Racial and Gender Bias, Stereotypes, and Out-Groups

Implicit biases in our public and private interactions remain an ongoing problem in American society. While the issue is complex and contentious, many scholars agree that group stereotypes, driven by limited information about and first-hand experience working with minority group members, are important to understanding bias and discrimination (Reference ArveyArvey 1979; Reference Christensen, Szmer and StritchChristensen et al. 2012; Reference Davison and BurkeDavison and Burke 2000; Reference Dovidio and GaertnerDovidio and Gaertner 2000; Reference Fiske and TaylorFiske and Taylor 2013; Reference QuillianQuillian 2006). Gender and race can serve as “status cues” that may “signal individuals to subconsciously assume that members of nonmajority classes are of different status than majority class members” (Reference Christensen, Szmer and StritchChristensen et al. 2012: 627). Inherent in this dynamic is a grouping process. Majorities, usually white men, constitute an “in-group,” while minorities, usually women and members of racial and ethnic minority groups, constitute an “out-group.” Bias occurs when this difference becomes salient and provides a foundation for stereotyping of out-group members. Those whose identities align with more than one out-group can face additional stereotyping challenges not based simply on either race or gender, but on the variable interplay of both (Reference Collins and MoyerCollins and Moyer 2008; Reference CrenshawCrenshaw 1992; Reference Rosette, Koval, Ma and LivingstonRosette et al. 2016). For women of color, for example, there are “sometimes-distinct ways that race may play out both between men and women and also among women” (Reference CrenshawCrenshaw 2012: 1438).

For members of the in-group—the majority group—“less is known about the out-group” then about fellow members of the in-group (Reference Davison and BurkeDavison and Burke 2000: 231). Consequently, in-group members are likely to view out-group members “as more similar to and more interchangeable with one another” (Reference Dovidio, Gaertner, Anastasio, Sanitioso, Knouse, Rosenfeld and CulbertsonDovidio et al. 1992: 170)—in short, as stereotypes. Those stereotypes then inform the in-group's understanding of the nature of out-group members (Reference Davison and BurkeDavison and Burke 2000). Out-group members are then judged consistently with the stereotypical, homogeneous group expectation (Reference Biernat and KobrynowiczBiernat and Kobrynowicz 1997), while in-group members are more likely to be trusted (Reference Collins, Dumas and MoyerCollins et al. 2017), evaluated individually (Reference Davison and BurkeDavison and Burke 2000), and “viewed as distinctive and positive on subjectively important dimensions” (Reference Fiske and TaylorFiske and Taylor 2013: 441).

Out-Group Assessment Bias in Employment Settings

In the employment hiring and management context, out-group theory predicts that implicit or explicit biases and stereotypes can have a significant negative impact on interviewers’ and employers’ evaluation of the professional competence of women and racial minorities, creating a “presumption of incompetence” (Reference ArveyArvey 1979; Reference Biernat and KobrynowiczBiernat and Kobrynowicz 1997; Reference Cuddy, Fiske and GlickCuddy et al. 2008; Reference Davison and BurkeDavison and Burke 2000; Reference Eagly and KarauEagly and Karau 2002). Competence in this context refers broadly “to the ability to do well on a task judged to be valuable” in the field of interest (Reference FoschiFoschi 2000: 22). This presumption of incompetence is likely to be present even when out-group applicants’ relevant application records and previous job performance are, objectively speaking, equivalent or even stronger than those belonging to in-group (male, white) applicants (Reference Eagly, Makhijani and KlonskyEagly et al. 1992; Reference FirthFirth 1982; Reference Inesi and CableInesi and Cable 2015; Reference LottLott 1985). Low-status, out-group members will need more evidence of their ability for the task (Reference Biernat and FuegenBiernat and Fuegen 2001). At the same time, out-group members will be allowed less latitude than in-group members in the face of evidence of ability or experience insufficiencies (Reference Biernat and KobrynowiczBiernat and Kobrynowicz 1997). To overcome this imbalanced assessment, lower status interviewees and employees must “jump through more hoops” and “work twice as hard” (Reference Biernat and KobrynowiczBiernat and Kobrynowicz 1997: 546) to prove their professional aptitude for the position or task. This use of double standards in the evaluation of competence is a “subtle exclusionary practice” (Reference FoschiFoschi 2000).

Experimental and observational studies confirm the strong presence of this competence bias against women and racial minorities in the employment setting. Women's likelihood of being devalued is particularly high when they occupy male-dominated roles or are evaluated by men (Reference Eagly, Makhijani and KlonskyEagly et al. 1992; Reference Heilman and HaynesHeilman and Haynes 2005; Reference Inesi and CableInesi and Cable 2015). Reference QuillianQuillian's (2006) review of race discrimination studies similarly uncovered that “subordinate groups have predominately negative stereotypic attributions when evaluated by dominant group members” (320).

Female and minority political elites seem to be as vulnerable to competence bias and stereotyping threats, as are other females and minorities, particularly when the evaluators themselves also are elites, such as lawyers, judges, and politicians.Footnote ¹ For example, state legislators rate their black colleagues’ legislative effectiveness much lower than that of white legislators (Reference HaynieHaynie 2002), even after controlling for relevant experience, bill introductions, committee membership, partisanship, and seniority factors. Female Cabinet Secretaries-designates are less likely to be granted insider status during their Senate confirmation hearings than their male counterparts with similar career paths (Reference BorrelliBorrelli 1997). Research on lawyers and law firms reveals that “[t]he perception of difference, sex stereotyping, and treating women as a category rather than individually, provide serious obstacles” to female lawyers’ advancement in the legal profession (Reference Epstein, Saute, Oglensky and GeverEpstein et al. 1995: 304).

Within the judiciary, male judges hold significantly more negative stereotypes about female lawyers than do female judges (Reference Martin, Reynolds and KeithMartin et al. 2002). Specifically, surveyed male judges were more likely than female judges to agree with statements, such as:

• • “By and large female attorneys lack the competence of their male colleagues.”

• • “Generally speaking, men are more credible than women.”

• • “A woman who is outspoken or strongly adversarial is obnoxious” (Reference Martin, Reynolds and KeithMartin et al. 2002: 693).

Similarly, Reference SenSen (2014) finds that female, racial, and ethnic minority federal district court judicial nominees were more likely to receive lower ratings from the American Bar Association than white and male nominees. This was the case even after controlling for key factors like party affiliation, education, and experience. As Sen notes, “we cannot rule out the possibility of implicit bias against these sorts of nominees” (63). In addition, Reference Gill, Lazos and WatersGill et al. (2011) demonstrate that female and minority state judges receive lower judicial performance evaluations than their male and white colleagues, providing further evidence for the possibility of unconscious biases in the evaluation of members of the judiciary.

At the U.S. Supreme Court, two recent studies find striking gendered effects in the presence of interruptions during the Court's oral arguments. Reference Feldman and GillFeldman and Gill (2016) find that female justices on the Court are interrupted more frequently by fellow justices than are their male colleagues, and Reference Jacobi and SchweersJacobi and Schweers (2017) find that female justices are interrupted at disproportionate rates by male lawyers. Reference Szmer, Sarver and KahenySzmer et al. (2010) also reveal that Supreme Court justices were nearly 9% less likely to vote in favor of legal positions favored by women attorneys than those favored by male attorneys. In addition, recent experimental work reveals that judicial decisions were perceived as more authoritative when labeled as authored by “Anthony Kennedy” than by “Sandra Day O'Connor” (Reference Boddery, Moyer and YatesBoddery et al. 2016). Nor is this phenomenon limited to the U.S. Supreme Court: a recent study demonstrated that some black male and white female state Supreme Court justices were found to be less likely than their traditional colleagues to receive opinion-writing assignments (Reference Christensen, Szmer and StritchChristensen et al. 2012).

Together, this evidence seems to suggest that some legal and political elites make biased evaluations of female and racial minority candidates. Within the legal system and judiciary, much of this lingering bias seems rooted in stereotyping of the legal profession as a male's field of work.Footnote ² Indeed, Justice Ruth Bader Ginsburg once recounted her understanding that: “men of the bench and bar” generally held “the unyielding conviction that women and lawyering, no less judging, do not mix” (Reference GinsburgGinsburg 2007: 1).

With this theory and prior research in mind, we now turn to our primary question: whether and to what degree these biases manifest themselves during U.S. Supreme Court confirmation hearings.

The Supreme Court Confirmation Hearings Connection

As most scholars of American politics and legal institutions know well, candidates nominated to the U.S. Supreme Court are now regularly called to testify before the Senate Judiciary Committee. At the heart of the confirmation hearings rests the question-and-answer sessions during which each Committee member questions the nominee for a set time on the topics of his or her choosing. Although some scholarly accounts of the hearings, based largely on anecdotes, question their value (e.g., Reference CarterCarter 1994), more rigorous evidence demonstrates that the hearings do play an important role in the Supreme Court selection process and the answers nominees give (or do not give) can have vast implications for nominees (Reference Collins and RinghandCollins Jr. and Ringhand 2013a; Reference Farganis and WedekingFarganis and Wedeking 2014).

If the confirmation process is subject to the type of out-group bias discussed above, the hearings are likely to be where it is most visible. Accounts from some nominees themselves suggest that such bias is present. Justice Sonia Sotomayor, the first Latina to go through the process, reported afterward her belief that female judicial candidates are treated differently than their male counterparts. “There are expectations about how women and men should behave,” she told a group of law students 2 years after her nomination. “I am probably a bit more aggressive, but to hear people describe me as brash, and rude, the language used suggests a difference in expectations about what's OK for people's behavior” (Reference WardWard 2011). Justice Clarence Thomas also famously spoke about bias at his hearing, particularly in relation to the allegations of sexual harassment brought by Anita Hill (Reference MorrisonMorrison 1992). In starkly racialized language, he referred to his experience before the Senate Judiciary Committee as a “high-tech lynching for uppity blacks” (United States Senate 1991: 157). Commentators have made similar observations over the years (Reference Collins and RinghandCollins Jr. and Ringhand 2013a: 169–170; Reference Kenney, Thomas and WilcoxKenney 2014: 230).

Out-group theory provides a useful way to quantitatively evaluate these claims. Given the historical and contemporary dominance of male and white senators in Congress generally and on the Senate Judiciary Committee specifically, male and white nominees appearing before the Committee for hearings are likely to be in most senators’ in-group and to benefit from in-group favoritism, including being evaluated in a subjectively positive way. By contrast, female and nonwhite nominees appearing before the Committee generally belong to the out-group and, as such, are likely to be evaluated consistently with their out-group's stereotypical expectations. As discussed above, these stereotypes can be quite negative: women nominees may be perceived as inappropriately aggressive or emotional, while racial minority nominees may be perceived as less smart or lazy.

By coupling out-group theory with the bias in hiring literature and applying both to the Supreme Court confirmation hearings process, we can begin to develop expectations regarding how the hearings may play out differently for in and out-group nominees. These theories, as well as the experimental and observational studies underlying them, give us good reason to believe that the “hiring”Footnote ³ senators are likely to view female and racial minority nominees as inherently “different” in some way from more traditional, in-group white and male nominees. This is particularly true given the above-discussed research finding that elites are particularly likely to be affected by out-group stereotypes, and that there are strong historically rooted stereotypes that judging is and should be dominated by white males. Even in the face of objective qualifications that are equal to or even exceed those of previous Supreme Court nominees, therefore, theory predicts that male or white senators will tend to lack relative confidence in out-group nominees’ professional competence to serve on the Court.

If such bias is present at the hearings, it is likely to emerge implicitly rather than explicitly. In many cases, the questioning senators may not even be aware that they are treating in-group nominees differently from out-group nominees. Therefore, instead of explicitly biased treatment, we would expect skeptical senators to use their questioning time to require out-group nominees to more elaborately prove that they have the professional competence required to sit on the high court.Footnote ⁴

Professional competence at this level cannot be captured by educational pedigree or job experience. All nominees to the Supreme Court are likely to be among the most talented and experienced lawyers of their generations. Consequently, we would expect perceptions of competence at Supreme Court confirmation hearings to be tied to something else: a nominee's ability, in the eyes of the questioning senator, to get the Constitution “right” by using the “correct” method of constitutional interpretation.Footnote ⁵ As Senator Orin Hatch (R-UT) stated at the Sotomayor hearing, “a judicial nominee's qualifications include not only legal experience but, more importantly, judicial philosophy” (United States Senate 2009: 11). If, as Foschi and others have asserted, the definition of professional competence is “the ability to do well on a task judged to be valuable” in the field of interest (Reference FoschiFoschi 2000: 22), then the core professional competence of a Supreme Court justice is to correctly interpret the U.S. Constitution. To do otherwise is to fail at the essential task required of the job.

The senators and nominees are quite clear on this point. When Senator Strom Thurmond (R-SC) wanted to make Thurgood Marshall look unfit to sit on the high court, he launched an elaborate attack on Marshall's ability to correctly understand and interpret the original understanding of the Fourteenth Amendment (United States Senate 1991: 523). Likewise, Senator Chuck Schumer (D-NY) opened the Alito hearing by describing a nominee's “most important qualification” as the nominee's “judicial philosophy” (United States Senate 2006: 37), and Senator Jeff Sessions (R-AL) skeptically informed Sonia Sotomayor in his opening statement at her hearing that the senators would “inquire into how your philosophy, which allows subjectivity in the courtroom, affects your decision-making” (United States Senate 2009: 8). Justice Alito, testifying before the Committee in 2006, captured the sentiment succinctly, stating simply that “the job of a Supreme Court Justice [is] to interpret the Constitution” (United States Senate 2009: 465).

There is unquestionably deep disagreement about what the “correct” method of constitutional interpretation is, and senators are likely to form their preferences on the matter in ways that advance their underlying policy goals (Reference Post and SiegelPost and Siegel 2006). Nonetheless, in the context of the confirmation hearings, questioning nominees about their judicial philosophies is a tool senators use to explore how, as future justices, nominees would execute that core task. When senators perceive a nominee as using the wrong interpretive methods, the senators will also view the nominee as unqualified to sit on the Supreme Court. In addition, while senators may well be skeptical of the judicial philosophy of all nominees named by opposing-party presidents, the theory here predicts that opposition party in-group nominees will enjoy much more of a presumption of mainstream, professional competence in this regard, while similarly situated female and minority nominees may enjoy no such presumption. Rather, stereotypes and negative perceptions of these nominees will lead them to be perceived as less willing or able to behave in a similarly professional manner.

In the Supreme Court confirmation hearing context, therefore, we would expect questioning senators to express this skepticism by requiring minority and female nominees to demonstrate their professional competence (or reveal their lack thereof) by spending more of their confirmation hearing time answering questions about their judicial philosophies. The out-group status of these nominees would tend to lead senators to implicitly or explicitly view them as less skillful and trustworthy than in-group nominees, and the senators would try to expose this tendency by probing more deeply in this area. Equally qualified in-group nominees, in contrast, might receive more latitude during confirmation questioning regarding judicial philosophy, as in-group senators will have more confidence in the professionalism of in-group nominees, or see casting doubt on the nominee's professional competencies as a less fruitful line of attack. In-group nominees therefore should expect to receive a lower proportion of questions in the areas of judicial philosophy than out-group nominees.

We formalize these expectations for in and out-group gender and racial minority nominees in Hypothesis 1.

1. Hypothesis 1a: Female Supreme Court nominees will receive a higher proportion of judicial philosophy questions from male senators than will male nominees.

2. Hypothesis 1b: Racial minority Supreme Court nominees will receive a higher proportion of judicial philosophy questions from white senators than will white nominees.

The Conditioning Effect of Politics on Out-group Bias

We turn now to the potential conditioning effect that politics can have on the magnitude of out-group bias during Supreme Court confirmation hearings. Whether out-group stereotyping leads to discriminatory behavior may depend on the presence of an attitude-behavior link (Reference Davison and BurkeDavison and Burke 2000). This opens the door for preferences to affect whether discrimination will take place. When an out-group member shares a person's values, ideals, and policy priorities, people may be able to control stereotypes and biased perceptions of out-group members and, to some degree, inhibit their activation or lessen their impact (Reference Kawakami, Dovidio, Moll, Hermsen and RussinKawakami et al. 2000). By contrast, when those preferences do not align, stereotypes are more likely to be adopted and acted upon. Indeed, the effects of these stereotypes may be negatively magnified, leading to much higher levels of discriminatory behavior toward out-group members.

Previous research finds a strong connection between political attitudes and preferences and the magnitude of activated out-group stereotypes. This includes evidence that, for Senate candidates, there is a strong link between political preferences and perceptions of women (Reference KochKoch 2000) and race (Reference Sigelman, Lee, Walkosz and NitzSigelman et al. 1995). At the U.S. Supreme Court, while justices were less likely to vote for legal positions favored by women attorneys than those favored by male attorneys, there is about a 10% higher likelihood of support for the female lawyer's position if it is consistent with the justice's ideology (Reference Szmer, Sarver and KahenySzmer et al. 2010).Footnote ⁶

In the context of Supreme Court confirmation hearings, we expect that partisan congruence between the questioning senator and the nominee will serve an important conditioning role on the presence and level of out-group bias. The modern Senate is highly polarized along party lines (Reference LeeLee 2008), and committee membership reflects the partisan composition of the full Senate (Reference Epstein and SegalEpstein and Segal 2007). Senate parties act as “teams,” with coordination on important issues like presidential agenda items and procedural control (Reference LeeLee 2009). Importantly, ideological proximity and policy agreement are not the only reasons that Senate partisans support one another; rather, some members will work with their party “because they understand the value of team play” (Reference LeeLee 2009: 47).Footnote ⁷ It would be “quite surprising if Democrats were not more likely to support Democratic presidents than were Republicans and vice versa” (Reference SegalSegal 1987: 1001). Moreover, this is exactly what the evidence indicates for the Senate's confirmation process: nearly all senators of the same party as the president support his Supreme Court nominees, but significantly fewer opposition party senators do so (Reference Epstein and SegalEpstein and Segal 2007; Reference Farganis and WedekingFarganis and Wedeking 2014).

How might this “partisan climate” in the Senate's deliberations over Supreme Court nominees (Reference Epstein and SegalEpstein and Segal 2007) differentially affect in-group and out-group nominees? Recall that out-group evaluation bias toward Supreme Court nominees often stems from a lack of specific information about and trust in the predictability of an out-group nominee's future behavior and decisions. In the case of out-group nominees (and their appointing president), who share the questioning senator's party, much of this uncertainty or distrust, will, inevitably, be lessened. Shared party provides a set of common expectations for the senator and nominee that may help avoid the strong activation of out-group biases. By contrast, when the nominee hails from the opposite party as the senator, stereotypes are likely to be magnified. In these cases, there is no shared set of expectations to stand in the way of gender and racial stereotyping leading to biased behavior. Moreover, because partisan divisions are piled on top of out-group status stereotyping, senators are likely to be particularly wary of the competence of these nominees. As a result, we expect to see a higher percentage of judicial philosophy questions asked of opposite party,Footnote ⁸ out-group nominees.

Therefore, our resulting conditional hypothesis includes:

1. Hypothesis 2a: Female Supreme Court nominees not from the senators’ political party will receive a higher proportion of judicial philosophy questions from male senators than same party female nominees.

2. Hypothesis 2b: Racial minority Supreme Court nominees not from the senators’ political party will receive a higher proportion of judicial philosophy questions from white senators than same party racial minority nominees.

Data and Methods

We test our hypotheses by utilizing the U.S. Supreme Court Confirmation Hearing Database developed by Reference Collins and RinghandCollins Jr. and Ringhand (2013b). This data set contains information on the subject matter of every statement made at every open, transcribed Supreme Court confirmation hearing held before the Senate Judiciary Committee from 1939 to 2010.Footnote ⁹ In our model that focuses on nominee gender, we begin our analysis in 1981, which was the first time a female nominee (Sandra Day O'Connor) appeared before the Committee. Our model that focuses on nominee race begins in 1967, which was the first time a nominee of color (Thurgood Marshall) testified before the Committee. Moreover, these years mark significant institutional developments related to an increase in the volume of questions about precedent (1967) and gavel-to-gavel television coverage of the hearings (1981), the latter of which resulted in lengthening the hearings overall (Reference Collins and RinghandCollins Jr. and Ringhand 2016). Thus, by focusing on these periods, we can make comparisons across nominees who faced the Committee in similar institutional environments.Footnote ¹⁰

Because we are interested in the relative amount of attention senators devote to questioning nominees on their judicial philosophies, the unit of analysis is the senator-nominee dyad. This means that there is one observation for every senator who questioned a nominee at their confirmation hearing (e.g., one observation for Senator Coburn during the Sotomayor hearing and one observation for Coburn during the Kagan hearing). Our empirical analysis is tailored to reflect out-group theory's focus on the status of both the senator and the nominee and the dominance of white men on the Senate Judiciary Committee over time. We limit our primary analysis to male senators’ questions (for gender-related tests) and white senators’ questions (for racial minority-related tests) during Supreme Court confirmation hearings. In the models that test for gender differences, the data set excludes the 10 observations pertaining to female senators. In the models that tests for racial differences, the data set excludes the 10 observations relating to minority senators. Thus, the gender models allow us to examine whether male senators treat female nominees differently than male senators treat male nominees, and the racial models allow us to investigate whether white senators treat minority nominees differently than white senators treat white nominees.Footnote ¹¹ After our main empirical analyses, we present a descriptive analysis of the questioning behavior of the limited number of female senators in the data.

While the relatively small number of female and racial minority nominees does have the potential to bias the results against finding statistically significant effects (e.g., Reference King, Keohane and VerbaKing et al. 1994), we note that although female and racial minority nominees appear before the Judiciary Committee far less frequently than more traditional white, male nominees, there is nonetheless sufficient variation to make statistical investigation appropriate. For example, female nominees make up 33.5% of observations in the 1981–2010 data and racial minority nominees constitute 13% of observations in the 1967–2010 data.

Our dependent variable represents the proportion of senatorial questions devoted to explorations of the nominees’ judicial philosophies.Footnote ¹² As noted above, questions about judicial philosophy include those that involve a nominee's judicial philosophy and constitutional interpretation as a concept, but do not include questions about constitutional interpretation that arise in the context of a specific issue area or are related to particular constitutional outcomes or cases. The coding of the dependent variable used here was found to be extremely reliable by Reference Collins and RinghandCollins Jr. and Ringhand (2013b: 77–80): there is 94% intercoder agreement on the issue area variable used to identify judicial philosophy questions (kappa = 0.921).

Because our dependent variable is a proportion that cannot take on negative values and cannot exceed 1, the use of ordinary least squares regression is inappropriate. Accordingly, we employ a fractional logit model. The fractional logit model is a quasi-likelihood method that is estimated as a generalized linear model and allows us to account for the bounded nature of the dependent variable (Reference Papke and WooldridgePapke and Wooldridge 1996). As nominees appear in the data more than once, we use robust standard errors, clustered on nominee.Footnote ¹³

Figure 1 reports the number of overall questions and judicial philosophy questions asked of nominees from 1967 to 2010, based on the data used in the statistical analyses that follow. The solid vertical line identifies the O'Connor hearing in 1981, which is the first hearing in the gender models. As this figure indicates, there was a slight increase in the overall questioning of nominees leading to William Rehnquist's hearing for the Chief Justice position in 1986. At that point, the overall number of questions evened out, although it is evident that Robert Bork received a particularly large number of questions. Overall, 12.4% of questions asked by senators involve querying nominees about their judicial philosophies. This makes this the second most common substantive topic of questioning at the hearings after civil rights (Reference Collins and RinghandCollins Jr. and Ringhand 2013a: 103). Indeed, both Alito (2006) and Kagan (2010) saw almost a fifth of their hearing devoted to this topic. In addition, Figure 1 also provides some preliminary evidence regarding gender and racial disparities among the nominees. For example, for female nominees, an average of 18.6% of the questions asked by senators were judicial philosophy questions, compared to 12.9% for male nominees (p = .006). Racial minority nominees received an average of 15.3% of judicial philosophy questions, compared to 13.5% for white nominees (p = .21). We now turn to analyzing these differences in a more rigorous manner.

Figure 1. The Number of Overall and Judicial Philosophy Questions Asked of Each Nominee, 1967–2010

To test our hypotheses, we set up a series of variables that allow us to capture the gender, race, and party affiliation of the questioning senator and the nominee. Female Nominee is scored 1 for questions asked by male senators to female nominees (Ginsburg, Kagan, O'Connor, and Sotomayor) and 0 for questions asked by male senators to male nominees, based on information in Reference Epstein, Walker, Staudt, Hendrickson and RobertsEpstein et al. (2013) and United States Senate (2016a). Male Nominee is coded in the opposite way (1 for male senator-male nominee questions; 0 for male senator-female nominee questions). Minority Nominee is scored 1 for questions asked by white senators to minority nominees (Marshall, Sotomayor, and Thomas) and 0 for questions asked by white senators to white nominees (Reference Epstein, Walker, Staudt, Hendrickson and RobertsEpstein et al. 2013; United States Senate 2016b). White Nominee is coded in the opposite way (1 for white senator-white nominee questions; 0 for white senator-minority nominee questions). Same Party is scored 1 if the senator questioning the nominee shares the partisan affiliation of the president who appointed the nominee and 0 if the senator and nominee have opposing party affiliations (Reference Carroll, Lewis, Lo, McCarty, Poole and RosenthalCarroll et al. 2015; Reference Epstein, Walker, Staudt, Hendrickson and RobertsEpstein et al. 2013). Different Party is coded in the opposite way.

Based on these constituent variables, our modeling below focuses primarily on the four possible combinations for a nominee's out-group status and party affiliation (for a similar modeling technique see Reference OwensOwens 2010). Doing so allows us to better understand the conditionality present between these statuses and how they compare with one another when it comes to the proportion of questions asked about judicial philosophy. For nominee gender, the four resulting combinations are: (1) Female Nominee/Same Party, (2) Male Nominee/Different Party, (3) Male Nominee/Same Party, and (4) Female Nominee/Different Party. Every senator-nominee pair within the data belongs to one (and only one) of these groups. For nominee race, the four similarly designed groupings are: (1) Minority Nominee/Same Party, (2) White Nominee/Different Party, (3) White Nominee/Same Party, and (4) Minority Nominee/Different Party.

Based on our hypotheses, we expect that the Female Nominee/Same Party variable will be negatively signed, indicating that female nominees receive a smaller proportion of questions regarding their judicial philosophies from same party senators, compared to opposite party senators. We expect the Male Nominee/Different Party variable will be negatively signed, revealing that male nominees receive a smaller proportion of questions regarding their judicial philosophies from different party senators, compared to female nominees. We expect that the Male Nominee/Same Party variable will be negatively signed, indicating that male nominees receive a smaller proportion of questions regarding their judicial philosophies from same party senators, compared to opposite party senators questioning female nominees. The Female Nominee/Different Party variable serves as the baseline in our statistical models. Our expectations for the variables pertaining the race and party of the nominee and senator also follow this logic.Footnote ¹⁴

We include three additional variables in the models to account for other factors that might affect the proportion of questions asked by senators about judicial philosophy. First, we control for each nominee's qualifications, which is intended to capture senators’ views of each nominee's professional preparation for the Supreme Court. Nominee Qualifications are based on the assessment of a nominee's professional credentials for the Court that appeared in newspaper editorials in four leading newspapers: Chicago Tribune, Los Angeles Times, New York Times, and Washington Post (Reference Cameron, Cover and SegalCameron et al. 1990). This variable ranges from 0 to 1, with higher values indicating more qualified nominees, and was obtained from Reference SegalSegal (2016).Footnote ¹⁵ We expect this variable will be negatively signed, indicating that senators ask nominees perceived to be more professionally qualified for the Court a smaller proportion of questions about judicial philosophy.

Second, we include a variable that indicates whether or not the nominee had prior judicial experience. Nominee Judge is coded 1 if the nominee served as a judge on a state or federal court and 0 otherwise, based on information in Reference Epstein, Walker, Staudt, Hendrickson and RobertsEpstein et al. (2013). We expect senators to ask nominees with judicial experience a lower proportion of questions regarding their judicial philosophies, as these individuals have established interpretation track records on this issue that are accessible to senators. This allows senators to focus their questioning on other issue areas at the hearings. Accordingly, we expect this variable to be negatively signed.Footnote ¹⁶

Finally, we include a Senator Reelection Year variable that captures whether the senator questioning the nominee was up for reelection during the year of the hearing, based on information in United States Senate (2016c). This variable is coded 1 if the senator was up for reelection and 0 otherwise. Because senators can use the questions they ask at the hearings to promote their reelection prospects, we expect them to ask more questions about judicial philosophy than senators who are not up for reelection. This is because senators can use these questions to demonstrate to their constituents that they take their position on the Committee very seriously and are willing to rigorously engage nominees on matters relevant to their fitness for the bench (Reference Collins and RinghandCollins Jr. and Ringhand 2013a; Reference Collins and RinghandCollins Jr. and Ringhand 2016). Accordingly, we expect this variable will be positively signed.Footnote ¹⁷ Table 1 reports the means and standard deviations of the dependent and independent variables, as well as the expected direction for each of the independent variables.

Table 1. Summary Statistics

Notes: Entries are the means of the dependent and independent variables. Entries in parentheses are the standard deviations of those variables.

Results

Table 2 reports the results of the statistical models that explore whether male senators ask female nominees a higher proportion of questions regarding their judicial philosophies. As described above, we test our hypotheses by assessing the proportion of judicial philosophy questions asked of nominees based on their combined out-group status and shared party affiliation. Table 2’s modeling sets female nominees receiving questions from opposite party senators (Female Nominee/Different Party) as the baseline for statistical comparison. This means that the proportion of judicial philosophy questions received by nominees in the other three possible groupings—Female Nominee/Same Party, Male Nominee/Different Party, and Male Nominee/Same Party—are being compared to female nominees from a different party than the questioning senator. If hypothesis 1a holds, we would expect to see negative signs on the Male Nominee/Different Party and Male Nominee/Same Party variables since, in these two situations, male nominees are receiving questions from male senators. Moreover, if hypothesis 2a holds and there is a conditioning partisan effect, we would expect the negative sign to also extend to the Female Nominee/Same Party variable where female nominees receive questions from partisan allies. Because the coefficients from a fractional logit model cannot be directly interpreted, Tables 2 and 3 provide marginal effects to permit substantive interpretations.

Table 2. The Proportion of Senate Questions Regarding Judicial Philosophy Asked by Male Senators, 1981–2010

^* p < .05 (one-tailed tests)

Notes. The dependent variable is the proportion of questions asked by male senators involving judicial philosophy. Entries are fractional logit regression coefficients. Numbers in parentheses are robust standard errors, clustered on nominee. The marginal effects of the coefficients of the statistically significant independent variables appear in brackets.

Table 3. Marginal Effects for Gender Models, 1981–2010

The results are striking. Female nominees being questioned by opposite party senators are asked 10% more questions about judicial philosophy than are male nominees being questioned by opposite party senators. In other words, preemptively skeptical (opposing party) senators grill female nominees much harder than they do similarly situated male nominees about the nominee's ability to competently perform the core professional task of constitutional interpretation.

Interestingly, as shown in Table 3, same party senators are not responding to the increased judicial philosophy questioning from opposite party senators by generating a corresponding increase of their own in friendly questioning in this area. This type of “tit for tat” ritual is readily observable in the hearings, but our data show that it is not happening here: same party senators do not respond to the increased skeptical questioning of opposite party senators in this area by providing their same-party female nominees with increased opportunities to rebut that skepticism under friendly questioning. Instead, the results show that female nominees receive 7% fewer judicial philosophy questions from same party senators than they do from opposite party senators. This is consistent with the shared values conditioning effect discussed above, which holds that stereotyping is most likely to be triggered when out-group members are seen as also having different values than the higher status individual. However, it may have the unfortunate effect of depriving female nominees of the opportunity to rebut increased skeptical questioning about their judicial philosophies through a compensatory increase in friendly questioning in this issue area.

Stereotypes regarding female nominees’ professional competence, reflected through questioning about judicial philosophy, thus may harm female nominees in three distinct ways. First, female nominees are subjected to more rigorous questioning in this area by opposite party senators who may see such questioning as revealing a perceived weakness. Second, their male counterparts enjoy the privilege of not being subjected to a similar rigorous questioning by opposing party senators, and therefore do not have their professional competence publicly challenged in the same way. In fact, male opposite party nominees receive 8% fewer judicial philosophy questions than male same party nominees, as illustrated in Table 3. Third, female nominees are not provided compensatory opportunities by same party senators to demonstrate their professional competence through a similarly friendly questioning from same party senators.

Turning now to the control variables, the model reveals that more qualified nominees receive a higher proportion of questions regarding their judicial philosophies than less qualified nominees. A one standard deviation increase in this variable, making a nominee more qualified, results in a 4% increase in questions regarding the nominee's judicial philosophy. This suggests that nominees who appear to be qualified are not immune from competency-based questions regarding their judicial philosophies. The model also reveals that, as expected, nominees with previous judicial experience receive about 6% fewer questions about their judicial philosophies. Since these nominees have a track record, via their opinions, that speaks to their judicial philosophies, senators apparently choose to focus more on other types of questions during their hearings.

Table 4 reports the results of the model that explores whether white senators ask minority nominees a higher proportion of questions regarding their judicial philosophies. Unlike the model that explores gender differences, this model fails to reveal racial differences, conditioned or unconditioned on shared party affiliation, on this measure. As others have demonstrated (and as even a cursory review of the hearing transcripts reveals) there is nonetheless evidence that minority nominees are subjected to unique confirmation hearing experiences (e.g., Reference Collins and RinghandCollins Jr. and Ringhand 2013a; Reference Davis and WildmanDavis and Wildman 1992; Reference DworkinDworkin 2009; Reference GreeneGreene 1989). For example, Reference Ringhand and CollinsRinghand and Collins Jr. (2011) show that minority nominees receive more questions on other issues areas, such as criminal justice. Why no difference appears on the measure examined here is a question warranting additional study.

Table 4. The Proportion of Senate Questions Regarding Judicial Philosophy Asked by White Senators, 1967–2010

^* p < .05 (one-tailed tests)

Notes. The dependent variable is the proportion of questions asked by white senators involving judicial philosophy. Entries are fractional logit regression coefficients. Numbers in parentheses are robust standard errors, clustered on nominee. The marginal effects of the coefficients of the statistically significant independent variables appear in brackets.

We do, however, have one possible explanation for this nonresult. As noted above, our research design anticipates senators’ unwillingness to use the hearing process to explicitly and directly call out minority nominees for their perceived lack of professional competence. We therefore use questioning about judicial philosophy as a vehicle through which this senatorial skepticism about the minority nominee's competence will be manifested. It may be the case, however, that senators do not require any such proxy concerning nominees who are racial minorities because the purported professional incompetence of these nominees is being discussed directly and explicitly.

This certainly was the case concerning Thurgood Marshall's confirmation hearing (Reference GreeneGreene 1989) and may have been in play at the hearings of Clarence Thomas as well (Reference Davis and WildmanDavis and Wildman 1992; Reference WillsWills 1995). The issue of competence repeatedly and directly came up at the Marshall hearing, including when Senator Strom Thurmond (R-SC) engaged in a series of not-so-thinly-veiled attacks on Marshall's fitness for the Court by grilling him on esoteric aspects of the history of the Thirteenth and Fourteenth Amendments (Reference Collins and RinghandCollins Jr. and Ringhand 2013a: 169–170). At the Thomas hearing, issues of competence likewise often came up directly, frequently in response to the view of two members of the American Bar Association's Standing Committee on the Judiciary that Thomas was not competent to serve on the Court (e.g., United States Senate 1991: 523).

Sonia Sotomayor's hearing provides another interesting example. Sotomayor is the only nominee in our data to be both a woman and a member of a racial minority, and her nomination was explicitly intended to diversify the Court (Reference Steigerwalt, Vining and StrikoSteigerwalt et al. 2013). As such, she may be subject to a unique set of biases and stereotypes and her competence may be particularly called into question (Reference Rosette, Koval, Ma and LivingstonRosette et al. 2016). Our findings support this.

Figure 2 reports the percentage of judicial philosophy questions that nominees received from opposite and same party senators from 1967 to 2010, based on the data used in our empirical models.Footnote ¹⁸ The data indicate that Sotomayor received a higher percentage of judicial philosophy questions than the average for all other categories nominees to the Court—white, minority, male, and white female. Overall, 19% of the questions asked of Sotomayor involved grilling her on matters pertaining to her judicial philosophy, more than any other nominee in the data, save for Justices Alito and O'Connor, who were asked the same percentage of judicial philosophy questions. These included questions related to whether it was appropriate for judges to exhibit “empathy” when deciding cases and those relating to Sotomayor's ability to be impartial in light of her comment regarding “wise Latinas” (Reference DworkinDworkin 2009; Reference KenneyKenney 2013). These results echo recent empirical findings that Sotomayor is the justice most commonly interrupted at Supreme Court oral arguments, providing another measure of disrespect directed disproportionately at the Court's only female, racial minority nominee in terms of race and gender (Reference Jacobi and SchweersJacobi and Schweers 2017).

Figure 2. The Percentage of Judicial Philosophy Questions Nominees Received from Opposite and Same Party Senators, 1967–2010

The Next Step: Female Senators’ Behavior

Our findings reveal that male senators ask in-group members—i.e., male nominees—a lower percentage of judicial philosophy questions than female nominees. This leads to the natural related question of whether female senators are similarly favorable to their in-group members (female nominees) in their questioning behavior.Footnote ¹⁹ While our robustness checks reveal that the inclusion of female senators in the analyses above has no substantive effect on the results, any plausible effect is almost certainly masked by the small number of female senators in the data. There were just three female senators on the Senate Judiciary Committee through 2010 (Senators Feinstein, Klobuchar, and Moseley-Braun),Footnote ²⁰ which amounts to only 10 senator-nominee dyad observations in the data set.

For female senators, out-group theory as strictly applied would predict that female senators would be more likely to favor and defer to those in their in-group (i.e., female nominees) than those in their out-group (i.e., male nominees). Simultaneously, the theory would expect that female senators would be more likely to implicitly call into question male nominees’ competence to serve on the Court. However, prior research indicates that when evaluators, employers, or interviewers are themselves of a lower status, they are much less likely than higher status evaluators to apply differing standards of competence to objectively similar candidates (Reference FoschiFoschi 2000). Applied to the Supreme Court confirmation hearings setting, these prior results would predict little-to-no difference in the questioning of male and female nominees by female senators. Finally, gender-based stereotype threat may lead female senators to adhere to the stereotypes of their group (Reference Latu, Mast and StewartLatu et al. 2015; Reference PinelPinel 1999). Within our data, stereotype threat could manifest itself through female senators, just like their male colleagues, questioning female nominees’ competence to serve on the Court more aggressively than male nominees’ competence.

To provide a preliminary assessment of the behavior of female senators on the Senate Judiciary Committee in their questioning of male and female nominees regarding their judicial philosophies, we turn to the descriptive data from this project. Because of the small number of female senators, systematic regression analysis would be unwise. Figure 3 reports the percentage of judicial philosophy questions female and male senators asked of female and male nominees from 1981 to 2010.

Figure 3. The Percentage of Judicial Philosophy Questions Nominees Received from Female and Male Senators, 1981–2010

Two key findings emerge. First, female senators ask both female and male nominees fewer judicial philosophy questions (5.5%) than male senators (12.7%). Second, like male senators, female senators ask more judicial philosophy questions of female nominees (7.7%), as compared to male nominees (2.1%). Although these findings are very preliminary (due to the small number of female senators on the Judiciary Committee and our corresponding inability to account for confounding factors), they nonetheless provide some support for theories of stereotype threat that suggest that female senators might invoke the same stereotypes as male senators when it comes to evaluating female nominees. It also is plausible that the qualitative content of female senators’ judicial philosophy questions to female nominees may be different from that of male senators, a topic that we leave for future research to fully assess. Regardless, it is evident that additional study into this area will shed important light on this important issue.

Conclusions

The female nominees who sit before the Senate Judiciary Committee are among the most accomplished lawyers of their respective generations. Nonetheless, as our results reveal, these female nominees are often subjected to a very different confirmation process than are their male colleagues. As predicted by out-group theory and prior studies of gender bias in hiring, male senators grill female nominees on their judicial philosophies—questions representing the core professional skill expected of U.S. Supreme Court justices—more so than they press male nominees. Further, this questioning comes disproportionately from opposite party—and presumably skeptical—senators. This effect was most acute for Sonia Sotomayor, although our statistical results do not indicate that the other minority nominees face a similarly difficult set of questions compared to white nominees. Finally, our preliminary descriptive analysis of the small number questions asked by female senators revealed that female senators, on average, tend to ask nominees, and particularly male nominees, a lower percentage of judicial philosophy questions than male senators.

The fact that female nominees are very publicly subjected to a different type of confirmation process by male senators than are male nominees serves to perpetuate negative stereotypes that male judges are more believably prepared to serve in the judiciary and, specifically, as Supreme Court justices. As more and more female and minority nominees participate in the confirmation process, this differential treatment could cast the legitimacy of the process and the Court itself into doubt, particularity if there is not a corresponding increase in female and minority senators (another high-profile and strongly white, male stereotyped profession) capable of reducing the out-group dynamics.

As the first study to rigorously explore the differential treatment of female and minority Supreme Court nominees at the hearings, this paper opens the door to a wide array of future research. Perhaps most obviously, it will be important to apply the theory and methods advanced here to the selection of other judicial and political actors. Future study of judicial nominees to the federal district courts and courts of appeals and executive branch nominees within the Department of Justice (such as U.S. Attorneys), for example, offers great potential to examine a larger number of female and racial minority nominees who testify before the Senate Judiciary Committee (e.g., Reference Dancey, Nelson and RingsmuthDancey et al. 2011; Reference Dancey, Nelson and RingsmuthDancey et al. 2013; Reference Martinek, Kemper and Van WinkleMartinek et al. 2002; Reference SchererScherer 2005; Reference SteigerwaltSteigerwalt 2010). Nominees to these institutions undergo somewhat similar scrutiny from the Judiciary Committee and tend to be more diverse in terms of their race and gender. They may also allow for a closer investigation into whether female, racial minority nominees face particularly intense scrutiny from the Judiciary Committee. We note, however, that this future area for research does not come without its own difficulties. Nominees to the lower federal courts and Department of Justice posts generally face a shorter and less salient confirmation process than Supreme Court justices. For example, lower federal court nominees’ hearings typically last just a few minutes (Reference O'BrienO'Brien 2002), with the average federal district court nominee receiving only 6.9 questions from the Senate Judiciary Committee (Reference Dancey, Nelson and RingsmuthDancey et al. 2013: 799). By comparison, Supreme Court nominees frequently testify for days with each nominee receiving an average of 717 questions (using the data in Figure 1 from 1967 to 2010).

It also will be useful to explore whether social role stereotypes (Reference Eagly, Wood, Diekman, Eckes and TraunterEagly et al. 2000) lead female and minority nominees to receive heightened questioning from senators beyond those questions involving their judicial philosophies. For example, black politicians are stereotyped as being empathetic and disproportionately good at handling issues of concern to the black community like civil rights and affirmative action while, at the same time, being perceived to be less strong than other politicians on issues related to the economy, military and national security, and taxes (Reference Schneider and BosSchneider and Bos 2011: 224, 229). Whether these stereotypes hold for an individual or not, the theory in this area expects that out-group members are judged consistently with the stereotypical, homogeneous group expectation (Reference Biernat and KobrynowiczBiernat and Kobrynowicz 1997). It is therefore plausible that senators might use questions about crime and criminal justice as a proxy for matters relating to racial issues, pressing minority nominees on this issue (e.g., Reference Ringhand and CollinsRinghand Collins Jr. 2011).

As our study indicates, the examination of the role of race and gender in the selection of legal and political actors continues to be relevant and important. We hope that our design and findings help to foster a robust line of inquiry into this topic across institutions, jurisdictions, decision-making stages, and time.