The Attorneys’ Gender: Exploring Counsel Success before the U.S. Supreme Court

Gender and Stereotyping

When making decisions, individuals often rely upon stereotypes and heuristics. Stereotyping—the assignment of “identical characteristics to any person in a group regardless of the actual variation among members of that group” (Aronson 1988, 233)—allows individuals to organize, process, and evaluate information. It provides an additional layer of structure to the world using easily accessed representations of objects, events, or people (Haslett, Geis, and Carter 1992). Stereotypes are frequently used as “shortcuts” when sorting through information is cognitively burdensome, or when there is too little information and a decision cannot easily be made (Conover 1984).

Gender is one of the most prominent and readily available stereotypes drawn upon when decisions are made. Further, different traits are frequently assigned to men and women based upon these stereotypes. Men are traditionally stereotyped as strong, rational, assertive, and competent, whereas women are often stereotyped as more emotional, empathetic, warmer, and less assertive (McKee and Sherriffs 1957; Valian 2005). This stereotyping has traditionally extended to political evaluations and decision-making, with women political elites often being stereotyped as more liberal (King and Matland 2003), more capable of dealing with social welfare issues, but less capable at dealing with foreign policy, defense, and crime (Alexander and Andersen 1993; Huddy and Terkildsen 1993; Lawless 2004), and often viewed as outsider candidates, compared with men (Fridkin and Kenney 2009). Often, as soon as one perceives an individual to be either a man or woman, they assign attributes to that individual via the stereotyping process, sometimes intentionally but often not (Greenwald, Klinger, and Schuh 1995).

These stereotypes generally assign women traits that make them less likely to be considered credible or competent in political or professional settings, and likely place women at a disadvantage in a variety of avenues. These stereotypes have generally been thought to disadvantage women in the business and political realms (Fox and Oxley 2003; Kahn and Fridkin 1996; Sanbonmatsu and Dolan 2009; but see, Brooks 2013; Hayes and Lawless 2015 on political matters; e.g., Cuddy, Fiske, and Glick 2004 on the business world). A growing literature has investigated the ways in which gender, gender stereotyping, and disadvantages often facing women appear in the judicial system.

Studying Gender and Judicial Politics

Research exploring gender and judicial decision-making has tended to focus on how a judge’s gender influences decisions from the bench (see, e.g., Kritzer and Uhlman 1977). For example, Gruhl, Spohn, and Welch (1981) show that women judges were considerably less likely to sentence women defendants to prison, than men judges. Songer, Davis, and Haire (1994) find that in employment discrimination cases, women judges were significantly more liberal than men. Turner (2015) explores the extent to which the majority opinion writer’s gender influences whether fellow justices file a concurrence. Still others have found mixed support for the influence of gender on decision-making, even for cases involving “women’s issues.” ⁵

Other avenues of inquiry into the intersection of gender and the Court have studied the underrepresentation of women in the Supreme Court Bar (Sarver, Kaheny, and Szmer 2007). Those exploring advocacy and the courts have often focused on how the justices engage with attorneys. For example, Phillips and Carter (2009) demonstrated that liberal justices engage in more information seeking behavior when the attorney is a woman, relative to when the attorney is a man. Conservative justices demonstrate the opposite effect. Still others (Gleason, Jones, and McBean 2019) have studied how adherence to gender norms surrounding argument presentation in party-briefs submitted to the Court can affect how men justices evaluate the arguments in those briefs. Looking at women’s treatment, generally, some have found high levels of incivility toward women at all levels and positions within the legal system (see, e.g., Cortina et al. 2002; Lonsway et al. 2002).

Recent scholarship has shown that legal advocacy shapes judicial outcomes (Johnson 2001; Ringsmuth, Bryan, and Johnson 2013). Attorney arguments—whether written or oral—are important sources of information for justices (Caldeira and Wright 1988; Johnson 2003; Johnson, Wahlbeck, and Spriggs 2006). Oral arguments provide the justices with an opportunity to interact with counsel, asking questions and assessing a lawyer’s real-time advocacy capabilities. In addition to providing a forum for the justices to ask questions and clarify issues, oral arguments may induce the use of cognitive heuristics on the part of the justices when assessing the “messenger.” During oral arguments, the justices are starkly presented with a lawyer’s individual traits, “including gender, race, and age, and often general ‘likableness,’ which immediately brings many social stereotypes into play” (Lau and Redlawsk 2001, 954). Cognitive heuristics, in this case attorneys’ individual traits, may trigger a range of emotional and other responses (Marcus and MacKuen 1993) on the part of the justices, coloring how they evaluate the arguments being presented and whether the Justices readily accept the information. ⁶

Oral arguments place an attorney physically before the justices making it the forum most likely to cue gendered stereotypes. Since attorney gender is arguably one of the most readily perceived individual characteristics, the justices—whether intentionally or not—may assess a lawyer’s gender, activating a variety of stereotypes, many of which were discussed above, influencing perceptions of credibility. Oral arguments, therefore, become the forum where automatic stereotyping based on gender is most likely to occur.

A burgeoning literature has demonstrated that men and women attorneys may be received differently during this stage of the judicial process. Women lawyers are interrupted earlier and more frequently during oral arguments (Patton and Smith 2017). Justices speak more frequently and for longer when a woman is arguing before the Court (Patton and Smith 2017, 2020; Phillips and Carter 2009). This effect is particularly pronounced among conservative justices (Patton and Smith 2020). Finally, Gleason (2020) has argued that attorneys who deviate from gender norms in their communication style may be less successful. These findings—although concerning, for a society claiming to seek gender parity—do not directly address the most important question for litigators: Does an attorney’s gender influence their likelihood of winning?

Scholarship investigating the influence of attorney gender on judicial outcomes is mixed, sometimes finding that attorney gender is related to outcomes and other times not. For example, some work finds that no differences emerge, generally, between the willingness of the justices to vote in favor of a party represented by men or women attorneys, except in the case of the most conservative justices (Szmer, Sarver, and Kaheny 2010). Gleason (2020) finds that litigants represented by women may be more likely to win when the attorney adheres to gender norms when addressing the Court, and less likely to win otherwise. This ultimately leads Gleason (2020, 1) to ponder about, “ . . . how effective women are at the Supreme Court.” Other analyses have investigated the influence of attorney gender on decision-making by the Supreme Court of Canada (Kaheny, Szmer, and Sarver 2011) and decision-making by U.S. Courts of Appeals judges (Szmer et al. 2013), demonstrating that women attorneys acting as oral advocates may actually be more likely to win in these forums in most types of cases. ⁷

Revisiting Gender, Oral Advocates, and Success before the Court

Although the literature covered above has begun to explore the role gender stereotypes play in influencing how attorneys are treated and fare before the Court, as well as the role of gender in the justice system more generally, more work is needed to fully assess accounts by contemporary Court observers about the challenges facing women attorneys. Notably, much of the work cited above suggests women attorneys may be less successful as advocates, however, few have answered the question of most importance to attorneys and their clients—do women advocates disproportionately lose before the Court? Furthermore, work that does consider this question focuses on narrow bands of time, making it difficult to understand the broader arc of women attorneys’ success before the Court. Considering this, we add to extant findings regarding attorney gender and decisions by the nation’s highest court. In addition to showing that an attorney’s gender does not influence their party’s success before the Court, we offer an expanded time period over which to study advocate success. Importantly, we conceptualize and operationalize attorney success as the overall decisions of the Court—that is, whether the party represented by an attorney wins or loses. ⁸

Reviewing prior work above on gender stereotyping, the importance of oral advocacy, and Supreme Court decision-making it is reasonable to expect that attorney gender may influence decision-making at the Supreme Court. Specifically, given the importance of oral advocacy to judicial decision-making, the opportunity oral arguments present for traits subject to stereotyping to become salient, as well as the prevalence of often-automatic stereotyping, it is uncontroversial to suspect that stereotypes may play a role in the decision-making process of justices. Considering that such stereotyping often assigns to women traits that make them less likely to be considered credible or competent, and the existence of evidence that these stereotypes are harmful to women in both the political and professional worlds, we suggest the following hypothesis:

Surprisingly, the limited research focused on this debate lacks consistent findings. These realities prove puzzling considering the numerous reasons discussed above to expect women attorneys may face challenges before the Supreme Court. We seek to provide clarity and context to this line of inquiry by considering reasons to expect that the gender of an attorney presenting oral arguments may not influence the Court’s decision-making.

Over the 20th century, society’s view of women has changed drastically, becoming more accepting, and even encouraging in some instances, of women as professionals, serving as top-tier political candidates, the primary income earner in households, and in many other roles once thought only the domain of men (Roper Center 2014). ⁹ Women’s place in politics and the legal profession has changed over the past 50 years. As views about the appropriate social and professional roles for women change, we may expect negative stereotypes about women in those roles to be less influential.

As women become more prominent in professional, legal, and political life, the use of negative stereotypes in evaluating them decreases (see, e.g., Craig and Jacobs 1985; Stevens and Gardner 1987). For example, women running for office are no longer being evaluated differently by the public (Brooks 2013; Hayes, Lawless, and Baitinger 2014) or other political elites such as the news media (Hayes and Lawless 2015). Researchers suggest that the increased prominence of women candidates has reduced the use of negative stereotypes to portray and evaluate them (Hayes and Lawless 2015).

Szmer et al. (2013, citing Craig and Jacobs 1985) expect a decrease in negative stereotyping of women attorneys as more diverse individuals come to occupy appellate court judgeships. The example they cite as evidence comes from a study of men firefighters. As men gained more experience working with women firefighters, they grew to hold their women peers in higher regard and seemed to shed some of their stereotypes of women. We expect this to be true for women attorneys arguing before the Supreme Court as well.

Moreover, women are becoming more prevalent in the legal profession, which may suggest a decrease in the use of harmful stereotypes. ¹⁰ As women matriculate into law school in greater numbers, argue more frequently before the Court, and come to occupy a greater number of seats on the Supreme Court, their presence in the Courtroom should be considered less novel. ¹¹ The justices, therefore, may be less likely to rely on potentially harmful stereotypes when evaluating arguments made by women attorneys in more recent times.

If it is the case that negative stereotypes are less frequently used in recent decades to evaluate women Supreme Court advocates, this might explain the puzzling findings of prior work, which have predominantly used data from recent decades. Evidence of this explanation would be an increase in women attorneys’ success before the Court the more recently one looks. Although litigants represented by women attorneys may have fared worse before the Court early-on, the relationship between attorney gender and party success may change over time, with women attorneys’ success before the Court reaching a point of no statistical or substantive difference from men. Although it is difficult to select a precise time at which we may expect this trend to take hold, considering most of the societal changes discussed above were underway throughout the 1970s through 1990s, we may expect that by 2000 there should be substantially different views of women attorneys than in the three or four decades prior. Thus, we suggest the following hypothesis:

In the next section, we discuss the data set and methods used to test our hypotheses. We then provide the results of our analysis, investigating the relationship between attorney gender and party success before the Court. Finally, we conclude with reflections upon this line of inquiry.

Data and Methods

We seek to answer three empirical questions: (1) Do women attorneys lose at a higher rate before the Supreme Court than men? (2) Has this changed over time? And, (3) what contextualizes these findings? To answer these questions, we employ an analytical strategy that accounts for attorney- and case-level characteristics. Our methodological decisions allow us to move beyond merely establishing whether a systematic bias against women advocates exists. Overall, this strategy helps to clarify the findings of previous work exploring gender biases affecting litigators before the Supreme Court.

We construct a data set consisting of every orally argued signed case before the Supreme Court from 1946 to 2016. ¹² We scripted a program that searched each case’s citation—as provided by the Supreme Court Database—in the Case Access Project and extracted information about the attorneys arguing before the Court. Following previous research (see, e.g., Black et al. 2016; Feldman 2016; Gleason 2020; Szmer, Sarver, and Kaheny 2010), we limit our universe of cases to those where one attorney argues for the petitioner and one for the respondent ¹³ —this returns 5,112 cases. ¹⁴ This is done to account for both interdependence and deterministic reasons. First, multiple attorneys arguing on the same side of a case violates the observational independence assumption inherent in most model specifications. Limiting the scope to cases where one attorney argues each side alleviates this concern. Second, when multiple attorneys argue for the petitioner(s), it becomes impossible to determine which attorney’s personal or professional attributes were influential. Finally, such an approach allows us to compare differences in success when opposing attorneys are of the same, as well as different, genders.

Since personal characteristics can cue a range of responses, our analysis is limited to attorneys who have stood and orally argued a case before the Court. Source cue research suggests that attorney attributes are likely to influence judicial decision-making in narrow ways. Oral arguments allow the justices to interact with attorneys, and it is in this forum where the justices are most likely to recognize an attorney’s attributes. Concretely, we expect the attributes of those attorneys who stood and orally argued a case to be most salient and have the most measurable influence on Court outcomes. ¹⁵

Results

As mentioned above, our dependent variable is dichotomous—capturing whether the petitioner secures a win before the Supreme Court. Given this choice, we focus our analysis on cases where one attorney argues for the petitioner and one for the respondent. This results in 10,224 attorneys nested in 5,112 cases from 1946 to 2016, spanning thirty-four natural courts. Before directly testing our hypotheses, we provide summary statistics and tabulations of these data.

Of the over 10,000 lawyer-appearances in our data set, 710 (6.9%) are women and 9,514 (93.1%) are men. Focusing on how attorney gender is distributed by party to the case, these data show that men are relatively evenly divided in terms of representing the petitioner or the respondent. However, of the 710 instances where a woman argued a case, more than half of the time (54.1%) it was on behalf of the respondent.

As expected, Figure 1 shows that women appear more frequently as time goes on, comprising a larger proportion of the attorneys who litigate before the Court. Starting in the late 1960s through the early-1990s, women gradually started solo-arguing more cases before the Court. By 1993, 15 percent of attorney-appearances were by women. Although 1994 proves curious, the overall upward trend continues into 2016. Importantly, the number of unique appearances for women attorneys—in any given year—never reaches beyond twenty-nine, indicating that the women who argue before the Court are limited to a small cadre of repeat attorneys.

OPEN IN VIEWER

Figure 1.

Percentage of women attorney solo-arguments before the Court.

Ultimately, our focus is to explore the influence of attorney gender on petitioner success. When cross-tabulating gender and petitioner success the chi-squared statistic fails to uncover any meaningful statistical association between these variables (χ² = 4.9; p = .2). ²³ Although exploratory statistics are useful for understanding underlying trends, we next present statistical models designed to test our expectations.

Given our binary dependent variable, whether the petitioner wins or not, we model these data using logistic regression with standard errors clustered by term. Since logit coefficients cannot be directly interpreted, we follow the work of Hanmer and Kalkan (2013) and present average marginal effects. ²⁴

Our interest is in understanding both the unconditional effects of gender on petitioner success as well as conditional effects of gender over time and when interacted with the Court’s ideological congruence with the petitioner. As such, we run two models. The first explores the unconditional effects of gender; the second controls for the interaction between gender and Supreme Court term, as well as the interaction between gender and the Court’s ideological congruence with the petitioner’s position. Figure 2 presents the average marginal effects calculated at observed values for both the unconditional and conditional models.

OPEN IN VIEWER

Figure 2.

Average marginal effects for models 1 and 2.

Focusing first on the unconditional effects of gender (unconditional model), we find that the petitioner is more likely to win when opposing counsel are both women, although this effect goes away in the conditional model. Interestingly, there is no statistical difference in the likelihood of petitioner success when opposing counsel are of different genders or when opposing counsel are both men.

Turning to the interaction between attorney gender and Supreme Court term, we find no moderating effects. Over time, women arguing for the petitioner are statistically no more likely to win than are men arguing for the petitioner against another man. Likewise, a man arguing for the petitioner facing a woman is no more likely to win or lose over time than if that man were facing another man. In cases in which a woman attorney faces a woman attorney, the petitioner appears to have become increasingly likely to win; however, that interaction is not statistically significant. ²⁵ Thus, regardless of the position of the gender of the attorneys, the likelihood of winning—relative to that of a man attorney representing the petitioner and facing another man—doesn’t statistically change over time. Figure 3 plots the marginal effects for the interaction between gender of the attorneys and Supreme Court term. Confidence intervals have been removed for visual purposes. However, in the Online Appendix (Figure A1) we present a faceted contrast plot for readers demonstrating our null findings.

OPEN IN VIEWER

Figure 3.

Marginal effects plot for interaction between attorneys’ gender and term.

Our findings indicate that professional stature (e.g., whether an attorney is from the Office of the Solicitor General) is far more important for understanding successful outcomes before the Supreme Court. For example, the petitioner is 0.12 points more likely to win when represented by an attorney from the U.S. government (p < .05). ²⁶ Conversely, the likelihood of petitioner success when represented by a state government attorney is not statistically distinct from a petitioner represented by a non-government attorney.

Next, we investigate whether previous experience in the Office of the Solicitor General influences petitioner success. We find that having worked in the SG’s Office decreases the likelihood of petitioner success. Attorneys who previously served in the SG’s office representing the petitioner are 0.12 points less likely to win, relative to attorneys who did not serve in the SG’s Office. Although surprising, upon further reflection and investigation, it is probable that this is a result of frequent appearance before the Court. As an attorney argues more frequently before the Court, the opportunity to lose increases as well.

As expected, repeat status increases petitioner success—corroborating previous research. We measure the difference in the number of appearances between the attorney arguing for the petitioner and the attorney arguing for the respondent. A positive value indicates that the petitioner’s attorney has more appearances before the Court than the respondent’s attorney. A negative value indicates that the respondent’s attorney has more appearances before the Court than the petitioner’s attorney. Zero indicates that opposing counsel have the same number of appearances before the Court.

In Figure 4, we show the overall trend. At a value of zero (opposing attorneys have the same number of appearances), the petitioner’s predicted probability of success is 0.65. As the discrepancy in prior appearances between the petitioner’s attorney and the respondent’s attorney increases, so too does the likelihood of petitioner success. Each appearance before the Court that a petitioner’s counsel has over the respondent’s increases the likelihood that the petitioner will win by 0.36 points. When the petitioner has five prior appearances more than the respondent, the likelihood of the petitioner winning increases by nearly 2 points, 1.8 percentage points. The inverse of this is that for every prior appearance before the Court the petitioner’s attorney has less than the respondent’s, the likelihood of the petitioner winning decreases by 0.36 points, or by just under two percentage points for every five less prior appearances the petitioner’s counsel has relative to the respondent’s counsel. This effect is statistically significant at traditional levels. This is a substantively significant relationship considering that this variable runs from −76 to 74, suggesting there are instances in which a substantial advantage or disadvantage in prior experience exists between attorneys.

OPEN IN VIEWER

Figure 4.

Difference in petitioner and respondent appearances before the court.

Considering the results from the unconditional model, we next turn to our control variables. ²⁷ First, we find no statistically significant effects on the likelihood of petitioner success when the attorney is Court appointed. The justices do not seem to lend preferential treatment—in terms of the likelihood of success—to those requested to argue a given position. Likewise, we find that neither case salience nor legal salience reach statistical significance in predicting petitioner success. ²⁸

We find that amici-support influences petitioner success. Each additional amicus brief filed in support of the petitioner over the respondent increases the petitioner’s probability of success by 0.2 points. This is a substantively significant relationship considering that the variable runs from −41 to 16. Moreover, this finding proves statistically significant. We find that petitioner success increases by 0.16 points when amici are filed by the United States in support of the petitioner, relative to no amici support. Amici filed on behalf of the petitioner by non-governmental groups or state governments fail to achieve statistically significant effects in the model.

Using our measure of ideological congruence discussed earlier, we find that the ideological congruence between the Court, measured as the ideology of the median justice, and the petitioner makes the petitioner more likely to succeed. Turning to the conditional model, the interaction between our measure of attorney gender for each party and ideological congruence is not statistically significant and an investigation of the contrasts of those variables demonstrates a similar finding. In addition, the coefficients on the attorney gender variable and ideological congruence variable are not statistically different between the conditional and unconditional models. This all suggests that ideological congruence and attorney gender do not have conditioning effects on one another.

Status as a former law clerk does not seem to have a significant effect on winning. This is likely for a few reasons. First, former clerks have a higher average number of previous arguments than non-clerks. It may well be that repeat appearances before the Court prove more influential than having been a clerk. Second, former law clerks are almost twice as likely to go on to appear before the Court representing the U.S. government in some capacity, which also leads to a higher level of success. Thus, although being a former law clerk may benefit attorneys in some instances, in most cases it is likely overwhelmed by other characteristics of those attorneys. We also interact petitioner former law clerk status with our attorney gender indicator. A model with these results can be found in the Online Appendix. Our substantive results for attorney gender do not change with the inclusion of this interaction. ²⁹

Finally, turning to women’s issues, controlling for women’s issues and an interaction between women’s issues and our attorney gender variable does not substantively change our results. Furthermore, the interaction is not statistically significant. We present the estimates from this model in the Online Appendix. Thus, whether women attorneys are arguing about a “women’s issue” or not does not appear to alter their success before the Court regardless of which side they’re arguing on or the gender of the attorney they are arguing against.

Discussion

When accounting for other factors influential to the decision-making process, parties represented by women during oral arguments are not substantively different from those represented by men when it comes to securing a win. We present these findings against the backdrop of a literature that has offered numerous results over limited time periods. Moreover, research centering on attorney gender and Supreme Court outcomes has generally focused on individual justice voting behavior rather than the behavior of the Court as a whole. Importantly, we build upon earlier work by presenting a more comprehensive understanding of the backgrounds and experience of the individuals arguing before the Court, and over a substantially longer period of time than prior work, which aids in contextualizing our results.

Ultimately, we find that attorney gender is a poor predictor of success before the Court. Other characteristics such as attorney experience and party support are more strongly associated with party success. One interpretation of these findings may be that women no longer face discrimination in the legal profession or in their appearances before the nation’s highest Court. This interpretation would be that these findings suggest that there is some “good news” when it comes to the question of whether the gender of an attorney affects their party’s success before the highest Court. Notably, the gender of an attorney does not seem to be systematically related to their success. This is good news in so much as it means that for those attorneys who make it before the Court to argue, justice is “gender blind”—a finding that should offer observers of the Court and citizens alike comfort.

However, we suggest that another, more nuanced interpretation of these findings is implicated. Although the inability to find evidence of gender bias before the Court is undeniably good, we believe these results also suggest that discrimination against women attorneys continues to be rampant. We demonstrate that once before the Court women fare just as well as men, controlling for other relevant factors. However, a brief investigation into the litigation background of the women arguing before the Court finds that more women argued their first case before the Court as government attorneys and/or holding significant prior experience. Although, on average, the men arguing before the Court possess similar attributes, there also exists far more variation in their backgrounds and prior experience. This highlights that while men attorneys with varying backgrounds and experience have been afforded the opportunity to appear before the Supreme Court, women have had to be, on average, more qualified and experienced to receive the same opportunity. To the extent gender discrimination occurs in oral advocacy before the Supreme Court, it appears to begin far before women even make it to the Court. This poses many questions left unanswered, many of which have significant normative implications for our understanding of the judicial branch and American politics. If the sorts of women who make it to argue before the Court are fundamentally different from women attorneys who do not, and perhaps are even different in important ways from the men who argue before the Court, then that may have implications for representation of interests, equality within the legal profession, and even the sorts of decisions the court makes.

We believe these results suggest at least three future avenues of research. First, future work should build off of these findings to begin to pinpoint more carefully where along the Supreme Court advocacy “pipeline” women face barriers to entry. This can be useful to identifying the causes and how to overcome the disparity. Second, although we find that attorney gender does not seem to affect the decisions of the Court as an institution, or often not the individual behavior of judges, we encourage future work to continue to identify the ways in which attorney gender may matter, as an already bourgeoning new literature has begun doing. Finally, further work should attempt to understand the ways in which those women attorneys who do make it before the Court differ, if at all, from those who do not. As noted above, this line of research may have particularly significant normative implications for the legal profession, as well as representation within our judicial system.

Supplemental Material