slightly modified Boolean searches. Appendix B presents the Boolean searches and number of cases in each issue area.
[134]. A regular expression is a sequence of characters specified to match a pattern within text-as-data. For example, the majority of opinions on Westlaw list the panel of judges that heard a case after the word “before” and a colon. Thus, to capture the panel,
I would write a regular expression specifying the return of the words after this pattern and before a new line. This would be the corollary regular expression: “(?<=Before)(.*?)(?=\\n)”. While there are sometimes errors in the opinion text that cause the regular expression to be unable to capture a judge (i.e., it just returns a blank space), a random verification on 50 cases found that the code never returned an incorrect (non-authoring) judge.
[135]. About 0.3% of cases (i.e., thirty-two cases) were decided by a two-judge “quorum” according to circuit-level internal operating procedures. In all such cases, a third judge was initially part of the panel. For eleven of these cases, the third judge passed away before the opinion’s disposition. For three cases, the third judge (namely, Sonia Sotomayor) was elevated to the Supreme Court prior to publication. For another three cases, the third judge retired before the opinion was published. For the remaining fifteen cases, the third judge either discovered there was reason to recuse themselves or was not able to participate in the opinion for unspecified reasons.
[136]. Note that there may be other permutations of opinions not present in this dataset that are thus unlisted on Table 1.
[137]. For the purposes of these designations, a partial concurrence, concurrence in judgment, concurrence dubitante, etc. are included in “concurrences” whereas a dissent in part is included as a “dissent.” “Partials” encompass those opinions that a judge specifies to be “concurring in part and dissenting in part” (or some variation on this language).
[138]. Less than 0.5% of cases in the dataset include a separate opinion written by a majority author (n = 44).
