Does Descriptive Representation of Women on the Bench Improve Institutional Trust?

Abstract

An often-used normative argument for increasing judicial diversity is that it will enhance public confidence in courts. This paper tests competing perspectives about whether the descriptive representation of women in the federal judiciary will improve institutional trust, using a nationally representative survey experiment. The findings suggest that, in the post-Dobbs era, descriptive representation for gender on low visibility courts yields a positive effect on institutional trust, but that the magnitude of the effect is quite small, time limited, and restricted to Democrats and to those who already have higher levels of trust.

An extensive scholarship on judicial diversity has shown how the inclusion of women as judges on the federal bench has influenced both the process and outcomes of cases, including disposition times for settlements by district court judges (Boyd 2013), panel decision-making in cases on civil rights and sex discrimination (Boyd, Epstein, and Martin 2010; Moyer 2013), circuit court dissent rates (Haire and Moyer 2015), and the depth and quality of majority opinions (Haire, Moyer, and Treier 2013; Moyer et al. 2021). But does the descriptive representation of women in prestigious federal judgeships have a broader effect on institutional trust for courts?

It is commonplace to see legitimacy, institutional trust, and enhanced public support invoked as rationales for increasing women’s representation on the bench (Malleson 2003; Grossman 2011, 2016). For instance, the organization Courts Matter says on its website, “Federal courts should reflect the communities they serve. Bringing diverse experiences and perspectives to the bench allows judges to make better-informed decisions, in addition to increasing public confidence in their rulings.”¹ Generally, the type of legitimacy that is meant in such statements is sociological legitimacy (Bodansky 2013); that is, the public’s perception that the court is legitimate and has “justified authority” to rule (Grossman 2011). Outside the judicial context, there is evidence that women’s representation in legislative and executive offices affects men and women’s confidence in government (Karp and Banducci 2008; Barnes and Taylor-Robinson 2018). However, the institutional features of courts differ in important ways from legislative bodies and executive offices, such that those findings might not extend to the judicial setting. Moreover, it is possible that backlash effects could be seen in response to the inclusion of underrepresented groups (Scherer and Curry 2010; Scherer 2023).

In this paper, I assess the competing perspectives about the relationship between descriptive representation for gender on the bench and public trust in the judiciary. Rather than focus on the Supreme Court, I instead expand my lens to include the courts that are the workhorses of the federal judiciary: US district and circuit courts. These institutions resolve tens of thousands of disputes each year and are the last word for the vast majority of litigants (Bowie, Songer, and Szmer 2014). Their decisions on many controversial issues, such as immigration policy, abortion, and gun control, increasingly receive national news coverage (Marimow 2023; Whitehurst and Richer 2023). They have also increasingly diversified across several dimensions, including gender and race (Solberg and Diascro 2020; Harris, Moyer, and Solberg 2022). Yet we know very little about how the gains in women’s judicial representation across the federal judiciary are assessed by the public.

To fill this gap in the literature, I leverage a survey experiment fielded in 2022, after Dobbs v. Jackson Women’s Health. The findings from this survey and a robustness test one year later indicate that descriptive representation can have a positive benefit on institutional trust in courts – but these effects are modest, are time bound, and may be limited to those who already hold high levels of trust. While I find no evidence that men and women respond to women’s descriptive representation differently, Democrats and Republicans do appear to respond differently to descriptive representation in the post-Dobbs era.

Women’s representation in political institutions and its impacts

Before developing an account of how gender representation on the federal judiciary affects public assessments of the courts, it is essential to acknowledge the rich literature on the impact and importance of women’s representation in political institutions (Pitkin 1967; Mansbridge 1999). Although this work considers elective and appointive institutions across the world (e.g., Sapiro 1981; Barnes and Burchard 2012; Jalazai 2018; O’Brien and Piscopo 2018; Stauffer 2021), the bulk of the research in this area examines legislatures and executives rather than courts (but see Nelson 2015; Shortell and Valdini 2021; Lee, Solberg, and Waltenburg 2022). Within this vast literature, several themes emerge with respect to how women’s representation in political institutions affects public support for or trust in those institutions.

One group of studies fails to find a clear relationship between women’s representation in political institutions and public attitudes or political engagement. For instance, using a regression discontinuity design, Broockman (2014) finds that women state legislative candidates have no effect on American women’s political engagement and ambition. This is similar to the conclusions drawn a decade earlier by other leading scholars of gender and politics (Lawless 2004; Dolan 2006).

Another group of studies finds positive effects on public assessments of government linked to women’s representation in government. This can further be disaggregated in terms of work that finds effects for both men and women versus work that finds effects only for women. Barnes and Taylor-Robinson (2018) find that women’s representation among cabinet ministers improves confidence in government for men and women; Karp and Banducci (2008) find a similar effect for women in legislatures with respect to satisfaction with government (see also Schwindt-Bayer 2010; Clayton, O’Brien, and Piscopo 2019; Stauffer 2021). Other studies show that increasing women’s presence in political institutions improves only women’s engagement in politics (Atkeson and Carrillo 2007; Reingold and Harrell 2010; Barnes and Burchard 2012). Pipeline effects, whereby “role models” in visible public offices encourage other women to seek office, have also helped shed light on the underpinnings of the political ambition gap and how to address it (e.g., Ladam, Harden, and Windett 2018).

What is the causal mechanism behind women’s officeholding on the one hand and public attitudes toward political institutions on the other? Stauffer highlights the crucial role of visibility, noting “research has shifted toward thinking about the symbolic effects associated with women serving in highly visible positions—such as governors, senators, heads of state, and cabinet posts—where citizens are far more likely to be familiar with the representative” (2021, 1229). Both Zetterberg (2009) and Morgan and Buice (2013) clarify in their theory that it is women’s visible presence in offices that animates the causal linkage between representation and political engagement. Having more women in visible posts may send a message to the public that the institution is open and responsive to many interests (e.g., Schwindt-Bayer and Mishler 2005; Schwindt-Bayer 2010). It may also trigger gendered stereotypes about women as leaders being honest and less corrupt than men in a way that extends these traits to the institution, enhancing public confidence that the institution will engage in fair processes when making decisions (e.g., Tyler 2003; Watson and Moreland 2014; Clayton, O’Brien, and Piscopo 2019; Kao et al. 2022)

Arguments for treating the judicial case differently

But there are real reasons to question whether the conclusions of research focused on descriptive representation in legislative and executive institutions could extend to the judicial context. First, to which courts are we referring? Courts vary greatly with respect to visibility; the public certainly has a very different awareness of the US Supreme Court and state courts, for instance (Benesh 2006). Indeed, with respect to visibility, research on women’s representation in legislative and executive posts emphasizes that the causal linkage between women’s representation and public attitudes turns on the high visibility of the posts occupied by women (Zetterberg 2009; Morgan and Buice 2013; Fridkin and Kenney 2015; Barnes and Taylor-Robinson 2018).² But because of the low visibility of US district and circuit courts to the general public, any positive effect of descriptive representation in those institutions would be unlikely to extend to the public at large, who would be unaware of the composition of those courts.³

Second, we should ask, support from whom? If the answer is the broader public, then it will also be important to assess whether backlash dynamics are in operation (Lee, Solberg, and Waltenburg 2022). If an increase in women’s representation results in a decrease in men’s presence in the judiciary, would this undermine positive assessments by men, similar to what Scherer (2023) documented in a 2010 survey experiment?⁴ Indeed, Lee, Solberg, and Waltenburg (2022) find that men’s diffuse support for state high courts declines when they are shown a female chief justice. In asking respondents about hypothetical high courts (i.e., not in a solely US context), Shortell and Valdini (2021) find that at higher levels of hostile sexism, respondents are less likely to agree that a court can exercise effective judicial review. And while men generally score higher on hostile sexism indices (Winter 2022), these views are also held by conservative women in ways that translate to political attitudes and behavior (Cassese and Barnes 2019).

Research on the (highly visible) US Supreme Court suggests that the impact of gender representation on the bench there may be limited at best. Extending a body of work that focuses on how source cues related to Supreme Court decisions affect public support for judicial rulings (e.g., Boddery and Yates 2014; Clark and Kastellec 2015), Boddery, Moyer, and Yates (2019) explore how the identity of the majority opinion writer affects public agreement with Supreme Court decisions.⁵ The effect of attributing a majority opinion (about criminal procedure) to Sandra Day O’Connor was positive in terms of support – but only for respondents from areas where women’s political representation and economic status was high.⁶ In another experiment focused on support for Supreme Court nominees, Kaslovsky, Rogowski, and Stone (2021) find evidence of ingroup favoritism related to race (and partisanship), but no effect for gender representation among any subgroup. Along similar lines, Chen and Savage (2024) find that, although Democratic and Independent women said that they preferred a more diverse and representative court when asked in a survey item, in a separate conjoint experiment neither women nor men evaluated same gender Supreme Court nominees any better, instead focusing more on political congruence. In summarizing a series of experiments about judge identity and court rulings, the authors conclude that the public is more concerned with policy outcomes than on the identity of the judge associated with the policy outcome (including female judges issuing pro-female decisions in gender-salient cases).

Descriptive representation for gender and support for federal courts

With this backdrop in mind, it is important to acknowledge that there is an ongoing debate in the judicial literature about the extent to which assessments of judicial legitimacy turn on ideological alignment with court decisions (Bartels and Johnston 2013, 2020; Christenson and Glick 2015) or are a function of our socialization to the positive symbols and attributes of legal institutions and processes (e.g., Gibson and Caldeira 2009; Gibson and Nelson 2016).

In this paper, I focus on a concept (institutional trust) that is related but distinct from “legitimacy” as it has been generally used in the political science literature.⁷ In doing so, I follow the lead of descriptive representation scholars outside of the judicial context (Jamieson and Hennessy 2006; Barnes and Beaulieu 2014; Barnes and Taylor-Robinson 2018; Clayton, O’Brien, and Piscopo 2019). Below, I lay out four hypotheses that allow us to assess distinct and sometimes competing claims suggested by the descriptive representation literature about the impact of women’s judicial presence on public trust in the federal courts.

First, if gendered associations about women being less corrupt than male politicians improve public trust in legislative and executive institutions, this could also extend to courts and suggests an across-the-board increase in trust for men and women alike. This is consistent with the procedural justice perspective (Tyler 2003), insofar as the public might view a gender representative bench as signifying fair processes for making decisions.

General Benefit Hypothesis: Trust in the federal courts will be higher when women are equitably represented on the federal bench relative to when women are underrepresented.

A competing perspective suggests that the public is less concerned with descriptive representation than with judicial outputs (e.g., Chen and Savage 2024), or is less concerned with descriptive representation for gender in particular (Kaslovsky, Rogowski, and Stone 2021). As such, women’s relative presence on the bench will not impact the public’s trust in the judiciary.

Indifference Hypothesis: The public’s trust in the federal courts will be unaffected by the relative presence of women on the federal bench.

Lastly, some work suggests that men and women will react differently to cues about the gender composition of the bench, with men more likely to negatively assess courts that include more women (Lee, Solberg, and Waltenburg 2022; Scherer 2023) and women more apt to respond favorably (Scherer 2023).

Backlash Hypothesis: Men’s trust in the federal courts will decrease when women are equitably represented on the federal bench compared to when they are underrepresented.

Targeted Benefit Hypothesis: Women’s trust in the federal courts will be higher when women are equitably represented on the federal bench relative to when women are underrepresented.

Data and methods

To test these expectations, I conducted a nationally representative survey in partnership with Qualtrics, who fielded the survey on panels of respondents aged 18 and older that the company maintains. Data used in Study 1 was collected between October 28, 2022 and November 7, 2022. A second wave of the survey was fielded between May 1, 2023 and June 16, 2023 to evaluate whether the findings from Study 1 were still evident.⁸ The quota sample was designed to match the characteristics of the population on gender, age, race/ethnicity, and educational attainment based on 2020 Census estimates. Following American Association for Public Opinion Research (AAPOR) best practices, survey rake weights were utilized to minimize any remaining discrepancies between sample and population.

A number of quality control measures were employed, such as the compensation of participants, the use of attention checks (Berinsky et al. 2021), guards against “speeding,” and overall length of the survey instrument (15 minutes).⁹ Qualtrics also uses reCAPTCHA and AI tools to detect fraudulent respondents and “flat-lining.”

The dependent variable, Trust in federal courts, is an adaptation of a long-running question about the Supreme Court on the Annenberg Public Policy Center survey (see also Bartels and Johnson 2013): “the federal courts can usually be trusted to make decisions that are right for the country as a whole.”¹⁰ It focuses attention on the federal judiciary, as opposed to state courts or courts in general.

Respondents were assigned to either the Women Underrepresented or the Equitable Representation conditions (see Appendix A for vignette text), which were adapted from the Scherer and Curry (2010) experiment on race.¹¹ The Women Underrepresented vignette told respondents that the percentage of all federal judges who are women was much lower (30.4%) than the percentage of women graduating from law school, while the Equitable Representation condition stated that the percentage of women in all federal judgeships was 49.4%, the same as the percentage of women in law school. To test whether the relationship between women’s presence on the federal bench and support for courts is conditioned on respondent gender, an interaction term is also included.

Those who know more about the Supreme Court, believe that the Court makes decisions fairly and objectively, and hold more positive attitudes toward that court are likely to have more positive assessments of judicial power and trust in the federal judiciary (Tyler 1988; Gibson and Caldeira 2009; Bartels and Johnston 2013; Boddery, Moyer, and Yates 2019). As such, controls were included for knowledge about the Supreme Court and Supreme Court support, measured by a feeling thermometer. Following past work, I also control for past court experience as a juror, litigant, witness, or other court role (Cann and Yates 2016), which can condition judicial support (Benesh 2006). Views about gender are likely to impact any assessment of women’s presence on the bench (Shortell and Valdini 2021), so controls for hostile sexism attitudes are included (Cassese and Barnes 2019; Archer and Clifford 2022).¹² Because the media attention to the Dobbs decision to overturn Roe v. Wade may have influenced attitudes about the Supreme Court or the judiciary in general, another variable measured the level of awareness about the Dobbs decision, asking respondents if they had heard “a lot,” “a little,” or “nothing at all” about the ruling.¹³ Lastly, controls for education, race/ethnicity, and partisanship are included. Higher educational attainment is associated with more favorable views of courts at the federal and state levels (Benesh 2006; Gibson and Caldeira 2009), so a positive relationship is expected. With respect to race and ethnicity, Black Americans generally hold less positive views about the American justice system than white Americans (Hurwitz and Peffley 2005), but race has generally not been found to be a significant predictor of legitimacy (Benesh 2006; Bartels and Johnston 2020). Latinos exhibit lower levels of Supreme Court knowledge than other groups (Pedraza and Ura 2021), but research is mixed on how Latinos respond to descriptive representation on the Supreme Court (Evans et al. 2017; Chen and Savage 2024). The impact of partisanship on Supreme Court approval has shifted in the post-Dobbs environment. To illustrate, Levendusky et al. 2024 find that post-Dobbs, partisanship has become a stronger predictor of a range of views on the Supreme Court, including trust and legitimacy, with Democrats’ views on legitimacy taking a nosedive and Republicans feeling more favorable. For more detail on coding of all variables, please see Table 1.¹⁴

Table 1. Description of Variables

Analysis

The survey experiment (N = 1,249) was fielded in the period just before the November 2022 election, as part of a larger survey on American political attitudes leading up to the midterm election. A quarter of respondents disagreed with the statement that federal courts can usually be trusted to make decisions that are right for the country as a whole, and about one-third of respondents appeared to be agnostic on the point.

To assess which of the competing hypotheses has more support, Figure 1 displays a coefficient plot of the results for the ordered logit model estimating support for trust in federal courts. Panel A shows the results from the baseline model, without the interaction term, allowing us to evaluate the competing hypotheses (General Benefit versus Indifference). For ease of interpretation, the equitable representation condition is the reference category, which means that the direction on the coefficient for the underrepresented condition should be positive, if the General Benefit Hypothesis is supported. Here, the coefficient for Equitable Representation is positive and significant, supporting the contention that increased women’s representation translates into enhanced trust in the federal courts. This allows us to reject the Indifference Hypothesis.

Figure 1. Ordered Logit Model of Trust in Federal Courts (Study 1) (see full table of results in Appendix B).

But what is the magnitude of the effect? Exploring this finding further, predicted probabilities were calculated at each of the outcomes of the dependent variable (support for the statement “federal courts can be trusted”), holding the other covariates at their means. This is graphed in Figure 2. By breaking out the probability of each dependent variable category across each experimental condition, this allows us to see more clearly where increasing women’s representation affects trust.

Figure 2. Predicted Probabilities of Trust in Federal Courts (Study 1).

Figure 2 shows that the increase in trust for equitable representation only occurs for those individuals who agree, or, to a lesser extent, strongly agree with the “federal courts can be trusted” statement. The other outcome categories show minimal change across conditions. In fact, when looking at the “disagree” line, we see that trust in federal courts is slightly higher in the underrepresented condition than in the equitable representation condition. When taking these nuances into account, this does not offer unqualified support for the General Benefit perspective.

Turning to the control variables, we see that Republican party identification and hostile sexism attitudes were negatively related to trust in the federal courts. While other work has found that Republicans are generally more supportive of the Supreme Court, here the question asks respondents about “federal courts,” which may be an important distinction.¹⁵ Like other work (Shortell and Valdini 2021), we see that higher levels of hostile sexism are associated with more negative views of courts. Support for the Supreme Court and belief in procedural fairness were positive and significant predictors (as expected), but Supreme Court knowledge and prior court experience failed to reach significance. Race, ethnicity, and awareness of Dobbs were also unrelated to trust in the federal courts.

Next, Panel B of Figure 1 displays the results of a model that interacts respondent gender with the experimental condition, to see whether men and women differ with respect to how descriptive representation affects trust. This allows us to assess whether there is support for either the Backlash or Targeted Support Hypotheses. However, the coefficient for the interaction term fails to reach statistical significance, offering no support for a differential effect that varies by gender.¹⁶ The results for the other control variables are unchanged from the baseline model.

To recap, the results from the 2022 survey show that, with respect to the impact of women’s descriptive representation in the federal courts, there is limited support for the General Benefit Hypothesis and no evidence of backlash from men or a positive effect for women. But given the turmoil and public opinion impacts immediately following the Dobbs decision (Levendusky et al. 2024), it is possible that respondents’ perspectives on federal courts during this time might not endure.¹⁷ As a robustness test, Study 2, fielded in May and early June 2023, replicated the survey experiment from Study 1 (N = 1569). (It is important to note that these were a different set of respondents from Study 1.)

Figure 3 displays a coefficient plot of the ordered logit results from Study 2, with Panel A showing the baseline model results and Panel B displaying the interaction model. Starting with Panel A, we can assess whether there is support for the General Benefit Hypothesis or the Indifference Hypothesis. Unlike the results from the previous year, we see no statistically significant relationship between the representation of women judges and enhanced trust in the federal courts, so the General Benefit Hypothesis is not supported.

Figure 3. Ordered Logit Model of Trust in Federal Courts (Study 2) (see full table of results in Appendix B).

Before we can conclude that the Indifference Hypothesis is supported, more is required. Because non-significance for the treatment conditions does not necessarily mean the absence of an effect, equivalence testing is a useful method to assess whether there is, in fact, the absence of a meaningful effect (Lakens, Scheel, and Isager 2018).¹⁸ Equivalence testing using the TOST procedure allows for the researcher to determine whether an effect is at least as extreme as the smallest effect size of interest (SESOI). Using a Cohen’s d value for a small effect size corresponds to a difference in Likert scores of .2 (slightly less than a quarter of a point on the scale). With this SESOI specified, the Welch modified two-sample t-test reveals that we can reject the null hypothesis of a meaningful effect (at p < .05). (Identical results are obtained with a Cohen’s d value for a medium effect size of 0.5.) This allows us to conclude with more confidence that respondents’ assessments of federal courts in the full sample were not meaningfully impacted by their understanding of how well women are represented on the federal bench. As such, this finding aligns with the perspective articulated by the Indifference Hypothesis.

Looking at the control variables, we see that Black respondents do show lower levels of trust than white respondents, unlike in the prior wave of the survey. Republican partisan identification is also negatively related to trust in the federal courts, similar to Study 1. Along with college education, Supreme Court support and higher beliefs in procedural fairness were positively associated with trust in the federal courts. With respect to gender, both Female and Hostile Sexism are negatively signed, but fail to reach significance.

Panel B of Figure 3 shows the results of a model that tests for a conditional relationship between descriptive representation and gender and reveals that the interaction term fails to reach significance. Across both waves of the survey, then, there is no support for either Targeted Representation Hypothesis or the Backlash Hypothesis.

The role of partisanship in evaluating gender equality on the bench

Given the widening partisan gap in assessments of the Supreme Court (Levendusky et al. 2024) and differing party positions on judicial diversity (Scherer 2023), it seems possible that Republicans and Democrats might respond in different ways to women’s equitable representation on the federal bench. To assess this possibility, I re-ran the baseline models from Study 1 and 2 on a subset of Republican and Democratic respondents, respectively. As the coefficient for Equitable Representation in Figure 4 shows, we can see that, in both waves of the survey, Democratic respondents react in ways consistent with the General Benefit Hypothesis, while there is no significant effect seen in the Republican only models. Interestingly, respondent gender is also not significant in any model for either party. (Full results appear in the Appendix.) Similar to Study 1, when predicted probabilities are calculated across the categories of the dependent variable (Figure 5), the only increase in trust comes from those who already agree that the federal courts can be trusted, with no effect or even a slight decrease in trust in the other response categories. Substantively, this suggests a less than rosy view about the ability for women’s representation to enhance trust in the courts, even for members affiliated with a political party that prioritizes descriptive representation for gender.

Figure 4. Selected Output from Ordered Logit Models by Partisanship (Studies 1 and 2) (see full table of results in Appendix C).

Figure 5. Predicted Probabilities of Trust in Federal Courts for Democrats.

Discussion

Drawing from work on legislative and executive institutions, scholars of judicial diversity regularly invoke enhanced public support as a normative rationale for the important goal of diversifying the bench. These arguments sometimes make empirical claims that point to an assumed general benefit for institutional trust across all members of the public and at other times, suggest a benefit will be seen only among the minoritized group being represented. However, the lack of visibility for lower federal court judges in particular, along with the absence of an electoral connection, raises serious questions about whether we should expect to find similar results in the judicial context.¹⁹

To assess these competing perspectives on how women’s presence on the bench affects the public’s trust in federal courts, I fielded one survey experiment in 2022 (after the Dobbs decision) and a follow-up in 2023. Overall, the evidence in support of a general benefit of gender diversity on institutional trust is modest at best. In the first study, equitable representation of women is associated with increased trust in the federal judiciary, but the positive effect on trust is only seen for those who are already very trusting of federal courts, suggesting that increasing women’s representation on the bench is unlikely to bring back disaffected citizens into the fold in terms of institutional support. A year later, this effect is gone. Delving deeper into the data, we see partisan asymmetry playing out, with Democrats responding positively to the equitable representation of women and Republicans not showing a significant response. Yet even among Democrats, the substantive effect is very small and limited to those who already agree that the federal courts can be trusted. This suggests that improving the representation of women on the bench will not be a panacea for enhancing public trust in the federal courts.

Unlike Scherer (2023), I do not find evidence of either a backlash effect among men or a positive effect among women. One possibility for the difference in findings is that the survey here was implemented in the post-Dobbs era, where partisan evaluations of the judiciary have diverged substantially beyond what was seen in the past.²⁰ My findings do align more closely with those from Chen and Savage (2024), whose experiments were fielded in 2020 and 2021.

Of course, there are limitations to this study. The survey experiment was not designed to assess the degree of support for individual judicial nominees, so it cannot address whether support for a particular judicial nominee is conditioned on the existing composition of the bench, though existing research (Solberg and Stout 2021) casts doubt about such a relationship.²¹ It also cannot address questions of whether the public would react differently to particular judicial nominees (e.g., Amy Coney Barrett versus Ketanji Brown Jackson) or substantive representation by female judges. In addition, respondents were not asked to evaluate a situation in which women were overrepresented on the bench relative to their share of the legal profession, which might be a more favorable context to observe differential effects for men and women. Lastly, the wording of the dependent variable (“federal courts”) could be tapping primarily into sentiment toward the Supreme Court, as we do not know what the general public understands “federal courts” to include and how much weight they might assign to the Supreme Court versus lower courts. The finding that Republicans hold a more negative view of “federal courts” than what previous research has found with respect to the Supreme Court also points to questions about how respondents might have interpreted the question.

Going forward, there are several interesting directions for future research. For instance, following Stauffer (2021), it may be useful to ask the public about what they think women’s representation on the lower federal courts is. This is because Stauffer finds that perceptions of women’s representation levels (regardless of the actual representation numbers) are what drive public support for the institution.²² Because the two studies reported here manipulated what respondents were told about the extent of women’s presence in the federal courts, it is not possible to assess how this compared to what respondents initially believed.

Another way of thinking about descriptive representation and public support for the judiciary is to move away from a focus on the general public and instead consider only litigants or affected parties. Perhaps the gender of the judge(s) would affect whether a losing litigant would accept the outcome as legitimate, though taskforce reports, interviews (e.g., Means 2023), and surveys (Gill, Lazos, and Waters 2011; Collins, Dumas, and Moyer 2017, 2018) of practicing attorneys and judges reveal discouraging trends about the prevalence of disrespect toward women judges (especially women of color) from the attorneys and parties who appear before them.²³

Future research could also explore whether descriptive representation in the federal courts has any impact on the political ambition of women attorneys, as researchers have found in other contexts. Indeed, attorneys who practice in federal court are the group most likely to have awareness of the composition of the bench and thus may form their perceptions about their “fit” for elected and judicial office because of what they encounter in their day-to-day practice. With respect to the public, it is possible that respondents’ perceptions are influenced by the sociopolitical environment for women in their state (Windett 2011). Previous work suggests that this environment may affect how the public evaluates women in the judiciary or women judges’ contributions (Boddery, Moyer, and Yates 2019).

Taken together, the findings about gender and judicial representation do not show an unqualified benefit on public attitudes toward the federal judiciary and cast doubt on the promise of descriptive representation to address public deficits in institutional trust. While more work needs to be done to untangle the causal mechanism at play, these results do suggest that, when it comes to low visibility judicial institutions, advocates for increasing women’s representation on the courts should consider exploring other normative rationales besides legitimacy, such as justice, equality, or representation (Kenney 2002, 2017; Torbisco-Casals 2016), when making the case for women on the bench. Indeed, there is a large body of work that documents how the inclusion of women on a range of courts – trial and appellate, domestic and international – has shaped their processes and outcomes (Boyd 2016; Johnson and Reid 2020; Leonard and Ross 2020; Voeten 2020). These empirical findings suggest that there are many good reasons to promote gender equity on the bench. In an environment in which women equal or outnumber their male peers in law school but still lag far behind in obtaining prestigious federal judgeships, a lack of public awareness or public indifference should not be an excuse to allow such a disparity to persist.