I Can’t See You; Can You Hear Me? Gender Norms and Context During In-Person and Teleconference U.S. Supreme Court Oral Arguments

Abstract

Female attorneys at the U.S. Supreme Court are less successful than male attorneys under some conditions because of gender norms, implicit expectations about how men and women should act. While previous work has found that women are more successful when they use more emotional language at oral arguments, gender norms are context sensitive. The COVID-19 pandemic prompted perhaps the most radical contextual shift in Supreme Court history: freewheeling in-person arguments were replaced with turn-based teleconference arguments. This change altered judicial decision-making and, I argue, justices’ assessments of attorneys’ gender performance. Using quantitative textual analysis of oral arguments, I demonstrate that justices implicitly evaluate gender performance with different metrics in each modality. Gender-normative levels of emotional language predict success in both formats. Function words, however, only predict success in teleconference arguments. Given gender’s salience at the Supreme Court and in broader society, my findings prompt questions about the extent to which women can substantively impact case law.

In March 2020, the COVID-19 pandemic prompted courts to suspend in-person oral arguments. Many conducted business over Zoom (Bergeron 2021), but the U.S. Supreme Court utilized teleconference arguments (Johnson et al. 2021). Under this new configuration, oral arguments were limited to an audio-only connection, meaning that justices and attorneys could not see each other. This altered many aspects of the Court’s regular operations (e.g., Houston, Johnson, and Ringsmuth 2023; Ringsmuth et al. 2023). Importantly, changing the context fundamentally alters how actors assess information (e.g., Chaiken and Eagly 1983; Ferran and Watts 2008; Guadagno and Cialdini 2007). This includes evaluations of gender norms, subtle expectations about how men and women should act. Gender norms are powerful predictors of outcomes (Eagly and Wood 2012; Rudman et al. 2012), but their manifestation is context dependent (Eagly and Karau 2002). I argue that the Court’s response to COVID-19 altered justices’ assessment of attorneys’ gender performance. This has normative consequences for how women contribute to the development of case law binding on lower courts throughout the United States and raises concerns about whether greater descriptive representation at the Supreme Court Bar translates into substantive representation in case law (e.g., Pitkin 1967).

Changes in women’s substantive impact on case law have far-ranging implications. People view institutions as more legitimate when their views and concerns are taken into account by decision makers (e.g., Pitkin 1967). Moreover, greater representation for women increases the overall quality of judicial outputs (Boyd 2015; Haire, Moyer, and Treier 2013). While the Bench is increasingly diverse, justices are legal generalists and depend on attorneys to provide information (Corley 2008). While the number of women at the Bar is slowly increasing, their greater presence does not necessarily translate into different case law (Gleason and Smart 2023). The extent to which justices draw on female attorneys’ arguments is dependent on gender-norm compliance (Gleason 2020).

Gender norms generally hold that women should emphasize communal traits, whereas men should exhibit agentic traits (see, e.g., Bakan 1966; Bauer 2017; Karpowitz and Mendelberg 2014). The legal profession generally, and the Supreme Court Bar specifically, is overwhelmingly male in composition and culture (Szmer, Kaheny, and Sarver 2021). Accordingly, its professional norms value adversarial and agentic behavior (Eagly and Karau 2002; Norgren 2018; Rhode 1994). Thus, women must often violate professional norms to comply with gender norms. Consequently, female attorneys are often less successful than their male colleagues under some conditions (Szmer, Sarver, and Kaheny 2010) and receive fewer professional opportunities (e.g., Szmer, Kaheny, and Sarver 2021; Tillman and Hinkle 2018). While scholars note that women must navigate a precarious double bind (Kanter 1977; Rhode 1994), I argue that the shift from freewheeling oral arguments with frequent interruptions to measured, turn-based teleconference arguments alters the context to the point that the underlying evaluative criteria for gender-norm compliance expand (e.g., Chung and Pennebaker 2007; Johnson et al. 2021; Ringsmuth et al. 2023). This should make it more difficult for women to substantively impact case law.

I examine the extent to which gender-normative language predicts oral argument success for (fe)male attorneys through a quantitative textual analysis of transcripts from all orally argued cases resulting in a signed opinion or judgment during the 2018–20 Supreme Court terms (Gleason 2020). The choice of these terms, centered on the Court’s March 2020 suspension of in-person oral arguments (Houston, Johnson, and Ringsmuth 2023; Ringsmuth et al. 2023), allows for a natural experiment assessing the operation of gender-normative language across two distinct argument formats. I find that changes in oral argument format expand the scope of verbal gender norms that shape attorney success.

Gender-Normative Language at Oral Arguments

Attorneys are, at the most basic level, tasked with communicating information to the Court (Johnson 2001, 2004). Yet communication is not uniformly evaluated across all speakers. Evaluations are based, in part, on whether speakers communicate consistent with socially constructed behavioral expectations (Gawronski 2003; Rosenthal and Rubin 1978), often dependent on sex (Eagly and Carli 2007; Karpowitz and Mendelberg 2014). To this end, arguments that are effective for men are not necessarily effective for women. This distinction is necessitated by juxtaposed gender and professional norms.

Gender norms, implicit expectations about how men and women should act in a myriad of contexts (Bauer 2017; Eagly and Carli 2007; Jones 2016), operate in numerous ways (Cashdan 1998). Typically, work on the Supreme Court examines how the presence of female attorneys impacts judicial behavior (see, e.g., Patton and Smith 2017; Szmer, Sarver, and Kaheny 2010). When a (wo)man approaches the lectern, justices implicitly note their sex and activate gender-normative behavioral expectations (e.g., Olivola and Todorov 2010; Shih, Pittinsky, and Ambady 1999; Witt and Wood 2010). More recent work examining the words attorneys use finds that female attorneys are more successful when they “sound like women” (Gleason 2020). But the professional norm of a good attorney is masculine and at odds with gendered expectations for female attorneys. Thus, women must carefully balance sounding like a “good woman” with sounding like a “good attorney” (Rhode 1994). The importance of this balancing act is growing; since the 1970s, the role of gender norms in predicting attorney success has become increasingly pervasive (Gleason and Smart 2023). This underscores that while gender norms are present in virtually all interactions, their precise manifestation is context dependent (e.g., Chung and Pennebaker 2007; Guadagno and Cialdini 2007). The COVID-19 pandemic ushered in perhaps the most profound context change in the Supreme Court’s history. I contend that this should also change the operation of gender norms.

The outbreak of COVID-19 prompted the Court to suspended in-person oral arguments and shift to teleconference arguments, in which justices could hear, but could not see, attorneys (Johnson et al. 2021). More than just a change in modality, the Court abandoned its traditional format whereby justices can interrupt attorneys at any time. In its place, justices took turns asking attorney questions (Jacobi et al. 2021; Ringsmuth et al. 2023). The shift was difficult for jurists and attorneys alike; it was challenging to know when to start speaking or to fully engage without seeing facial reactions or body posture (Bergeron 2021; McGaughey 2021). Importantly, changing the context can alter the way decision makers evaluate actors (Ferran and Watts 2008). This should extend to how justices evaluate attorneys’ gender performance.

Attorney Sex and the Performance of Gender

The Supreme Court Bar is the most elite group of lawyers in the United States (Black and Owens 2012; McGuire 1993). Their briefs and oral arguments shape the contours of judicial decisions binding on lower courts throughout the nation (see, e.g., Corley 2008; Johnson 2001). While the Bar is prestigious in its own right, there is remarkable variation in members’ success. ¹ Attorneys with prior clerkships (Feldman 2017b), more experience (McGuire 1995), or prestigious affiliations (Curry 2015) are typically more successful. All of these factors are mutable; attorneys can accrue more experience or get a new job. However, scholars increasingly note that attorney success is predicted by immutable characteristics, such as sex, under some conditions (see, e.g., Gleason 2020; Gleason and Smart 2023; Szmer, Sarver, and Kaheny 2010). These studies begin with presence: when a female attorney rises to speak, justices implicitly note her sex (Olivola and Todorov 2010; Shih, Pittinsky, and Ambady 1999). ² This sex cue calls forth gender norms.

Gender norms, socially constructed expectations about how men and women should act in a host of contexts (Bauer 2017; Eagly and Carli 2007; Heilman et al. 2004), are grounded in social role theory. Briefly, they dictate how gender should be performed. ³ These expectations operate at an implicit level (e.g., Eagly and Carli 2007; Uleman, Saribay, and Gonzalez 2008), and they are internalized as early as childhood (Cunningham 2001; Eagly and Wood 2012; Witt and Wood 2010). In general, gender norms hold that men should be agentic (e.g., assertive and persuasive), whereas women should be communal (e.g., conciliatory and interpersonally warm) (Bakan 1966; Cuddy, Fiske, and Glick 2008; Heilman et al. 2004). While gender norms can include manner of dress, hand gestures, vocal pitch, and language (de Lemus, Spears, and Moya 2012; Neumann, Fowler, and Ridout 2022), their precise operation is context dependent (see, e.g., Cheng, Chandramouli, and Subbalakshmi 2011; Chung and Pennebaker 2007; Gleason and Ivy 2021). This is particularly true in heavily gendered contexts.

Women are seen as less qualified in stereotypically male domains (Eagly and Karau 2002), such as the legal profession (e.g., Haire and Moyer 2015; Norgren 2018). To this end, female attorneys often report marginalization (Collins, Dumas, and Moyer 2017) and patronizing treatment from male counsel and jurists (Kearney and Sellers 1996; Winkle and Wedeking 2003). This is especially pronounced in American federal appellate courts, where, despite increased diversity in the legal profession overall, women remain underrepresented (Solberg and Diascro 2020; Szmer, Kaheny, and Sarver 2021). At the Supreme Court, female attorneys are less likely to secure justice votes under some conditions (Szmer, Sarver, and Kaheny 2010), and they are interrupted earlier and more often than male attorneys (Patton and Smith 2017, 2020). ⁴ While most studies rely on the presence of women, recent work takes a more performative conceptualization of gender and finds that women’s success varies depending on how gender is performed. That is to say, arguments that are effective for men are not necessarily effective for women (e.g., Butler 1999).

Because the common law tradition is adversarial, aggressive and assertive advocates are desirable. However, this is a stereotypically male trait (e.g., Eagly and Carli 2007). Accordingly, while attorneys writ large are more successful when they follow the Court’s directive to avoid emotional language (Black et al. 2016), female attorneys are more successful when they use more emotional language in briefs and oral arguments (Gleason 2020; Gleason, Jones, and McBean 2019). Thus, in order to be successful, women must violate the Court’s rules. This necessitates a careful balance for female attorneys that does not exist for men. However, much as in other institutional contexts in which women are a minority, women often downplay the salience of difference by adopting more stereotypically male behavior (e.g., Kanter 1977; Rhode 1994). To illustrate, female candidates for progressively higher and male-dominated elected offices use more masculine language (Jones 2016). This strategy is effective for legislators (Karl and Cormack 2023) but not for attorneys at the Court (Gleason 2020). Even still, female attorneys use less emotional language than their male counterparts on average (Gleason, Jones, and McBean 2017). That is to say, female attorneys argue in a more masculine way than men.

Context is key to outcomes at the Court (e.g., Black, Johnson, and Owens 2018). During the late 1970s and early 1980s, relatively few women appeared at the Court and gender-normative levels of emotional language did not predict attorney success. However, as more women argued at the Court and the conservatism of justices increased, gender-norm compliance became a significant predictor of attorney success (Gleason and Smart 2023). There is also evidence of this at other courts. Since more women argue at federal appellate courts than at the Supreme Court, women are as successful as their male colleagues under some conditions there (Szmer et al. 2013). The Supreme Court of Canada’s bench, bar, and law clerk corps are more diverse than the U.S. Supreme Court’s (Kaheny et al. 2015), and female attorneys are more successful there their male colleagues under some conditions (Kaheny, Szmer, and Sarver 2011). Case context is likewise important; female attorneys are more successful (Szmer et al. 2013; Szmer, Sarver, and Kaheny 2010) and less likely to be interrupted (Patton and Smith 2017, 2020) in “women’s issue” cases, in which they are presumed to be experts.

While scholars have extensively studied how context shapes female attorney success at oral arguments, these studies occur in the same in-person modality. An interdisciplinary literature stresses that the use of different modalities, such as in-person, video, audio, and text, changes which underlying aspects of speakers’ presentations resonate with decision makers (Chaiken and Eagly 1983; Ferran and Watts 2008). This includes how gender is evaluated (Areni and Sparks 2005; Guadagno and Cialdini 2007). Leaving the in-person modality makes it more difficult for speakers to hold evaluators’ attention (Chaiken and Eagly 1983), and evaluators are more likely to use heuristics (Ferran and Watts 2008). Even during in-person arguments, sex is frequently a heuristic for which attorneys should be interrupted (Patton and Smith 2017, 2020), win at the merits (Szmer, Sarver, and Kaheny 2010), and receive justice votes (Gleason 2020). Thus, even during in-person arguments, sex was shaping outcomes. Perhaps more importantly, gender was already highly salient when the Court shifted to teleconference arguments.

In the late 2010s, scholars and journalists highlighted the gender dynamics at the Court, particularly with respect to interruptions (see, e.g., Jacobi and Schweers 2017; Quinn 2021). Briefly, female attorneys and justices are interrupted earlier and more frequently than men (Feldman and Gill 2019; Patton and Smith 2017). Thus, the impact of gender extends beyond the hierarchical distinction between justices and attorneys; it extends to the dynamics on the Bench itself. By the start of the 2019 term, the Court was keenly aware of gendered differences in interruptions (Feldman 2019; Quinn 2021), and it adopted a new format whereby the first two minutes were reserved from interruption. While this change prompted multiple intriguing scholarly inquiries, any in-depth analysis was precluded by COVID-19. On March 16, 2020, the Court postponed scheduled in-person oral arguments and began teleconference arguments in May (Houston, Johnson, and Ringsmuth 2023).

During teleconference arguments, attorneys made an uninterrupted opening statement, and justices subsequently questioned attorneys directly for a set time period in order of seniority. Throughout the argument, Chief Justice John Roberts moderated the discussion and kept time (Jacobi et al. 2021). In the new format, justices and attorneys could hear but not see each other (Ringsmuth et al. 2023). This change was jarring; counsel and jurists alike were unsure of when to begin speaking without facial expressions indicating when a speaking turn was completed (Bergeron 2021; McGaughey 2021). As justices navigated this new format, gender remained on their minds (e.g., Quinn 2021). Yet gendered patterns quickly emerged. Chief Justice Roberts was more forgiving of male justices exceeding their time and more likely to interrupt female justices (Litman 2020–22). ⁵ Particularly as gender norms operate at an implicit level, it is intuitive that they should shape how justices evaluate attorneys during teleconference arguments.

While gender norms should shape justices’ evaluations of attorneys, it is likely they do not operate in the same way that they did in pre-pandemic in-person arguments. This is because the operation of gender norms is not static; they evolve and change at the Court along with the context (Gleason and Smart 2023). Moreover, work on the Court and in other fields demonstrates that shifting contexts and modalities has consequences for how gender-norm compliance is evaluated (see, e.g., Areni and Sparks 2005; Ferran and Watts 2008; Gleason and Ivy 2021; Guadagno and Cialdini 2007). COVID-19 altered the Court in an altogether new way that ushered in many changes (Johnson et al. 2021; Houston, Johnson, and Ringsmuth 2023; Ringsmuth et al. 2023), not the least of which should be the gender-normative metrics that justices use to evaluate attorneys. Accordingly,

H₁: The shift to teleconference oral arguments increases the importance of gender-normative arguments in predicting attorney success at oral arguments.

Methods

I test my expectations through a quantitative textual analysis of all oral argument transcripts from cases resulting in a signed opinion or judgment during the 2018–20 terms. I choose these terms because they span October 2018 through June 2021 and center on the Court’s March 2020 shift to teleconference arguments. Thus, the sample includes one term argued in person (2018), one by teleconference (2020), and one split between the two modalities (2019). I process transcripts with a Python script to produce separate text documents for each attorney’s utterances (e.g., Gleason 2020). This forms the basis for my unit of analysis: the attorney–justice vote in a given case. I combine these data with the Supreme Court Database, (Spaeth et al. 2021) resulting in 1,177 attorney–justice votes across 73 cases. ⁶ Of these, 766 votes occur during in-person arguments and 411 occur during teleconference arguments. The dependent variable is a binary marker noting whether the justice votes for the attorney’s position (Spaeth et al. 2021).

My main independent variables are two interaction terms examining the extent to which the attorney complies with gender norms in his or her argument. I first note attorney sex, constructed with the honorific used to address the attorney in the transcript (Mr. or Ms.). ⁷ I set this value to 1 for female attorneys and 0 otherwise. Next, I measure attorney gender-norm compliance. Many studies utilize the Linguistic Inquiry Wordcount (LIWC) software (Pennebaker et al. 2007) to extract the underlying properties of speech from political actors’ utterances (see, e.g., Black et al. 2016; Bryan and Ringsmuth 2016; Gleason, Jones, and McBean 2019; Jones 2016; Shaw 2000; Yu 2014). While there are many measures in the LIWC, affect (emotion) is used extensively to study sex and gender at the Court (Gleason 2020; Gleason and Ivy 2021; Gleason, Jones, and McBean 2019; Gleason and Smart 2023).

Lower affect scores are stereotypically more masculine, and higher scores are stereotypically more feminine. ⁸ This measure is constructed using a list of more than 900 affect words and word stems (Pennebaker et al. 2007) that are evaluated implicitly (Pennebaker 2011). To illustrate the affect measure in practice, consider the following excerpts. The first, delivered by Kristen Waggoner in Uzuegbunam v. Preczewski, ⁹ has high affect. The second, by Mithun Mansinghani in McGirt v. Oklahoma, ¹⁰ has low affect. Affect words appear in bold.

[I]n terms of the true concern about being—courts being flooded with frivolous claims for relief, protracted litigation, or avoiding a drain on government resources, the long-standing rule is the rule that best resolves those concerns.

Oklahoma has jurisdiction over the eastern half of the state because it never was reservation land and is certainly not reservation land today. To start, the land was not public land preserved from sale, where title remains with the United States, but instead patented in fee to the Creek Nation.

The affect words in the excerpts are seemingly innocuous and presumably should not impact evaluations. However, gender norms operate implicitly to the extent that even words that are part of standard courtroom vernacular, like “harm” or “certain,” are gendered. This has been noted in fields as diverse as psychology (de Lemus, Spears, and Moya 2012; Gawronski 2003; Rosenthal and Rubin 1978; Uleman, Saribay, and Gonzalez 2008) and the law (Lee 2015–16). Moreover, a number of works exploring sex and judicial behavior note that female attorneys and justices are evaluated differently because of implicit stereotypes about men’s and women’s roles (see, e.g., Feldman and Gill 2019; Gill, Kagan, and Marouf 2017; Patton and Smith 2017, 2020; Szmer, Sarver, and Kaheny 2010). Additionally, several studies have noted that affective language is a consistent predictor of attorney compliance with gender norms, although the precise operation is context dependent (Gleason 2020; Gleason and Ivy 2021; Gleason, Jones, and McBean 2019; Gleason and Smart 2023).

Since changing the context alters the operation of gender norms (e.g., Chaiken and Eagly 1983; Ferran and Watts 2008), I look beyond just affect. Function words are among the most versatile markers of gendered language (Chung and Pennebaker 2007), encompassing prepositions, articles, pronouns, conjunctions, and auxiliary verbs (e.g., “of,” “an,” “she,” “for,” “do”).

Function words carry little linguistic meaning in isolation and were historically considered “junk” words in text analysis (Pennebaker 2011). However, they are key markers of communication style (Phillips 2020; Rosenthal and Yoon 2011). Importantly, their usage varies by sex (Newman et al. 2008); higher levels of function words are female gender normative (Jones 2016; Pennebaker 2011), and people can typically identify an anonymous author’s sex via function word level (Chung and Pennebaker 2007). Accordingly, I also utilize LIWC’s standardized measure of function words. To illustrate function words, consider the same oral argument excerpts from earlier. Function words appear in bold.

[I]n terms of the true concern about being– courts being flooded with frivolous claims for relief, protracted litigation, or avoiding a drain on government resources, the long-standing rule is the rule that best resolves those concerns.

Oklahoma has jurisdiction over the eastern half of the state because it never was reservation land and is certainly not reservation land today. To start, the land was not public land preserved from sale, where title remains with the United States, but instead patented in fee to the Creek Nation.

As earlier, the highlighted function words are seemingly benign and carry little meaning in isolation. However, at an implicit level, justices evaluate attorney arguments based on the extent to which function word usage complies with gender norms (e.g., Jones 2016; Pennebaker 2011). Since my hypothesis is conditional, I create two interaction terms: attorney sex * affective language (function word usage) (e.g., Brambor, Clark, and Golder 2006). ¹¹

I include several control variables suggested by previous literature. Since more experienced attorneys tend to be more successful (McGuire 1995; Ringsmuth, Bryan, and Johnson 2013), I include the experience differential between the petitioner and the respondent by noting each attorney’s prior appearances and then subtracting the opposing attorney’s prior appearances from the attorney’s value. Not all attorneys are created equal; the Court privileges attorneys from the Office of the Solicitor General (Black and Owens 2012; McGuire 1998). Accordingly, I include a measure set to 1 when the attorney is from the Office of the Solicitor General and 0 otherwise. Since opposing the federal government is a daunting proposition, I include a binary measure of whether an attorney opposes the federal government. ¹² Because former clerks tend to be more successful advocates (Feldman 2017a), I include a dichotomous marker set to 1 for attorneys who previously served as clerks. ¹³ Since the Court more often rules for petitioners, I include a dichotomous marker noting whether the attorney represents the petitioner (Spaeth et al. 2021).

Outside parties can enhance attorney success by filing amicus curiae briefs (Box-Steffensmeier, Christenson, and Hitt 2013; Spriggs and Wahlbeck 1997). Accordingly, I include an amicus brief differential by taking the number of amici supporting each attorney and subtracting the number of amici supporting their opponent (Collins 2008; Gleason 2020). Since ideology is a powerful predictor of judicial behavior (Segal and Spaeth 2002), I multiply each justice’s Martin and Quinn (2002) score by –1 for attorneys representing the liberal position and 1 for attorneys representing the conservative position. This creates a measure of ideological congruence. Since female justices may have many of the same experiences as female attorneys and thus may be more sympathetic (Haire and Moyer 2015), I interact a binary marker for female justices with the female attorney measure. More cognitively complex oral arguments are more successful; I include a standardized measure of LIWC’s measure of cognitive complexity (Gleason 2020). Since briefs are an important part of the decision-making process (Corley 2008), I control for the extent to which party briefs are gender normative based on the counsel of record (Gleason 2020; Gleason, Jones, and McBean 2019). Attorneys who are interrupted more are less successful (Johnson et al. 2009). Accordingly, I create an interruption differential following the same logic used for experience differential earlier. Since the initial shift to teleconference arguments was jarring for justices and counsel alike (e.g., Bergeron 2021; Houston, Johnson, and Ringsmuth 2023; Litman 2020–22), I include a binary variable for cases argued in May 2020. Finally, since there are well documented sex differences across issue areas (see, e.g., Boyd, Epstein, and Martin 2010; Songer, Davis, and Haire 1994), I include binary variables denoting Spaeth et al.’s (2021) issue areas. ¹⁴

I run two models, one each for in-person and teleconference arguments. ¹⁵ Given my dichotomous dependent variable, I utilize logistic regression. Because of justice-level idiosyncrasies, I cluster standard errors on the justice. ¹⁶ Next, I turn to a detailed discussion of my results.

Results

The results demonstrate that shifting to teleconference oral arguments alters how gender-normative language impacts justices’ evaluations of attorneys. This is most evident with function words; their usage has little substantive impact on attorney success during in-person arguments, but it is a significant and substantive predictor of attorney success during teleconference arguments. Affective language, however, is significant both before and after the shift. Additionally, while some control variables are significant in both modalities, there are notable differences between models. This underscores context’s role in predicting justice votes and highlights the need to further explore how argument modality shapes attorney success. Before discussing the results of my models, however, it is worthwhile to examine the descriptive statistics.

Table 1 shows that male and female attorneys differ in important ways. Women constitute a distinct minority of oral advocates, making just 14.7% of appearances from 2018 to 2020. This is consistent with previous work examining women arguing in the twenty-first century (Szmer, Kaheny and Sarver 2021). While a greater percentage of women represent the federal government, a greater percentage of women are also opposed by the federal government. Likewise, more women than men represent parties with the lowest levels of party capability. Moreover, the women who do appear are less experienced and are less likely to have previously clerked. However, when women do argue against male attorneys, they are more successful, securing 62.5% of justice votes. This is consistent with previous work noting that women in male-dominated professions are often more qualified and effective than their male peers even if they face longer odds (see, e.g., Anzia and Berry 2011; Haire, Moyer, and Treier 2013).

Table 1. Male and female attorneys at the Supreme Court

While women are some of the most effective advocates, they make fewer appearances relative to their male counterparts. Table 2 shows the attorneys with the most appearances in the 2018–20 terms by sex. Erica L. Ross and Morgan L. Ratner are the most frequent female oral advocates, appearing seven times each. The most frequent male advocates, Paul D. Clement and Jeffrey B. Wall, have double the number of appearances (14). Underscoring the scarcity of female advocates, Ross and Ratner would not even make the top-10 list for men.

Table 2. Attorneys with the most appearances by sex, 2018–20 terms

Table 3 shows linguistic differences in male and female attorneys’ arguments. Ringsmuth et al. (2023) note that the average argument length increased after the shift to teleconference arguments; I find that this increase is largely driven by women. While men speak more than women in terms of word count and percentage of time before COVID-19, women utter more words during their arguments than men and have a greater percentage of speaking time relative to justices during teleconference arguments. In this sense, the format shift in oral arguments increased gender equity. Alternatively, this may indicate that justices are more sensitive to gendered interruptions. Yet in terms of gendered content, women actually use less affective language and function words than men in both modalities. This is consistent with previous work noting that minorities often adopt dominant group behavior to downplay the salience of difference (Gleason, Jones, and McBean 2017; Kanter 1977). Additionally, women tend to cluster around the mean levels of both affective language and function word usage. Since women do not reach high levels of gender normativity in oral arguments, they should be less successful in both formats. I now turn to the results of my predictive models.

Table 3. Linguistic differences between male and female attorneys, pre- and post-COVID

My models, one for in-person arguments and the other for teleconference arguments, appear in Table 4. Since my main argument is conditional, the expectations can only be properly assessed via interaction terms. Moreover, interaction terms cannot be evaluated in the same manner as additive terms; they are best evaluated graphically (Brambor, Clark, and Golder 2006).

Table 4. Predictors of attorney success

Note: Standard errors clustered on justice parentheses.

* p < .05

** p < .01

*** p < .001.

Figure 1 displays the interactions between female attorneys and affective language (function words) for in-person arguments. The first panel depicts affect; the second shows function words. In each panel, the x -axis represents the standardized value of affective language (function words) in attorney arguments. The y-axis shows the marginal effect of attorney sex on success at that particular level of affect (function words). Should the reference line at y = 0 fall within the dashed 95% confidence intervals, the interaction is not significant at that particular level. The interaction for affective language, in the left panel, tracks with previous work. The positive slope indicates that as female attorneys use more affective language, their arguments become more successful. At the mean level of affect, the predicted probability of female attorney success is 0.280. However, increasing to one standard deviation above the mean (e.g., becoming more gender normative) increases the predicted probability of success to .490. Function words, depicted in the right panel, are always significant. However, the effect is flat: moving from the mean value of function word usage to one standard deviation above the mean decreases the predicted probability of success from 0.217 to 0.209, a change of just 0.008. Ultimately, the difference between low function word usage and high function word usage has little substantive impact on female attorney success during in-person arguments.

Figure 1. Impact of affective language and function words on female attorney success during in-person arguments.

Figure 2 is structured the same as Figure 1. However, it depicts teleconference arguments. The left panel is in line with the in-person arguments results presented earlier; female attorneys are more successful when their arguments contain more affective language. Whereas female attorneys’ predicted probability of success at the mean level of affect is 0.168, this increases to 0.275 one standard deviation above the mean. In contrast with the in-person model, female attorneys are more successful when their arguments contain more gender-normative levels of function words. At the mean level of function words, female attorneys’ predicted probability of success is 0.357. However, increasing function word usage by one standard deviation increases the predicted probability of success to 0.588. Thus, while there is continuity from in-person to teleconference arguments for affect, function words only substantively predict female attorney success after the shift. This provides support for my argument that changing modality alters how justices evaluate attorneys’ gender performance.

Figure 2. Impact of affective language and function words on female attorney success during teleconference arguments.

A logical follow-up question is how gender-normative language impacts male attorney success across modalities. To assess this, I reestimate the models with a male attorney variable rather than a female attorney variable. In-person oral arguments are shown in Figure 3; male attorneys are progressively less successful as affective language increases. Recall that lower values of affect are gender normative for men. Moving from the mean level of affective language to one standard deviation above the mean drops the predicted probability of male attorney success by 0.210. Function word usage is always significant, but it is relatively flat and has little substantive effect on attorney success. Indeed, moving from the mean level of affective language to one standard deviation above the mean only shifts the predicted probability of success down by 0.01.

Figure 3. Impact of gender-normative language on male attorney success during in-person arguments.

Men’s results for teleconference arguments, shown in Figure 4, demonstrate that shifting the context alters the importance of function words. The left panel demonstrates that male attorneys are less successful as affective language increases. Moving from the mean level of affective language to one standard deviation above the mean reduces the predicted probability of success by 0.107. The right panel demonstrates that male attorneys are less successful as their function word usage increases. Moving from the mean level of function words to one standard deviation above the mean reduces the predicted probability of success by 0.231. Substantively, the results here are the same as for female attorneys: more gender-normative language leads to more success, although the metrics used to assess gender-normative arguments change with the modality. Critically, though, male attorneys do not need to strike a balance between gender and professional norms; they are coextensive for men. This means that it is easier for men to meet shifting gender-normative expectations than for women.

Figure 4. Impact of gender-normative language on male attorney success during teleconference arguments.

A number of control variables achieve statistical significance. While the results are largely consistent with previous work, there are notable differences between in-person and teleconference modalities. This underscores that the changing context impacts multiple predictors of attorney success. Since logit coefficients are unintuitive, I discuss substantive effects as predicted probabilities. ¹⁷ I begin with the differences between modalities before discussing common predictors.

Attorneys affiliated with the federal government have a predicted probability of success 0.351 higher than their peers during in-person oral arguments. However, during teleconference arguments, government attorneys’ predicted probability of success is 0.370 lower than their opponents. This suggests the federal government struggled to pivot modalities. Former clerks have a predicted probability of success 0.377 higher than their nonclerk counterparts, but only after the shift. This may bespeak a familiarity with the Court that makes pivoting easier. Female attorneys have a predicted probability of success 0.339 higher when arguing in front of female justices, but only during in-person arguments. This suggests that male justices took heed of the increased salience of gender dynamics (e.g., Feldman 2019; Jacobi and Schweers 2017; Quinn 2021). A one standard deviation increase in interruptions decreases the predicted probability of success by 0.109. This is only true before the shift to teleconference arguments, perhaps because the new format removed the “free-for-all” nature of oral arguments (Ringsmuth et al. 2023) and some of the underlying gendered elements therein (e.g., Ferran and Watts 2008). Underscoring the difficulty of shifting formats, attorneys arguing in May 2020 have a predicted probability of success 0.114 lower than their counterparts who did not argue in the immediate aftermath of the shift. This provides evidence for an acclimation period (Bergeron, 2021; Houston, Johnson, and Ringsmuth 2023). Collectively, shifting the context alters several predictors of attorney success.

There is also continuity between modalities. ¹⁸ Attorneys opposing the federal government have a predicted probability of success 0.437 lower than their federal counterparts. Attorneys representing petitioners have a predicted probability of success 0.177 higher than those representing respondents. A one standard deviation increase in amicus brief advantage increases the predicted probability of success by 0.198. A one standard deviation increase in ideological congruence increases the predicted probability of success by 0.144. Finally, a one standard deviation increase in argument cognitive complexity increases the predicted probability of success by 0.031. This demonstrates that while shifting the format resulted in changes, there is still continuity between modalities.

Discussion

While oral arguments have changed historically (e.g., Black, Johnson, and Owens 2018), few changes compare to the sea change ushered in by the COVID-19 pandemic (e.g., Jacobi et al. 2021; Litman 2020–22; Ringsmuth et al. 2023). Teleconference arguments did more than remove justices and attorneys from the room; they changed justices’ evaluative framework for attorneys’ gender performance. Prior to the pandemic, attorneys were more successful using gender-normative levels of affective language. Function words, long a mainstay of gender performance in other contexts (Chung and Pennebaker 2007), do not substantively impact jurists’ evaluations of attorneys. During teleconference arguments, however, affect still predicts success; but so, too, does function word usage. While both female and male attorneys are rewarded for gender-norm compliance, it is more difficult for female attorneys because they must navigate a double bind between competing gender and professional norms (Gleason and Smart 2023; Rhode 1994). The balance becomes more difficult during teleconference arguments. This raises questions about female attorneys’ ability to have a substantive impact on case law binding throughout the country. While this is normatively concerning in any context, it is particularly so at a time when gender is especially salient in the wake of the Dobbs v. Jackson decision ¹⁹ and the #MeToo movement. These results underscore the need for future research to arrive at a more complete understanding of how context shapes evaluations of gender performance.

While I focus here on female attorneys, it is also critical to understand men’s gender performance. I note that male attorneys are more successful when using gender-normative levels of affect and function words, depending on the context. However, male gender norms are more complex than the inverse of female gender norms (e.g., Moss-Racusin, Phelan, and Rudman 2010). A growing literature notes how masculinity shapes political outcomes (see, e.g., McDermott 2016; Palmer and Peterson 2020; Ralph-Morrow 2022). Moreover, context is likewise important in evaluations of masculinity (e.g., Karl and Cormack 2023). Specific to the judiciary, men are evaluated more harshly when litigating from a position of weakness or fear; women are more successful when doing so (Gill, Kagan, and Marouf 2017). More research on masculinity will perhaps shed light on additional layers of complexity about how multiple gender-normative expectations may be jointly at play.

Differences in institutional arrangements provide unique opportunities to further explore how context generally and gender norms specifically shape outcomes (e.g., Escobar-Lemmon et al. 2021; Goelzhauser 2018; Kaheny et al. 2015). Particularly since staffing impacts a host of judicial outcomes (see, e.g., Collins, Manning, and Carp 2010; Harris 2023; Kaheny et al. 2015; Leonard and Ross 2020), it is likely that replicating this study at state supreme, lower federal, and comparative apex courts will provide a more nuanced and complete understanding of how institutions shape gender evaluations during in-person and remote arguments. While this could take many forms, one of the most promising avenues is looking beyond the spoken word.

Like most oral argument studies, I focus on spoken language. However, approximately 60% of all communication is nonverbal (Schill 2012). Much like verbal communication, men and women face different nonverbal gender norms. For example, men typically use expansive and open postures, whereas women employ closed and contractive postures (Cashdan 1998). Women use less assertive hand motions than men (Everitt, Best, and Gaudet 2016), and they are often evaluated harshly for angry facial expressions (e.g., Boussalis et al. 2021). Because the Court bans cameras, it is difficult to impossible to explore most forms of nonverbal communication there. However, cameras were a mainstay at several courts before the pandemic (e.g., Black et al. 2023), and many utilized Zoom during COVID-19 (Bergeron 2021). In tandem with recent innovations in video processing allowing for frame-by-frame analysis of facial expressions (Boussalis et al. 2021) and postures (Everitt, Best, and Gaudet 2016; Neumann, Fowler, and Ridout 2022), scholars should examine how nonverbal gender norms impact outcomes across multiple courts. These studies should not occur in isolation; nonverbal and verbal communication often operate in tandem, particularly for women (Masters et al. 1987). A female candidate might deliver a masculine message while dressed in a stereotypically feminine manner with an image backdrop of families or education (e.g., Carpinella and Bauer 2021). Intuitively, female attorneys likely employ the same strategy (e.g., Cuddy et al. 2015; Tiedens and Fragale 2003).

While justice votes are consequential, they are the culmination of oral arguments. In order to arrive at a full understanding of how gender performance operates at the Court, it is important to look at the dynamics between justices and attorneys during oral arguments. This can be best accomplished via interruptions (Patton and Smith 2020). Interruptions are a key way justices gather information (Black, Johnson, and Wedeking 2012), and justices traditionally interrupt attorneys at will at any point (Wrightsman 2008). No attorney is immune to interruptions, but they are not evenly distributed. Women are generally interrupted more than men (Lindom, Gregory, and Johnson 2017; Patton and Smith 2017), often based on assumptions of competence (Patton and Smith 2020). ²⁰ Since interruptions take control of the narrative from the speaker before she fully articulates her ideas, the greater propensity for female attorneys to be interrupted not only limits success, but also her substantive impact on case law (e.g., Och 2020; Vera and Vidal 2022). While scholars know that women are interrupted more frequently than men at oral arguments (see, e.g., Patton and Smith 2017, 2020), little is known about how the words attorneys use impact interruptions. Future scholars should explore this further.

It is also worthwhile to explore gender norms in other modalities. Since the Court transitioned back to in-person arguments in October 2021, the traditional “free-for-all” argument is now followed by dedicated question time for each justice (Houston, Johnson, and Ringsmuth 2023). This represents a unique “hybrid” modality that, while in person, contains elements of two distinct approaches to oral arguments. At the same time, the dynamics on the Court itself continue to change. With the confirmation of Justice Ketanji Brown Jackson, the Court is at a historic high in terms of gender diversity on the Bench and in the law clerk corps. This in and of itself can change gender norms’ operation (Gleason and Smart 2023; Kaheny et al. 2015) and should be explored.

When attorneys present oral arguments, their words are filtered through a gendered lens. That lens, though, is context dependent. Examining both in-person and teleconference oral arguments, I demonstrate that justices evaluate female attorneys’ arguments using different gender-normative metrics across contexts. This poses few problems for male attorneys; the norms of a good man and a good attorney are coextensive. This is more difficult for women, as these expectations are competing. These results add a new wrinkle to the already delicate balance female attorneys must strike between gender and professional norms in order to substantively shape case law (Gleason 2020; Kanter 1977). My results also raise numerous normative and empirical questions, which I encourage future scholars to explore.