III. Sex, Race, and Other Characteristics
In the last Part we looked at the effects of ideology on judicial decisionmaking.
But ideology is not the only feature of judges’ identities that might shape how they
resolve difficult legal questions. In this Part we turn to studies that look at other personal
characteristics and life experiences to see whether they also influence judging.
A. Of Judicial Identity and “Bias”
It seems inevitable that the identities of the judges beneath those black robes will
affect their decisionmaking, at least in some cases. Isn’t that what Justice Cardozo’s
candid views on the subject would suggest? But which cases, and how much? If a
judge’s personal characteristics shape how she approaches a particular legal question,
should we worry that the judge is unable to decide the case fairly? If so, is there any way
to avoid the problem? If judges cannot help but see cases through their own eyes, does
that mean that all judges are improperly biased?
1. Personal Characteristics as Improper Bias
In any litigation, the parties have a right to ask the judge to step down on the
ground that the judge is unable for reasons of bias to judge the case fairly. That is the
subject of the following decision.
BLANK V. SULLIVAN & CROMWELL
418 F. Supp. 1 (S.D.N.Y. 1975)
MOTLEY, J.
This is an action brought under Title VII of the Civil Rights Act of 1964 (42
U.S.C. § 2000e et seq.). Plaintiff, who applied for a position as an attorney with
defendant firm and was rejected, claims that defendant discriminates in the employment
of lawyers on the basis of sex. The instant action has been certified as a class action by
this court. (See Memorandum Opinion and Order, July 14, 1975.)
Defendant has filed an application and affidavits pursuant to 28 U.S.C. § 144 and
28 U.S.C. § 455 requesting this Judge to disqualify herself on the grounds of personal and
extrajudicial bias against defendant and in favor of plaintiff and her cause. (Affidavit of
Arthur Dean at 4). The primary evidence of this purported bias is the court's ruling that
this action be maintained as a class and “more importantly the manner in which (this and
other) ruling have been issued.” (Affidavit of Arthur Dean at 3). These affidavits are, as a
matter of law, clearly insufficient to justify my disqualification. Defendant’s motion for
disqualification is, therefore, denied.
2
It is well settled that in determining a motion to disqualify under 28 U.S.C. § 144,
the facts stated in the affidavit as the basis for the belief that prejudice exists must be
accepted as true by the judge, even if he or she knows the statements to be false. Berger
v. United States, 255 U.S. 22, 41 S.Ct. 230, 65 L.Ed. 481 (1931); Hodgson v. Liquor
Salesmen’s Union Local No. 2 of State of New York, 444 F.2d 1344 (2d Cir. 1971); Rosen
v. Sugarman, 357 F.2d 794 (2d Cir. 1966). If, however, those facts as stated are
insufficient as a matter of law, the judge shall not be disqualified, for there is an equal
duty on the part of the judge not to recuse himself when there is no occasion to do so as
there is for him to do so when there is. Rosen, supra, at 799; Hodgson, supra at 1348;
Wolfson v. Palmieri, 396 F.2d 121, 124 (2d Cir. 1968).
Section 144 provides for disqualification if the affidavit “give(s) fair support to
the charge of a bent of mind that may prevent or impede impartiality of judgment.”
Berger, supra, 255 U.S. at 33-34, 41 S.Ct. at 233. In order to remove a judge, the bias
shown must be a “personal” prejudice, from an extrajudicial source and resulting in an
opinion on the merits not warranted by the facts or issues presented in the case. Wolfson,
supra at 124; U.S. v. Grinnell, 384 U.S. 563, 86 S.Ct. 1698, 16 L.Ed. 2d 778 (1966).
Defendant’s affidavits clearly fail to show the extrajudicial, personal bias, or even the
inference of such bias, required for disqualification by § 144.
[The court first discussed and disposed of Defendant’s claims that the court had
indicated bias in prior rulings, and in failing to give the Defendant an opportunity to be
heard. The Court pointed out that Defendant had been given ample opportunity to be
heard, and explained that prior rulings were all in accord with legal precedents.]
Defendant further seeks my disqualification on the ground that I “strongly
identified with those who suffered discrimination in employment because of sex or race”
. . . . This court has ruled against plaintiffs in civil rights cases and has denied class action
status to a woman plaintiff alleging discrimination in employment.*
It is beyond dispute that for much of my legal career I worked on behalf of blacks
who suffered race discrimination. I am a woman, and before being elevated to the bench,
was a woman lawyer. These obvious facts, however, clearly do not, ipso facto, indicate or
even suggest the personal bias or prejudice required by § 144. The assertion, without
more, that a judge who engaged in civil rights litigation and who happens to be of the
same sex as a plaintiff in a suit alleging sex discrimination on the part of a law firm, is,
therefore, so biased that he or she could not hear the case, comes nowhere near the
standards required for recusal. Indeed, if background or sex or race of each judge were,
by definition, sufficient grounds for removal, no judge on this court could hear this case,
or many others, by virtue of the fact that all of them were attorneys, of a sex, often with
*
See Marina Voustis v. Union Carbide Corp., Docket #70 Civ. 3435 (May 23, 1975); cf. Mullarkey v.
Borglum, 323 F. Supp. 1218 (S.D.N.Y. 1970); Wilson v. Follette, 438 F.2d 1197 (2d Cir. 1971), cert.
denied 402 U.S. 997, 91 S. Ct. 2182, 29 L. Ed. 2d 163 (1971); Coleman v. Mery, 72 CIV. 1455 (Nov.
1, 1972).
3
distinguished law firm or public service backgrounds. (Cf. Letter of Harriet Rabb,
attorney for plaintiff, dated April 17, 1975.) See lengthy opinion by Higgenbotham, J.
(E.D. Pa.) denying defendant’s motion to disqualify him in a race discrimination case
because he is black. Commonwealth of Pa. v. Local Union 542, Int. U. of Engineers, 388
F. Supp. 155 (E.D. Pa. 1974).
Nowhere in their affidavits do defense counsel or defendant indicate that I have
any relationship or personal association or interest in this litigation. They merely point to
my general background and the obvious facts of my race and sex as evidence of
extrajudicial prejudice.
. . .
As noted above, none of the facts included in the affidavits of defendant or
defense counsel are sufficient, under the statutes, for disqualification. Defendant’s motion
for disqualification is, therefore, denied.
So Ordered.
* * *
FOR DISCUSSION
1. Constance Baker Motley was the judge in Blank v. Sullivan & Cromwell. She was born
in New Haven, Connecticut to immigrant parents from Nevis.1
As she notes in the
opinion, she is both a woman and African-American. After attending Fisk University (a
historically black university) and New York University as an undergraduate, Motley
graduated from Columbia University School of Law in 1946. Following law school she
worked for the National Association for the Advancement of Colored Persons (NAACP),
first as a clerk for Thurgood Marshall and later as a leading litigator in her own right. She
wrote the first draft of the complaint in Brown v. Board of Education. She prevailed in
nine of ten cases she argued before the Supreme Court, including James Meredith’s
successful challenge of his denial of admission by the University of Mississippi. After
wide engagement in the civil rights movement, and a string of firsts in political positions,
she was appointed to the federal bench by President Lyndon Johnson in 1966, the first
African-American woman to take the bench.
1
These biographical notes are drawn from CONSTANCE BAKER MOTLEY, EQUAL JUSTICE UNDER
LAW: AN AUTOBIOGRAPHY (1998); Constance Baker Motley, Some Reflections on My Career, 6 LAW
& INEQ. 35 (1988); WOMEN AND THE CIVIL RIGHTS MOVEMENT, 1954-1965 (Davis W. Houck and
David E. Dixon, eds., 2009); WOMEN IN LAW: A BIO-BIBLIOGRAPHICAL SOURCEBOOK 181 (Rebecca
Mae Salokar & Mary L. Volcansek eds., 1996); and S. Res. 272, 109th Cong., 151 CONG. REC. 22642
(2005) (enacted) (resolution honoring the life of Judge Constance Baker Motley).
4
2. The defendant’s recusal motion in Blank took chutzpah, a Yiddish word meaning
“nerve.” But was it crazy? There are two ways to think of this: as a matter of trial
strategy, and on the merits. As a question of strategy, if the motion was unsuccessful, the
defendant faced further proceedings before a judge who he had accused of bias by reason
of her race and sex, and her prior background and career. In a part of the opinion we have
edited out, the judge explained that the defendant had raised questions of bias from the
outset, before any rulings were made, but had indicated he would not ask the judge to
remove herself. It was apparently only after receiving initial unfavorable rulings that he
decided to move forward.
How about the merits? Was it wrong to think that Judge Motley could not decide
a discrimination case fairly because she was a woman and a racial minority who had
made a career of litigating discrimination issues on behalf of plaintiffs? Of course, the
obvious question is: if not Judge Motley, then who? Every possible judge was going to
have both a gender and a racial identity. Would it necessarily have been fairer to both
parties to have a judge whose career had been in, say, insurance fraud?2
2. Life Experience and Judging: O’Connor, Sotomayor, and the Question of
a “Woman’s Perspective”
Blank v. Sullivan & Cromwell raises profound questions regarding the
relationship between the identity of a judge and that judge’s decisions in cases that come
before the court. At some level, most people (including most judges) acknowledge that
the identity of the judge will influence at least some of her decisions—we saw this in
Chapter Two. But which cases are most affected? Do different judges see all cases
differently? Or do certain types of legal issues trigger more disagreements than others?
Sandra Day O’Connor was the first woman nominated and confirmed to a seat on
the Supreme Court of the United States, in 1981. At the time, people wondered if having
a woman on the Supreme Court would change things. (One obvious change was that in
anticipation of a female Justice, in November 1980 the Justices stopped formally
referring to one another as “Mr. Justice ______.”) After O’Connor had been on the bench
five years, law professor Suzanna Sherry wrote an article entitled “Civic Virtue and the
Feminine Voice in Constitutional Adjudication,” in which she argued that women had a
different “world-view” from men, that woman lawyers were “a potentially innovative
force in the legal community,” and that although O’Connor voted regularly with her
2
John J. McConnell was nominated by President Obama in March 2010 to a judgeship in the federal
district court in Rhode Island, despite the strong objections of the U.S. Chamber of Commerce and
Republic Senators concerned that McConnell’s background as a plaintiffs’ attorney would give him
an unsuitable anti-business bias. The 111th
Senate did not confirm him. President Obama renominated
him, and the Senate confirmed him, in 2011.
5
conservative colleagues, she “approaches legal issues from a different perspective than
that of her male colleagues.”3
Justice O’Connor dissented. In a speech given in 1991 at New York University
School of Law, O’Connor dismissed the entire line of inquiry as pernicious.4
This view
“is interesting, but troubling,” said O’Connor, “precisely because it so nearly echoes the
Victorian myth of the ‘True Woman’ that kept women out of law for so long. It is a little
chilling to compare these suggestions to Clarence Darrow’s assertion that women are too
kind and warm-hearted to be shining lights at the bar.”5
She then took up the question
“Do women judges decide cases differently by virtue of being women?” Her response: “I
would echo the answer of my colleague, Justice Jeanne Coyne of the Supreme Court of
Oklahoma, who responded that ‘a wise old man and a wise old woman reach the same
conclusion.’”6
Fast-forward another ten years. When giving a speech at the Berkeley School of
Law, then-Judge Sotomayor said that, “[w]hether born from experience or inherent
physiological or cultural differences, . . . our gender and national origins may and will
make a difference in our judging.”7
In other words, Judge Sotomayor asserted that factors
like sex and race do in fact matter to judging. (In that same speech, she went on to claim,
provocatively, that “a wise Latina woman” like herself would reach better legal
conclusions precisely because of her life experience.)8
Although as a nominee to the
Supreme Court, Judge Sotomayor backed away from this strong statement under
3
Suzanna Sherry, Civic Virtue and the Feminine Voice in Constitutional Adjudication, 86 VA. L. REV.
543, 580, 581, 613 (1986). Sherry’s views were based in part on research performed by Carol Gilligan
suggesting that women approach moral questions differently than men do. See CAROL GILLIGAN, IN
A DIFFERENT VOICE: PSYCHOLOGICAL THEORY AND WOMEN’S DEVELOPMENT (1982). Gilligan’s
empirical methodology has been criticized, and Gilligan herself later softened her thesis somewhat.
For criticism of Gilligan, see, e.g., John M. Broughton, Women’s Rationality and Men’s Virtues: A
Critique of Gender Dualism in Gilligan’s Theory of Moral Development, 50 SOC. RES. 597 (1983)
(suggesting that some of Gilligan’s own subject interviews actually problematize the notion of a strict
male/female duality); Owen J. Flanagan Jr. & Jonathan E. Adler, Impartiality and Particularity, 50
SOC. RES. 576 (1983) (highlighting Gilligan’s focus on individual dilemmas of fact rather than
abstract moral reasoning about justice); Debra Nails, Social-Scientific Sexism: Gilligan’s Mismeasure
of Man, 50 Soc. Res. 643 (1983) (looking critically at Gilligan’s interview procedures and data
presentation); Lawrence J. Walker, Sex Difference in the Development of Moral Reasoning: A Critical
View, 55 CHILD DEV. 677 (1984) (pointing out that Gilligan’s claim is based almost entirely on
anecdotal evidence, without any experimental, longitudinal, or cross-sectional evidence). For
Gilligan’s later, somewhat softer version of the thesis, see, e.g., Carol Gilligan & Jane Attanucci, Two
Moral Orientations, in Mapping the Moral Domain: A Contribution of Women’s Thinking to
Psychological Theory and Education 73, 82-85 (Carol Gilligan et. al., eds. 1988).
4
The speech, delivered at New York University School of Law on October 29, 1991 as the twentythird
James Madison Lecture on Constitutional Law, was later reproduced in the NYU Law Review.
Sandra Day O’Connor, Portia’s Progress, 66 N.Y.U. L. REV. 1546 (1991).
5
Id. at 1553.
6
Id. at 1558.
7
Sonia Sotomayor, A Latina Judge’s Voice, 13 BERKELEY LA RAZA L.J. 87, 92 (2002).
8
Id. at 92.
6
pressure, she is hardly alone in her conclusion. Bertha Wilson, the first female justice of
the Canadian Supreme Court, claims: “Studies show overwhelming evidence that genderbased
myths, biases and stereotypes are deeply embedded in the attitudes of many male
judges…Researchers have concluded that gender difference has been a significant factor
in judicial decision-making.”9
So, here is the question: If Justice Sotomayor is correct that personal
characteristics like gender and race will “make a difference in . . . judging,” do you think
that is because women and racial minorities have a different “world-view” than white
men? Do you expect the differences (if any) to show up across all classes of cases, or
only those that touch directly on issues of race and gender?
B. Empirically Studying Race, Sex, And Other Characteristics
Before we can present what we know—and don’t know—about the effects of sex,
race, and other characteristics on judging, let’s think about how we would go about
studying these questions. Do you think this will be easier, or more difficult, than studying
ideology?
To begin, we need a dataset—a set of cases to study. We might choose to study
different types of cases, or a particular issue area where we think personal characteristics
might matter, such as discrimination cases. Either way, the method we’ll use to select
cases will essentially be the same as we saw in the previous Part.
Then we’ll code a dependent variable of interest. This, too, is exactly the same as
before. (Including the fact that studies overwhelming tend to focus on dispositions rather
than written opinions.) As an example, in most studies of sex discrimination cases, the
dependent variable is whether or not the party alleging discrimination (the plaintiff) gets
relief or not.
The next step is to measure the independent variable. This was really a challenge
when studying ideology. Because ideology is unobservable, proxies must be used instead.
Measuring the independent variable is much easier in most “personal characteristics”
studies because the characteristics in question—race, gender, etc.—are relatively
observable and can be measured directly.
Once we have our dependent variable and our independent variable, the final
thing to do is to compare behavior. For example, we could compare the percentage of
times when male judges found for the plaintiff in discrimination cases with the
percentage of times when female judges found for the plaintiff, and then use statistical
9
Bertha Wilson, Will Women Judges Really Make a Difference?, 28 OSGOODE HALL L.J. 507, 512
(1990).
7
analysis to determine whether or not the differences are significant. The task seems
straightforward. But—as we saw with ideology—there are many ways a study like this
can go awry.
Consider the study by Professors Pat Chew and Robert Kelley in the Washington
University Law Review.10
These authors were interested in studying how the race of
judges affected the outcomes of workplace racial harassment cases. Here’s how they
describe what they did:
Noting the absence of research on the subject, we conducted the first
study of the relationship between judges’ race and the outcomes in
racial harassment in the workplace cases... We designed a research
process to identify representative cases and judges so that any
inferences we drew from the data would be as generalizable as
possible. We randomly selected forty percent of all reported racial
harassment cases from six federal circuits between 1981 and 2003.
From these judicial opinions, we collected information on each case
(including characteristics of the parties, the nature of the alleged
harassment, and the outcome of the proceeding). We also collected
detailed biographical information on the presiding judge in each case
(including her or his race, gender, and political affiliation). A single
judge presided in district court cases, and a panel of three judges
typically sat in appellate court cases. Our study included a total of 256
different judges, some of whom heard more than one case. Pairing
each case with each judge hearing the case yielded a total sample of
428 judge/case pairs.11
For each judge vote in the dataset they coded whether or not the person alleging
racial harassment (the plaintiff) was successful. This was the dependent variable.
One independent variable in the study is the race of the judge. Here is what the
authors found when they looked at that variable:
Our multiple statistical analyses clearly indicate that judges of
different races do exhibit significantly different decision-making
patterns. White judges are more likely than their African American
counterparts to favor employer defendants, often granting their
motions for summary judgment. African American judges are more
10
Pat K. Chew and Robert E. Kelley, Myth of the Color-Blind Judge: An Empirical Analysis of Racial
Harassment Cases, 86 WASH. U. L. REV. 1117 (2009).
11
Id. at 1129-1138.
8
likely to believe that employees have credible grievances of racial
harassment.12
To see this difference in practice, look at Figure XX. This figure summarizes the
observed difference between Black and White judges in this study. Take a look at the top
panel. The difference in probabilities of finding for the plaintiff is around twenty-five
percent. The horizontal line represents the uncertainty we have about the finding. Since
this line doesn’t cross zero, we can be confident that the effect is statistically significant.
When just comparing Black and White judges, there are statistically significant
differences. That is the bottom-line conclusion that Chew and Kelley reach in their study.
But is it the end of the story? You hopefully remember something called “omitted
variable bias” from the last Part. If you don’t control for relevant variables, you
sometimes can get the wrong answers. This is especially the case when two independent
variables are related to one other. Is there something else that might affect judging?
Ideology, of course! We know that ideology is probably related to race and sex
since minority judges are more likely to be liberal, as are female judges. Moreover, many
(though not all) female judges in the federal courts were appointed by Democratic
presidents; so, too, for minority judges. Any study of the effects of a background
characteristic must, therefore, simultaneously control for political ideology in some
fashion. If not, any findings might just be an artifact. Suppose, hypothetically, that a court
had a large number of female judges, all but one of whom was liberal, and further
suppose that the men who sat on the same court were almost all conservative. If we just
compared the behavior of the men and the women, we might want to conclude that their
sex explained the difference. But without controlling for their ideologies, it’s impossible
to know whether sex or ideology was truly the factor that mattered most.
What happens to the findings from the Chew and Kelley study when you control
for ideology? It turns out that in their article they estimated models that included the
party of the appointing President. These models were reported in the Appendix. What do
they find? Take a look at the bottom panel in Figure XX. Two things should be apparent
in that panel. First, the effect size is smaller. Second, and more important, the black line
that shows how confident we are in the findings crosses zero. This means that the
relationship is statistically insignificant. In other words, there is no reliable evidence in
the data that suggests that Black and White judges behave differently. Once we control
for ideology, the effects of race are no longer significant.
So if the race of the judge isn’t doing the work, something else is. In Figure XX
we show the estimated difference in behavior between Democrat and Republican
appointees, controlling for their race. We see a roughly twenty-percentage-point
difference that is statistically significant. This is a classic example of omitted variable
12
Id. at 1161.
9
bias. Had the party of the appointing President not been controlled for, we would have
reached the wrong conclusion about the effects of race and judging. Indeed, the
conclusion reached by Professors Chew and Kelley that relies solely on the comparison
of Black and White judges without considering ideology is not the conclusion supported
by the data.
Figure XX. Estimated difference in probability of finding for plaintiff in racial
harassment cases between black and white judges. Pat K. Chew & Robert E. Kelley,
Myth of the Color-Blind Judge: An Empirical Analysis of Racial Harassment Cases, 86
WASH. U. L. REV. 1117-1166 (2009). The findings come from results tables in Appendix
B (Models 1 and 5). The black line is a 95% confidence interval for the difference.
Difference between Black and White Judges
0.0 0.1 0.2 0.3
●
Controlling for Ideology
●
Not Controlling for Ideology
10
Figure XX. Estimated difference in probability of finding for plaintiff in racial
harassment cases between Democrat and Republican appointees. Pat K. Chew and Robert
E. Kelley, Myth of the Color-Blind Judge: An Empirical Analysis of Racial Harassment
Cases, 86 WASH. U. L. REV. 1117 (2009). The findings come from a results table in
Appendix B (Models 5). The black line is a 95% confidence interval for the difference.
FOR DISCUSSION
1. Why did we see a difference when we just compared the two groups, and then saw it
go away when we controlled for ideology? One thing to remember is that this couldn’t
happen if ideology and race were completed unrelated to one another. So this means
these two variables are related. How do you think the race of the judge and her ideology
might be related?
2. One thing you will notice in the next section is that there are no studies about the
effects of race and sex on judging on the Supreme Court. Why do you suppose that is the
case? (Hint: Justices Thurgood Marshall and Clarence Thomas; Justices O’Connor,
Ginsburg, Sotomayor, and Kagan.)
C. What We Know
The best empirical studies recognize all too well the importance of controlling for
ideology when studying sex and race. What do these studies tell us about whether—and
when—such personal characteristics affect judicial decisionmaking?
Difference between Democrats and Republicans
0.0 0.1 0.2 0.3
●●
Effect of Political Party
11
1. Sex
Most studies of the effects of gender on judicial decisionmaking look at the
Courts of Appeals. For example, Professor Jennifer Peresie looked at all sexual
harassment and sexual discrimination cases in 1999, 2000, and 2001, including both
published and unpublished opinions.13
Her dependent variable was whether or not the
outcome was pro-plaintiff (the person alleging discrimination or harassment). After
controlling for ideology and other characteristics of both judge and case, she found that
the likelihood of a pro-plaintiff outcome increases from 22% to 41% in sexual
harassment cases, and from 17% to 28% in sexual discrimination cases, when the judge is
a woman.
Professor Christina Boyd and her colleagues looked at decisionmaking in the
circuit courts in thirteen areas of law: abortion, Americans with Disabilities Act,
affirmative action, campaign finance, capital punishment, contract clause, EPA,
federalism, piercing the corporate veil, Title VII sex discrimination, sexual harassment,
takings clause, and Title VII race discrimination.14
The time periods run from the 1970s,
1980s, or 1990s through 2002, depending on the issue area.
Once they controlled for ideology, Professor Boyd and her colleagues found that
in all areas of law except Title VII sex discrimination cases, male and female judges
behaved exactly the same. They found that, “[o]n average, the probability of female
judges voting in favor of the plaintif in a sex discrimination case is around .10 higher
than it is for male judges—a difference with meaning.”15
This suggests that sex itself
might be doing little work, except in the one area where past professional experience as a
female lawyer might lead a female judge to view a particular case differently.
Here is a sampling of some other findings across courts and issue areas.
• Professors Lisa Baldez, Lee Epstein, and Andrew Martin studied sex
discrimination cases in state courts of last resort from 1960-1999. Controlling for
case characteristics and ideology, their analysis demonstrates that, as the number
of female justices on the court increases, the likelihood that the court finds in
favor of the plaintiff increases as well.16
• Professors Elaine Martin and Barry Pyle studied the decisions of state courts of
last resort in divorce cases 1998-1999. After controlling for ideology, their results
13
Jennifer Peresie, Female Judges Matter: Gender and Collegial Decisionmaking in the Federal
Appellate Courts, 115 YALE L. J. 1759 (2005).
14
Christina L. Boyd, et al., Untangling the Causal Effects of Sex on Judging, 54 AM. J. POL. SCI. 389
(2010).
15
Id. at 401.
16
Lisa Bladez, Lee Epstein, & Andrew D. Martin, Does the U.S. Constitution Need an ERA?, 35 J.
LEGAL STUD. 243 (2006).
12
indicate that female judges tend to side with female litigants; furthermore, if a
female justice is present on the court, the male members also become more likely
to side with the female litigant.17
(We discuss the issue of inter-judge effects on
multi-member courts in more detail in Chapter XX.)
• Professors Thomas Walker and Deborah Barrow studied the effects of gender
diversification on the decisions of the U.S. District Courts. Focusing only on
judges appointed by President Carter, they found statistically different sex
differences in personal liberties and federal economic regulation cases; however,
they found no difference between male and female judges in cases involving
women’s issues.18
• Professor Sue Davis and her coauthors examined the decisions of the U.S. Courts
of Appeals in three issue areas: search and seizure, obscenity, and employment
discrimination. Their regression analysis controls for the political party of the
appointing president, and they, like Boyd et al., conclude that female judges are
more likely to support the plaintiff in employment discrimination cases while
finding no significant differences in other issue areas.19
In short, except for issues like sex discrimination where female judges—
regardless of their ideology—are likely to have direct personal experience, the
differences in behavior between male and female judges are either nonexistent or small in
magnitude once we take ideology into account.20
2. Race
The findings in the literature with regard to race closely mirror those for the sex
studies. As one recent study reported, “[o]ver the past several decades, extensive
empirical research has concluded that a judge’s race has little, if any, impact on
decisionmaking in a variety of criminal and civil rights litigation.”21
As with sex,
however, some studies have found race-effects in areas where race may be particularly
salient, or where Black judges may have distinctive experience. Such studies tend to
focus on the courts of appeals, rather than the district courts. Here is a sampling:
17
Elaine Martin & Barry Pyle, State High Courts and Divorce: The Impact of Judicial Gender 36 U.
TOLEDO L. REV. 923 (2005).
18
Thomas G. Walker & Deborah J. Barrow, The Diversification of the Federal Bench: Policy and
Process Ramifications, 47 J. POL. 596 (1985).
19
Sue Davis, Susan Haire & Donald R. Songer, Voting Behavior and Gender on the U.S. Courts of
Appeals, 77 JUDICATURE 129 (1993).
20
See, e.g., Orley Ashenfelter, et al., Politics and the Judiciary: The Influence of Judicial Background
on Case Outcomes. 24 J. LEGAL STUD. 259 (1995); Gregory Sisk, et al., Charting the Influences on the
Judicial Mind: An Empirical Study of Judicial Reasoning, 73 N. Y. U. L. REV. 1377 (1998).
21
Adam B. Cox & Thomas J. Miles, Judging the Voting Rights Act, 108 COLUM. L. REV. 1 (2008). At
4-5.
13
• Professors Adam Cox and Thomas Miles studied a set of dispositions in cases
arising under Section 2 of the Voting Rights Act since 1982, which prohibits
states from adopting any “voting qualification[s] or prerequisite[s] to voting . . .
which result[] in a denial or abridgment of the right of any citizen of the United
States to vote on account of race . . . .” The study includes votes in cases by
circuit court judges, district court judges, and a handful of cases with district court
judges sitting on three-judge panels (a unique institutional feature of the Voting
Rights Act). After controlling for a variety of factors, Cox and Miles found
significant differences between Democratic and Republican appointees. But even
after controlling for the political party of the appointing President, they found that
“an African-American judge is more than twice as likely as a non-AfricanAmerican
judge to vote for section 2 liability. While the number of AfricanAmerican
judges in our dataset is relatively small, the size of this effect is
striking…. These findings are perhaps the first to document a strong connection
between a judge’s race and her judicial decisions—and certainly are the first to do
so in the area of minority voting rights litigation.”22
• Professor Jonathan Kastellec examined the effects of racial diversity on the U.S.
Courts of Appeals. After controlling for ideology, he found that Black judges are
more likely than non-Black judges to support affirmative action programs.23
• Sean Farhang and Greg Wawro examined the effect of diversity on the decisions
of the U.S. Courts of Appeals. Their data includes 400 published employment
discrimination cases decided in 1998 and 1999. While their analysis reveals some
significant differences between male and female judges, they do not find any
statistically discernable race effects.24
Just as with the studies about sex, the race studies show that except for a few
highly salient areas, like voting rights or affirmative action, any effects of the race of the
judge are swamped when ideology is taken into account.
3. Social Backgrounds and Career Experience
Sex and race are two characteristics that we might expect to be related to judicial
behavior. There is another literature that looks at the effects of other types of background
characteristics and previous professional experiences on decision-making.
22
Adam B. Cox & Thomas J. Miles, Judging the Voting Rights Act, 108 COLUM. L. REV. 1 (2008). At
4-5.
23
Jonathan P. Kastellec, Racial Diversity and Judicial Influence on Appellate Courts (unpublished
paper) (2011). Kastellec’s study also finds strong panel effects: His results indicate that, if a single
black judge is placed on a panel, that panel is very likely to support the affirmative action program.
We discuss panel effects in more detail in Chapter XX.
24
Sean Farhang & Gregory Wawro, Institutional Dynamics on the U.S. Court of Appeals: Minority
Representation Under Panel Decision Making 20 J. LAW, ECON., & ORG 299 (2004).
14
The most-cited political science study on the question is Professor Neal Tate’s
article, Personal Attribute Models of the Voting Behavior of U.S. Supreme Court
Justices: Liberalism in Civil Liberties and Economics Decisions, 1946-1978, 75 AM.
POL. SCI. REV. 355 (1981). Professor Tate posited a social background model of
decisionmaking for justices on the Supreme Court. He looked at the voting behavior of
the Justices in both civil liberties and economics cases from 1946 (the start of the Vinson
Court) to 1978 (the end of the first decade of the Burger Court). He collected a host of
information, including: birth, upbringing, and education; past career experience; age and
tenure; and political partisanship. Then, he performed an analysis to determine which of
these factors were most predictive of judicial behavior.
Professor Tate’s results are quite interesting. For both economics and civil
liberties cases, he found strong relationships between judicial behavior and the political
party of the appointing President: Justices appointed by President Johnson were likely to
have higher percent liberalism scores than those appointed by Presidents Truman or
Nixon in both civil liberties and economics cases. But the backgrounds of the Justices
matter as well. In civil rights cases, background as a prosecutor led to more conservative
decisions, while prior judicial experience led to more liberal ones. In economics cases,
justices who previously served in elective office or had previously served on the bench
tended to reach more liberal outcomes; those with prosecutorial experience tended to
reach more conservative ones.
Professor James J. Brudney and his colleagues Sara Schiavoni and Professor
Deborah J. Merritt used a similar social background model to understand decisionmaking
in the federal circuit courts.25
Their study focused on both published and unpublished
cases cases reviewing unfair labor practices under the National Labor Relations Act from
1986-1993. Just like Professor Tate’s study, Professor Brudney and his colleagues looked
several different characteristics. The authors summarize their findings as follows:
A number of judicial attributes . . . also make a significant difference
in predicting outcomes. . . . [J]udges appointed by Democratic
Presidents were significantly more likely to support the union than
were their Republican colleagues. Older judges and judges who had
graduated from more selective colleges were significantly more likely
to reject the union’s claims…Gender [also] makes a difference in
NLRA cases—but only for Republican appointees…[The model]
indicate[s] that Democratic men and women both are more likely than
Republican men to vote in favor of the union, but that Democratic men
and women do not differ significantly among themselves. Republican
women, on the other hand, are significantly more likely than
Republican men to support the union. The votes of Republican women,
25
James J. Brudney, et al., Judicial Hostility Toward Labor Unions? Applying the Social Background
Model to a Celebrated Concern, 60 OHIO ST. L.J. 1675 (1999).
15
in fact, are at least as pro-union as those of as those of Democratic
men or women.26
There was one particularly surprising finding in the Brudney, et al. study:
[J]udges with NLRA management experience were significantly more
likely than their colleagues to vote for the union—but only in cases in
which the Board had issued a decision for the employer. When the
Board supported the union, most judges showed a high probability of
voting for the union as well. . . . When holding other characteristics
constant in our judicial profile, in other words, a judge with NLRA
management experience was more than twice as likely as a judge
without that experience to reverse a pro-employer Board outcome and
support the union.27
Can you think of any reason this might be? Here is what the authors of the study posit:
A possible explanation for the noteworthy effects of NLRA
management experience stems from the distinctive nature of the NLRA
itself. This New Deal era statute rests on two legal paradigms that have
become anomalous in our contemporary legal culture. The first
involves the primacy accorded to group action. The Act's emphasis on
protecting collective bargaining and other concerted mutual aid
activity by workers entails a subordination of traditional individualistic
perspectives. Statutory recognition of collectively bargained terms and
conditions of employment means that individuals give up their
freedom to negotiate their own job conditions. . . . [T]his policy
perspective has all but disappeared in recent decades. . . . The second
anomalous aspect of the NLRA is its furtherance of anticompetitive
and interventionist policy goals. By protecting unions in their effort to
remove wages from competition, the Act promotes the cartelization of
labor markets. This approach runs counter to basic norms of our legal
system, norms that oppose restraints on private competition and accept
the efficiency of markets as a means of maximizing consumer and
public welfare. . . . As union power and visibility have declined, the
anticompetitive thrust of NLRA union protections has become harder
to reconcile with the predominantly procompetition aspirations of
federal law.
26
Id. at 1717-19.
27
Id. at 1720.
16
Given the distinctive nature of basic policies and history underlying
the NLRA, the impact of management-side NLRA experience may
signify that familiarity with the Act breeds greater respect for its
protective doctrinal scope--even if the familiarity is developed while
representing employer interests. Attorneys who perform substantial
work representing management before the Labor Board and the courts
come to understand the Act's practical realities and modest policy
goals. These attorneys may be better prepared to breathe life into
NLRA purposes and priorities as an interpretive matter once they are
separated from a client-based ideological perspective.28
These findings suggest that, at least in areas where judges have some discretion,
social background matters. Just as with Tate’s study, the political party of the appointing
President matters—a lot. But so too does the age of the judge, the sex of the judge, where
the judge went to college, and—perhaps most importantly—whether the judge had
relevant NLRA management experience.
Here is some additional evidence that social background affects decisionmaking:
• Professors Theodore Eisenberg and Sheri Lynn Johnson studied all of the
published federal appellate and district court opinions involving constitutional
race-based discrimination claims between 1976 and 1988. In addition to controls
for ideology, sex, and race, their analysis included a number of judicial
background factors including prosecutorial experience, prior judicial experience,
prior elected experience, the judge’s age, and her level of experience.
Interestingly, the results indicate that judges with prior experience as prosecutors
and judges with prior judicial experience are more likely to side with the litigant
bringing the discrimination claim.29
• In an early study of the United States Courts of Appeals, Sheldon Goldman
examined voting behavior in a number of issue areas. His analysis included a
number of background variables including political party of the appointing
President, religion, age, and years on the Court of Appeals. He concluded: “party
and age seemed to have some limited importance in explaining the variance in
judicial behavior, and the other background variables appeared negligible (with
the possible exception of religion).”30
28
Id. at 1742-45.
29
Theodore Eisenberg & Sheri Lynn Johnson, The Effects of Intent: Do We Know How Legal
Standards Work?, 76 CORNELL L. REV. 1162 (1990-1991).
30
Sheldon Goldman, Voting Behavior on the United States Courts of Appeals Revisted, 69 AM. POL.
SCI. REV. 491, 505 (1975).
17
These studies are interesting, but they are just about all we have in terms of
positive findings. In studies of the circuit courts, the effects of social background and
career experience are dwarfed by political party or ideology. A few studies look at similar
factors in the district courts, but the effects just are not there. Many studies look at things
like age, past professional experience, and even the type of law school attended, and
systematically find little differences in the behavior of judges. As an example, the study
of all filed cases in three federal districts by Orley Ashenfelter, et al. finds no effects of
social background whatsoever.31
The literature on the characteristics and background of judges leaves us with a set
of conclusions. In contentious areas of law, and especially when we look at published
opinions—whether in the Supreme Court or circuit courts—the social background of
judges does seem to be related to decisionmaking at the margins. But the effect size
seems fairly insubstantial when we compare it to other factors, like the ideology of the
judge. The findings with regard to race and sex paint a clear picture. On issues like sexual
discrimination or minority voting rights—where female or minority judges may have
unique experiences and insights—we find some sex and race effects. But in just about
every other place we look, there is scant evidence that race and sex affects judging.
FOR DISCUSSION
1. The areas where there seem to significant effects of personal characteristics are those
where individuals might have some personal experience or knowledge, e.g., women and
sex discrimination in the workplace, or African-Americans and voting rights. Why might
this be the case? Can you think of other areas of law where women, Blacks, or judges
with various backgrounds might have law-relevant personal experiences?
2. There are a number of personal characteristics that haven’t been studied very much.
One is geography. Can you think of a time in our nation’s history where the state from
which the judge came might have mattered a lot? Do you think geography is less
important today than in the past? Why or why not?
3. Religion is another area that hasn’t been studied much. Why do you think that is? Can
you think of areas of law where the religion of the judge might play an important role? In
what areas of law might it matter that six of the nine current justices on the U.S. Supreme
Court are Catholic?
4. Can you think of other personal characteristics that might we be worth studying? What
areas of law would you want to look at? Why hypotheses do you have?
31
Orley Ashenfelter, et al., Politics and the Judiciary: The Influence of Judicial Background on Case
Outcomes. 24 J. LEGAL STUD. 259 (1995).
18
5. All of the studies reviewed in this Part look at dispositions rather than opinions, just as
with all of the studies of ideology. How does this limit what we know? Can you think of
reasons why male or White judges might craft opinions differently from their female or
minority colleagues? The idea of a “different voice” by female judges in particular might
suggest a different use of language, which could be picked up in opinions.32
To date no
one has studied this empirically, although the tools of computational linguistics seem
very promising.33
6. In an interesting study using data from the Courts of Appeals, Professors Maya Sen
and Adam Glynn find that for male judges, having daughters causes them to approach
sexual discrimination and sexual harassment cases differently from their colleagues who
don’t have daughters.34
Why do you suppose that is?
32
Suzanna Sherry, Civic Virtue and the Feminine Voice in Constitutional Adjudication, 86 VA. L. REV.
543, 580, 581, 613 (1986).
33
Moshe Koppel, Shlomo Argamon, & Anat Rachel Shimoni, Automatically Categorizing Written
Texts by Author Gender, 17 LITERARY & LINGUISTIC COMPUTING 401 (2002).
34
Maya Sen & Adam Glynn, Like Daughter, like Judge: How Having Daughters Affect Judges'
Voting on Women's Issues (Havard University working paper).