The legal profession’s history of discrimination against women and people of color is well documented. But women of color face unique hurdles to equal professional treatment (see, e.g., Blackburne-Rigsby 2010; Burleigh 1988; Collins, Dumas and Moyer 2017; Smith 1997). An American Bar Association (2006) survey reports, for example, that lawyers who are women of color are more likely to face workplace harassment, receive insufficient professional mentoring, be denied high-profile client assignments, and receive negative performance evaluations. As one state judge put it, “Women of color in the justice system of our nation, whether judge, attorney, or court staff suffer a double disadvantage—gender discrimination and ethnic bias” (Aranda 1996, 29).

Although the study of state judicial diversification is thriving (e.g., Arrington 2018; Bratton and Spill 2002; Goelzhauser 2016; Graham 1990; Hurwitz and Lanier 2003; Reddick, Nelson, and Caufield 2009), much of it emphasizes single-axis representation, particularly the separate seating of women and people of color. In contrast, intersectionality research “emphasizes the interaction of categories of difference (including but not limited to race, gender, class, and sexual orientation” (Hancock 2007, 63–64). Scholars have noted in other contexts that the double disadvantage women of color experience conditions the relationship between institutions and representation.^[1] As Scola concludes in a study of intersectional legislative representation, “The process seems to be more complex than what is captured by race/ethnicity or gender separately” (2013, 344). Thus it is imperative to examine whether institutional design choices differentially impact intersectional representation.

This chapter considers the relationship between judicial selection institutions and the representation of women of color on state supreme courts. It begins with an overview of intersectionality and the law and continues with a theoretical consideration of the connection between selection institutions and diversification. The subsequent empirical analysis offers two contributions. First, I highlight the groundbreaking women of color who diversified state supreme courts—a group of people who have largely not been recognized for their achievement. Second, using data from 1960 through 2016, I examine whether selection institutions are associated with intersectional differences in seating new state supreme court justices. The results suggest that women of color are more likely to be seated through appointment mechanisms. The results are similar for men of color, but white men are more likely to be seated through elections. Selection system differences are not associated with changes in the probability of seating white women. These findings contribute to the nascent literature on intersectional judicial representation (see, e.g., Collins and Moyer 2008; Haire and Moyer 2015; Hurwitz and Lanier 2008, 2017; Solberg and Diascro 2019). They also have important policy implications for designing judicial selection institutions.

Diversity, Intersectionality, and the Courts

Understanding judicial diversification is important for several reasons. As an initial matter, the descriptive representation of women and people of color in political institutions can increase trust, engagement, and perceptions of legitimacy (see, e.g., Bobo and Gilliam 1990; Broockman 2014; Gay 2002; Mansbridge 1999; Reingold and Harrell 2010; Wolak 2015).^[2] With respect to the courts, for example, support for the judiciary increases among Black respondents with the number of Black judges (Scherer and Curry 2010). And regarding intersectional judicial representation in particular, scholars have suggested that increasing the number of women of color on the bench will “hold strong symbolic meaning, instill greater confidence in the courts for all the litigants who come before them, and increase general confidence in the system of democracy in the country” (Fricke and Onwuachi-Willig 2012, 1542).

Greater representation can also influence judicial decision-making. Crenshaw’s (1989) seminal article on the difficulty of proving intersectional discrimination claims highlights how a lack of perspective can shape legal doctrine. In one case involving alleged employment discrimination brought “on behalf of black women,” a federal district court held that plaintiffs were “not . . . allowed to combine [gender- and race-based] statutory remedies to create a new ‘super-remedy’ which would give them relief beyond what the drafters of the relevant statutes intended” (DeGraffenreid v. General Motors 1976, 143). Thus plaintiffs could raise separate gender and race discrimination claims but could not base a cause of action on intersectional discrimination. As a result, evidence of nondiscrimination against women (even if that evidence involved, for example, only white women) could be used to defeat gender discrimination claims, and evidence of nondiscrimination against people of color (even if that evidence involved, for example, only Black men) could be used to defeat race discrimination claims.

Empirical evidence also suggests that women of color fare worse in court. One early study of employment discrimination claims in California finds that win rates and damage awards are lower in cases brought by Black women than by Black men or other women (Oppenheimer 2003, 544–45). Analyzing a national sample of federal equal employment opportunity cases from 1965 through 1999, another study finds that intersectional discrimination claims are increasing but only about half as likely to succeed; moreover, women of color are less likely to win (Best et al. 2011). Furthermore, the results indicate that women of color are the least likely gender-race intersection to win. Although win rates are similar for men and women generally, the “intersectionality penalty” (1009) results in women of color faring worse than expected when examining the data through a single-axis framework.

There is mixed single-axis evidence on whether judges who are women (e.g., Gryski, Main, and Dixon 1986; Martin and Pyle 2005; Songer, Davis, and Haire 1994) or people of color (e.g., Farhang and Wawro 2004; Scherer 2005; Welch, Combs, and Gruhl 1988) decide cases differently. At the federal level, Black circuit court judges are more likely to support affirmative action policies, as are white judges assigned to a panel with a Black judge (Kastellec 2013). A study exploring the relationship between gender and judicial decision-making on federal circuit courts finds few differences between men and women overall but notes that women are more likely to support plaintiffs in gender discrimination cases, as are white judges assigned to a panel with a woman (Boyd, Epstein, and Martin 2010). At the state level, supreme court justices of color are more likely than white justices to overturn convictions (Bonneau and Rice 2009).

Race and gender differences are also manifest in other aspects of judicial behavior. For example, there are gender-based differences in the types of experience men and women bring to the bench (Martin 1990), state apex court justices who are women or people of color are more likely to dissent in cases that are salient to members of those groups (Szmer, Christensen, and Kaheny 2015), men and women have different opinion-writing styles (Davis 1992), federal circuit court panels that include women or people of color produce opinions with different characteristics regardless of whether a woman or person of color writes the opinion (Gill and Hall 2015), and judges of color are more likely “to adopt a nonmainstream approach” and “experiment with alternative theories” in sentencing guideline cases (Sisk, Heise, and Morriss 1998, 1459). Although much of this literature adopts a single-axis perspective, findings of difference are prominent.

As more women of color join the bench, scholars are increasingly studying the link between intersectionality and substantive judicial representation.^[3] One study of judicial behavior in federal circuit court criminal cases finds that women of color are more likely than judges with other gender-race combinations to vote for defendants (Collins and Moyer 2008). Another study reports that Black women exhibit distinct tendencies to dissent from majority dispositions on state high courts (Szmer, Christensen, and Kaheny 2015). And an analysis of opinion content finds that intersectional representation has important implications for how many points of law judicial opinions cover (Haire, Moyer, and Treirer 2013). Furthermore, Black women are more likely than white men to vote liberally in gender discrimination cases decided by federal circuit courts (Haire and Moyer 2015).

Selection Institutions and Judicial Diversification

The literature on state court diversification emphasizes the potential importance of judicial selection institutions. While Article III judges are seated by presidential nomination and Senate confirmation, state judicial selection mechanisms vary considerably. The primary selection methods include unilateral gubernatorial or legislative appointment, commission-aided gubernatorial appointment, and contestable nonpartisan or partisan elections. Along with diversity, selectors emphasize characteristics such as experience, temperament, and ideology. Given that selection institutions are designed for myriad reasons, there is no strong a priori theoretical justification for rank ordering them with respect to success in promoting diversity—much less the representation of women of color in particular. The existing empirical evidence on the relationship between selection institutions and diversification offers some a posteriori guidance but is mixed.

Under commission-aided gubernatorial selection, interested lawyers apply to fill a judicial vacancy, a commission typically comprising lawyers and nonlawyers selects a short list of nominees from the applicant pool, and the governor appoints one of the nominees. Early proponents of “merit selection” argued that it would prioritize qualifications over political connections. After President Carter experimented with commissions to help diversify the federal bench, supporters argued that use of the system would enhance representation for historically disadvantaged groups at the state level. The core argument is that emphasizing qualifications over political connections increases opportunities for marginalized groups. Capturing the prevailing sentiment among many merit selection proponents, one scholar argues, “There is no question but that the merit selection system affords greater opportunities for women and minorities to find their way to the bench” (Krivosha 1987, 19). Although arguments concerning merit selection rarely focus on intersectional representation in particular, perhaps the broader emphasis on diversity increases opportunities for women of color.

Proponents of other selection systems are typically less explicit about whether or how their favored mechanisms impact diversity. A possible benefit of unilateral elite appointment systems is that selectors are free to seek out prospective judges from the broader pool of qualified attorneys. Whereas choice is constrained under merit selection by applicant and nominee pools, and in elections by who runs for office, unilateral appointers can reach out to individuals who may not otherwise seek judgeships. This may be particularly important for attracting members of doubly disadvantaged groups. Unilateral appointment systems also facilitate holding stakeholders accountable for a lack of diversity. In merit selection systems, for example, it may be difficult to observe the extent to which a lack of diversity is due to pool imbalance, nomination decisions, or appointment decisions. In election systems, responsibility for a lack of diversification is likely to be more diffuse. Conversely, critics of unilateral elite appointment contend that unconstrained selectors use their power for patronage, in which case women of color may be disfavored due to their historical political underrepresentation.

Numerous arguments have been made for (e.g., Bonneau and Hall 2009) and against (e.g., Geyh 2003) judicial elections. However, this debate typically emphasizes issues such as voter knowledge, campaign dynamics, and legitimacy rather than diversity (see, e.g., Bonneau and Cann 2015; Cann and Yates 2016; Gibson 2012; Hall 2015). Two features of the broader political landscape drive concerns about elections and diversity. First, women and people of color may be more reluctant to run for office (e.g., Fox and Lawless 2005). Second, implicit bias may suppress support for underrepresented candidates (e.g., Weaver 2012). One scholar offered a hypothetical rebuttal to diversity arguments against judicial elections, suggesting that appointments “will often be white-shoe affairs subject to capture and cronyism and liable to scant diversity along any number of dimensions,” while “elections, at least, are covered by the Voting Rights Act” (Pozen 2008, 295). More practically, judicial elections may be uniquely suited to generate quick and sizeable increases in diversity. In 2018, for example, seventeen Black women were elected to the bench in Harris County, Texas (encompassing Houston), representing an 850 percent increase in the number serving (Schneider 2019).

The existing empirical literature offers insight into the relationship between selection institutions and state court diversification. Using a variety of samples and modeling strategies, numerous studies find no relationship between selection system choice and judicial diversity (e.g., Alozie 1990, 1996; Graham 1990; but see Martin and Pyle 2002). For example, selection system choice is largely unassociated with the pace at which states first diversified their apex courts (Goelzhauser 2011). And there is no consistent relationship between selection institutions and the number of women or people of color serving on state supreme courts in 1985, 1999, or 2005 (Hurwitz and Lanier 2003, 2008).

The most comprehensive studies leverage decades of data on state supreme court seatings to better understand the relationship between selection institutions and diversification. Examining seatings from 1980 through 1997, Bratton and Spill (2002) find that appointment systems are more likely to produce justices who are women. And data on seatings from 1960 through 2014 suggest that women are more likely to be seated under unilateral elite appointment than merit selection, with no meaningful difference between merit selection or unilateral appointment and elections; furthermore, people of color are more likely to be seated under unilateral elite appointment or merit selection than elections, with no discernable difference between appointment systems (Goelzhauser 2016).

Other research streams yield important insights about specific selection mechanisms and diversity. Under merit selection, for example, there is evidence that commissions are less likely to nominate women (Goelzhauser 2018a). Survey evidence reveals that attorneys who are women express more interest in seeking judgeships under merit selection while there is no race differential (Jensen and Martinek 2009). Evidence from submitted applications, however, finds no gender differential with respect to expressive ambition and no gender or race differential with respect to progressive ambition (Goelzhauser 2019). Furthermore, the extent to which the design of merit selection systems facilitates commission capture is not associated with diversification (Goelzhauser 2019).

Under elections, survey evidence reveals that attorneys who are women express more interest in running for office while there is no race differential (Williams 2008). Evidence from other institutional contexts suggests that members of historically underrepresented groups exhibit more election aversion (e.g., Kanthak and Woon 2015), receive less elite encouragement to run for office (e.g., Lawless and Fox 2010), and suffer from stereotyped perceptions that discourage entry (e.g., Dolan 2010). Contrary to appointers, though, voters are no less likely to support the selection of women with other women in office (Solberg and Stout n.d.; Bratton and Spill 2002). And there is evidence that once on the bench, members of underrepresented groups do not suffer an electoral penalty (Frederick and Streb 2008; Gill and Eugenis 2019; Hall 2001; Solberg and Stout n.d.). Members of marginalized groups do, however, receive biased performance ratings leading up to retention elections (Gill 2014; Gill, Lazos, and Waters 2011).

Empirical Analysis

Are women of color more likely to be seated under certain judicial selection systems? The empirical analysis proceeds in two steps. First, I provide descriptive information on the first women of color seated on state apex courts. Although the first women of color to serve in the federal judiciary are well known, no systematic information exists about women of color on state courts. As a result, these groundbreaking women have not been collectively recognized, and many of their stories are largely unknown. To recover this history, I gathered information on all state supreme court justices seated from 1960 through 2016 from a variety of sources, including biographies, press releases, and newspaper articles (see Goelzhauser 2016). Second, I analyze the relationship between selection institutions and the seating of women of color on state supreme courts. While emphasizing women of color, the analysis provides comparable results for other gender-race combinations.

The First Women of Color on State Supreme Courts

The history of women of color on the federal bench is comparatively well known. Constance Baker Motley became the first woman of color to hold an Article III judgeship and serve on a district court when she was nominated to the US District Court for the Southern District of New York by President Johnson in 1966—seventeen years after the first woman (Burnita Shelton Matthews) and twenty-nine years after the first person of color (William Henry Hastie Jr., who was also the first person of color to hold an Article III judgeship). Amalya Lyle Kearse was the first woman of color seated on a federal circuit court after being nominated to the US Court of Appeals for the Second Circuit by President Carter in 1979—forty-five years after the first woman (Florence Ellinwood Allen, also the first woman to hold an Article III judgeship) and twenty-nine years after the first person of color (William Henry Hastie Jr.). And Sonia Sotomayor became the first woman of color seated on the US Supreme Court after being nominated by President Obama in 2009—twenty-eight years after the first woman (Sandra Day O’Connor) and forty-two years after the first person of color (Thurgood Marshall). It is notable that in each instance, seating the first woman of color took considerably longer than seating either the first white woman or man of color.

In the states, Dorothy Comstock Riley, who was Hispanic, may have been the first woman of color seated on a state apex court when governor William Milliken appointed her to the Michigan Supreme Court in December 1982. For comparison, Florence Ellinwood Allen, who as noted previously was the first woman to hold an Article III judgeship, was the first woman to serve as a state supreme court justice with her election by Ohio voters in 1922. And Jonathan Jasper Wright, an African American appointed to the South Carolina Supreme Court by the state legislature during Reconstruction in 1870, may have been the first man of color seated on a state apex court. The next Black justice was not seated until Otis M. Smith was appointed to the Michigan Supreme Court in 1961 (see Goelzhauser 2011, 762–763). Thus as with federal firsts, the first woman of color to sit on a state high court did so well after the first white woman and man of color.

Justice Riley spent her career in private practice before serving as a trial and appellate court judge. Riley lost an electoral bid for the Michigan Supreme Court in November 1982, but outgoing Republican governor William Milliken appointed her to an interim position when an incumbent justice named Blair Moody died a few weeks after being reelected. To prevent the incoming Democratic governor James Blanchard from filling the seat, Milliken appointed Riley to serve Moody’s new term even though it had not started. After this maneuver was challenged by the Blanchard administration, the Michigan Supreme Court (with Justice Riley recused) ruled that Justice Riley’s interim appointment was only valid through the end of Blair’s previous term (Attorney General v. Riley 1983). As a result, Riley’s initial tenure lasted from December 9, 1982, through 12:00 p.m. on January 1, 1983. In 1984, however, Riley was elected to the court for a full term and served until 1997, when she resigned due to the onset of Parkinson’s disease.^[4] Despite perhaps being the first woman of color seated on a state high court and the only woman of color seated twice during the sample period, her racial identity was seldom mentioned at the time of initial appointment, subsequent election, or death (see, e.g., Associated Press 2004).

Bernette Joshua Johnson was the only other woman of color to be initially seated on a state apex court through election during the sample period. In 1988, the US Court of Appeals for the Fifth Circuit held that Louisiana violated the Voting Rights Act by combining Orleans Parish and its large percentage of Black voters with three predominantly white parishes for the purpose of electing two state supreme court justices while other parishes elected a single justice each; the court found it “particularly significant that no Black person has ever been elected to the Louisiana Supreme Court” (Chisholm v. Edwards 1988, 1058). To remedy Black vote dilution in Orleans Parish, Louisiana entered into a consent decree, deferring redistricting but agreeing to create an intermediate appellate court judgeship for Orleans Parish that would immediately be assigned to the state supreme court, thereby increasing the latter’s size from seven to eight justices (see Perschall v. Louisiana 1997). Johnson, who had been a trial court judge, finished second in a three-way Democratic primary race, which would have resulted in a runoff election except that the white first-round winner withdrew, claiming her presence “threatened to ‘permanently scar’ race relations in New Orleans” amid arguments by Johnson and another Black candidate that the position should be filled by a Black judge in light of the reason for its creation (Orlando Sentinel 1994). After redistricting in 2000, Justice Johnson ran unopposed for reelection to a regular seat, and the court was reduced in size back to seven seats. Johnson was reelected unopposed again in 2010.^[5]

Juanita Kidd Stout became the first Black woman to sit on a state’s highest court when governor Robert Patrick Casey Jr. appointed her to the Pennsylvania Supreme Court in 1988. Stout was a music teacher in her home state of Oklahoma before pursuing a career in law after working as a legal stenographer during World War II (Thomas 1998). Stout’s legal experience included private practice and service as an assistant district attorney before receiving an interim appointment to the municipal bench in 1959 (Adams 2015). By retaining her position in a subsequent election, she became the first Black woman elected to any court in the United States (Thomas 1998). After Stout’s 1988 appointment to the Pennsylvania Supreme Court, she was forced off the bench in 1989 upon reaching the mandatory retirement age. In 1998, Justice Stout died of leukemia.

Joyce Kennard may have been the first Asian American woman to sit on a state’s apex court when governor George Deukmejian appointed her to the California Supreme Court in 1989. Kennard was born in Indonesia, where her family was detained in a Japanese internment camp before moving to the Netherlands after World War II (Dolan 2014). With English as a third language, Kennard later immigrated to the United States, where after law school she worked as a deputy attorney general and research attorney for a state appellate court. Her judicial career progressed swiftly; she served as a municipal court judge, superior (trial) court judge, and appellate court judge before joining the California Supreme Court (Kort 1993). In 2014, at the age of seventy-two, Justice Kennard retired on the twenty-fifth anniversary of her appointment to California’s high court, declaring, “I’d like to finally make time for my long-neglected friends” (Egelko 2014).

Overall, women of color accounted for 2 percent of state supreme court seatings from 1960 through 2016. During the same period, women and men of color, respectively, accounted for 17 percent and 9 percent of seatings. After 1982, when the first woman of color was seated on a state apex court, 4 percent of the seatings through 2016 were women of color compared to 26 percent and 13 percent for women and men of color, respectively. Figure 1 plots the name and year seated for each woman of color to join a state high court during the sample period. In total, twenty-two states seated women of color on their apex courts during this time. Women of color secured 2 percent of seatings in the 1980s, 3 percent in the 1990s, 1 percent in the 2000s, and 11 percent from 2010 through 2016. For comparison, figure 2 plots the percentage of seats obtained by decade for women of color, men of color, white women, and white men. Of the seatings on state supreme courts involving women of color, 57 percent of the justices were Black, 24 percent were Hispanic or Latina, 16 percent were Asian or Pacific Islander, and 3 percent were Asian and Latina.

Quantitative Analysis

The quantitative analysis employs justice-level data on all new seatings to state supreme courts from 1960 through 2016. The primary dependent variable is scored 1 when a newly seated justice was a woman of color and 0 otherwise. For comparison, I fit separate models for seating men of color, white men, and white women. Combining women from different race and ethnic categories is inconsistent with the overarching emphasis of intersectionality research on treating members of different disadvantaged statuses separately. However, this approach is necessary at this time due to the few women of color to be seated on state high courts. As Scola explained when combining race and ethnic categories in a study of intersectional legislative representation, “While this justification may not be wholly satisfying (for the reader or the author), collapsing the data was necessary to have a sufficient number of cases to test the model” (2013, 341). The key explanatory variables are indicators for whether justices were seated under unilateral elite appointment (gubernatorial or legislative), merit selection, or contestable elections (partisan or nonpartisan).^[6] This categorization reflects core institutional similarities but is also driven by the paucity of women of color seated across more finely delineated classifications.^[7]

Omitted variable bias occurs when excluding predictors that are correlated with the dependent variable and key explanatory variables. As a result, the models include several control variables that may be associated with state selection system choice and diversification. Previous research indicates that members of underrepresented groups may be more disadvantaged as position value increases. To capture position value, I include an index of state court professionalism (Squire 2008), term length in years, number of court seats, and an indicator for use of a mandatory retirement rule.^[8] While each measure captures position value, having shorter terms, more seats, and mandatory retirement may also induce turnover, thereby increasing opportunities for members of disadvantaged groups. Eligibility pool controls account for the number of attorneys in a state who are women or people of color.^[9] Since the effect of pool size may be nonlinear, with increases at the low end in the number of attorneys from underrepresented groups mattering more than at the high end, I take the natural log transformation of the raw counts. Given that women of color are generally more likely to secure political office as state liberalism increases (e.g., Scola 2006), I include a dynamic measure of citizen ideology (Berry et al. 1998). Last, because of changes in the likelihood of seating women of color over time, I account for time dependence by including variables counting the number of years since the start of the sample period, the number of years squared, and the number of years cubed (Carter and Signorino 2010).

Logistic regression can be used to model binary outcomes, such as whether a woman of color is seated on an apex court. Due to the infrequency of seating women of color, which can bias substantive effects downward, the models are fit using a rare event correction for logistic regression (King and Zeng 2001a, 2001b). Standard errors are clustered by state to account for the fact that observations within a state are not independent, which can lead to incorrect standard errors (Primo, Jacobsmeier, and Milyo 2007). Table 1 in the appendix displays the full results. Table 2 in the appendix displays the results using a multinomial choice model as an alternative estimation strategy. The results are consistent across specifications. Focusing on substantive effects, figure 3 plots first differences for pairwise comparisons of gender and race across selection methods. The effects seem small, but this is often the case with rare event data. To better examine substantive impact with rare event data, King and Zeng suggest examining relative risks, which “are typically considered important in rare event studies if they are at least 10–20 percent” (2001a, 152).

 Overall, the results indicate that selection system choice is associated with the likelihood of seating women of color on state supreme courts. As an initial matter, the difference between elite appointment and merit selection is not statistically distinguishable from zero. However, both appointment systems are more likely than elections to produce justices who are women of color. Moving from elite appointment to election is associated with an −88 percent (−97 percent, −40 percent) decrease in the relative risk of seating a woman of color.^[10] Moving from election to merit selection is associated with a 316 percent (5 percent, 1,692 percent) increase in the relative risk of seating a woman of color.

Compared to other gender-race combinations, the results for women of color are most similar to those for men of color. The results are more distinct when compared to white men and women, which in turn are distinct from each other. Moving from elite appointment to election is associated with an −82 percent (−92 percent, −65 percent) decrease in the relative risk of seating a man of color. Unlike with women of color, elections are also associated with a change in the likelihood of seating men of color relative to merit selection, with the relative risk increasing by 327 percent (78 percent, 901 percent) moving from the former to the latter. No pairwise selection institution comparison is associated with a statistically significant difference in the likelihood of seating white women. For white men, moving from unilateral elite appointment to election is associated with a 14 percent (7 percent, 22 percent) increase in the relative risk of being seated. Moving from unilateral elite appointment to merit selection is associated with a 9 percent (3 percent, 18 percent) increase in the risk of being seated. While moving from election to merit selection is associated with a −4 percent (−8 percent, 1 percent) decrease, the confidence interval overlaps zero, indicating that we cannot reject the null hypothesis of no relationship.

Pool size and time are the most notable control variables. Increasing the natural log of the number of attorneys of color in a state from its twenty-fifth to seventy-fifth percentile is associated with a 488 percent (55 percent, 1,953 percent) increase in the relative risk of seating a woman of color. However, the same increase in the natural log of the number of attorneys who are women is associated with a −69 percent (−90 percent, −14 percent) decrease in the relative risk of seating a woman of color. The results are similar for men of color. Increasing the natural log of the number of attorneys of color from its twenty-fifth to seventy-fifth percentile is associated with an 860 percent (329 percent, 2,291 percent) increase in the relative risk of seating a man of color. The same increase in the natural log of the number of attorneys who are women, however, is associated with an −82 percent (−92 percent, −61 percent) decrease in the relative risk of seating a man of color. For white women, increasing the natural log of the number of attorneys of color from its twenty-fifth to seventy-fifth percentile is associated with a −34 percent (−50 percent, −11 percent) decrease in the relative risk of being seated. The same increase in the natural log of the number of attorneys who are women is associated with a 47 percent (−1 percent, 109 percent) increase in the relative risk of being seated, though the confidence interval overlaps zero. For white men, increasing the natural log of the number of attorneys of color from its twenty-fifth to seventy-fifth percentile is associated with a −7 percent (−15 percent, 3 percent) decrease in the relative risk of being seated, but this confidence interval also includes zero.

The eligibility pool results have important implications. Consistent with existing research on political institutions and intersectionality, eligibility pools for women and people of color can have disparate effects on representation. Relatedly, there appears to be a tradeoff in seating women and people of color on state apex courts. An increase in the supply of attorneys who are women, for example, decreases the likelihood of seating a person of color. Last, many of the substantive effects are large relative to the effects associated with selection institution choice, suggesting that stakeholders who value diversity should emphasize increasing candidate pools and doing more to encourage and facilitate inclusive access to the bench.

Figure 4 plots the probability of seating justices from various gender-race combinations over time. The probability of seating women of color remains flat and near zero through much of the sample period, with a slight uptick this decade. After remaining relatively flat and near zero through the 1960s, there is a noticeably higher probability of seating men of color and white women beginning in the mid- to late 1970s, corresponding with President Carter’s efforts to diversify the federal judiciary. The uptick for white women, however, is more persistent. In recent years, the probability of seating men of color drops below that of seating women of color. White men held a virtual monopoly in obtaining state supreme court positions at the beginning of the sample period with a declining probability of being seated as members of underrepresented groups joined the bench more regularly. Overall, in the sample period’s later years the probability of being seating is increasing for women of color and white men but decreasing for men of color and white women.

Conclusion

Facing the double disadvantage of gender and race bias, women of color endure discrimination in the legal profession, manifested in part by a lack of representation. This chapter makes two contributions to the existing literature on state court diversity. First, it offers a descriptive account of the women of color who have served on state high courts from 1960 through 2016. While historic federal firsts are widely recognized, this group of groundbreaking women has received comparatively little attention. Second, I examine the relationship between selection mechanism choice and seating women of color. The results indicate that newly seated women of color are less likely to arrive through election than unilateral elite appointment or merit selection. The results are similar for men of color, while newly seated white men are more likely to arrive through election. Selection system choice is not associated with differences in seating white women.

The results have important policy implications. Scholars have amassed a wealth of empirical evidence on the performance of various judicial selection and retention institutions (see, e.g., Bonneau and Cann 2015; Bonneau and Hall 2009, 2017; Cann and Yates 2016; Geyh 2019; Gibson 2012; Goelzhauser 2016, 2019; Hall 2015; Kritzer 2015; Streb 2007). Albeit just one part of the broader debate, the extent to which institutional design choices impact judicial diversification has important consequences for descriptive and substantive representation. With respect to seating women of color in particular, the results presented here suggest that appointment systems outperform elections. While the mechanism is unclear, evidence from a variety of institutional contexts suggests that members of underrepresented groups may be disadvantaged in election systems due to factors such as contest aversion, lack of party support, having fewer political connections, and implicit stakeholder bias. Given increasing evidence that appointment systems tend to outperform elections with respect to diversification, one option for promoting diversity while enjoying the benefits of elections is to increase the use of short-term appointments.

By emphasizing women of color on state supreme courts, this chapter advances our understanding of how marginalized groups secure political representation. But it is important to stress that the broader study of intersectionality is complex, requiring scholars to leverage varied analytical approaches to shed light on the relationship between intersectionality and political action (see, e.g., Cho, Crenshaw, and McCall 2013; Hancock 2013; Holman and Schneider 2018; McCall 2005). These varied approaches are necessary in part because it is important to address other intersections of disadvantage in future research. Discrimination based on marginalized statuses such as religion, sexual orientation, and social class interacts with dimensions such as gender and race in complex ways (see, e.g., Beckwith 2014; Htun and Ossa 2013). One recent study, for example, examined evaluations of political candidates on the basis of gender and sexual orientation (Doan and Haider-Markel 2010). Moving forward, it is imperative to better map this complexity onto theories and empirical tests concerning intersectionality and the study of law and courts.