Public Opinion and The Rehnquist Court - Marshall, Thomas R. (Author)

Public Opinion
and the Rehnquist Court
Thomas R. Marshall
Public Opinion and the Rehnquist Court
This page intentionally left blank.
Public Opinion and the
Rehnquist Court
Thomas R. Marshall
S t a t e U n i v e r s i t y o f N e w Yo r k P r e s s
Cover photo credit, istockphoto/Jim Pruitt
Published by
State University of New York Press, Albany
2008 State University of New York
All rights reserved
Printed in the United States of America
No part of this book may be used or reproduced in any manner whatsoever

without written permission. No part of this book may be stored in a retrieval
system or transmitted in any form or by any means including electronic,
electrostatic, magnetic tape, mechanical, photocopying, recording, or otherwise
without the prior permission in writing of the publisher.
For information, contact State University of New York Press, Albany, NY

www.sunypress.edu
Production by Kelli W. LeRoux

Marketing by Michael Campochiaro
Library of Congress Cataloging-in-Publication Data
Marshall, Thomas R., 1949

Public opinion and the Rehnquist court / Thomas R. Marshall
p. cm.
Includes bibliographical references and index.
ISBN-13: 9780791473474 (hardcover : alk. paper)
1. United States. Supreme CourtPublic opinion. 2. Judicial process
United StatesPublic opinion. 3. Public opinionUnited States.
4. Rehnquist, William H., 19242005 I. Title.
KF8748.M288 2008
347.7326dc22
2007024544
10 9 8 7 6 5 4 3 2 1
Contents
Preface vii
List of Illustrations ix
List of Tables xi
Chapter One: Public Opinion and Supreme Court Policy-Making 1
A Representative Court? 1
How Pollsters View the Court 3
How the Justices View Public Opinion 4
Judicial Theories of Public Opinion 8
Linking Public Opinion and the Rehnquist Court 14
Conclusion 21
Chapter Two: Public Opinion and the Rehnquist Court 23
Studying Representation 24
Data Results 35
The State of Public Opinion Model 38
The Federal Policy Model 40
The State/Local Policy Model 41
Alternative Linkages 42
Conclusion 49
Chapter Three: Judicial Norms, Representation,

and Public Opinion 51
Ideology and Representation 52
Poll Correction and the Lower Courts 55
Judicial Activism, Judicial Restraint, and Representation 59
Caseload and Representation 62
Conclusion 65
Chapter Four: Interest Groups, Representation,
and Public Opinion 67
Interest Group Litigation and Public Opinion 68
Public Opinion and Litigant Success Rates 69
Conclusion 74
v
vi Contents
Chapter Five: Justices and Representation 77

The Political Parties and Ideology Model 80
The Political Socialization Model 84
The Appointments Process Model 90
The Judicial Roles Model 92
The Tenure and Aging Model 95
The Realignment Model 99
Combining the Predictors 102
Conclusion 104
Chapter Six: Symbolic Representation and the Court 107

The Politics of Symbolic Representation 107
Symbolic Representation on the Rehnquist Court 112
Conclusion 119
Chapter Seven: Did the Rehnquist Court Influence

Public Opinion? 123
Public Impressions of the Court 123
The Short-Term Manipulation Model 131
The Long-Term Manipulation Model 134
Conclusion 136
Chapter Eight: Public Opinion and the Test of Time 139

Surviving the Test of Time 142
Data Analysis 147
Conclusion 151
Chapter Nine: An Empirical Model of Representation 153

The Majoritarian Court 153
Reconsidering the Evidence 154
An Empirical Linkage Model 157
Appendix One: Poll-to-Ruling Matches 163
Appendix Two: List of Cases 171
Notes 181
Bibliography 215
Index 259
Preface
P ublic Opinion and the Rehnquist Court asks two simple questions.
First, did the Rehnquist Courts decisions typically reflect American
public opinion? Second, to the extent its decisions did, why? To pre-
view this books findings, between three-fifths and two-thirds of the
Rehnquist Courts decisions agreed with American public opinion.
At least since the New Deal era, most Supreme Court decisions
have represented American public opinion, well-known exceptions
notwithstanding. Some justices agree with American public opinion
much more often than do others. The Courts pattern of representation
is sensitive to ideological shifts on the Court, caseload, the justices the-
ories of judicial decision-making, the justices backgrounds and roles,
and the attentiveness of American public opinion. The Rehnquist Court
enjoyed relatively high popularity, at least compared to the other polit-
ical branches. Yet it could not influence American public opinion
through its decisions.
Sixteen years ago, during the last Burger Court term, I wrote an
account of Supreme Court decision-making and American public opin-
ion, Public Opinion and the Supreme Court. That book relied heavily
on 146 matches between public opinion poll questions with Supreme
Court decisions since the mid-1930s. Since that time, judicial scholars
have made great advances in studying judicial politics and representa-
tion. I gratefully acknowledge the many authors work cited in this
book, as well as the polling data used herein, without which Public
Opinion and the Rehnquist Court would not have been possible. I hope
this book both updates and expands on my previous work. For their
comments, I thank Charles Hadley and Steve Wasby. For assistance in
producing this manuscript, I thank Erika Contreras, Lydia Villegos,
Shelby Henderson, Kenneth Zuercher, and Leah Cook. This book is for
Ruth, Aram, Erica, Deacon, Sara, Jack, Claire, Mark, and Bertie.
vii
List of Illustrations
Figure 1.1 Number of poll questions, per year, during

the Rehnquist Court 5
Figure 2.1 Percentage of Supreme Court decisions consistent
with public opinion, five-year moving averages,
19702005 37
Figure 3.1 Percentage of Rehnquist Court decisions
consistent with public opinion, with differing
vote changes, by ideology 54
Figure 3.2 Percentage of Supreme Court decisions
consistent with public opinion, prior to
the Rehnquist Court, with differing vote changes,
by ideology 56
Figure 7.1 Public confidence in the Supreme Court,
the Executive Branch, and Congress,
NORC/GSS positive ratings 126
Figure 7.2 Short-term poll shifts during the Rehnquist Court 132
Figure 7.3 Long-term poll shifts during the Rehnquist Court 135
Figure 8.1 Probability that a Rehnquist Court decision
will survive, by whether the decision is
consistent with public opinion or not 151
ix
List of Tables
Table 1.1 Frequency of Direct Mentions of Public Opinion

in Supreme Court Decisions, by Era 6
Table 1.2 Evidence of Public Opinion Used in Supreme
Court Opinions, by Era 7
Table 1.3 Type of Case Involved in Direct Mentions of
Public Opinion, by Era 9
Table 1.4 Direct Mentions of Public Opinion, by Type of
Opinion and Era 9
Table 1.5 Theories of Public Opinion, by Frequency and Era 10
Table 2.1 Number of Poll Matches, by Issue, Including
or Excluding Denials of Certiorari 36
Table 2.2 Supreme Court Agreement with Public Opinion,
by Court, Including All Poll Matches 36
Table 2.3 Supreme Court Agreement with Public Opinion,
by Court, Excluding Instances of Unclear Polls 37
Table 2.4 Percentage of Consistent Decisions, by Explanatory
Variables, for the Rehnquist Court 44
Table 2.5 Percentage of Consistent Decisions, by Issue 47
Table 2.6 Percentage of Consistent Decisions for the
Rehnquist Court and Earlier Courts since
the mid-1930s 48
Table 4.1 Interest Group Involvement and Success
during the Rehnquist Court 71
Table 5.1 Rehnquist Court Justices Ranked by Agreement
with Public Opinion 78
Table 5.2 Rehnquist Court Justices Voting Patterns,
Percentage of Conservative Votes, by
Ideology of Public Opinion 79
xi
xii List of Tables
Table 5.3 Rehnquist Court Justices Voting Patterns, by

Public Opinion Ideology, by Groups of Justices 80
Table 5.4 Percentage of Justices Votes Consistent with Public
Opinion, by Political Party and Ideology 83
Table 5.5 Percentage of Votes Consistent with Public
Opinion, by Justices Backgrounds 85
Table 5.6 The Appointments Process Model 91
Table 5.7 The Judicial Roles Model 93
Table 5.8 The Length-of-Tenure Model 96
Table 5.9 The Realignment Model 101
Table 5.10 Predicted versus Actual Agreement With Public
Opinion, by Justice 103
Table 6.1 Percentage of Rehnquist Court Decisions That
Agreed with Group Attitudes 114
Table 6.2 Agreement Scores, by Gender 114
Table 6.3 Agreement Scores, by Race 115
Table 6.4 Agreement Scores, by Religion 116
Table 6.5 Agreement Scores, by Political Party 116
Table 6.6 Agreement Scores for Types of Justices, Only for
Decisions in which Groups Disagree on a Ruling 117
Table 6.7 Agreement Scores for Selected Core Cases Only 118
Table 7.1 Explanations for Short-Term and Long-Term
Poll Shifts 133
Table 8.1 Percentage of Rehnquist Court Decisions Classified
as Prevailing 148
CHAPTER ONE
Public Opinion and Supreme Court

Policy-Making
A REPRESENTATIVE COURT?
S hould U.S. Supreme Court decisions reflect American public opin-

ion? Do most of its decisions actually reflect American public opin-
ion? Since the late 1700s, these questions have divided politicians,
judicial scholars, ordinary Americans, and the justices themselves.
Alexander Hamiltons well-known essay, Federalist 78, argued
that a life-tenured federal judiciary would protect minorities constitu-
tional rights by serving as an excellent barrier against the
encroachments and oppressions of the representative body and the
occasional ill humors in the society. In Hamiltons view, the federal
courts willand sometimes shoulddefy legislators, presidents, and
majority public opinion. This argument was not unique. Many, but
not all of Hamiltons contemporaries held a negative view of majority
public opinion.1 Antifederalist critics of the newly proposed Consti-
tution feared Hamilton would be correct, and that the federal courts
would often defy majority public opinion (Friedman 2002; Storing
1985: 197).
Until the 1930s there was no direct test by which to tell whether or
not Supreme Court decisions agreed with American public opinion. Did
Marbury v. Madison (1803) agree with American public opinion? Did
McCullough v. Maryland (1819)? Gibbons v. Ogden (1824)? Dred
Scott v. Sanford (1857)? Plessy v. Ferguson (1896)? Lochner v. New
York (1905)?
Regrettably, no public opinion polls are available on these land-
mark decisions. True, some public opinion polls were conducted in the
United States as early as the 1824 presidential election (Converse 1987,
Robinson 1932, Smith 1990b). However, these early straw polls
were conducted with very informal and unscientific methods, and only
in selected localities, states, or regions. Until the 1930s, few polls were
1
2 Public Opinion and the Rehnquist Court
conducted on a nationwide basis. Further, until the 1930s, most polls

only measured the attitudes of targeted or local groups, such as farmers
or consumers or likely voters, rather than the general public. When poll
results could be compared to actual behavior, such as the vote returns
in presidential elections, these early polls often had very large margins
of error. By the early 1900s, nationwide pollsthe Literary Digest poll
being the best-known examplefocused on voter choices in upcoming
presidential elections, not on attitudes toward Supreme Court deci-
sions. As a result, one cannot unearth polls of nationwide public opin-
ion on prominent Supreme Court decisions before the 1930s.
During the 1930s, pioneering pollsters such as George Gallup,
Elmo Roper, and Archibald Crossley developed new sampling methods
that allowed cheaper, faster, and more accurate nationwide polls than
those in earlier years (Converse 1987). For the first time nationwide
polls included questions on Supreme Court decisions. That pollsters
surveyed attitudes on well-known Supreme Court decisions is not sur-
prising. Many important controversies eventually become lawsuits and
are appealed to the Supreme Court (Cushman 2002). Indeed, nearly a
century before modern public opinion polling, the French journalist,
Alexis de Tocqueville, noted that (t)here is hardly a political question
in the United States which does not sooner or later turn into a judicial
one (Lawrence 1966: 248).
Modern public opinion polls made it possible to determine whether
Supreme Court decisions reflected American public opinion. During the
Rehnquist Court, for example, poll majorities agreed with the Good
News Club v. Milford Central School (2001) decision allowing after-
hours use of school facilities by student religious groups.2 Poll majori-
ties disagreed with the Penry v. Lynaugh (1989) decision upholding the
death penalty for mentally retarded convicted murders.3 Multiple polls
yielded closely divided and conflicting results on the Bush v. Gore
(2000) ruling, ending the Florida presidential recount and effectively
giving the 2000 presidential election to George W. Bush.
Prior to the Rehnquist Court, if a clear poll majority (or plurality)
existed, three-fifths (63%) of Supreme Court decisions agreed with
public opinion (Marshall 1989: 7879). Supreme Court decisions typi-
cally agreed with public opinion when public attention was closely
focused on a dispute (such as during crisis times) or when a federal
law or policy was involved (most of which laws and policies were
themselves consistent with public opinion polls). Chief justices, ideolog-
ically moderate justices, and justices from prestigious law schools most
often cast votes agreeing with public opinion. Supreme Court rulings
themselves did not usually greatly affect public opinion. Unpopular
Public Opinion and Supreme Court Policy-Making 3
Supreme Court decisions were more often overturned by constitutional

amendments, Congress, the president, or by the Court itself.
Were the Rehnquist Courts decisions as often consistent with
American public opinion as those of earlier Courts since the 1930s?
What best explains which Rehnquist Court decisions agreed with
American public opinion and which did not?
The best way to answer these questions is to compare nationwide
public opinion polls with Supreme Court decisions. This book sug-
gests that the Rehnquist Court was consistent with public opinion in
three-fifths to two-thirds of its decisionsroughly as often as were
earlier Courts since the 1930s. Yet the Rehnquist Court differed in
several important ways from earlier Courts in how it achieved that
level of representation.
HOW POLLSTERS VIEW THE COURT
American pollsters often write questions about Supreme Court rulings,

the justices and nominees, or public knowledge of or attitudes toward
the Court itself. During the Rehnquist Court, pollsters wrote well over
two thousand poll questions tapping attitudes on Supreme Court deci-
sions, the Court as an institution, or individual justices or nominees.
On the average, major pollsters wrote roughly 122 questions a year
thereby providing a rich source of data by which to examine attitudes
toward the Court
To investigate how pollsters view the Supreme Court, the Roper
Archives poll questions on the Supreme Court were reviewed. All poll
questions directly mentioning the U.S. Supreme Court were counted,
as were questions that mentioned individual justices or nominees
(whether confirmed or not), and questions on an upcoming or previ-
ously announced decision.4
A total of 2,310 poll questions were identified for the Rehnquist
Court era. The most common category (44%) of these questions mea-
sured attitudes toward specific decisions. Nearly all these were
approval questions on specific cases, mostly pending or recently
decided, but also including a few long-past decisions, for example, Roe
v. Wade (1973).5 Only a few (about 2%) of all poll questions tapped
the publics factual knowledge of decisions. Later chapters rely heavily
on these matches between attitudes and specific rulings.
Nearly as many questions (41%) measured public opinion
toward Supreme Court nominees or sitting justices. Most of these
questions were on controversial nominees, particularly Clarence
Thomas and Robert Bork. A small number of questions tapped

either factual knowledge or attitudes toward sitting justices (about
1% each of all questions).
Fifteen percent of poll questions tapped attitudes toward the
Supreme Court as an institution. The most common such questions
measured trust, confidence, or approval of the Court as an institution
(6% of all questions). Pollsters also wrote a few questions to tap factual
knowledge of the Court (2% of all questions), perceptions of the
Courts ideology and preferences for ideological change (3% of all
questions), or the importance of the Court as an election issue (3% of
all questions). The remaining 1% of all questions fell into the miscella-
neous category.
Pollsters attention to the Rehnquist Court varied considerably
from year to year, with the number of poll questions averaging 122 a
year. However, this average is deceptive. Only two types of events trig-
gered an outburst of poll questions. The first was a controversial deci-
sion such as the flag-burning and abortion rulings in 1989 or the
Florida presidential vote dispute in 2000. The other was a controversial
or important nominee, such as Robert Bork in 1987, Clarence Thomas
in 1991, or John Roberts in 2005. During these five years the number
of poll questions rose sharplyto an average of 273 per year. For the
remaining years the number of questions averaged only 53 a yearor
about one poll question a week. Given that most pollsters archived in
the Roper Archive are financed by the news media, this current
events view of the Supreme Court is not surprising. Yet even with this
caveat, these poll questions provide a rich source of information on
American attitudes toward the Court. Figure 1.1 tracks the number of
poll questions annually.
HOW THE JUSTICES VIEW PUBLIC OPINION
At least 123 Rehnquist Court opinions directly mentioned public opin-

ion in a majority, concurring, dissenting, or per curiam opinionan
average of about six to seven opinions per term. A direct mention of
public opinion either uses the term public opinion, a close rewording
(such as opinion of the public), or a close synonym that clearly refers
to the mass publics attitudes and beliefs (such as poll data, survey
results, the common perception, public reputation, public confi-
dence, publics consciousness, a national consensus, public dis-
approval, public confidence, public trust, or public attitudes
and beliefs).6
Figure 1.1
Number of poll questions, per year, during the Rehnquist Court
As Table 1.1 indicates, over two centuries the number of direct

mentions of public opinion, per term, rose slowly. The Rehnquist Court
continued the long upward historical trend in direct mentions of public
opinion. In fact, the number of direct mentions during the Rehnquist
Court era about doubled from the Warren and Burger Courts.7 The
over-time rise in the number of direct mentions of public opinion
resulted, in part, from a growing acceptance of public opinion as rele-
vant to several key legal theories, discussed further below. The rising
number of direct mentions is probably also an artifact of the growing
length of written opinions and the growing number of dissents and
concurring opinions (Ginsburg 1990; Johnson 1999; OBrien 1999).
Each of the fourteen justices who served on the Rehnquist Court
made at least one direct mention of public opinion, and several justices
frequently referred to public opinion. Not surprisingly, the justices who
most frequently referred to public opinion in their authored opinions
were also those who served throughout the entire Rehnquist Court
eraJustices Stevens (thirty-five direct mentions), Scalia (twenty-three
direct mentions), Kennedy (nineteen), OConnor (eighteen), and Chief
Justice Rehnquist (sixteen). On a per-term basis the fourteen justices
mentioned public opinion about equally often; none of the justices
mentioned public opinion more than an average of twice per term in a
majority, concurring, or dissenting opinion.8 Liberal justices and con-
servative justices mentioned public opinion about equally often.
Table 1.1
Frequency of Direct Mentions of Public Opinion in
Supreme Court Decisions, by Era
Average Annual Number of Direct Mentions
17921859 .3
18601933 .5
19341959 2.4
19601986 3.1
19862005 6.5
Although the justices often discussed public opinion, they rarely

referred to specific polls. As Table 1.2 indicates, two-thirds (70%) of
direct mentions to public opinion were normative, theoretical, or
abstract, and contained no empirical description at all. The next most
common reference was to very indirect measures of public opinion,
such as general knowledge or election results, statutes, or lower court
decisions. Only a small percentage (15%) of direct mentions actually
cited any specific polls.
As Table 1.2 indicates, few direct mentions of public opinion actu-
ally cited specific public opinion polls, either during the Rehnquist
Court (when poll results are more often and more easily available) or in
earlier time periods. In part, this inattention to polls occurs because
polls were apparently not introduced as evidence in many cases that
reach the Supreme Court. The Rehnquist Courts citations to specific
poll questions most frequently appeared in two types of cases: first,
cases involving the death penalty and information given to juries during
sentencing; and second, in cases involving trademarks, secondary mean-
ings, and commercial free speech and advertising claims. Examples of
the former cases include Atkins v. Virginia (2002), In re Stanford
(2002), Penry v. Lynaugh (1989), Simmons v. South Carolina (1994),
Brown v. Texas (1997), and Ramdass v. Angelone (2000). The latter
cases rarely win Supreme Court review; for examples that did, see Wal-
Mart Stores, Inc. v. Samara Bros (2000), Thompson v. Western States
Medical Center (2002), Lorillard Tobacco Co v. Massachusetts (2001),
or Borgner v. Florida Board of Dentistry (2002).9
Perhaps the lengthiest discussion of specific poll results to date is in
Atkins v. Virginia (2002), a landmark case in which the Court, 63,
ruled that mentally retarded convicted murderers could not be exe-
cuted. Justice Stevens majority opinion cited a list of national poll
questions submitted in an amicus brief as evidence of a national con-
sensus among Americans against executing mentally retarded convicted
murderers.10 Chief Justice Rehnquist, dissenting, criticized the same
polling data as too inadequately explained and poorly presented to be

credible. Justice Rehnquist also argued that legislatures and juries were
better indicators of American attitudes.
Table 1.2
Evidence of Public Opinion Used in Supreme Court Opinions, by Era
1792-1859 1860-1933 1934-1986 1986-2005
Direct evidence: 9% 16%
specific poll results (8%) (15%)
poll results mentioned, but
no specific polls cited (1%) (1%)
Indirect evidence: 73% 81% 39% 28%
Elections, referendums (2%) (2%)
Community actions, word-of-
mouth, news stories, editorials,
mob actions) (9%) (14%) (4%) (2%)
Elite or specialized opinions (3%) (14%) (2%) (2%)
General knowledge (32%) (48%) (17%) (15%)
Statutes, lower court decisions (29%) (5%) (14%) (7%)
Normative, theoretical, or
abstract (non-empirical)
mentions only 44% 19% 70% 70%
Note: Percentages sum down and may exceed 100%, by column, due to multi-
ple sources indicated in an opinion. Percentages in parentheses indicate subcat-
egories.
In part, the inattention to specific polls may occur because several

justices believe public opinion should not play a role in Supreme Court
decision-making, or else these justices base their readings of American
public opinion on statutes or jury verdicts, rather than poll results.11
For example, in Planned Parenthood of Southeastern Pennsylvania v.
Casey (1992), at 958959, Justice Rehnquist, dissenting, wrote that
the Courts duty is to ignore public opinion and criticism on issues
that come before it.12 He noted that Supreme Court justices were poor
judges of contemporary public opinion. Similarly, in Stanford v.
Kentucky (1989), a death penalty case, Justice Scalia wrote that statutes
and jury decisions, not public opinion polls, were the best indicators of
contemporary public opinion.13
The justices often write about public attitudes and beliefs without
referring to any empirical evidence. As discussed at greater length later
in this chapter, Justices OConnor and Scalias lengthy debate over
stare decisis, public confidence, and the Courts legitimacy in Planned
Parenthood of Southeastern Pennsylvania v. Casey (1992) is a well-

known example of discussing the proper role of public opinion without
any polling data at all.
Other justices set a high standard for the type of survey research
they would consider. In practice, this demanding standard disallows
many polls sponsored by litigants and polls with small or inappropriate
samples, including many polls reported in law journal articles that
might otherwise be cited in an opinion.14 In Ramdass v. Angelone
(2000) Justice Kennedy wrote that courts should disregard polls that
had inadequate sample sizes, samples drawn from the wrong area,
poorly drafted questions, inappropriate interviewing methods, or
biased sponsors.15 By contrast, Justice OConnor set out a much more
lenient standard for introducing polls and surveys as evidence in
Florida Bar v. Went For It, Inc (1995).16
During the Rehnquist Court, direct mentions of public opinion
most often occurred in criminal disputes. The plurality (41%) of the
Rehnquist Courts direct references to public opinion involved a jury
trial, a death penalty case, a criminal statutes interpretation, or some
other court proceeding. Courtrooms, trials, sentencing, and other crim-
inal matters have long been an area in which the Supreme Court evalu-
ated American public opinionin large part because these cases often
involve the Eighth Amendments ban on cruel and unusual punish-
ment, and many justices interpret this clause in light of changing
public opinion (Robinson 2004). No other topic elicited even half as
many references to public opinion as did crime and trials. Table 1.3
compares the Rehnquist Courts direct mentions of public opinion to
earlier time periods.
The Rehnquist Courts direct mentions to public opinion included
a mix of majority, concurring, and dissenting opinions, as reported in
Table 1.4. A larger percentage appeared in dissenting opinions, as com-
pared to earlier time periods. The growing percentage of direct men-
tions to public opinion in dissenting and concurring opinions may
simply be an artifact of the growing number of concurring or dissenting
opinions, sometimes very lengthy ones.
JUDICIAL THEORIES OF PUBLIC OPINION
Long before modern public opinion polling developed during the

1930s, the Supreme Court developed four theories to explain what role
public opinion should play in judicial policy-making (Roesch 2006).
None of these theories requires a very precise reading of public opin-
ion.17 The Rehnquist Court did not originate any new or novel theories
of the role of public opinion. However, as Table 1.5 indicates, it relied
on the theory of evolving or contemporary public opinion much more
often than did earlier Courts, and it almost ceased to use the judicial
restraint theory of public opinion. Overall, nearly four-fifths of the
Rehnquist Courts direct mentions of public opinion reflected a gener-
ally positive view of public opinion. Only a fifth of its direct mentions
reflected the negative view that public opinion was a threat to rights.
That the justices themselves apparently hold a generally positive view
of American public opinion may, in part, explain why most Supreme
Court decisions agree with public opinion.
Table 1.3
Type of Case Involved in Direct Mentions of Public Opinion, by Era
1792-1859 1860-1933 1934-1986 1986-2005
Crime, trials, prisoners,
courtrooms 19% 24% 35% 41%
Press, media 6% 21% 2%
Dissent, speech, religion 18% 11%
Labor, strikes, bargaining 6% 14% 2%
Elections, campaign,
campaign finance 6% 8% 11%
Business regulations 48% 50% 8% 16%
Foreign, military policy 33% 9% 5% 1%
Civil rights, race 5% 12% 4% 9%
Privacy, sex, obscenity 3% 11%
All others 3% 6% 6%
Note: Percentages sum down and may exceed 100%, per column, because mul-
tiple topics are involved in a single case.
Table 1.4
Direct Mentions of Public Opinion, by Type of Opinion and Era
1792-1859 1860-1933 1934-1986 1986-2005
Direct mention of public
opinion in
Majority opinion 76% 79% 52% 44%
Concurring opinion 5% 0% 13% 14%
Dissenting opinion 19% 21% 35% 42%
Note: Percentages sum down, by column, to 100%.
As a caveat, most direct mentions of public opinion are not based

on any close reading of public opinion polls. Indeed, several justices are
quite skeptical that polls should carry much weight in judicial decision-
making. While the Rehnquist Court was as sensitive to public opinion
as were earlier Courts since the mid-1930s, it was clearly not poll
driven.18 The next section briefly reviews how the Rehnquist Courts
justices applied each of these four theories of public opinion.
Table 1.5
Theories of Public Opinion, by Frequency and Era
1792-1859 1860-1933 1934-1986 1986-2005
Speech or action influences
or informs public opinion, and
merits protection 6% 41% 34%
In press cases (6%) (17%) (3%)
In election, religion, speech
and dissent cases (16%) (10%)
all other cases (8%) (21%)
Public opinion alone is an
adequate check on policy 24% 50% 21% 7%
For government actions (14%) (44%) (11%) (6%)
For nongovernment groups (10%) (6%) (10%) (1%)
Law and policy should reflect
contemporary or evolving
public opinion 43% 29% 21% 64%
Public opinion threatens rights,
should be restrained 29% 18% 19% 20%
In fair trial rights cases (6%) (8%) (11%)
In speech and dissent cases (3%) (8%) (2%)
In economic rights (29%) (9%) (3%) (1%)
All other cases - (7%)
All other usages 5% 3% 6% 14%
Government influences public
opinion (3%) (2%)
Miscellaneous uses (5%) (3%) (3%) (11%)
Note: Percentages sum down and may exceed 100%, by column, due to multi-
ple usages in some opinions.
An Informed Public Opinion

Since the early 1900s, the Supreme Court has often held that certain
types of speech inform public opinion, and that even controversial or
unpopular speech deserves legal protection. Earlier Courts typically
used this theory on behalf of the news media, political dissidents, labor
union actions (such as picketing), or commercial speech (such as adver-

tising). Justices Oliver Wendell Holmes and Louis Brandeis are well
known for advancing this theory during the early 1900s. From the mid-
1930s through the mid-1980s, it became the Supreme Courts single
most frequent theory of public opinion. Well-known examples include
Schenck v. U.S. (1919), Whitney v. California (1927), Near v. Min-
nesota (1931), Thornhill v. Alabama (1940), In re Oliver (1948), and
New York Times Company v. Sullivan (1964).
The Rehnquist Court continued to use this theory, although not
quite as heavily as did earlier Courts since the 1930s. A third (34%) of
the Rehnquist Courts direct mentions of public opinion relied on this
theory. Examples include protecting speech that informs or influences
trials and lawsuits, in Simmons v. South Carolina (1994), at 170171
(holding that juries should be informed about the possibility of parole
in a death penalty case), or Gentile v. State Bar of Nevada (1991), at
1043, 1064 (allowing a defense attorneys out-of-courtroom efforts to
influence public opinion on behalf of his client).19 This theory was also
used in defense of speech in elections,20 by dissenters,21 by the media,22
or by interest groupsfor example, that expressive private groups
can limit their membership despite prevailing public opinion, in Boy
Scouts v. Dale (2000), at 648, 650, 661. Elsewhere, the theory was
used in the defense of commercial speech that informs consumer opin-
ion, in Greater New Orleans Broadcasting Association v. U.S. (2000),
Thompson v. Western States Medical Center (2002), or 44
Liquormart, Inc., v. Rhode Island (1996), and in defense of campaign
disclosure laws23 or secondary picketing in a labor dispute in
Burlington Northern Railroad Co. v. Brotherhood of Maintenance of
Way Employees (1987).
Judicial Restraint
A second theory is that public opinion alone is an adequate check on
abuses by government, and that the courts need not become involved in
a dispute. This theory became popular during the early 1900s when
several judges (most notably, Justices Holmes and Stone) urged the
Court to defer to Congress or state legislatures in economic policy-
making. Justice Frankfurter often argued that the Court should defer to
public opinion and elected officials in criminal, dissent, and redistrict-
ing cases.
Decisions in favor of judicial restraint were quite common during
the Rehnquist Court, but few such decisions were based on public
opinion. This theory apparently fell out of favor with the justices
during the Rehnquist Court. Examples include Justice Breyers argu-

ment that the Court should defer to a congressional statute protecting
persons with disabilities, in Board of Trustees v. Garrett (2001), at
964, or Justice Souters argument that the Court should defer to con-
gressional findings that rape was a serious national problem, in U.S. v.
Morrison (2000), at 628.
Contemporary Public Opinion

Since the late 1800s,24 the Supreme Court advanced a third theory of
public opinion by arguing that the Court should strike down laws and
policies inconsistent with contemporary or evolving public opinion.
This theory became the Rehnquist Courts most popular theory, with
two widely used versions.25 In the first version the Court may hold that
a practice was permissible at an earlier time, and perhaps even widely
practiced when the U.S. Constitution was signed. However, since then
American public opinion has changed, and the practice is no longer
acceptable.26 Weems v. U.S. (1910)27 and Trop v. Dulles (1957)28 are
landmark decisions holding that changing public opinion should deter-
mine what constitutes a cruel and unusual punishment. Earlier
Courts used this theory most frequently in criminal cases, particularly
in death penalty cases.29
The Rehnquist Courts justices frequently debated this theory both
in criminal and noncriminal cases. In criminal cases the Rehnquist
Court used the theory in arguing that the Miranda warning should con-
tinue to be required, in part because it had become part of our
national culture.30 This theory was also found in several other crimi-
nal cases, particularly death penalty cases, for example, Gomez v. U.S.
District Court (1992),31 Penry v. Lynaugh (1989),32 Stanford v. Ken-
tucky (1989), Thompson v. Oklahoma (1988),33 Campbell v. Wood
(1994),34 Atkins v. Virginia (2002), and Roper v. Simmons (2005). In
noncriminal cases the theory of evolving public opinion appears in Boy
Scouts v. Dale (2000), in dissent, at 700702, and in National
Endowment for the Arts v. Finley (1998).
An alternative, albeit less common version of this theory argues
that the Court should follow public opinion because public confidence
is important to the Courts legitimacy. The Rehnquist Courts best-
known debate over this theory was in Planned Parenthood of
Southeastern Pennsylvania v. Casey (1992), involving several state
restrictions on abortions. Justice OConnors opinion, in part, argued
that in extremely important cases the Court should consider public
opinion in deciding whether to overturn precedents. OConnor argued
that Brown v. Board of Education (1954, 1955) and West Coast Hotel
v. Parrish (1937) correctly overturned earlier rulings, partly because
American public opinion, American social life, and Supreme Court
decisions had greatly changed since the earlier decisions in Plessy v.
Ferguson (1896) or Lochner v. New York (1905). OConnor argued
that overturning Roe v. Wades (1973) view that abortion was a funda-
mental right would call into question public confidence in the Supreme
Court: The Courts power lies . . . in its legitimacy, a product of sub-
stance and perception that shows itself in the peoples acceptance of the
Judiciary as fit to determine what the Nations law means and to
declare what it demands, at 865. In Justice OConnors view, the
Court could protect its legitimacy by upholding a well-known, recent
precedent that still enjoyed widespread legal acceptance, especially
when public opinion was still sharply divided over the decision.35
Predictably, other justices vigorously criticized this argument.
Justice Scalia, in dissent, argued that there was no fundamental right to
an abortion, and that Roe v. Wade (1973) was wrongly decided; he
also argued that later decisions sharply limited Roe. Justice Scalia then
criticized OConnors argument that the Court should resist public
opinion criticism of landmark rulings: I am appalled by the Courts
suggestion that the decision whether to stand by an erroneous constitu-
tional decision must be strongly influenced . . . by the substantial and
continuing public opposition the decision has generated . . . (W)hether it
would subvert the Courts legitimacy or not, the notion that we would
decide a case differently from the way we otherwise would have in
order to show that we can stand firm against public disapproval is
frightening . . . (T)he notion that the Court must adhere to a decision for
as long as the decision faces great opposition and the Court is under
fire acquires a character of almost czarist arrogance, at 998999.
Several other decisions argued that public perceptions should
influence Supreme Court rulings. Most of these cases involved criminal
trials, and suggested that an appeals court should correct a lower
courts plain errors if there was evidence of actual innocence, or if
the error seriously affects the fairness, integrity, or public reputation
of judicial proceedings.36 This argument about the public reputa-
tion of the courts reappeared in later criminal cases.37 The Rehnquist
Court also considered the importance of public perceptions in other
areas, arguing that public opinion should be given serious weight in
campaign finance regulations,38 gambling regulations,39 regulations
concerning payments to or paid appearances by public employees,40
financial and banking regulations,41 and government-funded legal ser-
vices to the poor.42
Public Opinion as a Threat

In the fourth theory, public opinion can at times threaten basic consti-
tutional rights, and in these instances, should be ignored. Prior to the
New Deal period this argument was most often made when justices
voted against economic regulations, such as minimum wage laws,43 or
when criminal defendants had been convicted amidst an atmosphere of
outraged local community opinion or threatened mob violence.44
During the mid-1900s, this theory was used to support political dis-
senters, especially suspected Communists45 or civil rights groups.46
The Rehnquist Court applied this theory in a fifth (20%) of its
direct mentions of public opinion, most frequently in disputes over fair
trial rights, such as California v. Brown (1987), allowing a jury instruc-
tion to disregard mere sentiment, conjecture, sympathy, passion, prej-
udice, public opinion, or public feeling. Several justices argued
(usually in dissent) that public opinion favoring the death penalty
threatens a defendants ability to receive fair (i.e., nondeath) sentence
see, for example, Harris v. Alabama (1995), at 519, or Payne v.
Tennessee (1991), at 866. 47 Elsewhere, this theory was used in business
regulation cases,48 in First Amendment cases,49 or on behalf of mentally
disabled persons,50 gays and lesbians,51 or families of suicide victims.52
LINKING PUBLIC OPINION AND THE REHNQUIST COURT
By this books count, three-fifths to two-thirds of Rehnquist Court

decisions agreed with American public opinion. How can this pat-
tern best be explained? From the 1930s through the mid-1980s,
Supreme Court decision-making was not equally likely to agree with
public opinion on all types of cases. The Court most often agreed
with public opinion in interstate commerce and civil rights decisions.
On criminal rights, abortion, morality, privacy, and poverty deci-
sions, the Court least often agreed with public opinion. What
explains such patterns?
Court-watchers offer fifteen linkage models, described briefly
below, to explain the relationship between Supreme Court decision-
making and American public opinion. Some models focus on public
opinion itselffor example, whether public attention is closely focused
on an issue before the Court. Other models focus on popular judicial
theories, such as federalism, judicial activism and restraint, or the pre-
ferred position argument. Other models focus on the justices back-
grounds, length of tenure on the Court, or on-the-Court roles. Still
other models focus on public reactions to Court decisions, or on key

changes in public opinion, such as political realignments.
The fifteen linkage models come from a variety of different sources:
well-known judicial theories, famous Supreme Court cases, case stud-
ies, judicial biographies, on-the-Court bargaining strategies, well-
known scholarly essays, or empirical studies from other fields in
political science, such as political socialization or interest groups.
Individually, each model may seem logical and persuasive, but whether
it is actually a good description is an empirical question. Because the
models vary so widely, they must be tested in different ways: some by
examining Supreme Court decisions, others at the level of the individ-
ual justice, and still others by examining events such as realignments or
public opinion changes outside the Court.
The remainder of chapter one briefly describes the fifteen linkage
models, the source of each model, and the predictions each model
generates. Later chapters reexamine, document, and test each model.
Just as for earlier Courts, several models receive strong empirical sup-
port during the Rehnquist Court, but for other models, little or no
support appears. Chapter nine summarizes the evidence and offers an
empirical model linking American public opinion and Supreme Court
decision-making.
The State of Public Opinion Model

In the first model, the distribution and intensity of public opinion affect
the Supreme Courts level of agreement with public opinion. Supreme
Court decisions should most often agree with public opinion under
three conditions. First, decisions should usually agree with one-sided
public opinionthat is, when very large poll majorities exist on an
issue. Second, decisions and public opinion should agree when public
opinion is closely focused on an issue, such as during crisis times.
Third, decisions should agree with public opinion when an issue is
highly visiblethat is, when nearly all Americans express an opinion
on an issue before the Court. When any or all of these conditions exist,
Court decisions should significantly more often agree with American
public opinion than a random-behavior model would predictthat is,
more often than half of the time. Under any of these three conditions,
the justices may themselves most easily sense and share current
public opinion (Rehnquist 1986). In this first model, justices are not
necessarily coerced by American public opinion. Rather, the justices are
themselves attentive to and influenced by current issues, events, and
media coverage, and the justices thereby sense and share prevailing
American attitudes. As one Court-watcher wrote: (T)he justices have

often agreed with the main current of public sentiment because they
were themselves part of that current, and not because they feared to
disagree with it . . . (T)he Court has seldom lagged far behind or forged
far ahead of America (McCloskey 1960: 224).
When public opinion is one-sided, closely focused on an issue, and
well informed on an issue, the justices would not necessarily need to see
any public opinion polls.53 The justices could gauge public attention
and interest from news reports, public statements, or interest group
activities. At the extreme, the justices may sense that a highly unpopu-
lar decision might be widely criticized, evaded, or overturned, or that
Congress and the president might even try to limit the Courts jurisdic-
tion. Chapter two tests this model at the decision level.
The Federal Policy Model

For most of its history the Supreme Court was extremely deferential
toward federal laws and policies. Indeed, for the first seventy years of its
history the Court struck down as unconstitutional only two federal laws
Marbury v. Madison (1803) and Dred Scott v. Sanford (1857). Although
the Supreme Court subsequently struck down many more federal laws and
policies, it still does so far less often than for state-level laws and policies
(Epstein et al. 1996: 144174). Historically, on the average the Court
struck down as unconstitutional only one or two federal laws or policies,
per term, while striking down on constitutional grounds roughly ten times
that many state or local laws per term (Baum 1998a: 199204).54 Taking
this argument a step further, about three-fourths of challenged federal laws
and policies reviewed by the Supreme Court are consistent with American
public opinion. The federal policy model predicts that the Supreme Court
will be very consistent with American public opinion because it simply
upholds most challenged federal laws and policies.
From the mid-1930s through the mid-1980s, this model accurately
described the Supreme Courts behavior in challenges to federal laws
and policies. Was this linkage model as successful for the Rehnquist
Court, which less often deferred to federal laws and policies than did
earlier Courts? Chapters two and three examine this model at the deci-
sion level.
The State/Local Policy Model

Historically, the Supreme Court was not as deferential toward chal-
lenged state and local laws and policies as it was toward those of the
federal government. Further, challenges to state/local laws and policies
constitute a very large part of the Courts annual docketin recent

terms, well over half of its full, written decisions. The state/local policy
model makes three assumptions. First, challenged state/local laws or
policies do not reflect nationwide public opinion very closely. Second,
the Supreme Court shows little deference toward state/local laws and
policies. Third, when nationwide polls disagree with state/local laws and
policies, the Court will typically support nationwide public opinion. As
a result, the Court strikes down many state/local laws and policies that
disagree with nationwide public opinion. However, given that the
Rehnquist Court was often more deferential toward states and localities
than were earlier Courts, this linkage model merits reexamination.
Chapters two and three examine this model at the decision level.
The Business-as-Normal Model

The modern Supreme Court clearly has routine customs and norms
that influence its decision-making. For example, the Rehnquist Court
heard only a small percentage of all appeals and it handed down a
dwindling number of full, written opinions each term. Typically, the
Court reverses most of the lower court rulings it hears. Further, most
justices are fairly predictable in their liberalconservative ideologies,
and do not greatly change that position over time. Most justices bring
their favorite judicial theories with them to the Court. Do these rou-
tine norms and practices affect the level of representation the Court
provides? Chapter three examines the impact of several important
norms and customs on representation.
The Interest Groups Model

Interest groups play an important role in initiating, sponsoring, publiciz-
ing, and financing lawsuits that reach the Supreme Court. Yet there is no
evidence that interest groups litigation strategies closely reflect nationwide
public opinion. The interest groups model makes two assumptions. First,
there is no significant relationship between American public opinion and
the positions interest groups take in their lawsuits. Second, public opinion
influences Supreme Court decisions (at least at the full opinion stage), but
interest groups have no effect on decisions independent of public opinion.
Chapter four examines this null model at the decision level.
The Political Parties and Ideology Model

Presidents typically pick justices from their own political party and,
with varying success, who share their ideological views. Liberal
Democratic presidents typically pick liberal Democratic justices; con-

servative Republican presidents typically pick conservative Republican
justices. Chapter five tests several measures of the justices political
party and ideology to see if these are linked to representing American
public opinion.
The Political Socialization Model

Supreme Court justices bring a wide variety of background experiences
with them to the Court. In the political socialization model, personal
and career backgrounds that expose a justice to a broad, diverse, or
majority-oriented culture, and political and career experiences that
make a justice more sensitive to the nuances of public opinion predis-
pose a justice, once confirmed, to vote more often with majority public
opinion. Chapter five tests a large number of personal, familial, educa-
tional, political, and career experiences at the individual level.
The Appointment Process Model

Most Supreme Court justices are easily confirmed, but a few nominees
face very divisive and controversial hearings, and receive numerous
negative votes in the U.S. Senate. Typically, ideologically moderate
nominees with outstanding credentials most easily win Senate confir-
mation. This model predicts that the appointments process provides
valuable clues to a justices later behavior. Nominees with broad, bipar-
tisan support will later be the most consistent with majority public
opinion. Chapter five tests this model at the individual justice level.
The Judicial Roles Model

Upon joining the Court, a justice may take on many roles. The best-
known role is the chief justices position, which carries greater manage-
ment and lobbying duties. Some justices play a role as intellectual
leaders. Others are particularly active in oral argument. Others are task
leaders. Historically, a few justices even sought a political career
beyond the Court. Some justices take on no particular leadership role at
all. Chief justice status, intellectual or argument or task leadership, and
ambitions beyond the Court are all predicted to sensitize a justice to
public opinion, and to lead to a higher level of agreement with public
opinion. Chapter five tests this model at the individual justice level.
The Length of Tenure Model

Supreme Court justices now average about a quarter of a century on
the Court from their initial appointment (usually during their forties or
fifties) until their retirement (usually in old age and poor health) or
death. Over their entire Court tenureand certainly by the time they
retirethey represent an increasingly old and less numerous cohort,
and perhaps move further and further away from the experiences and
ideas of most Americans. This model predicts that the longer a justice
sits on the Court, the less often his or her votes will represent majority
public opinion. Stated otherwise, recently appointed justices should
most often vote consistently with public opinion, and then become
steadily less consistent as their Court tenure lengthens. Very senior jus-
tices should least often vote consistently with public opinion. Chapter
five tests this model at the individual justice level.
The Realignment Model

Periodic crises, new issues, the coming of new generations and the
passing of older generations, or unusually skillful (or inept) politicians
can cause a political realignment in American political life. Political
realignments cause millions of voters to change their political party
loyalties and voting patterns. Within a few years this changes the out-
come of state, congressional, and presidential electionsand, eventu-
ally, appointments to the Supreme Court itself. In the realignment
model, public opinion reacts much more quickly to realignment events
than do sitting Supreme Court justices or decisions. During realign-
ment periods, sitting justices will be less consistent with majority
public opinion than will be newly appointed justices. Unfortunately,
realignments occur very infrequently in American life, and it is diffi-
cult to test this argument adequately. Chapter five tests the realign-
ment model at the individual justice level both for the Reagan era of
the 1980s and the early 1990s.
The Symbolic Representation Model

In this model the justices do not simply represent American public
opinion as a whole; some justices best represent specific groups. The
symbolic representation model argues that presidents often pick justices
who represent an important group. Once confirmed, a justice who
symbolically represents a group will more often vote in agreement with
that groups policy attitudes than do the remaining justices. Over half
of all Supreme Court justices now symbolically represent some group,
such as women, blacks, Catholics, or Jews. Chapter six examines
whether symbolic nominees, once confirmed, also represent their
groups policy views through their votes.
The Short-Term Manipulation Model

The short-term and long-term manipulation models are quite different
from the other linkage models. Both assume that Supreme Court deci-
sions themselves favorably influence public opinion. Once a decision is
announced, public opinion should move at least modestly toward the
Courts position. In the short-term model, public opinion changes
should follow Court decisions very quickly, perhaps within a few days
or weeks. The level of media publicity that a decision receives, the type
of decision handed down, and the Courts degree of unanimity affect
the publics response. Chapter seven tests this model at the decision
level.
The Long-Term Manipulation Model

This model also assumes that Supreme Court decisions change atti-
tudes, but assumes that public opinion responds to decisions very grad-
uallyover many months and years, not over a few days or weeks.
Decisions become accepted by being incorporated into daily American
life, and are then slowly and gradually accepted over time. Over-time
polls should show a slow and growing acceptance of Supreme Court
decisions. Chapter seven tests this model at the decision level.
The Test-of-Time Model

In the final model, not all Supreme Court decisions endure over time.
The Court itself reverses a few decisions, sometimes only a few terms
after the decision was announced. Congress can overturn a statutory
construction decisions by rewriting the legislation in question. A presi-
dential or gubernatorial pardon or a new regulatory rule can mean that
a decision no longer has any real-world effect. A few constitutional
amendments overturn a Supreme Court decision. The test-of-time
model predicts that unpopular decisions will more frequently and more
quickly be overturned than popular decisions. Stated otherwise, public
opinion has a second chance at Supreme Court decisions. Chapter
eight tests this model at the decision level. Chapter nine then reconsid-
ers all fifteen models and offers an empirical model between American
public opinion and the Rehnquist Court.
CONCLUSION
This chapter began with a long-standing debate in democratic theory:

should Supreme Court decisions reflect American public opinion? That
debate has now gone on for over two centuries, and very likely will
continue. The question here is whether most Supreme Court decisions
actually do reflect public opinion and, if so, why? From the 1930s to
the mid-1980s, most Supreme Court decisions reflected majority public
opinion. Was that also true for the Rehnquist Court?
During the Rehnquist Court era, pollsters wrote over two thousand
poll questions tapping American attitudes toward the Court, its nomi-
nees and justices, and its decisions. The Rehnquist Courts justices often
considered the role that public opinion should play in judicial decision-
making. On the average, about six Rehnquist Court decisions a year
directly referred to public opiniona higher figure than for earlier
Courts. The Rehnquist Courts direct mentions of public opinion
ranged widely in their underlying theory, with a growing focus on
evolving and contemporary public opinion. Direct mentions of public
opinion were most common in cases involving criminal rights, court-
rooms, and juries. Over three-quarters of the Rehnquist Courts direct
mentions of public opinion reflected a generally positive view of public
opinion.
This chapter briefly outlined fifteen models to help explain which
Supreme Court decisions agree with American public opinion, and
which do not. The remaining chapters test these fifteen models.
Chapter nine then offers an empirical linkage model of public opinion
and judicial policy-making.
CHAPTER TWO

Rehnquist Court
T he U.S. Supreme Court might seem unlikely to represent American

public opinion, since the Court is quite insulated from public opin-
ion in its day-to-day decision-making, in the justices backgrounds and
lives, and even in the Courts historical background. Although a few
nominees face lively confirmation hearings in the U.S. Senate, most jus-
tices easily win confirmation. Once confirmed, the justices never face
reelection campaigns, and in recent years none have been seriously
threatened with impeachment for making unpopular decisions.1 The
justices discuss and vote on cases in closed conferences, and do not
allow themselves to be lobbied individually on pending cases. The
Courts sessions are not, as yet, televised, although oral arguments are
sometimes made available shortly after the hearings. Aside from their
written opinions, the justices make little effort to justify, explain, or
spin their decisions to win over public opinion. True, the justices fre-
quently speak to groups of attorneys, judges, law students, or other
audiences, and several have written extensively on legal issues and legal
history.2 However, the justices rarely grant interviews or appear at con-
gressional hearings to explain or justify their decisions (Mauro 2004).
Very few of the Courts constitutional law decisions are overturned
through a constitutional amendment, although Congress more often
overturns or revises statutory law decisions.3
Taken as a group, the justices are far from a representative cross
section of American life. The justices are much more highly educated,
older, and more affluent than an average American.4 Most justices hail
from affluent family backgrounds (Epstein, Segal, Spaeth, and Walker
2003). Very few recent justices ran for elective office prior to coming to
the Supreme Court,5 although several previously lobbied, worked in
political campaigns, or held high-level appointive public offices. Most
justices now come to the Court after serving as a federal appeals judge.
No recent justice demonstrated any ambition to leave the Court and
23
run for public office. Typically, the justices now remain on the Court
until they die or retire at an advanced age, often in ill health.6 Because
they usually remain so long on the Court, few justices were actually
appointed by the current president.
The justices might also ignore contemporary public opinion for
another reason. Several well-known theories of how the Court should
make its decisions bear no obvious relationship to contemporary public
opinion, or else encourage the justices to disregard public opinion. For
example, the theory of original intent encourages the justices to look
back at the practices of much earlier times. Stare decisis encourages jus-
tices to look back at the Supreme Courts own earlier decisions. The
preferred position doctrine encourages justices to value minority
rights over majority preferences. Historically, many American politi-
cians saw the Supreme Court as a barrier against majority public opin-
ion. As Alexander Hamilton wrote in Federalist 78, the Supreme Court
would be an excellent barrier against the legislature and public opin-
ion, and an essential safeguard against the effects of occasional ill
humors in the society. Very few decisions actually cite polls or use
agreement with the polls as a justification for decision-making.
Some well-known Supreme Court decisions are clearly at odds with
American public opinion, as measured in nationwide polls. In these
instances the Court may hand down an unpopular decision, and then
continue to uphold the precedent for years. Two well-known examples
are the school prayer and the flag-burning decisions. To date, public
opinion poll majorities still oppose the Engle v. Vitale (1962) and
Abington School District v. Schempp (1963) school prayer and bible
reading decisions.7 In the flag-burning cases, Texas v. Johnson (1989)
and U.S. v. Eichman (1990), the Rehnquist Court struck down both a
state law and a federal law outlawing flag burning as a political protest
despite vocal, overwhelming, and consistent public opinion opposition.8
All this notwithstanding, it may seem surprising that the Supreme
Court is typically responsive to public opinion. Several studies,
reviewed below, indicate that the Supreme Courts decision-making
usually follows liberal-versus-conservative poll trends, and agrees with
specific public opinion poll majorities. Indeed, some justices, most
notably Chief Justice Rehnquist (1986), argue that justices often sense
and share prevailing public opinion.
STUDYING REPRESENTATION
All these arguments suggest that there is little reason to expect that the
justices will consider public opinion when they decide cases. However,
Public Opinion and the Rehnquist Court 25
regardless of whether the justices consciously consider public opinion,

there might be actually a match between decisions and polls. Whether
the Courts decision-making actually represents American public opin-
ion can best be examined as an empirical question by comparing deci-
sions with available polls. Court-watchers use two approaches to
compare decisions with public opinionthe over-time trend method
and the pairwise method (Friedman 2003). Perhaps surprisingly, results
from both methods suggest that the Supreme Court is typically respon-
sive to public opinion trends and poll majorities.
The Trend Method

Since the 1970s, political scientists have compared Supreme Court deci-
sions with trends in American public opinion. Devine (1970) first com-
pared poll trends (based on presidential election year polls in the 1950s
and 1960s) with policy-making, and reported that policy-makers were
more responsive to the attentive public than to the general public.
Weissberg (1976) compared several instances of national policy-making
to national poll trends. Both Devine and Weissberg reported that
Supreme Court policy-making typically, although not invariably, fol-
lowed poll trends.
Page and Shapiro (1983) compared several hundred instances of
public policy-making with public opinion polls in cases in which an
identically worded poll item had been repeated at least twice. Typically,
policy-making became more consistent with public opinion over time.
Policy-making was especially consistent with poll trends if most
Americans expressed an opinion, or if large poll changes occurred. The
federal courts reflected public opinion about as often as did Congress
or the executive branch. Barnum (1985) compared poll trends to four-
teen areas of Supreme Court decision-making, and reported that major
decisions usually reflected poll majorities, or at least poll trends.9
These early studies typically compared Supreme Court decisions
with poll trends in specific policy areasfor example, school prayer or
busing cases. Since the 1990s, several studies used Stimsons (1991,
1999) innovative measure of the public mood,10 a general trend mea-
sure or index of changes in the American publics liberal-versus-conser-
vative leanings since the 1950s. The public mood is based on a large
number (145) of specific poll items. Since each poll item was reported
two or more times, the public mood measure is based on 2,056 poll
item administrations (Stimson 1999: 38). Over-time changes in the
public mood are then compared in a times series analysis either with
the overall trend in the Supreme Courts liberal-conservative rulings, or
with the Courts ideological decision-making in specific areas, such as
civil liberties, economic, or criminal procedure cases (Flemming and

Wood 1997; Link 1995; Stimson, MacKuen, and Erikson 1995). Some
studies (Flemming and Wood 1997; Mischler and Sheehan 1996) exam-
ined how closely individual justices reflect the public mood.
Trend studies (Flemming and Wood 1997; Mischler and Sheehan
1993, 1994, 1996; Norpoth and Segal 1994) agree that Supreme Court
decisions typically respond indirectly to the public mood through presi-
dential elections and the appointment of new justices. On the average, a
president now appoints a new justice every three or four years, and
newly appointed justices often bring the Court back into line with cur-
rent public opinion.11 Occasionally, as during the 1980s, the rapid
replacement of several older justices (at that time, more liberal justices
replaced by more conservative Reagan and Bush appointees) can lead
the Court to become ideologically one-sided and ignore public mood
trends for a decade or longer (Mischler and Sheehan 1993).
Some studies (Flemming and Wood 1997; Mishler and Sheehan
1993, 1994, 1996) go further, arguing that public mood trends also
have a direct impact on sitting justices, especially ideologically moder-
ate justices, such as Justices White or Stewart.12 Other scholars dispute
this view and argue that public opinion has only a very modest direct
effect on sitting justices (Stimson, MacKuen, and Erikson 1995), or else
none at all (Norpoth and Segal 1994). The Supreme Court appears to
react much more slowly to public mood changes than do the president,
the U.S. House, and the Senate.13 Supreme Court decisions that reverse
a lower court decision are much more responsive to the public mood
than decisions that simply affirm a lower court decision (McGuire and
Stimson 2004).
The Pairwise Method

The second approachand the approach used in this bookis to com-
pare (or match) specific Supreme Court decisions to poll questions.
The pairwise method of comparing a specific poll item with a specific
policy decision is not new. The earliest examples of such studies date
back to the 1970s.
Alan Monroes two studies (1979, 1998) pioneered this method
and included poll-to-policy matches across many different policy areas
and national institutions.14 Based on comparing several hundred spe-
cific poll questions and specific policy decisions, Monroe reported that
poll-to-policy consistency declined during the 19801993 time period,
compared to the 19601979 period. Energy, environment, and foreign
policy decisions most often reflected public opinion. Highly salient
issues (as measured by the Gallup Polls most important problem

question) much more often agreed with public opinion than did less
salient issues. Civil rights and liberties policies (those most likely to be
decided in federal courts) were consistent with public opinion about as
often as the average for all poll-to-policy matches (Monroe 1998).
A previous study of Supreme Court decision-making (Marshall
1989) matched 146 U.S. Supreme Court decisions with available polls
from the mid-1930s through the last (1985/1986) Burger Court term.
The 146 matches captured many high-profile decisions as well as many
less visible ones; a wide range of rulings by issue; and a diverse group
of full written opinions, per curiam rulings, and cert denials.
Overall, three-fifths to two-thirds of Supreme Court decisions
agreed with public opinion poll majorities or pluralities. This rate of
agreement was fairly consistent over time, and was roughly similar to
the rate for decisions made by elected policy-makers (Monroe 1979,
1998). Supreme Court rulings were most often consistent during crisis
times, when public opinion was closely focused on an issue, or when
the Court upheld federal laws or policies (most of which were them-
selves consistent with the polls). Chief justices, justices from prestigious
law schools, and justices who were presidential confidants were more
consistent with the polls than were other justices. Rulings that were
consistent with the polls were less often overturned than were inconsis-
tent rulings.
Matching specific public opinion polls with Supreme Court deci-
sions is possible for two reasons. First, as Alexis de Tocqueville
remarked a century and a half ago, many political disputes in America
eventually go to the courts, reach the U.S. Supreme Court, and result in
a written opinion, a per curiam opinion, a denial of certiorari, a denial
of an application for a stay, a denial of habeas corpus, or another pos-
sible decision. In other words, on many top controversies in American
life the Supreme Court makes some sort of a decision.
Second, pollsters often write questions to tap public attitudes on
these issues. Since modern polling originated during the mid-1930s, a
steadily increasing number of poll questions can be matched with a
Supreme Court ruling. During the 1930s and 1940s, pollsters wrote
an average of only two or three poll-to-decision matches a year.
During the Burger and Rehnquist Courts, that number rose to about
five or six matches a year, on the average, reflecting both the growing
number of polls in recent decades and pollsters greater interest in
legal controversies.15
Compared to the trend method, the pairwise method has both advan-
tages and disadvantages. Among the advantages is greater specificity. The
pairwise method compares specific Supreme Court decisions and spe-

cific public opinion polls, rather than matching a general index of the
publics ideology (the public mood) with the overall liberal-conservative
trend of Supreme Court rulings. While the public mood is based on a
large number of repeat poll questions, many of these poll questions are
little related to issues that actually reach the Supreme Court (Best
1999). Further, few of the lawsuits that reach the Supreme Court are
among the poll questions that comprise the public mood index
(Stimson 1999: 143149).
The pairwise method avoids discarding useful polling information.
Pollsters often write only a single poll question on a Supreme Court
cases. In other instances, pollsters write several poll questions on a
Supreme Court decision, but the wording varies from question to ques-
tion. Too seldom are identically worded poll questions available both
before and after a Supreme Court decision. If a single poll question
closely matched the issue in a decision, the pairwise method could use
the poll-to-decision match, but the trend method could not.
The pairwise method also allows modeling how changes in the
Supreme Courts behavior or its membership would affect how well it
would represent American public opinion. In other words, how well
would the Court represent American public opinion if the Court
behaved differently than it actually does? For example, would the
Court better represent public opinion if it always upheld the decisions
of state and local governments? What if the Court refused to hear any
appeals at all, and always voted to deny certiorarithereby effectively
upholding a lower courts ruling? What if the Court had one more con-
servative vote than it actually had? Alternatively, what if the Court had
one more liberal vote than it actually had? Chapter three considers sev-
eral such what if questions.16
While all these arguments can be given in favor of the pairwise
method, this method also has several practical limits. Pollsters do not
always write poll items to match Supreme Court cases, and they some-
times ignore even prominent cases. As a result, the sample of Supreme
Court decisions that can be matched with available poll questions can
be reweighted, but it is clearly not a random sample of all decisions.
Poll matches are too seldom available for some types of decisions, such
as the Courts economic cases, and especially for cert denials. A poll
questions wording or its timing (e.g., whether the poll was taken
before, versus after, the Supreme Courts ruling) may affect poll results.
Because the pairwise method is usually based on only a single poll item,
per match, it is difficult to use some useful statistics, such as time series
analysis (Mischler and Sheehan 1993:88; Stimson, MacKuen, and
Erikson 1995). Most obviously, no poll questions are available before

the mid-1930s.
The trend method and the pairwise method often, but not always,
report similar results. Both methods report that politically moderate
justices are especially responsive to public opinion (Marshall 1989:
110; Mischler and Sheehan 1996). Yet while trend studies suggest that
the Supreme Court was unresponsiveness to public opinion after 1981
(Mischler and Sheehan 1993), the pairwise method, as described later
in this chapter, shows only a short-term decline. This book uses the
pairwise method to report results based on 111 poll-to-decision
matches from the Rehnquist Courts first (1986/87) term through its
last (2004/2005) term.
Data Sources
Unfortunately, no published index of scientific, nationwide polls that
match Supreme Court decisions exists. The most useful source of polling
information is the Roper Archive, now available online, by searching
under the key word Supreme Court. Of all the questions using this
phrase during the Rehnquist Court, many ask respondents their views
on a specific controversy currently being considered or else previously
decided by the Court.17 However, many poll questions raising issues in
cases before the Court do not actually use the phrase Supreme Court.
These poll questions were identified by searching for key words from
Court decisions in the online Roper Archive, or by checking published
polls, such as the Gallup Poll, The Polls section of Public Opinion
Quarterly, and various newspaper or magazine polls.
For each poll-to-decision match considered here, the poll responses
were used to classify the Supreme Court decision as consistent,
inconsistent, or as unclear with public opinion, and to compute
the poll margin, described further below.18 Most poll questions
report only three responses (e.g., favor, oppose, and no opinion), as is
typical in modern polling. A few poll questions that report more
responses (e.g., strongly favor, somewhat favor, undecided, somewhat
oppose, strongly oppose) were collapsed to indicate the percentage of
respondents who favored, versus those who opposed the Court deci-
sion, or who reported no response. Very few poll questions used filter
questions. The percentage of respondents who responded no opinion,
or who gave no response, was usually small, averaging about 5%
across these poll-to-decision matches.
The most frequent source of poll matches during the Rehnquist
Court era was the Gallup Poll (54 matches); followed by Princeton
Survey Research Associates (16 matches); CBS/New York Times (7

matches); the Harris Poll and Washington Post/ABC (6 matches each);
Los Angeles Times (5 matches); NBC/Wall Street Journal (4 matches);
Roper (2 matches); and various other polls (11 matches). Because most
polling organizations are tied to newspapers, news magazines, or net-
works, these poll questions represent a diverse, but nonrandom sample
of Rehnquist Court decisions. Topics of great national interest more
often elicit poll questionsparticularly abortion, crime, and social
issues. Pollsters less often write questions if they see little public interest
in the topic, such as business or labor regulations, taxation, bank-
ruptcy, or securities disputes. Pollsters seldom write poll questions on
disputes resulting in a denial of certiorari. Because these poll-to-deci-
sion matches are not a random sample of Rehnquist Court decisions,
the inferential statistics reported below should be interpreted cau-
tiously. This sample is later reweighted to approximate the actual mix
of Rehnquist Court decisions.
Just as there is no listing of polls related to Supreme Court deci-
sions, so, too, there is no previously published listing of Court decisions
for which a poll question is available. The Supreme Courts majority,
concurring, or dissenting opinions rarely cite poll questions even when
a poll question is available before the decision is announced. Decisions
that might be matched with a poll question were identified from a vari-
ety of sources, among the most helpful of which were annual reviews of
key decisions in the New York Times, Congressional Quarterly
Almanac, or Harvard Law Review; a Lexis computer search on key
words in poll questions; various news accounts, law journal articles,
and casebooks; and U.S. Reports, Supreme Court Reporter, or for
recent decisions, slip opinions.19
Coding Rules
These 111 poll-to-decision matches were coded to follow the same
rules from an earlier study (Marshall 1989: 7577). Briefly, the coding
rules are:
1. A poll could be matched with a Supreme Court decision

if the poll wording closely matched the substance of an
issue raised in a decision. Only nationwide samples of
American adults were used. A consistent Court deci-
sion agreed with a poll majority (or plurality). For
example, shortly after the Courts decision in Dickerson
v. U.S. (2000), a nationwide survey asked respondents:
(As I read you a list of some recent decisions by the

Supreme Court, please tell me if you generally agree or
disagree with each one.) What about . . . the recent deci-
sion upholding Miranda Rules requiring police to
inform arrested suspects of their rights to remain silent
and to have a lawyer present during any questioning?
Do you generally agree or disagree with this deci-
sion?20 An 86-to-11% poll majority agreed with the
ruling, with 3% reporting no opinion, thereby yielding a
+75% poll margin (i.e., 86% 11% = +75%).
2. A decision that disagreed in substance with a poll
majority (or plurality) was classified as inconsistent.
For example, shortly before the Clinton v. City of New
York (1998) decision, striking down the line item veto
as an unconstitutional violation of the Presentment
Clause, Gallup Poll asked respondents: As you may
know, Congress recently approved legislation called the
line item veto, which for the first time allows the
President to veto some items in a spending bill without
vetoing the entire bill. Do you generally favor or oppose
the line item veto?21 A 65-to-24% majority favored the
line item veto, and Clinton v. City of New York was
coded as inconsistent.
3. If available poll questions either reported closely divided
results or conflicting results, the decision was classified
as unclear since no clear poll majority or plurality
existed. In defining a clear poll margin, the .05 statisti-
cal significance level was usedusually a poll margin of
three to five percentage points for the polls used here.
For example, the Mitchell v. Helms (2000) plurality
opinion upheld federal grants for materials and com-
puter equipment to private and parochial schools.
Shortly after the decision was announced, a survey
asked respondents whether they agreed or disagreed
with the recent decision that taxpayer money may be
used to buy computers and textbooks for religious and
other private schools. Nearly as many respondents
agreed (47%) as disagreed (49%),22 and the Mitchell v.
Helms (2000) ruling was coded as unclear.23
4. If an earlier decision was coded, later decisions follow-
ing, citing, or applying that precedent were not coded
as a separate match, even if later poll questions were
available.24 However, a later decision that overturned

an earlier precedent could be separately counted as a
match. For example, the Courts 63 decision in Payne
v. Tennessee (1991), allowing victim impact statements
at the sentencing stage, overturned the earlier decisions
in South Carolina v. Gathers (1989) and Booth v.
Maryland (1987), both of which were also combined
and counted as a separate match. Similarly, Atkins v.
Virginia (2002), holding unconstitutional the death
penalty for mentally retarded convicted killers, was
coded as a separate match from Penry v. Lynaugh
(1989), an earlier ruling overturned by Atkins.
5. If a later decision followed an earlier precedent, but a
poll was available only for the later decision, then the
later ruling was counted even if the precedent was better
known. If the Rehnquist Court declined to rule on a
controversy, usually by denying certiorari, but later
accepted another case on that same issue and issued an
opinion, then the latter case was coded, but not the
former. Thus, Raines v. Byrd (1997), vacated and
remanded on grounds that individual members of
Congress had no standing to bring suit against the line
item veto, was not coded, but Clinton v. City of New
York (1998) was.
6. If the Court upheld and applied a specific application of
a more general law or policy, it was counted as uphold-
ing the law or policy itself. For example, in Muscarello
v. U.S. (1998) the Court interpreted and upheld a
statute providing a mandatory prison sentence for carry-
ing or using a firearm during a drug trafficking crime.25
Similarly, a Court decision was considered to uphold a
practice by permitting it under frequently arising,
although not all possible circumstances. For example,
Kansas v. Hendricks (1997) upheld Kansas civil incar-
ceration statute for violent sexual offenders, although
the 54 majority opinion at least implied that some
statutes or practices might be voided in the future.26
7. A single Court ruling that involved two (or more) sub-
stantive issues could be divided and treated as two (or
more) separate decisions if it could thereby be better
compared with poll questions. For example, Planned
Parenthood of Southeastern Pennsylvania v. Casey
(1992) was coded as four separate decisionshusband

notification, informed consent, the twenty-four-hour
waiting rule, and one-parent consent. McConnell v.
Federal Election Commission (2003) produced two dif-
ferent matches, one on soft money and the other on the
requirement that candidates identify themselves during
and approve their own television ads.
8. Similarly, two (or more) closely related decisions could
be combined and treated as a single decision. For exam-
ple, three decisions announced on the same day
Graham v. Collins (1993), Herrera v. Collins (1993),
and Lockhart v. Fretwell (1993)were all combined
and counted as a single match on the issue of limiting
death penalty appeals.
9. Companion cases (typically handed down on the same
day) were combined as a single match and are entered in
appendices one and two under the lead case.
10. If a poll question showed that a public opinion majority
(or plurality) favored a policy through a more general
means (such as a constitutional amendment), then
public opinion was also assumed to favor that same
policy through a less general means (such as a federal,
state, or local law).
11. As in a prior study (Marshall 1989), decisions were
counted only if the poll item was within five years of the
decision. If multiple poll questions were identified, the
poll question closest in time to the ruling was chosen as
the matching poll question, whether it occurred before
or after the Court decision. The median time lag
between a poll-to-decision match was 151 days. In most
(85%) instances the poll date was within two years of
the decision. Poll questions preceded the Courts deci-
sion about as often as they followed the decision (53%
versus 47% of the matches, respectively).
Coding Denials of Certiorari

Whether denials of certiorari should be included at all is open to dis-
pute, since they present both theoretical and practical problems. The
most serious theoretical issue is whether a denial of certiorari has any
substantive meaning. As several justices have written in cases such as
Hopfmann v. Connolly (1985) or Equality Foundation of Greater
Cincinnati v. Cincinnati (1998), a denial of certiorari neither has prece-

dent value nor is a ruling on the merits. The most commonly given
explanation for a cert denial is that it simply indicates that fewer than
four justices voted to hear the case,27 due to lack of ripeness, mootness,
narrow technical issues raised, lack of a final review by a lower court,
lack of standing, failure to meet the Courts deadlines, or an inadequate
record or petition (Baum 1998a:105127).
At the practical level, the substantive issues in a case in which cer-
tiorari is denied are often difficult to identify, since they are rarely dis-
cussed in the media or in law journals. Pollsters seldom write poll
questions for cases where certiorari is denied. Poll matches could be
identified in about 5% of the Rehnquist Courts full, written opinions,
but in less than 1% of cert denial decisions. Justices rarely write to
explain their reasons for denying certiorari (or a denial of habeas
corpus or a denial of an application for a stay); even when they do,
their explanations typically fall short of explaining how they would
have decided the appeal.28 As a result, it is seldom possible to code an
individual justices vote in cert denials as either consistent or inconsis-
tent with public opinion.
All these problems notwithstanding, there are four reasons to
include denials of certiorari in this study. First, at a minimum, cert
denials have the practical effect of letting a lower court ruling stand,
albeit without precedent value. Since the dispute may not soon (or ever)
again reach the Supreme Court, a cert denial may be the Courts only
decision in a dispute. This is similar to the practical effect of a denial of
habeas corpus in a death penalty appeal.29 Second, the Rehnquist Court
sometimes relied heavily on cert denials, even in high profile cases or
where appeals were repeatedly brought from the lower courts.
Thomasson v. Perry (1996) on dont ask, dont tell policies toward
gays in the military or the series of cert denials on involuntary AIDs
testing are well-known examples included here. Third, justices occa-
sionally use dissents from denials of certiorari to clarify their positions
on issues. Justice Brennan and Marshalls frequent dissents from denials
of certiorari in death penalty decisions are the best-known,30 but by no
means only example.31 Fourth, including denials of certiorari allows a
comparison of whether the Rehnquist Court was more consistent with
public opinion in cert denial decisions than in full, written opinions.
The twenty-three denials of certiorari (or denials of habeas corpus
or denials of an application for stay) identified here were classified by
their effectas if the Rehnquist Court had upheld the lower court deci-
sion. Because of the concerns discussed just above, they are reported
separately from full, written decisions. Unless a justice wrote a dissent
from a denial of certiorari sufficiently clear to explain how he or she

would actually have voted, cert denials are never counted in the votes
of individual justices in chapters five and six.
DATA RESULTS
Table 2.1 breaks down the 111 poll-to-decision matches by the type of
issue presented. Although not a random sample of the Rehnquist
Courts decisions, they are a diverse sampling of topics. Pollsters most
often wrote questions on criminal rights, and privacy, morality, or
abortion decisions. The mix of poll matches also greatly changed from
earlier Courts, in large part reflecting changes in the Courts docket
(Pacelle 1991). Labor and workplace disputes or national security deci-
sions were often matched with poll questions during earlier Courts, but
now rarely elicit a matching poll question (Marshall 1989: 73). By
comparison, more health and family law decisions cases can be
matched than in earlier periods.
Overall Results
About three-fifths of the Rehnquist Court decisions considered here
were consistent with American public opinion. Depending on whether
cert denial decisions are included or excluded, some 61 or 60%, respec-
tively, of Rehnquist Court decisions were consistent with the polls
when a clear poll majority or plurality existed. If the unclear cases
(where public opinion polls were closely divided or inconsistent) are
excluded, then 64% of Rehnquist Court decisions were consistent with
poll majorities or pluralities.
This level of agreement with the polls is strikingly similar to that of
earlier Courts since the 1930s. Earlier Courts were consistent with
American public opinion in about three-fifths of their decisions. So was
the Rehnquist Court. Indeed, excluding cases of unclear polls, the
records of the Warren, Burger, and Rehnquist Courts are virtually iden-
ticalranging only from 61 to 64% consistent (Marshall 1989:
7879). Apparently, the modern Supreme Court reached a very stable
plateau in representing American public opinion over the last half-cen-
tury. Roughly three of five decisions agree with the polls, while two of
five do not.
The Supreme Courts over-time stability in representation is also
apparent in yet another way. The Rehnquist Court was more often con-
sistent with public opinion in its cert denial decisions (77% consistent)
Table 2.1
Number of Poll Matches, by Issue, Including or Excluding Denials of Certiorari
Including Cert Excluding Cert
Denials Denials
Number Percentage Number Percentage
of Cases Matches of Cases Matches
Labor unions, strikes 3 3% 2 2%
Criminal rights, courts, police 29 26% 25 28%
Education, schools 16 14% 10 11%
National security, foreign policy,
communism, immigration 6 4% 3 3%
Privacy, morality, abortion 23 21% 18 20%
Civil rights, race, integration 8 7% 6 7%
Business regulation, taxes 7 6% 6 7%
Elections, campaign finance 9 8% 8 9%
Intergovernmental relations, federalism 10 9% 8 9%
Free speech, dissent, obscenity, media 7 6% 7 8%
Social welfare, poverty 1 1% 0 0%
Transportation, commerce 4 4% 4 5%
Religion 12 11% 9 10%
Health, hospitals, HMOs 16 14% 10 11%
Families, family law 10 9% 6 7%
Women and gender claims 4 4% 4 5%
Note: Numbers and percentages sum down, by column, in excess of 111 or 88
cases (or 100%) because several decisions are coded in two or more issue cate-
gories.
than in its full, written opinions (60% consistent). Earlier Courts from
the mid-1930s to the mid-1980s were also more often consistent in cert
denial cases (70% consistent) than in full, written opinions (62%)in
each comparison excluding unclear polls.
Table 2.2
Supreme Court Agreement with Public Opinion, by Court,
Including All Poll Matches
Hughes Stone Vinson Warren Burger Rehnquist
% consistent 65% 54% 68% 52% 53% 61%
% inconsistent 24% 46% 32% 33% 33% 35%
% unclear 12% 14% 15% 4%
Percent of cases 101% 100% 100% 99% 101% 100%
Number of cases (17) (13) (19) (21) (76) (111)
Table 2.3
Supreme Court Agreement with Public Opinion, by Court,
Excluding Instances of Unclear Polls
Hughes Stone Vinson Warren Burger Rehnquist
% consistent 73% 54% 68% 61% 62% 64%
% inconsistent 27% 46% 32% 39% 38% 36%
Percentage of cases 100% 100% 100% 100% 100% 100%
Number of cases (15) (13) (19) (18) (65) (107)
These results show considerable over-time stability in the Courts

pattern of representation. Yet they also show striking differences
between these results and those reported elsewhere. Monroe (1998)
reports that consistency between public opinion and policy-making
markedly declined during the 19801993 time period. Mischler and
Sheehan (1993) and Hurwitz, Mischler, and Sheehan (2004) suggest
that the Court was less responsive to the national mood during the
1980s and mid-1990s. Although these decades do not match up pre-
cisely to Supreme Court eras, Tables 2.2 and 2.3 and Figure 2.1 all
indicate more stability over time.
Throughout the Rehnquist Court, the Courts decisions were typi-
cally more consistent with public opinion than not. Figure 2.1 depicts
the five-year moving averages in terms of how often the Rehnquist
Courts decisions agreed with public opinion, along with figures for the
earlier Burger Court. The Rehnquist Courts five-year running averages
of consistency remain at or well above the 50% level, as was mostly
true for the Burger Court.
Figure 2.1
Percentage of Supreme Court decisions consistent with public
opinion, five-year moving averages, 19702005
These figures do not change greatly if this sample of 111 decisions

is reweighted to the docket reported in a leading data base of Supreme
Court decisions, the Spaeth data base. When this sample of decisions is
adjusted to the type of cases the Rehnquist Court actually decided,
some 62% of its decisions were consistent, and 38% were inconsistent
with American public opinion.32 Nor do these figures change greatly
when reweighted against the numbers of federal versus state/local dis-
puted laws from the Harvard Law Reviews annual reports on Supreme
Court decisions.33 This remainder of chapter two breaks down poll-to-
decision matches to examine the first three linkage models between
public opinion and Supreme Court decision-making and look for other
patterns and variations.
THE STATE OF PUBLIC OPINION MODEL
The state of public opinion model assumes that Supreme Court deci-
sions will be most consistent with the polls under three conditions: first,
when public opinion itself is very one-sided; second, when public atten-
tion is closely focused on the issue in the case, such as during crisis
times or on a high-profile controversy; or third, when nearly all
Americans hold opinions on a dispute. Under these circumstances the
justices can more easily sense and share prevailing public opinion
(Rehnquist 1986). From the mid-1930s to the mid-1980s, Supreme
Court decisions significantly more often agreed with the polls when
public attention was closely focused upon the decisions key issue
(Marshall 1989).
One-sided Poll Margins

Three versions of the state of public opinion model were tested. The
first predicts that the Court will be most consistent with the polls when
public opinion is very one-sided. When a clear (or landslide) public
opinion majority exists, the justices may easily sense a poll majority,
even if no polls are actually available at the time of the decision.
Table 2.4 tests this hypothesis by comparing the Courts agreement
with public opinion with the size of the poll margin. For example, in
Clinton v. City of New York (1998), only 24% of Americans agreed
with the Courts decision to strike down the line-item veto, while 65%
of Americans disagreed with the Courts ruling.34 In this case the poll
margin was 41% (or 24%65% = 41%). Poll margins exceeding
30% (whether positive or negative) were considered to be landslide
poll margins. Contrary to the hypothesis, the Rehnquist Court was

slightly less often consistent with public opinion in landslide deci-
sions.35 This suggests that the Rehnquist Court did not simply respond
to large, one-sided poll margins.
Crisis-Times Cases
In a second version, the state of public opinion model predicts that the
Court will more often represent public opinion when public attention is
closely focused on an issue. The crisis-times thesis dates back to the
1940s and 1950s, when the Court seldom supported controversial or
unpopular fundamental freedoms claims during World War II or the
subsequent anticommunist Red Scare period (Becker 1970: 229).36
Admittedly, few Rehnquist Court decisions rose to the high level of
public visibility for the crisis-times cases during the 1940s and 1950s.
However, several controversies were closely related to the top national
concern, as measured by Gallup Polls question, What do you think is
the most important problem facing this country today? Table 2.4
reports figures by whether the single, top national concern was
involved, or not.37
Just as for earlier Courts, the Rehnquist Courts decisions were
more often consistent with public opinion during crisis-times cases. The
Rehnquist Court was consistent in 74% of its crisis-times decisions,
versus only 61% of decisions when the issue was not the top national
concern. This pattern is also evident from the mid-1930s to the mid-
1980s, when earlier Court decisions were consistent with the polls for
76% of crisis-times cases, but for only 59% of other cases.
In an alternative test, the Court might be more consistent with
public opinion when the controversy is an especially important one,
and likely to garner attention from the public or the legal commu-
nity. Here, the decisions were split according to whether Court
scholars (Biskupic and Witt 1997, updated) consider the case unusu-
ally important or not. No strong findings resulted, with 67% of less
important, versus 60% of more important decisions being consistent
with the polls.
Issue Clarity
A third test breaks down these 111 rulings according to how many
respondents reported no opinion to the poll question. When very few
respondents (5% or fewer) indicated no opinion on an issue, the jus-
tices might sense that the issue is one on which public opinion is atten-
tive. The Rehnquist Courts rulings were more consistent with public
opinion when nearly all poll respondents give an opinion70% consis-
tent, versus only 49% consistent when over 5% of respondents answer
no opinion to a poll question a pattern similar to earlier Courts
since the mid-1930s.
Overall, there was some support for the state of public opinion
model. Most of the results in Table 2.4 were in the expected direction,
although only when nearly all Americans held opinions on an issue
were the Rehnquist Courts decisions significantly more often in agree-
ment with the polls. Only the test for one-sided or landslide poll
margins failed to muster any supportive evidence.
THE FEDERAL POLICY MODEL
The federal policy model makes three key assumptions. First, most fed-
eral laws and policies agree with American public opinion. Second, the
Supreme Court usually upholds a challenged federal law or policy. As a
result, the Court is very consistent with American public opinion in
these disputes. Third, extending this argument a bit further, challenged
federal laws and policies are a major part (roughly a third) of the
Courts docket. Thus, the federal policy model helps explain why the
Court so often reflects public opinion.
Earlier research (Marshall 1989: 8388) indicated that this linkage
model was very important. Most (72%) of the federal laws and policies
that the Supreme Court reviewed were consistent with nationwide polls
between the mid-1930s and the mid-1980s. Further, earlier Courts
upheld 81% of those federal-level laws or policies that were consistent
with public opinion and even 63% of those that were not.
The Rehnquist Court was less deferential to federal laws and poli-
cies than were earlier Courts since the New Deal, however. Several
well-known Rehnquist Court rulings gave the federal governments
decisions less deference than earlier Courts might have done, including
the FDA v. Brown Williamson Tobacco Corp (2000), Reno v. ACLU
(1997), Printz v. U.S. (1997), Rasul v. Bush (2004) and Padilla v.
Rumsfield (2004) rulings counted here.38 As a result, the federal-level
linkage model may need to be reevaluated.
Tables 2.4 and 2.6 report how the Rehnquist Court and earlier
Courts treated challenged federal laws and policies, broken down by
whether or not the challenged federal laws themselves agreed with
public opinion. Challenged federal laws and policies were as often con-
sistent with nationwide polls during the Rehnquist Court as during ear-
lier Courts73% were, versus 72% in earlier years (in both cases
excluding instances of unclear or evenly divided polls). However, the
Rehnquist Court less often deferred to consistent federal laws and poli-
ciessupporting the federal law (or policy) only 67% of the time when
the challenged law was consistent with the polls, versus 81% of the
time in earlier courts. When the federal law or policy was inconsistent
with nationwide polls, the Rehnquist Court upheld it 60% of the time,
versus 63% of the time for earlier courts. As a result, the Rehnquist
Court was only consistent with nationwide public opinion 59% of the
time when a federal law or policy was before itversus 69% of the
time for earlier courts since the mid-1930s.
In short, the federal policy model was not as good a linkage model
for the Rehnquist Court as it was for earlier Courts since the mid-
1930s. The Rehnquist Courts dwindling deference toward federal
policy-making should not be simply of interest to legal scholars and
attorneys. It also weakened the linkage between Supreme Court deci-
sion-making and American public opinion. Chapter three returns to
this issue.
THE STATE/LOCAL POLICY MODEL
The state/local policy model is quite different than the federal policy
model, and makes three key assumptions. First, the Supreme Court
upholds state/local laws or policies less often than it upholds federal
laws and policies. Second, state/local laws and policies themselves less
often agree with nationwide public opinion than do federal laws and
policies. Third, when a state/local law or policy disagrees with nation-
wide public opinion, the Supreme Court usually agrees with nationwide
public opinion. Extending this argument, disputes over state and local
laws and policies comprise the majority of the modern Courts docket.
From the mid-1930s through the mid-1980s, this model accu-
rately described Supreme Court decisions involving state or local laws
and policies. However, the Rehnquist Court gave more deference to
state/local policy-making than did earlier Courts since the 1930s.
Does this model still accurately describe the Courts relationship to
public opinion?
How does the Rehnquist Court compare to earlier Courts? First, the
Rehnquist Court usually upheld state/local laws and policies that agreed
with nationwide public opiniondoing so 71% of the time if the
state/local law or policy was consistent with nationwide polls, but uphold-
ing only 43% of inconsistent state laws and policies. By comparison, from
the mid-1930s to the mid-1980s, earlier Courts upheld state or local

laws and policies only 56% of the time when these laws and policies
agreed with nationwide polls, and 43% when they did not (Marshall
1989: 8485).
Second, the state/local laws and policies considered by the
Rehnquist Court were themselves much more often consistent with
nationwide public opinion than was true for earlier Courts. During the
Rehnquist Court, 70% of challenged state and local policies agreed
with nationwide poll majorities. For earlier Courts since the mid-1930s,
that figure was only 58%.
Third, when state/local laws and policies actually disagreed with
nationwide public opinion, the Rehnquist Court (like earlier Courts)
slightly more often decided in favor of national public opinion than for
the state/local laws and policies in dispute. When a state/local law or
policy inconsistent with nationwide public opinion was being consid-
ered, the Rehnquist Court decided in favor of nationwide public opinion
57% of the time, a figure identical to the 57% rate for prior Courts.
In short, the Rehnquist Court was more deferential than were ear-
lier Courts to state/local laws and policies that agreed with nationwide
polls. Overall, Rehnquist Court decisions were consistent with nation-
wide polls 67% of the time when a state or local law or policy was
involveda figure almost identical to the percentage (70%) of
state/local laws and policies themselves that agreed with nationwide
polls. By comparison, earlier Courts decisions were consistent with
national polls only 57% of the time in disputes from the state/local level.
ALTERNATIVE LINKAGES
Did the Rehnquist Court reflect public opinion in some other circum-
stances, rather than others? Table 2.4 reports several results, including
fundamental freedoms disputes, First Amendment claims, economic
claims, polling artifacts, the publics liberalconservative views, whether
the decision was a full written decision or not, and shorter Court eras.
Fundamental Freedoms Decisions

During the 1930s, the Supreme Court developed a rationale for ignor-
ing majority public opinion in favor of controversial or unpopular
groups. The argument that civil liberties and civil rights claims should
enjoy a preferred position under the Constitution was first set out in a
footnote in an otherwise obscure milk regulation case, United States v.
Carolene Products Corporation (1938), and later expanded upon by

numerous decisions.39 The argument is that the courts should take a criti-
cal look at laws or policies that infringe upon Bill of Rights or Fourteenth
Amendment protections for unpopular or controversial groups, such as
discrete and insular minorities. In these cases, the Court should gener-
ally ignore majority public opinion. However, the Court should ordinar-
ily uphold legislation involving economic regulations.
To be sure, not all Rehnquist Courts justices strongly supported
the preferred position doctrine, and, after the early 1990s, few justices
seemed deeply committed to this position. At its core, the fundamental
freedoms hypothesis predicts that the Court will less often agree with
majority public opinion when Bill of Rights or Fourteenth Amendment
claims are raised.40
Little support appears for the fundamental freedoms argument.
The Rehnquist Courts decisions nearly as often agreed with American
public opinion in fundamental freedoms cases as in other cases.41 These
results are similar to those for earlier Courts (Marshall 1989:8789).
Simply put, polling data offers little empirical support for the funda-
mental freedoms thesis.
First Amendment Claims

Is the fundamental freedoms argument too broadly stated? Is there a
subset of preferred position claims that better fits the original preferred
position argument? A more narrowly defined group of First
Amendment cases was examined, including freedom of religion, press,
speech, or assembly claims. Here the Court might least often reflect
majority public opinion. However, in deciding these First Amendment
claims, the Rehnquist Court agreed with nationwide public opinion
only a little less often than it did in other cases57% versus 66%,
respectively. Further, the Rehnquist Court was likely to disagree with
nationwide public opinion in a liberal (or pro-rights claim) position. In
First Amendment cases when public opinion was conservative (or anti-
claim), the Rehnquist Court agreed with the polls only 50% of the
time, but favored the First Amendment claim 50% of the time. When
nationwide public opinion was liberal (and favored the claim), the
Rehnquist Court agreed with the claim 77% of the time.
Economic Cases
Economic cases, such as workplace, taxes, lawsuits, business regula-
tions, transportation, or commerce disputes, have long been a major
Table 2.4
Percentage of Consistent Decisions, by Explanatory Variables,
for the Rehnquist Court
Percentage Consistent Number of Cases
Overall results 64% (107)
By size of poll margin
Landslide margin (over 30%) 62% (76)
Smaller poll margin (29% or less) 68% (31)
By state of public opinion
Crisis-times cases 74% (19)
Non-crisis times cases 61% (88)
By importance of decision
Most important cases 60% (52)
Less important cases 69% (55)
By level of issue clarity*
Undecided is 5% or less 70% (76)
Undecided is over 5% 49% (31)
By level and context
Consistent federal law 67% (27)
Inconsistent federal law 40% (10)
Consistent state/local law 71% (52)
Inconsistent state/local law 57% (22)
By type of issue
Fundamental freedoms case 62% (77)
Non-fundamental freedoms case 67% (30)
First Amendment claims 57% (28)
Non-First Amendment claims 66% (79)
Economic cases 53% (17)
Non-economic cases 66% (90)
By timing of the poll*
Pre-decision poll 71% (65)
Post-decision poll 52% (42)
By ideology of public opinion*
Conservative public opinion 70% (59)
Liberal public opinion 55% (44)
No clear ideology 75% (4)
By type of decision*
Full, written opinion 60% (85)
Cert denials, others 77% (22)
By Court period
Pre-Justice Thomas 59% (41)
Post-Justice Thomas 67% (66)
Note: The maximum number of decisions in this table is 107; instances of
evenly-divided or inconsistent polls are excluded from this table. To compute
the percent of inconsistent decisions, subtract the percent of consistent deci-
sions from 100%.
* = stat. signif. at .05
part of the Supreme Courts docket. In economic cases, the Courts deci-
sions were predicted to be more often consistent with public opinion
than in non-economic decisions, largely because of the modern Courts
greater deference to elected officials in economic disputes. However,
unlike earlier Courts, the Rehnquist Courts economic decisions less
often agreed with public opinion than non-economic decisions.
Polling Artifacts
Some accounts argue that the Supreme Court is highly regarded as an
institution and that its decisions at least modestly influence public
opinion.42 If this is true, then the Courts decisions should more often
agree with the polls if the poll was taken after, not before, the decision
was announced.
To test for this effect, these 111 decisions were divided into two
groups: those in which the matching poll was taken before, versus
after the decision was announced. No support appeared for this argu-
ment, as reported in Table 2.4. Nor was there any support in an ear-
lier study (Marshall 1989: 90). In fact, the Rehnquist Courts
decisions were more often consistent with public opinion when the
poll was taken before the decision was announced.43 As chapter seven
indicates, Supreme Court rulings do not have an endorsement
effect on public opinion.
Ideology
Many scholars (Friedelbaum 1994; Gottlieb 2000; Savage 1992;
Yarbrough 2000) describe the Rehnquist Court as a relatively conserv-
ative Court, at least compared to the Burger and Warren Courts.
Further, the justices have a great deal of freedom in deciding cases and
the justices votes usually reflect their own liberalconservative views
(Segal and Cover 1989; Segal and Spaeth 1993; Segal, Epstein,
Cameron, and Spaeth 1995). As a result, the justices ideological lean-
ings may be closely tied to how often the Court represents public opin-
ion. Here, the hypothesis is that the Rehnquist Courts decisions will
more often reflect public opinion when poll majorities themselves are
politically conservative, rather than when poll majorities are liberal.44
Considerable support exists for this hypothesis. When American
public opinion was politically conservative, the Rehnquist Courts
decisions agreed with the polls 70% of the time. When public opin-
ion was politically liberal, the Rehnquist Courts decisions agreed
with the polls only 55% of the time. Further, public opinion itself
was relatively conservative on cases that reached the Rehnquist
Court; conservative poll majorities considerably outnumbered lib-

eral ones, by nearly a 32 ratio.
Type of Decision
Sometimes the Supreme Court sets a case for oral argument and then
hands down a full, written opinion, but much more frequently, it
simply denies certiorari and declines to hear the case at all.45 When did
the Rehnquist Court more often agree with public opinion?
The Rehnquist Courts decisions were divided in two groups: its
full, written opinions, versus its denials of certiorari, denials of an appli-
cation for a stay, and denials of habeas corpus. Just as in an earlier
study (Marshall 1989: 7879), the Rehnquist Courts full, written deci-
sions were less often consistent with public opinion. Sixty percent of the
Courts full, written opinions agreed with public opinion, while 77% of
its cert denial (or denial of a stay, or denial of habeas corpus) decisions
did. Chapter three reconsiders this finding and the implications of the
Rehnquist Courts dwindling number of full, written opinions.
Court Eras and Judicial Appointments

New justices can greatly change the Courts voting patternsespecially
when a new justice joins an otherwise closely divided Court, replacing
an older justice who had different ideological leanings (Smith 1993).
The Rehnquist Court was divided into two time periods: the earlier
terms when the Court still included several older, more liberal justices
such as Justice Blackmun, Brennan, or Marshall, versus the later terms
after Justice Clarence Thomass 1991 confirmation solidified the
Rehnquist Courts conservative majority.46
Justice Thomass confirmation bolstered the Courts conservative
bloc of justices and often provided a critical fifth vote in closely divided
cases (Devins 2004; Gottlieb 2000; Yarbrough 2000). It also had a
modest impact on how often the Rehnquist Court reflected public opin-
ion. After Justice Thomass confirmation, the Courts agreement with
the polls increased slightlyrising from 59% (prior to Justice Thomas)
to 67% consistent (after Justice Thomass confirmation).
Issues and Decisions

Table 2.5 breaks down the Rehnquist Courts decisions by the type of
issue involved, when at least six cases could be matched with the polls.
The Rehnquist Court typically agreed with nationwide public opinion in
most of its more commonly decided casesbeing particularly consistent in

two common types of decisions, criminal rights decisions and privacy and
morality decisions. In only a few areas, such as free speech, dissent, media,
and obscenity cases, did the Rehnquist Court typically disagree with the
polls. The modern Courts record of agreeing with the polls is quite wide-
spread across the docket, and small changes in its docket would not
greatly affect the Courts level of agreement with public opinion.
Table 2.5
Percentage of Consistent Decisions, by Issue
Percentage Consistent Number of Cases
Type of issue
Criminal rights, courts, police 76% (29)
Education, schools 64% (14)
National security, foreign policy,
communism, immigration 50% (6)
Privacy, morality, abortion 68% (22)
Race, integration 63% (8)
Business regulation, taxes 57% (7)
Elections, campaign finance 75% (8)
Intergovernmental relations,
federalism 60% (10)
Free speech, dissent, media,
obscenity 29% (7)
Religion 50% (10)
Health, hospitals, HMOs 56% (16)
Families, children, family law 66% (9)
The Rehnquist and Earlier Courts

Overall, the Rehnquist Courts decisions agreed with American public
opinion about as often as did earlier Court decisions since the mid-
1930s. Yet the Rehnquist Court reached that level of representation
differently than did the Hughes, Stone, Vinson, Warren, and Burger
Courts. Table 2.6 compares the Rehnquist Court, versus earlier
Courts, in terms of how often their decisions agreed with public opin-
ion for the categories reported in Table 2.4 (except for the pre/post-
Justice Thomas measure).
For most types of decisions, the Rehnquist Courts record was
fairly similar to that of earlier Courts since the mid-1930s. A few
double-digit differences appear. The Rehnquist Court was less deferen-
tial to consistent federal laws, more deferential to consistent state laws,
more consistent with public opinion in nonfundamental freedoms cases,
less consistent in economic disputes, less consistent when American
public opinion itself was politically liberal, and more consistent in cert
denial decisions, compared to earlier Courts. Popular impressions of

the Rehnquist Courts legacy find some support here, although only the
finding on liberal poll majorities reaches statistical significance.
Table 2.6
Percentage of Consistent Decisions for the Rehnquist Court and Earlier
Courts since the Mid-1930s
RehnquistCourt EarlierCourts Difference
Overall results 64% 63% +1%
By state of public opinion
Crisis-times cases 74% 76% -2%
Non-crisis times cases 61% 59% +2%
By level of issue clarity
Undecided is 5% or less 70% 79% -9%
Undecided is over 5% 49% 57% -8%
By size of poll margin
Landslide margin (over 30%) 62% 65% -3%
Smaller poll margin (29% or less) 68% 60% +8%
By level and context
Consistent federal law 67% 81% -14%
Inconsistent federal law 40% 37% +3%
Consistent state/local law 71% 56% +15%
Inconsistent state/local law 57% 57% 0%
By type of issue
Fundamental freedoms cases 62% 66% -4%
Nonfundamental freedoms cases 67% 56% +11%
First Amendment claims 57% 63% -6%
Non-First Amendment claims 66% 63% +3%
Economic cases 53% 71% -18%
Non-economic cases 66% 60% +6%
By timing of the poll
Pre-decision poll 71% 67% +4%
Post-decision poll 52% 57% -5%
By ideology of public opinion
Conservative public opinion 70% 65% +5%
Liberal public opinion 55% 72% -17%*
No clear ideology 75% 80% -5%
By type of decision
Full, written opinion 60% 54% +6%
Cert denial, other 77% 67% +10%
Note: The total number of decisions is 130 for earlier Courts and 107 for the
Rehnquist Court; decisions with evenly divided or unclear polls are excluded
from this table. To compute the percent of inconsistent decisions, subtract the
percent of consistent decisions from 100%, by column.
* = stat. signif. at .05 for the difference, by time periods.
Reevaluating the Models
So far, the results are described simply in percentages and cross-tabula-

tions. To test how well these many variables predicted the Rehnquist
Courts agreement with public opinion, all the predictors described ear-
lier in this chapter were used individually and in various combinations
in a binary logistical (logit) model. A logit model tests the impact of
explanatory variables when the dependent variable (whether the
Courts decision agreed with public opinion or not) is dichotomous.
As expected, very few predictors were statistically significant in a
logit model. Only a single-predictor model improved upon the base rate
of 64% of the decisions correctly predicted. That predictor was
whether a large (over 5%) number of Americans were undecided on the
case, or not. Although not outstanding, this model correctly predicted
whether 66% of the Rehnquist Courts decisions would agree with
American public opinion, a significant improvement (at the .05 level)
over the 64% base rate. No other statistical model improved upon the
base rate. The Rehnquist Courts agreement with public opinion was
fairly widespread, and the state of public opinion model was the most
closely tied to consistent rulings.
Equally important, the prediction model developed for earlier
Courts since the mid-1930s failed to reach statistical significance for
the Rehnquist Court. In this earlier model (Marshall 1989: 93),
whether a Court decision agreed with American public opinion was
higher in crisis times, lower for inconsistent federal laws, and lower for
both consistent and inconsistent state/local laws and policies. This
model failed to predict when the Rehnquist Court would agree with the
pollsindicating that the Rehnquist Court relied less on the state and
federal policy models, and more on the nature of public opinion, com-
pared to earlier Courts. Apparently, different Courts find different
routes to reach agreement with public opinion.
CONCLUSION
This evidence points to both continuity and change. First, the

Rehnquist Court continued an earlier pattern of agreeing with
American public opinion about three-fifths of the time. Like earlier
Courts since the 1930s, between three-fifths and two-thirds of
Rehnquist Court decisions agreed with the polls. This is a remarkable
pattern of stability, given the Supreme Courts changing docket and
turnover among the justices themselves. The Rehnquist Court was a
majoritarian Court in that its decisions typically reflected American

attitudes (Friedman 2005; Peretti 1999).
At the same time, there are clearly differences in how the Rehnquist
Court, versus earlier Courts, represented American public opinion. In
large part, these changes are a by-product of the Rehnquist Courts
declining deference to the federal government and its increasing defer-
ence to state and local decision-makers. The Rehnquist Court less often
upheld consistent federal laws, but more often upheld consistent
state/local laws than did earlier Courts since the 1930s.
Of the three models considered in this chapter, the state of public
opinion model stands out as the strongest model. Nearly all its predic-
tions were in the expected direction, and one reached statistical signif-
icance. By comparison, the federal and state/local policy models lost
much of their explanatory power, compared to earlier Courts. Only
the state of public opinion model is included in chapter nines empiri-
cal model.
CHAPTER THREE
Judicial Norms, Representation,

and Public Opinion
T his chapter tests four examples of the business-as-normal model

the fourth model described in chapter one. This model asks what
impact the Supreme Courts routine norms, customs, and practices have
on representation. The pairwise method makes it possible to test how well
Supreme Court decisions would represent American public opinionif
only the Court behaved differently than it actually did.
The first example is ideology. What if the Rehnquist Court had
been a more politically conservative (or, alternatively, a more liberal)
Court than it actually was? The second example examines the
Rehnquist Courts pattern of overturning most of the lower court deci-
sions it heard. The third example tests the impact of judicial activism
and restraint against nonjudicial policy-makers. The fourth example
tests the impact of the Rehnquist Courts dwindling number of full,
written opinions.
Each what if example asks whether the Rehnquist Courts agree-
ment with public opinion was simply a by-product of other norms.
Perhaps the simplest example is ideology. Nearly all the Rehnquist
Courts decisions had a clear liberal-versus-conservative dimension.
Presidents try to pick like-minded justices with fair, albeit not perfect,
success. Once confirmed, most justices are usually fairly consistent in
their ideological voting patterns over time. Indeed, a justices on-the-
Court ideology can be predicted fairly accurately by their pre-appoint-
ment behavior as a judge, an attorney, or a legal scholar.
What if there had always been one more conservative vote (or,
alternatively, one more liberal vote) on the Rehnquist Court? Would
the Rehnquist Courts decisions have better represented American
public opinion? Or what if there had always been two or three more
conservative (or, alternatively, more liberal) votes? Or what if the
Rehnquist Court had always voted in a conservative (or, alternatively, a
liberal) direction?
51
In the second what if example, the Rehnquist Court, like earlier

Courts, typically reversed most of the lower court decisions it heard.
Did the Rehnquist Court more often agree with public opinion than the
lower court decisions under review? Or is the reverse true? By the evi-
dence below, the Rehnquist Court often reversed lower court rulings
that disagreed with nationwide polls. In so doing, the Rehnquist Court
engaged in poll correction, explained further below, against lower
federal and state court decisions.
The third example of the business-as-normal model examines the
impact of judicial activism and restraint. Assume, hypothetically, that
the justices do not care a great deal whether or not their decisions
agree with public opinion. Yet the justices want to hand down rulings
that consistently favor either the federal government or the states and
localities. If a majority of justices consistently upholds state, local, and
federal laws, and these laws themselves usually agree with nationwide
public opinion, then the Court would itself usually also agree with
nationwide public opinion. Yet the Courts agreement with public
opinion would simply be a by-product of the justices deference to
elected officials.
In the fourth what if example, the Rehnquist Court cut the
number of full, written decisions, per term, by roughly half, compared
to the Warren or Burger Courts. What impact did the Rehnquist
Courts dwindling caseload have on representation? If the justices had
simply denied certiorari (or denied writs of habeas corpus or appeals
for a stay) on all appeals, would the Rehnquist Court have more often
agreed with public opinion? Or less often?
Chapter three discusses these four examples, one by one. The
results suggest that the Courts business-as-normal norms have a great
impact on representation.
IDEOLOGY AND REPRESENTATION
Many authors (Baum 1998a; Gottlieb 2000; Savage 1995; Yarborough

2000) attribute the Rehnquist Courts conservative leanings to the
appointment of several conservative justices during the twelve-year-
long Reagan-Bush (Sr.) presidencies, coupled with the retirements of
older, liberal justices, particularly Justices Blackmun, Marshall, and
Brennan. Together, the Reagan-Bush justices often formed a conserva-
tive voting majority on the Rehnquist Court. Further, most justices are
relatively consistent in their ideological voting patterns over time. Only
a few justices sharply change their ideological voting patterns, even
Judicial Norms, Representation, and Public Opinion 53
though they typically sit on the Supreme Court for two decades or
longer (Epstein et al. 1998, Segal and Cover 1989, Segal et al. 1995).
Once appointed, most justices vote consistently with their pre-appoint-
ment views as demonstrated by their voting behavior on lower courts,
now usually federal appeals courts, or by their political and legal back-
ground, or by the arguments they made as a legal scholar.
To be sure, not all the Rehnquist Courts decisions were politically
conservative. Several well-known rulings (such as the flag-burning deci-
sions) were politically liberal (Clayton 1999; Kahn 1999). Nor did any
of the justices inevitably vote in the same ideological direction.1
Nonetheless, the Rehnquist Courts generally conservative record has
been widely commented upon. Some critics charge that the Rehnquist
Courts conservative bloc (most frequently, Justices Rehnquist,
OConnor, Scalia, Kennedy, and Thomas) reached politically conserva-
tive goals even at the expense of upholding precedents (Brisbin 1997:
90-92; Chemerinsky 1999a; Gottlieb 2000: 66; Maltz 1994).
The Rehnquist Courts conservative record, coupled with its fre-
quent 54 or 63 ideological divisions, raises an important question.
Does ideology affect how well the Court represents American public
opinion? Under the business-as-normal model, assume that each jus-
tice usually simply votes his or her own liberal-versus-conservative
leanings in a predictable and straightforward manner (Segal and
Spaeth 1993, 2002). If so, then how often the Court agrees with public
opinion depends on two issues. First, is American public opinion lib-
eral or conservative on cases that reach the Supreme Court? Second,
do liberal or conservative justices happen to hold a Court majority at
the time of the ruling?
Through the pairwise method it is possible to test whether one
additional conservative vote (or, alternatively, one additional liberal
vote) would have made the Rehnquist Court more representative of
public opinion. Or, would two or three additional conservative (or,
alternatively, liberal) votes have affected how well the Rehnquist Court
represented public opinion?
As an example, consider the 54 Roper v. Simmons (2005) decision,
overturning the death penalty for convicted murderers who were
younger than eighteen at the time of their crime. Roper v. Simmons
agreed with nationwide public opinion2 and was classified both as a lib-
eral decision and as consistent with the polls. If there had been one, two,
or three additional liberal votes, the Courts 54 decision would simply
have been 63, 72, or 81, and still consistent with public opinion.
However, had there been one (or two or three) additional conservative
votes, the Roper decision would have come out as conservative and as
inconsistent with public opinion by a 54, 63, or 72 margin, respec-

tively. Close votes (such as 54 or 63) are the most easily changed by
a small shift in the Courts voting patterns. Unanimous (90) votes
would not be changed by anything except a wholesale shift in the
Courts ideological makeup (such as five new presidential appoint-
ments) or by an unprecedented shift in the current justices votes.3
Figure 3.1 indicates how often the Rehnquist Courts decisions
would have agreed with public opinion if its decision-making had been
ideologically different than it actually was. The top line in Figure 3.1
indicates the percentage of decisions that would have been consistent
with public opinion if there had been one, two, or three additional
conservative votes, or (at the extreme) if the Rehnquist Court had
always decided cases in a conservative ideological direction. The
bottom line indicates the percentage of decisions consistent with public
opinion had there been one, two, or three additional liberal votes, or
(again at the extreme) if the Rehnquist Court had always decided cases
in a liberal direction.
Figure 3.1
Percentage of Rehnquist Court decisions consistent with public
opinion, with differing vote changes, by ideology (Note: Figure 3.1
includes only full, written decisons, and excludes unclear decisions.)
As Figure 3.1 suggests, adding one more conservative vote would

have increased the percentage of full, written opinions that was consis-
tent with public opinion from 58% to 63% consistent.4 Beyond that
point, one or two additional conservative votes would have slightly
decreased the Rehnquist Courts level of agreement with public opin-
ion. By contrast, one additional liberal vote would have dropped the
Rehnquist Courts level of agreement with public opinion only from

58% to 56%. However, additional liberal votes beyond that point
would have dropped the Rehnquist Courts level of agreement with
public opinion sharplydown to 46% consistent with two additional
liberal votes, and to 40% with three additional liberal votes. At the
extreme, if the Rehnquist Courts decisions had always been ideologi-
cally liberal, only 40% of its decisions would have agreed with public
opinion poll majorities.5
Clearly, the Supreme Courts ideological makeup has an impact on
representation. These results may bolster the view that the Courts
responsiveness to public opinion occurs chiefly through new appoint-
ments to the Court (Dahl 1957). Since few justices greatly change their
liberal-versus-conservative views, once appointed, major ideological
shifts on the Court typically result from new appointments. President
Bush Sr.s appointment of Justice Clarence Thomas to replace Justice
Thurgood Marshall was the most notable, yet not the only such exam-
ple during the Rehnquist Court (Epstein et al. 1998). New justices lib-
eral-versus-conservative views often differ considerably from those of
their predecessors, thereby also affecting the Courts responsiveness to
public opinion (Perine 2005). Chapter five returns to this question.
This pattern is also apparent for the five earlier Courts since the
1930s, as Figure 3.2 indicates. On decisions where vote totals were
reported, earlier Courts decisions agreed with public opinion 62% of
the time.6 One additional liberal or conservative vote would not have
much changed that figure, but more dramatic changes (in either a lib-
eral or conservative direction) would have sharply reduced earlier
Courts level of representation. Having either a completely liberal or a
completely conservative Court would have reduced earlier Courts level
of representation from 62% to only 49% or 50%in other words, to
a random-behavior level.
POLL CORRECTION AND THE LOWER COURTS
Typically, if the Supreme Court grants certiorari on an appeal from a

lower court decision, it then reverses that ruling. Does this practice of
reversing two-thirds to three-quarters of the lower court rulings also
bring judicial decision-making more closely into line with public opinion?
As an example, consider Kansas v. Hendricks (1997). There the
Rehnquist Court reversed a Kansas Supreme Court decision that had
struck down (on due process grounds) a state law allowing civil con-
finement for violent sexual offenders.7 In reversing the Kansas state
supreme court decision, and upholding the Kansas state law (the
Sexually Violent Predator Act), the Rehnquist Court agreed with a
large (82-to-17%) Gallup Poll majority. In so deciding, the Rehnquist
Court poll corrected the Kansas state supreme court decisionthat
is, it reversed an inconsistent lower court decision.
Figure 3.2
Percentage of Supreme Court decisions consistent with public opinion,
prior to the Rehnquist Court, with differing vote changes, by
ideology (Note: Figure 3.2 includes only full, written
decisions and excludes unclear decisions.)
If poll correction actually exists, it is an important concept in

democratic theory. Through poll correction, higher-level courts bring
judicial policy-making more closely into line with public opinion
(Marshall 2003). Historically, many scholars (Bickel 1962; Choper
1967, 1974; Kurland 1970; Mace 1972; Wellington 1982) believed the
Supreme Court was a countermajoritarian institutionone that was
little concerned with majority public opinion. Yet, by engaging in poll
correction, the Supreme Court is actually bringing judicial decision-
making more closely into agreement with public opinionhence, acting
as a majoritarian institution.
To test whether poll correction frequently occurs, the Rehnquist
Courts full, written opinions were compared to both public opinion
polls and the immediate lower court ruling.8 Overall, 61% of the
Rehnquist Courts full, written decisions agreed with nationwide public
opinion. By comparison, only 53% of the last lower courts decisions in
these cases agreed with nationwide polls. On balance, then, the net
effect of granting certiorari and then reversing, vacating, or reversing
and remanding so many lower court decisions was to raise from 53%
to 61% the number of these judicial decisions that agreed with public
opinion. The Rehnquist Court thereby had a poll correction rate of
+8% (or 61% 53% = +8%).
Whether or not a lower court decision agrees with American public
opinion also helps to predict whether the Supreme Court will uphold or
overturn a lower court decision, once certiorari has been granted. If a
lower court decision agreed with nationwide polls, the Rehnquist Court
upheld it 60% of the time. However, if a lower court decision disagreed
with nationwide polls, the Rehnquist Court upheld it only 38% of the
time.9 In other words, if the Supreme Court granted certiorari, a lower
court decision that agreed with nationwide polls had a much better
chance of being affirmed than a lower court decision that did not.
The figures are based only on cases for which the Rehnquist Court
granted certiorari, then handed down a full, written opinion. In reality,
much like earlier Courts, the Rehnquist Court denied certiorari on
nearly 99% of all the thousands of appeals it received each term. What
would a comparison of the much smaller cert grant pool and the
much larger cert denial pool show?
In the Rehnquist Courts much larger cert denial pool, some 77%
of the last lower court decisions were consistent with nationwide polls.
By denying certiorari on these lower court decisions, the Rehnquist
Court was thereby also consistent with nationwide public opinion 77%
of the time. By comparison, in the much smaller cert grant pool,
barely half (only 53%) of the last lower court decisions agreed with
nationwide polls. In the cert grant pool, the Supreme Court then
reversed, vacated, or reversed and remanded enough lower court deci-
sions to raise that 53% figure to 61% consistent with public opinion.
Restated differently, the Rehnquist Court denied certiorari on a pool of
cases 77% of which were consistent with American public opinion, and
granted certiorari on a pool of cases only 53% that were consistent with
public opinion. Although based on small samples of decisions, it seems
that the Rehnquist Courts certiorari grants also brought judicial deci-
sion-making more closely into line with public opinion.
The Rehnquist Court did not invariably engage in poll correction.
Its highest poll correction rates occurred on high profile (as measured
by Gallup Polls most important problem question) decisions (a
+13% poll correction rate), criminal issues (+24%), and morality and
abortion issues (+18%). Decisions involving health care and economic
issues represent the opposite extreme, with a negative 30% and
15% poll correction rate, respectively. By level of last lower court,
the Rehnquist Court had a +8% poll correction rate against federal
appeals courts and a +26% poll correction rate against state courts
(typically, state supreme courts).
Some justices more often engaged in poll correction against the lower
courts than did others. During the Rehnquist Court, Justices Rehnquist,
Kennedy, and OConnor had the highest poll correction rates+18%,
+17%, and +13%, respectively. Justices Ginsburg, Blackmun, Stevens,
Thomas, and Souter did not engage in poll correction at allreceiving
8%, 5%, 7%, 3%, and 5% rates, respectively.10
Whether poll correction is intentional on the justices part is an
intriguing, yet difficult question to answer. According to Chief Justice
Rehnquists (1986) sense and share argument, the justices may some-
times pick unpopular lower court decisions to review and reverse.
Alternatively, poll correction may simply be an inadvertent by-product
of the justices other goals, such as producing politically conservative
(or liberal) decisions, reducing demands on their time to a manageable
level (Baum 1997), or supporting their favorite judicial theories. Higher
level judges may be more sensitive to nationwide public opinion
because they are more closely scrutinized during the nominations
process, make decisions that receive media or congressional scrutiny, or
feel that they should reflect national, not statewide or regional atti-
tudes. Higher level courts are also collegial courts, thereby limiting the
influence of a single judge. Higher level judges typically receive more
information about public opinion through amicus briefs and the media
(Epstein 1999). Higher level judge may also follow contemporary
public opinion to avoid having a decision overturned or (admittedly
less likely) having the Courts jurisdiction curbed. Supreme Court jus-
tices are relatively free to disregard precedents that reflect long-out-
dated popular opinion, although lower level judges are not so free to
do so. Whichever is the case, both at the certiorari stage and at the full
opinion stage, the Rehnquist Courts business-as-normal patterns of
poll correction brought judicial decision-making more closely into line
with public opinion and, in so doing, served as a majoritarian check on
the lower courts.
Poll correction is not unique to the U.S. Supreme Court. Federal
appeals courts also engage in poll correction against federal district
courts. In a longer over-time analysis since the 1930s (Marshall 2003),
when both a federal district court and a federal appeals court made a
decision that received Supreme Court review, some 61% of federal
appeals court decisions, but only 49% of federal district court decisions
agreed with nationwide polls. Thus, federal appeals courts had a +12%
poll correction rate against federal district courts. Federal district court
decisions that reached the U.S. Supreme Court directly were consistent
with nationwide polls only 24% of the time, but the Supreme Courts
decision agreed with the polls 51% of the time in these disputesa
+27% poll correction rate.11 In short, higher-level courts consistently

engage in poll correction against lower courts.
Did previous Supreme Courts also engage in poll correction? The
evidence is that they didand to about the same extent as did the
Rehnquist Court. For the Hughes, Stone, Vinson, Warren, and Burger
Courts, when certiorari was granted, the last lower court had agreed
with nationwide public opinion 53% of the time. By comparison,
these five Courts since the mid-1930s agreed with nationwide polls
60% of the timea poll correction rate of +7%.12 This +7% figure for
five earlier Courts is nearly identical to the +8% figure for the
Rehnquist Court.
The Hughes, Stone, Vinson, Warren, and Burger Courts also much
more often upheld a consistent, rather than an inconsistent lower court
decision. For full, written decisions, these five Courts upheld 59% of
consistent lower court decisions, versus only 40% of inconsistent lower
court decisions. As well, some 67% of cases in the cert denial pool
were consistent with public opinion for these five earlier Courts, versus
only 53% consistent in the cert grant pool. These results provide
consistent evidence that a pattern of poll correction has existed at least
since the mid-1930s.
JUDICIAL ACTIVISM, JUDICIAL RESTRAINT, AND

REPRESENTATION
The terms judicial activism and judicial restraint are used in sev-
eral different ways, perhaps the most common of which is to refer to
the Supreme Courts willingness (or reluctance) to strike down dis-
puted federal, state, or local laws and policies, including agency deci-
sions (Canon 1982; Halpern and Lamb 1982; Todd 2004). Here, a
judicial activist decision is one that strikes down a law or policy
either on constitutional or other grounds, or else adopts such a
narrow statutory interpretation that the law or policy is effectively
nullified. The term judicial restraint means that the Supreme Court
upholds a federal, state, or local law or policy, or defers to an
agencys statutory interpretation.
The Rehnquist Courts decisions on the proper role of the federal,
state and local governments are among its most significant and contro-
versial legacies (Chemerinsky 2001; Fallon 2002; Gottlieb 2000;
Massey 2002; Yarbrough 2000). The Rehnquist Court upheld many
state and local laws, actions, and policies that might well have been
struck down by the Burger Court or the Warren Court. Some authors
(Davis 1989; Shapiro 1976, 1980) describe the Rehnquist Courts incli-
nation to make decisions favoring states and localities as resulting pri-
marily from changes in the Courts membershipespecially the
pro-states views of Justices OConnor, Kennedy, Scalia, and Rehnquist.
While it certainly did not uphold all state/local laws or policies, the
Rehnquist Court was much more favorable toward state and local
policy-making than earlier Courts were.
For the federal government, the picture is more complex, and the
Rehnquist Courts record of judicial activism versus restraint varied
from area to area. The Rehnquist Courts skepticism toward federal
policy-making was especially evident in cases involving relations
between the executive and legislative branches, toward claims by the
president, and in decisions that involved federal control or supervi-
sion over the states and localities. In a string of rulings, mostly since
the early 1990s, the Rehnquist Court reexamined long-standing fed-
eralism doctrines, struck down several key federal laws, and set out
new theories of state-federal relations. These rulings were often
decided narrowly, by a 5-4 vote, with the Courts conservative bloc
(typically, Justices Rehnquist, OConnor, Scalia, Kennedy, and
Thomas, and, in earlier years, sometimes Justice White) prevailing.
These decisions ranged widely in their constitutional logicfrom the
Guarantee Clause of Article IV, to the Fourteenth Amendment, to
the Tenth Amendment, to the Eleventh Amendment, to the
Commerce Clause.
For example, several well-known Rehnquist Court rulings
expanded a states sovereign immunity standing under the Eleventh
Amendment.13 The Rehnquist Court handed down several federalism
rulings unfavorable to the federal government, striking down or nar-
rowly construing federal laws to make them less widely applicable.14
The Rehnquist Court also narrowed Fourteenth Amendment or
Commerce Clause suits against states.15 Taken together, these rulings
often tilted the scales in favor of states or localities and against the fed-
eral government.
Not all the Rehnquist Courts federalism decisions favored the
states (Ponnuru 2004). Several key decisions upheld federal regulations
or allowed lawsuits against the states to proceed.16 Nor did dissenters
(in the Rehnquist Courts last decade, typically, Justices Stevens, Souter,
Ginsburg, and Breyer) accept either the logic or the precedent value of
these federalism rulings.17 Nonetheless, the trend in favor of state gov-
ernment decisions is an important Rehnquist Court legacy.
Elsewhere, the Rehnquist Courts record on federalism was very
mixed. In criminal cases, the Rehnquist Court often favored federal and
state/local governments at the expense of accused or convicted crimi-

nals (Yarbrough 2000). Well-known decisions generally favorable to
prosecutors and police include those on the exclusionary rule, searches,
some interrogations and confessions, or the earlier death penalty deci-
sions. On the other hand, its drug testing, excessive fines, ex post facto,
peremptory challenge decisions, and later death penalty decisions were
less often favorable to prosecutors or police. In its abortion, civil rights,
religion, and civil liberties cases, the Rehnquist Courts record of judi-
cial activism and restraint was also mixed (Belsky 2000; Yarbrough
2000). While upholding a basic right to abortion in Planned
Parenthood v. Casey (1992), the Rehnquist Court allowed several
restrictions on abortion, but struck down others. On gay rights claims,
assisted suicide, the right to travel interstate, loitering laws, religion and
education, indecency and obscenity, political speech, commercial
speech, regulation of political parties and interest group activities, and
affirmative action in the workplace or in contracts, the Rehnquist
Courts record was very mixed.
The key question here is what if the Rehnquist Court had always
ruled for either the states or localities, or for the federal governments
position. If the Rehnquist Court had done so (and, of course, it did
not), this would be pure judicial restraint. By comparison, if the
Rehnquist Court had (once again, hypothetically) always struck down
every state, local, or federal policy, that would be pure judicial
activism. At either extreme, would the Rehnquist Courts decisions
have more often been consistent with American public opinion?
The Supreme Courts views of judicial activism influence how
closely it represents public opinion for two reasons. First, Supreme
Court decisions routinely involve challenges to state, local, or federal
laws or policies. In this sample of decisions, two-thirds (71%) involved
a challenge to state or local law or policy, and a third (34%) involved a
challenge to a federal law or policy.18 Second, state, local, and federal
laws and policies do not themselves always agree with American public
opinion. In these Rehnquist Court decisions, most (70%) challenged
state or local laws or policies agreed with nationwide polls, while 30%
did not. Nearly the same percentage (73%) of challenged federal laws
or policies agreed with nationwide polls, while 27% did not.
If the Rehnquist Court had adopted a pure judicial activism
position (overturning all state, local, and federal laws and policies),
only 29% of its decisions would have agreed with American public
opinion. At the other (hypothetical) extreme, had the Rehnquist Court
adopted a pure judicial restraint position, some 71% of its decisions
would have agreed with public opinion. By comparison, the Rehnquist
Courts actual record was that 64% of its decisions agreed with the
polls.19 Restated, by adopting a pure judicial restraint model, the
Rehnquist Court would have agreed 7% more often with American
public opinion. However, adopting a pure judicial activism model
would have reduced from 64% to 29% the frequency with which the
Rehnquist Court agreed with American public opinion.20
These findings hold up even if federal and state/local laws or poli-
cies are considered separately. In both areas, by simply practicing judi-
cial restraint, the Rehnquist Court would have better reflected
American public opinion. On challenges to federal laws and policies,
the Rehnquist Courts decisions agreed with public opinion 59% of the
time - but if it had always practiced judicial restraint, that figure
would have jumped to 73% consistent. In its decisions on state and
local laws and policies, the Rehnquist Courts decisions agreed with
public opinion 65% of the timebut if it had always practicing judicial
restraint, that figure would have increased to 70% consistent.21
To summarize: the Supreme Courts views on judicial activism and
restraint affect how often its decisions reflect American public opin-
ion. As Rosen (2006) argued, by adopting a pure judicial restraint
stance, the Rehnquist Court would have better represented public
opinion. Stated otherwise, the Rehnquist Court did not poll correct
federal, state, and local actions in the same way as it poll corrected the
lower courts.
CASELOAD AND REPRESENTATION
One of the Rehnquist Courts most dramatic trends was the declining
number of full, written opinions. From its first to its last term, the
number of full, written decisions declined substantiallyfrom 152 to
79, with a record low number of only 77 full, written decisions during
the 1999/2000 term, according to the Harvard Law Review.22 By com-
parison, during the last two terms prior to the Rehnquist Court, the
Burger Court handed down an average of 155 full, written decisions
per term. The number of full, written opinions declined, although there
was a steady increase in the number of appeals filed per term.
Journalists and judicial scholars have debated both the reasons for,
and the importance of the Rehnquist Courts light docket (Banks
1995) or shrunken docket (Hellman 1996). In 1988 Congress elimi-
nated nearly all the Courts mandatory jurisdiction, thereby modestly
decreasing the Courts plenary docket (Hellman 1996: 412). Declining
conflict between circuits modestly contributed to the decline in the
Courts full, written opinions, as did the declining number of requests

for review filed by the U.S. Solicitor General. The relatively conserva-
tive federal appeals bench produced fewer liberal decisions that the
Rehnquist Courts majority wanted to overturn (Hellman 1996). Or,
the lower courts less frequent straying from Supreme Court views,
along with lower courts greater ability to predict the views of a Court
with little turnover may both have limited the number of appeals
(Posner 2005: 7071). The Rehnquist Courts liberal justices may have
been unwilling to vote to grant certiorari if they believed that the
Courts conservative majority would simply outvote them at the merits
stage (Biskupic 1996; Caldeira, Wright, and Zorn 1999). Similarly, lib-
eral interest groups, such as the ACLU or the NAACP, may have
decided not to file petitions for certiorari, even if they lost at the
appeals courts (Savage 1995). A lack of sweeping new federal statutes
also contributed to a dwindling number of path-breaking decisions
(Lazarus 1999: 30).
Further, some justices were simply unwilling to grant certiorari to
cases they believed to be of limited importance as a precedent. Justice
Ginsburg, Kennedy, Souter, and Scalia are less concerned either with
error correction or in building doctrine on a case-by-case basis
(Hellman 1996). An older generation of justices who sometimes
granted review to correct what they believed to be wrongly decided
lower court rulings was replaced by new justices with less expansive
views (Tushnet 1999). Most justices experience as a federal appeals
judge may also have disposed them to accept fewer appeals (Posner
2005: 71). Since the early 1990s, the Court more narrowly focused on
decisions of sweeping national importance, cases that police the bound-
aries between branches of government, cases with disagreement among
the appeals courts, or cases overturning precedents no longer supported
by a majority of justices (Hellman 1996; Tushnet 1999). Finally, fewer
justices were willing to join three and grant certiorari that otherwise
lacked the required four votes (OBrien 1997, 2005).
Critics also debated the importance of the Rehnquist Courts
shrinking docket. Some critics believe that the reduced number of full,
written opinions leaves more inconsistent rulings between circuits
making it more difficult for large businesses or government agencies to
operate consistently nationwide (Baum 1998a: 123127). Others com-
plain that when the chances that the Supreme Court will review a deci-
sion are so low, attorneys simply engage in forum shopping to find a
favorable circuit court of appeals (Biskupic 1996). Some critics argue
that lower court judges are left with little guidance on emerging issues,
and that the Supreme Court will lose its educational role in American
life (Chemerinsky 1996; Closen 1998). On the other hand, other schol-
ars argue that the Rehnquist Courts shrinking docket and the limited
scope of many decisions yield positive results in that states, Congress,
and the lower courts enjoy more flexibility in dealing with modern
problems (Dorf 1998; Sunstein, 1996).
Whatever its causes or consequences, what impact had the
Rehnquist Courts shrinking docket on representing public opinion?
What if the Rehnquist Court had not reduced by nearly half the
number of full, written decisions?
Consider that the Rehnquist Courts business-as-normal pattern
was to select lower court rulings only half (53%) of which agreed with
nationwide polls, and then to reverse enough of these decisions to raise
that figure from 53% to 61%a poll correction rate of +8%.
Lowering the number of full, written opinions by about seventy-five a
term means that the Rehnquist Court thereby failed to poll correct
about six lower court decisions a term. Over the entire Rehnquist
Court, this would be nearly a hundred lower court decisions that were
not poll corrected.
By this estimate, the dwindling number of full, written decisions,
per term, had a modest, negative impact on representation. If the
Rehnquist Court (hypothetically) had denied certiorari on all appeals,
the net number of consistent judicial decisions would have dropped by
yet another six decisions, per term. By contrast, if the Rehnquist Court
had decided an extra seventy-five or so full, written decisions per
termthereby keeping its output at the Burger and Warren Court
levelsthe number of consistent judicial decisions would have
increased by six decisions, per term.
These numbers may seem quite modestindeed almost trivial
in light of the thousands of appeals the Court denies each term.23 Yet
even a modest change in poll correction may have an important
impact. First, some Supreme Court decisions are of sweeping
national significance. When the Court overturns an inconsistent
lower court ruling on a highly visible issue, it demonstrates its
responsiveness to public opinion, occasionally even increasing or
reducing its own popularity. Second, Supreme Court decisions are
precedents for the future disputes. One additional decision that is
consistent with public opinion may lead to many more consistent
lower court rulings in the future. Third, Supreme Court rulings send
a signal to interest groups and potential litigants that the Court
might hear more appeals from the lower courts in the near future
(Baird 2004; Richards and Kritzer 2002).
CONCLUSION
The Supreme Courts changing judicial norms, practices, and customs

clearly affect how well the Court represents American public opinion.
Those who study representation should not ignore the impact of these
practices. Perhaps the strongest evidence is for ideology. Even a small
change in the Courts conservative-versus-liberal balance can greatly
affect how often its decisions agree with American public opinion.
Much like earlier Courts since the 1930s, the Rehnquist Court also did
a discernible amount of poll correction against lower courts. The
Rehnquist Court more often granted certiorari on unpopular lower
court decisions, and, having granted certiorari, more often overturned
unpopular decisions. Norms of judicial restraint and activism also
affect representation. Had the Rehnquist Court always deferred to leg-
islatures, executives, and agencies, more of its decisions would have
agreed with American public opinion. Finally, the Rehnquist Courts
dwindling number of full, written opinions reduced the number of con-
sistent judicial decisions, per term.
Admittedly, this chapter examined only four contemporary busi-
ness-as-normal customs that are easily quantified, measured, and
tested. More difficult to measure is the impact of other judicial theories,
such as textualism or original intent (Ring 2004; Scalia 1997, Weizer
2004), stare decisis (Spaeth and Segal 1999), clear statement rules
(Eskridge and Frickey 1994: 78), a contextualizing approach
(Maveety 1996), the preferred position (Miller 1987; Rogers 1999),
higher or natural law (Moore 2001, Thomas 1989), or substantive due
process. Nor is it clear how specific constitutional provisions such as
the ex post facto clause, long-standing views of constitutional provi-
sions such as the dormant commerce theory reading of the interstate
commerce clause (Eskridge and Frickey 1994:51), views of judicial
review such as the clear error standard (Shapiro 1986) or subsequent
legislative history tests (Eskridge and Frickey 1994: 65), or contempo-
rary legal practices (such as the increasing federalization of crime)
affect representation. Equally difficult to measure are how various bal-
ancing tests (Maveety 1996, Van Sickel 1998), the numerous canons of
statutory construction used by the Rehnquist Court (Eskridge 1989;
Eskridge and Frickey 1994), or ideas such as effective risk manage-
ment (Breyer 1993), active liberty (Breyer 2005), legal pragmatism
(Posner 2003, 2005), or intertextualism (Amar 1998) would affect
representation. Nor is the impact clear of intra-Court practices, such as
the justices reliance on high-quality oral arguments (Johnson,
Wahlbeck, and Spriggs 2006; Johnson 2004), high-quality attorneys

(McGuire 1993, 1994), bargaining over how opinions are written
(Maltzman, Spriggs, and Wahlbeck 2001; Spriggs, Maltzman, and
Wahlbeck, 1999; Wahlbeck, Spriggs, and Maltzman 1998), or strategic
behavior and voting (Boucher and Segal 1995; Schoen and Wahlbeck
2006). Even with these caveats, the business-as-normal model clearly
merits inclusion in chapter nines empirical linkage model.
CHAPTER FOUR
Interest Groups, Representation,

and Public Opinion
A t one time, interest groups were seldom involved in Supreme Court

nominations or lawsuits, but today interest groups play an impor-
tant role in the Courts decision-making. Interest groups sometimes
mount major lobbying and publicity campaigns for or against a nomi-
nee (Bronner 1989; Caldeira, Hojnacki, and Wright 1996; Caldeira and
Wright 1998; DeGregorio and Rossotti, 1994, 1995; Lichtman 1990;
Toner 2001). By supporting friendly nominees or by blocking
unfriendly nominees, an interest group can make a long-term impact on
the Supreme Courts decisions.
Interest groups also frequently initiate, support, and finance law-
suits that reach the Supreme Court (McIntosh 1985; OConnor and
Epstein 1984; Owens and Epstein 2005). At the certiorari stage, filing
an amicus brief can signal a cases importance and increase the chances
that an appeal will win a full Court review (Caldeira and Wright 1988,
1990; McGuire 1994; McGuire and Caldeira 1993).1 If the Court
agrees to hear a high-profile case, dozens of groups may file an amicus
brief (Behuniak-Long 1991; Epstein 1993; Krislov 1963; ONeill 1985;
Schlozman and Tierney 1986: 150).2 If the Court indicates an interest
in hearing a new issue, interest groups can provide more appeals within
the next two to five years (Baird 2004).
At the full opinion stage, interest groups often file amicus briefs in
hopes of persuading the justices to vote with the group and to adopt
the groups arguments (Behuniak-Long 1991; Collins 2004; Epstein
and Knight 1999; Epstein and Kobylka 1992; Ivers and OConnor
1987; ONeill 1985; Phillips 2004; Spriggs and Wahlbeck 1997;
Sungaila 1999).3 Amicus briefs on behalf of poor or disadvantaged liti-
gants help to offset the advantages that advantaged litigants (or
upperdogs) otherwise enjoy (Kearney and Merrill 2000; Songer,
Kuersten and Kaheny 2000).
Interest groups are not, of course, always successful, either in lower
courts or at the Supreme Court (Epstein and Rowland 1991; McGuire
67
and Caldeira 1993). On balance, the literature on interest group success

offers mixed results. Most groups participate in only a few cases when
their immediate interests are at stake (Kobylka 1991). Yet a few groups
play a key role by picking test cases to appeal,4 providing money and
legal assistance for appeals, advancing legal arguments, and sometimes
winning key victories (Bentley 1908; Cortner 1988; Epstein 1985; Ivers
1992, 1995; Kearney and Merrill 2000; Lawrence 1990; Manwaring
1962; Medalie 1968; OConnor 1980; OConnor and Epstein 1984;
OConnor and McFall 1992; Olson 1984; Sorauf 1976; Vose 1957,
1959; Wasby 1995; Way and Burt 1983; Wilcox 1992, 1996).
INTEREST GROUP LITIGATION AND PUBLIC OPINION
Do interest groups also mediate the linkage between American public

opinion and Supreme Court decision-making? The interest group link-
age null model makes two key assumptions. The first is that interest
group positions are not closely related to public opinion on lawsuits
reaching the Supreme Court. Interest groups seldom consider whether
their positions actually represent nationwide public opinion in deciding
whether or not to participate in a case. A groups decision to partici-
pate as a direct party, by offering financial support, by providing legal
assistance, or by simply filing an amicus brief is chiefly influenced by
other considerations. These include the cases importance; the likeli-
hood of a legal victory; the ability to set an important precedent; the
availability of funds; the expertise of the lawyers involved in the appeal;
a desire to raise the groups visibility; or pressure to take a position
from members, active supporters, or donors (Caldeira and Wright
1989; Cortner 1970a, 1970b, 1988; Epstein 1985; Hansford 2004a,
2004b; Kluger 1976; Lawrence 1990; Manwaring 1962; OConnor
1980; ONeill 1985; Sorauf 1976; Tushnet 1987; Vose 1959).
The desire to represent American public opinion is typically only a
marginal concern for groups, if it is a concern at all. In some ways,
interest groups do consider public opinion. If possible, groups may seek
out a sympathetic and upstanding plaintiff, file a lawsuit against an
unsympathetic opponent, and file in a friendly court. Groups often
engage in spin control by putting out news releases, encouraging
friendly law journal articles, and seeking favorable publicity during the
appeal or after the decision is announced (Cortner 1988, 1993, 1996;
Epstein 1985; Mezey 1996; OConnor 1980; ONeill 1985; Sorauf
1976; Vose 1959). Favorable publicity may help persuade Congress to
Interest Groups, Representation, and Public Opinion 69
overturn an unfavorable decision (OConnor 1980: 121), an issue

reconsidered in chapter eights test-of-time model.
Even so, litigation-minded groups often cannot control what cases
are filed, who the plaintiffs and defendants are, where the cases are
filed, or what appeals are pursued (Cortner 1970a, 1970b, 1993;
Kobylka 1991; Lawrence 1990; Manwaring 1962: 33; OConnor 1980;
Sorauf 1976: 95114; Wasby 1995). At the extreme, some groups
enjoy so little public opinion support that a litigation strategy is a prac-
tical necessity (Kobylka 1991; Manwaring 1962; Sorauf 1976; Vose
1959; Wasby 1995). Litigation-minded groups seldom pay great atten-
tion to public opinion polls in pursuing cases, or rely on public opinion
polls in planning their legal strategy.5
The interest group null models second assumption is that, at
least at the full opinion stage, group positions have no significant effect
on winning or losing a decision, independent of public opinion.
Restated more simply, a group that takes a position agreeing with
public opinion will usually win at the Supreme Court, but a group that
disagrees with public opinion will usually lose. Beyond the impact of
public opinion, interest group positions have no independent effect on
how the Supreme Court decides cases.
PUBLIC OPINION AND LITIGANT SUCCESS RATES
An earlier study (Marshall 1989: 9497) found considerable support

for this null model for the Hughes Court through the Burger Court.
Overall, the litigant or amicus brief positions of nine frequently active
types of groups agreed with American public opinion only at the
random-behavior levelthat is, 50% of the time. A group that agreed
with public opinion won 67% of the time. A group that disagreed with
public opinion won only 47% of the time. Including interest group
positions in a probit model did not help to explain whether or not
Supreme Court rulings would agree with public opinion.
Whether interest groups limited success during earlier Courts still
holds true for the Rehnquist Court is open to question. Compared to
earlier years, more interest groups now litigate cases and file amicus
briefs. The Rehnquist Courts generally conservative orientation may
also have affected interest groups success rates. To test whether inter-
est group positions mediate the linkage between public opinion and the
Rehnquist Courts decisions, the positions of twelve commonly active
types of groups were examined.6
Two measures were coded for each type of group. First, did the
groups position agree with public opinion? Second, did the Rehnquist
Courts decision agree with the groups position? Because groups
seldom participate by filing amicus briefs at the certiorari stage, due to
time and strategic concerns (Caldeira and Wright 1988; Epstein 1993),
only full, written decisions are included in the analysis.
As a benchmark, the federal governments involvement through the
solicitor generals arguing a case or filing an amicus brief was also
coded in the same manner. The solicitor general is a very successful liti-
gant, both as a direct party and as an amicus participant (Caplan 1987;
Deen, Ignagni, and Meernik 2003; Kearney and Merrill 2000; Salokar
1992; Sheehan, Mishler, and Songer 1992; Strauss 1998; Yates 2002),
in part due to that offices litigation experience (Days 1998; McGuire
1998), restraint in filing appeals, and careful selection of cases. The
solicitor generals position, which also typically reflects current federal
law, policy, and politics, also benefits from the long-standing norm of
judicial deference to Congress, the White House, and federal agencies
(Johnson 2003).7 Often, but by no means always, the solicitor generals
position is close to the Courts own preferences (Bailey, Kamoie, and
Maltzman 2005).
As Table 4.1 reports, interest group positions were not closely
related to American public opinion. These twelve types of frequently
active interest groups, active either as a direct party or much more fre-
quently through an amicus brief, agreed with public opinion only half
(51%) the time.8 Nearly as often (49% of the time) interest groups dis-
agreed with public opinion. These results are almost identical to those
from the 1930s through the mid-1980s when only 50% of interest
group positions agreed with nationwide polls (Marshall 1989). Interest
groups much less often agreed with public opinion than did the solici-
tor general or federal agencies, whose positions agreed with public
opinion 80% of the time.9 Taken individually, six of the twelve groups
mostly took positions agreeing with American public opinion, and six
did not. At the extreme, Catholic and pro-life groups were the most
likely to agree with public opinion, and liberal, religious groups were
the least likely to do so.
Interest groups that agreed with public opinion were much more
successful than those that did not. All twelve types of groups suffered
a drop in their success rates when they disagreed with public opinion.
Still excluding the federal government, if an interest groups position
agreed with public opinion, the group won 67% of the time. If a
group disagreed with majority public opinion, however, its success
rate dropped to 41%a 26% penalty. This penalty differed greatly
between groups. Catholic and pro-life groups had the smallest penalty
(only 1%), while feminist and pro-choice groups had the largest
penalty (55%).10
Table 4.1
Interest Group Involvement and Success during the Rehnquist Court
Groups Agreement withPercentage Success
Public Opinion When
Agree Disagree
with Pub. with Pub.
% Agree % Disagree Opinion Opinion
Type of group and
number of positions
taken
Civil liberties, free speech,
media (41) 41% 59% 65% 29%
States, counties, cities (50) 62% 38% 61% 47%
Civil rights, race (23) 35% 65% 50% 33%
Education (12) 33% 67% 75% 63%
Corporations, businesses,
associations (26) 42% 58% 73% 33%
Legal, bar associations (14) 57% 43% 75% 67%
Labor unions (7) 57% 43% 75% 33%
Feminist, pro-choice (18) 44% 56% 75% 20%
Liberal religious (22) 32% 68% 57% 47%
Catholic, pro-life (22) 86% 14% 68% 67%
Conservative (54) 57% 43% 65% 48%
Law enforcement (11) 64% 36% 100% 50%
All 12 types of groups (300) 51% 49% 67% 41%
Federal government (56) 80% 20% 69% 73%
Note: The category for all twelve types of groups excludes the federal govern-
ments involvement.
How does this 26% penalty for disagreeing public opinion com-
pare to the impact of an interest group disagreeing with the federal gov-
ernment? Overall, these twelve types of interest groups agreed with the
federal governments position roughly half (48%) of the time.11 If a
groups position agreed with the federal governments position, the
group won 77% of the time. However, a group that took a position
disagreeing with the federal government won only 36% of the timea
penalty of 41%. In percentage terms, then, the penalty for disagreeing
with public opinion was about two-thirds the penalty for disagreeing
with the federal government.
As expected, the federal government had a very different experi-

ence with the Rehnquist Court than did these twelve types of interest
groups. The federal governments position (either as an amicus or as
a party) overwhelmingly (80% of the time) agreed with public opin-
ion. Further, unlike interest groups, the federal government typically
won whether or not it agreed with public opinion. The federal gov-
ernment won 69% of the time when it agreed with public opinion,
and just as often (73% of the time) when it did not.12 Unlike interest
groups, the federal government did not pay a penalty for disagreeing
with public opinion.
In short, the interest group null linkage model remains a good
description of public opinion and Supreme Court decision-making.
First, interest group positions agree with public opinion no more often
than a random model would predictthat is, half the time. Second,
interest groups win much more often when they agree than when they
disagree with public opinion. An interest group that took a position
disagreeing with the polls could expect that its success rate would drop
sharply during the Rehnquist Court.
Court Eras and Public Opinion

The impact of public opinion on interest group success varies over time.
Consider, for example, the five types of very liberal interest groups:
free speech and media groups, civil rights groups, labor unions, liberal
religious groups, and feminist and pro-choice groups.13 From the mid-
1930s, until the mid 1980s, these groups, together, took eighty-five
positions that could be compared with American public opinion, and
won 73% of the time when they agreed with public opinion, but only
50% of the time when they did not (Marshall 1989: 96).
Now, compare these five liberal groups success rates during the
Rehnquist Court. Again taken together, these five types of groups took
111 positions that could be compared with public opinion. During the
Rehnquist Court these liberal groups won 64% of the time when they
agreed with public opinion, but only 33% of the time when they did
not. As expected, these groups paid a penalty during the more conser-
vative Rehnquist Court. Their winloss ratios dropped by 9% when
the groups agreed with public opinion, and by 17% when they group
disagreed with public opinion, comparing earlier Courts to the
Rehnquist Court.
Perhaps surprisingly, for liberal groups the Rehnquist Court
penalty is only about half as much as the penalty for disagreeing
with public opinion. Prior to the Rehnquist Court, disagreeing with
public opinion cost the group 23% in their winloss ratios. During the
Rehnquist Court, disagreeing with public opinion cost these groups
31% in their win-loss ratios.
Now, consider the success rates of the three most politically con-
servative types of groups: Catholic and pro-life, conservative, and law
enforcement groups. Prior to the Rehnquist Court, these groups took
twenty-one positions that could be compared to public opinion polls;
they were successful 54% of the time if they agreed with public opin-
ion, and 50% of the time if they did not. During the Rehnquist Court
these groups took eighty-seven positions that could be compared with
public opinion polls. As expected, these groups enjoyed a reward
during the Rehnquist Courtwinning 70% of the time when they
agreed with public opinion, versus 50% of the time when they did not.
Over time, then, conservative groups winloss ratios improved by
16% if they agreed with public opinion, although not at all if they did
not. By comparison, the penalty for disagreeing with public opinion
was 4% prior to the Rehnquist Court, versus 20% during the
Rehnquist Court.
Some groups winloss ratios are much more sensitive to public
opinion than to Court eras. Consider the experience of corpora-
tions, professional, and trade associations. Prior to the Rehnquist
Court, these groups took nineteen positions that could be compared
to nationwide polls, winning 57% of the time when they agreed
with the polls, but only 30% of the time when they disagreed with
the polls. During the Rehnquist Court, corporations and profes-
sional and trade associations took twenty-seven positions that could
be compared with nationwide polls, winning 73% of the time when
they agreed with the polls, but only 33% of the time when they did
not. This group experienced a small reward from the Rehnquist
Court. Their winloss ratio improved by 16% when agreeing with
public opinion, although only by 3% when disagreeing with public
opinion. By contrast, the penalty for disagreeing with public opin-
ion was much greater. Before the Rehnquist Court era, the penalty
for disagreeing with public opinion was 27%, and during the
Rehnquist Court the penalty for disagreeing with public opinion
was 40%.
To summarize: just as expected, some interest groups suffered a
penalty from the Rehnquist Court, while other groups enjoyed a
reward. Even so, all three types of groups experienced a penalty for dis-
agreeing with public opinion, both before and during the Rehnquist
Court. Since the 1930s, American public opinion has significantly
mediated the success rates of interest group litigants.
Rethinking Interest Group Influence

The interest group models second prediction is that interest group
positions do not significantly influence Supreme Court decisions, at
least at the full, written decision stage. This can be tested in a logit
model for the Rehnquist Court. The model was run (only for full, writ-
ten opinions) with each interest groups positions entered one by one.14
Each group could either take a position that agreed with public opinion
(coded as +1), disagreed with public opinion (coded as 1), or else take
no position (coded as 0). Whether the Supreme Court decision agreed,
versus disagreed with nationwide public opinion is the dependent vari-
able, also coded 1, 0.
When each interest groups position were entered, group by group,
in a logit equation, only the federal governments position significantly
predicted Supreme Court decisions (at the .05 significance level), indi-
cating that only the solicitor generals position mediates the relation-
ship between Supreme Court decision-making and American public
opinion.15 None of the interest group positions, taken alone or in com-
bination, significantly helped to predict when the Rehnquist Court
agreed with public opinion.
Described otherwise, the odds were 61% that a full, written deci-
sion agreed with public opinion during the Rehnquist Court.16 If the
solicitor general agreed with public opinion, those odds rose to 83%. If
the solicitor general took no position, the odds fell to 55%. If the solic-
itor general disagreed with public opinion, the odds fell to 27%.17
None of the positions of the 12 types of groups considered here signifi-
cantly changed those odds.
CONCLUSION
This chapter adds to the growing list of caveats about interest group
influence on Supreme Court decision-making. The U.S. solicitor gen-
erals position significantly affected whether or not the Rehnquist
Courts full, written decisions agreed with American public opinion.
None of the twelve types of interest groups considered here had such a
significant impact. On the average, disagreeing (versus agreeing) with
public opinion decreased an interest groups chances of winning a
Rehnquist Court decision from 67% to 41%. All types of interest
groups suffered a penalty for disagreeing with public opinion, although
the size of the penalty varied greatly from group to group.
For several interest groups, the penalty for disagreeing with public
opinion was even greater than the size of their penalty (or reward) that
resulted from the shift from earlier Courts to the Rehnquist Court. Not
a single type of interest group significantly mediated the relationship
between American public opinion and the Rehnquist Courts full, writ-
ten decisions. Why this occurs is an intriguing question, but one that
cannot be fully answered from this data. Under the attitudinal model, it
may be that the justices usually vote their own values, and interest
group positions add relatively little to this process. Perhaps the justices,
if strategically motivated, fear specific interest groups much less than
public opinion. Because most state, local, and federal decisions being
challenged at the Supreme Court do reflect contemporary public opin-
ion, interest groups challenging these policies also suffer from whatever
judicial deference exists among the justices. Conceded, this sample of
poll-matched decisions is a small one, and a much larger sample of
decisions might uncover different findings. However, results from ear-
lier Courts are quite similar to those for the Rehnquist Court.
In sharp contrast to interest groups, the U.S. solicitor general did
not pay a penalty for disagreeing with public opinion. The solicitor
general also significantly mediated the linkage between public opinion
and Rehnquist Court decision-making. The same results occurred for
earlier Courts since the 1930s. For the modern Court, then, the null
interest group model accurately describes the relationship between
public opinion and Supreme Court decision-making, at least at the full,
written stage. Accordingly, only the solicitor general is included in
chapter nines empirical model.
To be sure, these results do not suggest interest groups are insignif-
icant at the Supreme Court. Interest groups play a critical role in many
confirmation battles over Supreme Court nominees, and they often ini-
tiate, finance, and support appeals to the Supreme Court. By filing
amicus briefs at the certiorari stage, groups signal an appeals impor-
tance, and may persuade the Court to grant certiorari. Some amicus
briefs succeed in influencing dicta. All this notwithstanding, at the
Supreme Courts full, written stage, contemporary public opinion con-
strains interest groups ambitions much more than the reverse.
CHAPTER FIVE
Justices and Representation
J ust as some Supreme Court decisions agree with public opinion, but
others do not, so, too, can the individual justices votes either agree
with public opinion, or not. Beginning in the 1930s, modern polling
made it possible to measure how closely the justices votes actually rep-
resent public opinion. During the Hughes, Stone, Vinson, Warren, and
Burger Courts, Justices Hughes, Brandeis, and Vinson voted with poll
majorities over three-quarters of the time. At the other extreme, ten jus-
tices did so less than half the time. Chief justices, justices from presti-
gious law schools, and justices who had been close presidential advisers
most often agreed with public opinion (Marshall 1989: 106).
Were the same patterns true for the Rehnquist Court? As an exam-
ple, consider the Lee v. Weisman (1992) ruling. By a 54 decision, the
Supreme Court affirmed a First Circuit Court of Appeals ruling ban-
ning prayers as invocations and benedictions during a public schools
graduation ceremony. Justices Kennedy, Blackmun, Stevens, OConnor,
and Souter joined or concurred in the majority opinion. Justices Scalia,
Rehnquist, White, and Thomas dissented.
The Lee v. Weisman ruling was matched with a 1993 Gallup Poll
that asked: The U.S. Supreme Court has ruled that conducting reli-
gious prayers at any public school graduation ceremony is unconstitu-
tional because it violates the First Amendment, which concerns the
separation between church and state. Do you, yourself, believe that
prayers should or should not be part of public school graduation cere-
monies? A large majority (74 to 23%) of those surveyed responded
that prayers should be allowed, and the Lee v. Weisman ruling was
coded as inconsistent with public opinion.1 The five majority justices
votes (Kennedy, Blackmun, Stevens, OConnor, and Souter) were
counted as inconsistent. The four dissenters votes (Scalia, Rehnquist,
White, and Thomas) were counted as consistent with public opinion.
By this standard, the Supreme Courts full, written decisions yield
votes for the individual justices that are either consistent with American
public opinion, inconsistent, or unclear (if the polls themselves are
77
inconsistent or evenly divided).2 The fourteen justices who served on

the Rehnquist Court differed considerably in how often their votes
agreed with public opinion. Excluding cases of unclear polls, a total
of 764 individual votes could be counted for these fourteen justices, of
which 55% agreed with the polls and 45% did not.3
Just as for earlier Courts, some justices agreed with public opin-
ion much more often than did others. Justices White and Rehnquist
most often agreed with public opiniondoing so 73% or 69% of the
time, respectively. At the other extreme, Justices Brennan and
Marshall least often agreed with the pollsdoing so only 35% or
38% of the time, respectively.
Table 5.1
Rehnquist Court Justices Ranked by Agreement with Public Opinion
Rank Justice % Consistent Number of Votes
1 White 73% 37
2 Rehnquist 69% 86
3 OConnor 63% 84
4 Breyer 60% 47
5 Scalia 58% 85
6 Kennedy 56% 77
7 Thomas 53% 60
8 Souter 51% 63
9-10 (tie) Ginsburg 50% 50
9-10 (tie) Powell 50% 6
11 Stevens 45% 86
12 Blackmun 41% 37
13 Marshall 38% 26
14 Brennan 35% 20
Table 5.2 extends Table 5.1 by showing how the justices vote,
depending on whether public opinion itself is liberal or conservative. As
many Court-watchers note, the justices typically vote their ideological
preferences, which are relatively stable over time. If this is true, then a
justices votes might be very consistent with public opinion, but only
because the polls agree with the ideological position the justice was
already likely to vote for. For example, if public opinion was always
conservative on issues that reached the Court, Justice Thomas or Justice
Scalia (the Rehnquist Courts two most conservative justices) would be
quite consistent with public opinion. However, this would simply be an
artifact of the state of public opinion.
This is a different issue than whether a justice simply votes consis-
tently with public opinion. Table 5.1 addresses only consistency, and
Justices and Representation 79
shows that the Rehnquist Courts justices differed considerably on this

dimension. In Table 5.2 each justice who served on the Rehnquist
Court is listed, ranked (with that percentage indicated) from the most
conservative to the most liberal. The next column indicates the percent-
age of conservative votes a justice cast when poll majorities were con-
servative. The next column indicates the percentage of conservative
votes a justice cast when poll majorities were liberal. The final differ-
ence column indicates the percent change between these two columns.
For example, Justice Scalia (the most conservative justice) voted conser-
vative 85% of the time when poll majorities were conservative, and
77% of the time when poll majorities were liberala difference of 8%.
Table 5.2
Rehnquist Court Justices Voting Patterns, Percentage of Conservative
Votes, by Ideology of Public Opinion
Conservative Poll Liberal Poll
Justice Majorities Majorities %Difference
(82%) Scalia 85% 77% 8%
(82%) Thomas 83% 79% 4%
(80%) Rehnquist 88% 69% 19%
(74%) White 84% 50% 34%
(68%) Kennedy 76% 58% 18%
(59%) OConnor 72% 41% 31%
(33%) Breyer 52% 16% 36%
(29%) Souter 38% 20% 18%
(20%) Powell 25% 0% 25%
(16%) Ginsburg 23% 12% 11%
(15%) Stevens 19% 11% 8%
(6%) Blackmun 8% 0% 8%
(6%) Brennan 7% 0% 7%
(4%) Marshall 6% 0% 6%
Note: Justices are listed, in descending order, according to their percentage of
conservative votes, indicated in parentheses. Entries in the first and second data
columns indicate the percentage of conservative votes a justice cast in decisions
for which public opinion was conservative, or alternatively, liberal. The differ-
ence column indicates the percentage change in the two prior columns.
Because votes without a clear ideological content are excluded, the number of
votes, per justice, in Table 5.2 is slightly lower than in Table 5.1.
All the justices show some difference in their voting patterns, the
size of which varies greatly by justice. Justices Breyer, White, and
OConnor show the largest difference. Justices Thomas, Marshall,
Brennan, Scalia, Blackmun, and Stevens show the smallest differences.
Overall, the justices often voted differently, depending on the ideology

of public opinion.
The fourteen justices were then divided into four groups by their
ideology scoresincluding six very liberal justices (who voted less than
25% conservative), two moderately liberal justices (who voted from
26% to 50% conservative), three moderately conservative justices (who
voted 51% to 75% conservative), and three very conservative justices
(who voted over 76% conservative). Table 5.3 presents the results for
justices, by quartiles. All four groups of justices voted somewhat differ-
ently, according to the prevailing ideology of public opinion.
Ideologically moderate justices were the most likely to vote differently,
depending on the publics ideology. The average difference was 15%
for the Rehnquist Court as a whole.
Table 5.3
Rehnquist Court Justices Voting Patterns, by Public Opinion Ideology,
by Groups of Justices
% Conservative Votes % Conservative Votes Difference
When Public Opinion When Public Opinion
is Conservative is Liberal
( 3) Most conservative
justices 86% 72% 14%
( 3) Moderate-to-conserv-
ative justices 76% 49% 27%
( 2) Moderate-to-liberal
justices 43% 18% 25%
( 6) Most liberal justices 15% 8% 7%
(14) Total justices 56% 41% 15%
Note: Results in Table 5.3 are significant at the .01 level.
In short, the Rehnquist Court was as often consistent with the polls
as were earlier Courts since the mid-1930s. Further, both the Rehnquist
Court, taken as a whole, and all its fourteen justices, taken individu-
ally, voted differently, depending on the publics ideology. The remain-
der of this chapter describes six models linking American public
opinion to the justices voting patterns.4
THE POLITICAL PARTIES AND IDEOLOGY MODEL
The political parties and ideology model predicts that a justices votes
are closely tied to political party and ideology. Both party and ideology
are important in the nomination and confirmation process, and are pre-
dicted to be tied to whether or not a justice represents public opinion.
A justices political party leaning would be known prior to his or her
confirmation, and for most justices, so, too, would be the justices ideo-
logical leanings.
Political Party
All the Rehnquist Courts justices had clear political party ties. With
two exceptions (Justice Powell and Brennan), all were from the
appointing presidents party. Presidents now typically choose federal
judges almost exclusively from their own party, and political party is
often closely tied to a judges ideology. Democratic judges more often
support liberal economic claims, civil rights and civil liberties claims,
and criminal defendants and prisoners claims (Baum 1998b; Goldman
and Slotnick 1997; Stidham, Carp and Songer 1996).
From the 1930s through the 1970s, the Democratic Party claimed a
plurality, and sometimes a majority of Americans party loyalties.
Accordingly, Democratic justices and justices appointed by Democratic
presidents might have been predicted to be more consistent with public
opinion. Yet since the mid-1980s, the grassroots balance between
Republicans and Democrats has grown much closer,5 and no prediction
is offered about which partys justices or which presidential party
appointees will be more consistent with public opinion. The appointees
of recent presidents of either political party (Reagan, Bush, and
Clinton) are predicted to be more consistent with public opinion than
earlier presidents appointees.
Justices Own Party Preferences. The fourteen Rehnquist Court justices
were first divided by the justices own party preferences. Overall, only
slight differences appeared. Republican justices votes were only slightly
more consistent with public opinion than were Democratic justices
votes56% versus 54% consistent, respectively, reported in Table 5.4.
Neither were significant differences by justices party reported for ear-
lier Courts since the 1930s (Marshall 1989: 108).
Party of the Appointing President. Typically, a president now picks jus-
tices from his own party. Only Justices Brennan and Powell were
crossover appointeesboth Democrats appointed by a Republican
president. As a separate test, all Republican presidents appointees were
compared to all Democratic presidents appointees. Again, there was
little impact of political party in terms of how often a justices votes
reflected public opinion. Appointees of Republican presidents voted
consistently with public opinion 55% of the time, nearly identical to

the 56% figure for justices appointed by Democratic presidents.6
Era and Party of the Appointing President. The Rehnquist Courts
fourteen justices were appointed over a long period. The first (Justice
Brennan) joined the Court in 1956; the last (Justice Breyer) joined the
Court over a third of a century later. Appointees of Presidents
Eisenhower, Nixon, and Ford were examined together (early
Republicans), and compared to the appointees of Presidents
Kennedy and Johnson (early Democrats), Presidents Reagan and
George H. W. Bush (later Republicans), or President Clinton
(later Democrat)yielding groups of five, two, five, and two jus-
tices, respectively.7 The percentage of votes consistent with public
opinion varied only slightlyfrom 59% consistent (for early
Democratic presidents), to 57% (for later Republican presidents), to
52% or 55% consistent (for early Republican or later Democratic
presidents, respectively).8
Just as in an earlier study (Marshall 1989), political party does not
explain which justices will vote consistently with public opinion. While
political party helps to explain many other voting patterns, it does not
explain a justices agreement with public opinion.
Ideology
Most justices are consistent, over time, in their ideology. For example,
Justices Thurgood Marshall and William Brennan were described as
liberals at the time of their nomination, and later consistently voted in a
liberal direction. Justices Antonin Scalia and Clarence Thomas were
quite conservative when nominated, and then compiled conservative
voting records on the Court. Some justices, however, are less pre-
dictable, and change their voting behavior over time while on the Court
(Epstein et al. 1998; Martin and Quinn 2002). In two earlier studies
(Marshall 1989: 109110; Young 2001), ideologically moderate jus-
tices more often reflected public opinion in their votes than did liberal
or conservative justices.
Judicial scholars measure ideology in several different ways
through the opinions of journalists at the time of a justices nomina-
tion; or, alternatively, through scales, clustering techniques, bloc
analysis, or other measures based on justices actual votes on the Court.
Table 5.4 reports two measures of ideologyfirst, journalists impres-
sions of nominees ideology, and second, the justices actual voting
behavior while on the Court.
Pre-confirmation Ideology. A first way to measure a justices ideology

is to content analyze newspaper editorials at the time of the justices
nomination (Segal and Cover 1989; Segal et al. 1995). The Rehnquist
Courts justices were divided into three groups by ideology: conserva-
tive, moderate, and liberal justices.9 The justices agreement with public
opinion was then examined. As in an earlier study, justices thought to
be politically moderate at the time of their confirmation significantly
more often agreed with public opinion.
Table 5.4
Percentage of Justices Votes Consistent with Public Opinion,
by Political Party and Ideology
Percentage of Number of
Consistent Votes (Justices/Votes)
Overall average consistent 55% (14/764)
By political party
Justices own party
Republican 56% (8/578)
Democratic 54% (6/186)
Appointing presidents party
Republican 55% (10/604)
Democrat 56% (4/160)
Party and era of appointing president
Earlier Republicans 52% (5/235)
Earlier Democrats 59% (2/63)
Later Republicans 57% (5/369)
Later Democrats 55% (2/97)
By Ideology
Preconfirmation ideology**
Conservative 54% (7/423)
Moderate 62% (4/245)
Liberal 44% (3/96)
On-the-Court ideology**
Conservative 61% (3/231)
Moderate 60% (5/307)
Liberal 45% (6/222)
Note: Variables listed in Tables 5.4 through 5.9 are not significant at the .05
level unless so indicated. **= signif. at .01. Ratio-level variables reported in the
following tables include all fourteen justices.
On-the-Court Ideology. Although most justices voting records on civil

liberties and economic issues are well predicted by journalists pre-con-
firmation impressions, some justices vote quite differently than journal-
ists (Epstein et al. 1998) or their appointing presidents had expected

(Segal, Timpone, and Howard 2000; Szmer and Songer 2005.). An
alternative measure of ideology reclassifies these fourteen justices by
their actual liberal-versus-conservative voting behavior on the Court.
These on-the-Court results differ from those for pre-confirmation ideol-
ogy. Conservative justices slightly more often agreed with public opin-
ion (61% of their votes did) than did moderate justices (60%
consistent). Liberal justices votes least often agreed with the polls (only
45% consistent).10
THE POLITICAL SOCIALIZATION MODEL
The socialization model predicts that a justices pre-Court experiences

precondition a justice to follow majority public opinion. Majority-ori-
ented, broad, and diverse experiences, and exposure to situations that
require attention to public attitudes will sensitize a justice to majority
public opinion. Several hypotheses were described and reported in
Table 5.5. All the political socialization variables would be known even
before a justice joined the Court.
Judicial Backgrounds
Justices come to the Court with a wide variety of backgrounds.

Experiences that expose a justice to more diverse or majority-oriented
groups, to a contemporary culture, or to a politically active lifestyle are
predicted to make a justice more sensitive to public opinion and more
likely to agree with public opinion.
Region. The justices were compared according to the region or urban

setting in which they grew up, in which they lived at the time of their
nomination, or as to whether the justice ever lived in different regions.
Non-southern justices, justices who lived in different regions, and jus-
tices with an urban childhood were predicted to be most consistent
with public opinion.
Childhood Region. As a first test, the justices childhood region was

divided into four regions. Justices from the western states were the
most consistent with public opinion, although, contrary to expecta-
tions, eastern justices ranked last. Overall, no significant difference
appeared between southern and non-southern justices in how often
they agreed with public opinion.
Table 5.5
Percentage of Votes Consistent with Public Opinion,
by Justices Backgrounds
By personal background
Childhood region
East 50% (5/244)
Midwest 54% (3/209)
West 62% (4/245)
South 53% (2/66)
Region at nomination
East 57% (9/474)
Midwest 44% (2/123)
West 60% (2/161)
South 50% (1/6)
Mobility between regions**
Yes, moved between regions 62% (6/399)
No, did not 48% (8/365)
Childhood urban setting
Solely urban 53% (10/520)
Rural, small town, mixed 59% (4/244)
Childhood social class
Upper or upper-middle 58% (6/386)
Middle 55% (6/282)
Lower-middle or lower 49% (2/86)
Adult wealth
Above average 54% (5/330)
Below average 56% (4/434)
Minority status, by gender
Women 58% (2/134)
Men 55% (12/630)
Minority status, by race
Black 49% (2/86)
White 56% (12/678)
Minority status, by religion
Jewish 55% (2/97)
Catholic 54% (4/242)
Protestant 56% (8/425)
Combined minority status
Some minority status 55% (8/449)
No minority status 56% (6/315)
Age cohort
Born before 1930 54% (7/298)
Born 1930 or later 56% (7/466)
Table 5.5 (Continued)
Percentage of Votes Consistent with Public Opinion,
by Justices Backgrounds
Veteran status
Full, part-time, or related 57% (8/469)
None 52% (6/295)
By law school prestige
Top national law school 56% (13/738)
Less prestigious 38% (1/26)
By clerkship experience
Supreme Court clerk 60% (4/256)
No Supreme Court experience 53% (10/508)
Occupation upon nomination
Lower court judge 53% (10/609)
Appointive 64% (3/149)
Private practice 50% (1/6)
Prior judicial experience**
Some judicial experience 52% (11/635)
No prior judicial experience 69% (3/129)
Top administrative post*
Yes 61% (4/209)
No 53% (10/555)
Top executive or congressional post
Yes 57% (7/427)
No 53% (7/337)
Law school teaching experience
Yes 52% (6/382)
No 58% (8/382)
Perceived qualifications
High (.90 or higher) 52% (7/365)
Medium (.89 to .75) 59% (3/189)
Low (.74 or lower) 57% (4/210)
By political background
Candidate experience
Yes 63% (1/84)
No 55% (13/680)
Campaign experience**
Yes 60% (7/373)
No 51% (7/391)
Close presidential confidant*
Yes 73% (1/35)
No 54% (13/727)
* = signif. at .05 ** = signif. at .o1
Region at Nomination. By a second test, justices were reclassified as to

where they lived when nominated to the Court.11 Often this was a dif-
ferent region than a justices childhood region. By the same four-region
classification, western justices were the most (60%) consistent and mid-
western justices the least (only 44%) consistent.
Mobility between Regions. By a third test, the justices were reclassified
as to whether they had lived in different geographic regions.12 Because
geographic mobility exposes a justice to different groups and customs,
justices who moved between regions were predicted to vote more often
with public opinionand, indeed, were significantly more consistent
(62% versus 48% consistent).
UrbanRural. Most justices on the Rehnquist Court grew up in urban
areas, rather than in small towns or rural areas. Because urban areas
contain greater diversity and perhaps a greater tolerance for diversity,
and since the United States is now heavily urban, justices who grew up
in urban areas were predicted to be more consistent with public opin-
ion. Unexpectedly, however, justices who grew up solely in urban areas
less often cast votes that were consistent with public opinion.
Childhood Social Class. Supreme Court justices hail from widely differ-
ent childhoodsranging from privileged, wealthy, or upper-class back-
grounds (e.g., Powell and Stevens), to solidly middle class (Blackmun or
White), to relatively poor (Thomas). Justices from middle-class families
were predicted to vote most often with public opinion. However, the
differences were slight and justices from upper-class or upper-middle-
class family backgrounds actually voted the most consistently with
public opinion.
Adult Wealth. By the standards of average Americans, all the justices
are extremely well-educated and well-paid. Nonetheless, their reported
assets vary widely; some are quite wealthy, while others have much
more modest assets. Using a recent report (Epstein et al. 2003) to clas-
sify the justices wealth, nine could be classified as higher wealth, versus
lower wealth. As predicted, lower-wealth justices slightly more often
voted consistently with public opinion.13
Minority Status. Historically, most justices were white Protestant
males. In recent years, however, symbolic appointments have become
common; for the first time in history, a majority of justices on the late
Rehnquist Court were from minority religious (Catholic or Jewish),
gender (women), or racial (black) groups. Because whites and
Protestants are more numerous than blacks, Catholics, or Jews, white
and Protestant justices were predicted to vote more consistently with

public opinion. Women justices were predicted to vote more consis-
tently with public opinion than men.
Admittedly, the numbers of women (only two), black (only two),
Catholic (only four) or Jewish justices (only two) on the Rehnquist
Court is small. The differences between groups, though not significant,
are in the predicted direction. Women justices are slightly more often
consistent than men; black, Catholic, and Jewish justices are slightly
less often consistent than whites or Protestants.
Generation and Age. From oldest to youngest, the Rehnquist Court jus-
tices spanned a large age rangeby birthdates, from 1906 (Brennan) to
1948 (Thomas). As a test for generation effects, the seven justices born
before the 1929 Great Crash were compared to the seven justices born
after. As predicted, justices born after the Great Crash were slightly
more consistent with public opinion.
Veteran Status. Several justices served in the militarysome, such as

Justice Powell, with great distinction. Because the military exposes
many adults to wider social circles, new tasks, and distant locales, jus-
tices with military experience were predicted to agree more often with
public opinion. Five justices were veterans; three had part-time or
related experience; six had no military experience.14 As predicted, jus-
tices with some type of military experience were slightly more consis-
tent with public opinion than were justices without this experience.
Career Backgrounds
Prior to joining the Court, these justices had a variety of career experi-
ences. Careers that expose a future justice to more diverse, majority-
oriented experiences are predicted to dispose a justice to vote with
majority public opinion.
Law School Prestige. Historically, Supreme Court justices had a wide
range of law school training. Several never attended law school at all,
and were entirely self-taught. In recent years, a prestigious law school
background has apparently become a prerequisite for nomination to
the Court; only one Rehnquist Court justice (Marshall) did not attend a
top law school (Yale, Harvard, Stanford, or Northwestern). In an ear-
lier study (Marshall 1989), attending a prestigious law school signifi-
cantly predicted agreement with public opinion. The same pattern
occurred for the Rehnquist Court, although the lack of variation
among justices now decreases this predictors usefulness.
Clerkship Experience. Top law school graduates are sometimes asked

to spend a year or two clerking for a justice. While a clerks pay may be
relatively low, a clerkship is often an entree to a better position and a
more prestigious career (McGuire 1993, 1995). The four justices who
clerked for a Supreme Court justice were the most consistent with
public opinion.15
Occupation Upon Appointment. In an earlier study (Marshall 1989),

justices serving as a lower court judge were less consistent with public
opinion than those who were either in an appointive post or in pri-
vate practice at the time of their confirmation. The same result
appeared here, at least for lower court judges, versus those in
appointive posts.16 Again, however, the contemporary pattern of
picking sitting federal court judges makes this a less useful predictor
than during earlier Courts.
Previous Judicial Experience. All but three justices on the Rehnquist

Court had prior judicial experienceusually at the time of their nomi-
nation, but sometimes earlier in their careers. As predicted, prior judi-
cial experience was negatively related to a justices agreement rate with
public opinion. Justices with any prior judicial experience agreed with
public opinion only 52% of the timeversus 69% for the three justices
(Rehnquist, White, and Powell) without prior judicial experience.
Overall Career Patterns. Two other career positions were examined.

First, did the justice ever hold a top-level administration postinclud-
ing solicitor general (Justice Marshall), a top Justice Department post
(White, Rehnquist), or the Equal Employment Opportunity
Commission (Thomas)? As predicted, justices with top-level adminis-
trative experience were more consistent with public opinion.
As a broader test, justices who had held either a top-level adminis-
trative post (Marshall, White, Rehnquist, Thomas), a high-level con-
gressional post (Breyer, Stevens), or a high-level executive post (Scalia)
were compared to the remaining justices. As predicted, justices with
high-level Washington experience were somewhat more consistent with
public opinion than the remaining justices.
Law School Teaching. Several justices taught at a law school, even if

intermittently or early in their careers. In an earlier study (Marshall
1989), such justices were neither more nor less consistent with public
opinion than other justices. For the Rehnquist Court, however, those
with law school teaching experience were somewhat less consistent
with public opinion than the remaining justices.
Perceived Qualifications. Some Supreme Court nominees, such as

Justice Powell or Scalia, won praise as exceptionally well-qualified,
while others were described in less flattering terms. Highly praised
nominees were predicted to be more consistent with public opinion, but
the reverse was true. Nominees who received mixed or less favorable
editorial comments were the more consistent with public opinion while
serving on the Court
Political Backgrounds
Historically, many justices had experience in electoral politics, such as
serving as a U.S. senator or governor. In recent years, however, this
type of experience is no longer common. On the Rehnquist Court, only
Justice OConnor had any elective experience, and, as predicted, she
was above average in her agreement with public opinion (63% consis-
tent, versus 55% for the remaining justices).
Several justices had worked actively in an election as a campaign
advisor or strategist, as a political party activist, or as a candidates vol-
unteer supporter. Examples include Justice Kennedys role for a
California ballot proposition, or Justice Whites role as a Colorado
coordinator for President Kennedys 1960 Democratic presidential race.
As predicted, justices with such experience were significantly more con-
sistent with public opinion than the remaining justices60% versus
51%, respectively.
Presidential Confidant. Historically, several justices served as a close
presidential confidantan experience significantly related to high levels
of agreement with public opinion (Marshall 1989: 114). That practice
(sometimes criticized as cronyism) died out with President Johnsons
failed nominations of Homer Thornberry to the Court and Justice Abe
Fortas to chief justice. On the Rehnquist Court, only Justice White
would have so qualified, and, as predicted, White was more often con-
sistent with public opinion than other justices.17
THE APPOINTMENTS PROCESS MODEL
Presidential Choices
Most Supreme Court nominees were a presidents first and only choice
to a vacancy. Yet a few justices were nominated and confirmed only
after one or more other presidential choices refused, were rated so
poorly by the American Bar Associations review committee that they
were dropped from consideration, were rejected by the Senate, or
withdrew. Two justices (Powell and Kennedy) were nominated only

after two other nominees were rejected by the Senate or else withdrew
their nomination. Two other justices (Ginsburg and Breyer) were not
the presidents first or even second choice (Abraham 1999). Third
choice justices were predicted to be less often consistent with public
opinion, but no differences appeared. For earlier Courts since the
1930s, second (or lower) choices were slightly more often consistent
with public opinion than were first choices62% versus 57%, respec-
tively, reported in Table 5.6.
Table 5.6
The Appointments Process Model
Presidential choice
First choice 55% (9/584)
Not first choice 55% (4/180)
Crossover appointee
Yes 38% (2/26)
No 56% (12/738)
Confirmation margin
Unanimous or voice vote 55% (7/426)
Fewer than 12 negative votes 52% (5/188)
Over a dozen negative votes 62% (2/146)
Crossover Appointments
Historically, a few presidents appointed opposition party supporters to
the Supreme Courtusually to appeal to voters in an upcoming elec-
tion, to give an air of bipartisanship, or to secure a confirmation when
the Senate was in the opposite partys hands. Crossover appointees
were never common, and the practice last occurred over a third of a
century ago. Only two justices on the Rehnquist Court (Brennan and
Powell) were crossover appointees. Although crossover appointees were
predicted to be more often consistent with public opinion than other
justices, they proved to be somewhat less so. The crossover appoint-
ment thesis not only finds little empirical support; it is also outdated,
since no such appointments have been made in over three decades.
Senate Confirmation Margin

While most Supreme Court nominees are easily confirmed, a few have
difficult confirmation battles and only narrowly win Senate approval.18
Justice Clarence Thomas is an example of a difficult confirmation

battlewinning only after lengthy hearings and by a narrow 5248
margin. These fourteen justices were divided into three groupsthose
confirmed unanimously or by a voice vote, those who received fewer
than a dozen negative votes, and those who received a dozen or more
negative votes. Perhaps surprisingly, the two justices (Rehnquist and
Thomas) with the most negative confirmation votes also more often
agreed with public opinion. The same pattern existed for earlier jus-
tices. Justices who received more than a dozen negative votes were
more often consistent with public opinion than justices who received
one to eleven negative votes, or who were confirmed unanimously or
by voice vote (62%, versus 50% or 59% consistent, respectively).19
In short, the appointments process yields few clues as to which jus-
tices will be the most consistent with public opinion. None of the three
predictors consistently predicted whether a justice, once confirmed, will
vote consistently with American attitudes. This may be surprising,
given that nominees are now closely scrutinized by the president, the
media, and the U.S. Senate. Perhaps the model would find some sup-
port if there were some way to compute a voting record for failed nom-
inees; since they never cast votes on the Court, however, this is not
possible. The appointments process model cannot be included in chap-
ter nines empirical model.
THE JUDICIAL ROLES MODEL
Once confirmed, a justice can take on several roles on the Court. Most
notably, the chief justice is usually in the position to assign opinions,
assuming that he is in the majority in a vote (Maltzman and Wahlbeck
1996). The chief justice also fulfills an institutional role as the Courts
spokesperson before Congress, the executive branch, and the public
(Davis 1999). Some justices are task leaders or play-makers; others
are intellectual leaders; still others act as social leaders, or as argument
leaders during oral argument (Danielski 1972, 1986; Haynie 1992;
Mason 1958; Murphy 1964; Walker, Epstein, and Dixon 1988; White
1976: 200-215). Historically, a few justices pursued off-the-Court
goals, such as running for president or leading important commissions.
Some justices take on none of these roles.
Roles that lead to greater visibility or institutional responsibilities
(such as the chief justice), or involve bargaining skills (such as being a
task leader or writing landmark opinions) were predicted to lead to
greater agreement with public opinion. Unlike the socialization and
appointments process models, judicial roles (except chief justice status)

would only be known after a justice joins the Court.
Chief Justice
Since the mid-1930s, chief justices agreed with public opinion more
than other justicesa significant and consistent pattern over time
(Marshall 1989: 124).20 The same pattern reappeared. Chief Justice
Rehnquist was 15% more often consistent with public opinion than
other justices. Further, after Justice Rehnquist became chief justice, his
level of agreement with public opinion rose by 9% (from 60% to 69%)
reported in Table 5.7. By the weight of the evidence, the chief justices
role as the Courts chief lobbyist and spokesperson21 encourages a
greater sensitivity to public opinion.
Table 5.7
The Judicial Roles Model
Chief justice**
Yes 69% (1/86)
No 54% (13/678)
Intellectual leadership*
Yes 63% (3/177)
No 53% (11/587)
Task leadership
Yes 62% (3/112)
No 54% (11/652)
Argument leadership
Yes 55% (2/135)
No 55% (12/629)
Swing votes and opinion assignment
Yes 57% (5/353)
No 54% (9/411)
Career significant opinions
authored per term
Rehnquist Court (only), significant
opinions authored per term
* = signif. at .05
Reputation
Some justices are well-regarded by historians, judicial scholars, judges,
journalists, and attorneys. Yet most justices rank as only average,
below average, or even as a failure. Judicial scholars use two techniques
to rank justices. The first is a reputational method, based on scholars
views of the justices behavior on the Court, writing of opinions, intel-
lectual breath, legal influence, personal integrity, and bargaining skills
(Pederson and Provizer 2003, Schwartz 1997). The second approach is
based on more easily quantified measures, such as number of landmark
opinions written, or the total number of cases decided (Choi and Gulati
2004; Epstein, George, Giles, and Walker 1992).
Admittedly, it is premature to rank the Rehnquist Courts justices,
since some have already died or retired, but others are barely past their
first decade on the Court. Even so, some comparisons are tentatively
possible. Justice Brennan sometimes appears on lists of great or near-
great justices (Abraham 1999: 369372; Pederson and Provizer 2003:
215231; Schwartz 1997: 1113). However, contrary to predictions,
Brennan was among the justices who least often agreed with public
opinion.22 Chief Justice Rehnquists accolades as an outstanding chief
justice sometimes place him on lists of outstanding justices (Pederson
and Provizer 2003: 232245). Rehnquist also ranks as among the jus-
tices who most often agreed with the polls, particularly after he became
chief justice. The evidence from past Courts is also mixed.23
Leadership on the Court

Some justices win praise for providing outstanding intellectual leader-
ship, task leadership, or, occasionally, both. While such lists are subjec-
tive, three justices (Powell, Rehnquist, and Scalia) are cited as providing
intellectual leadership. Three justices (Brennan, Powell, and Rehnquist)
are cited as providing task leadershipthat is, skillful leadership in
piecing or keeping together a majority beyond simply supplying a fifth
vote. Two justices (Scalia and Ginsburg) are described as the most
aggressive during oral arguments, and can be styled as argument lead-
ers (Carp and Stidham 1998: 266268).24
If these descriptions of intellectual, task, and argument leaders are
accepted, leadership is often associated with greater agreement with public
opinion. Task leaders agreed with the polls 62% of the time, versus 54%
for other justices. Intellectual leaders agreed with the polls 63% of the
time, versus a 53% figure for the other justices.25 Argument leaders agreed
with the polls 55% of the time, the same rate as for other justices.
Another account (Edelman and Chen 2001) treats leadership as

the likelihood of being the critical fifth vote, or as assigning the
majority opinions author. While different measures and terms pro-
duce different leaders, five justices often led this ranking: Justices
Kennedy and OConnor for providing the critical fifth vote, and
Justices Rehnquist, Stevens, and Brennan for assigning authorship of
the majority opinion. Together, these five justices agreed with public
opinion 57% of the time, versus 54% for other justices. In short, the
most dangerous justice agreed with public opinion only a little
more often than did other justices.
Significant Opinion Authorship

Some justices author many landmark decisions, while other justices
seldom do so. During the Rehnquist Court through 1996 (Epstein et al.
2003), Justice Marshall authored the fewest, while Justices Brennan
and Rehnquist authored the most significant opinions, per term. Chief
justices and senior justices can most often assign decisions, including
landmark decisionsand some often assign important decisions to
themselves, perhaps giving them an advantage in this ranking. Junior
justices usually score lower than they would in their later years. Even
so, authoring landmark opinions is an important role on which justices
vary considerably.
The fourteen justices were divided into two groups in terms of how
many significant opinions they authored over their entire careers.
Justices who authored more than one significant decision, per term,
were somewhat more consistent with public opinion than those who
authored fewer significant opinions.26 As an alternate test, the justices
were ranked on significant opinion authorship for the Rehnquist Court,
and ranked as above-average or below-average. Again, above-average
justices were more often consistent with public opinion than were
below-average justices.
THE TENURE AND AGING MODEL
Justices now average about a quarter century on the Court before dying
or retiring (Ward 2003). The justices increasing years on the Court
often leave them as the last political survivors of their generation, and
increasingly distant, at least in age, from most Americans. This pattern
is likely to continue, due to medical advances and presidents preference
for relatively young justices who can survive two or three decades on
the Court. Some critics also cite the Courts comfortable schedule, the
assistance of law clerks, and the reduction of the Courts mandatory
docket in explaining growing longevity. Law professors, in particular,
debate the pros and cons of longer tenure, perhaps as a result of the
Rehnquist Courts record eleven terms without a new appointment. To
some critics, longer tenure reduces the number of new openings on the
Court, denies some presidents the ability to nominate any justices, and
increases the number of ineffective or even decrepit justices
(Atkinson 1999; Calabresi and Lindgren 2006; Cramton and
Carrington 2006; Farnsworth 2005, 2006; Garrow 2000; Gruhl 1997).
The length-of-tenure model predicts that newly appointed justices
will be more consistent with public opinion than are longer-tenured jus-
tices. Further, the justices will most often vote consistently with public
opinion during their early years, rather than during their later years on
the Court. Very senior justices (e.g., those who have been on the Court
over twenty years) will be less consistent with public opinion than are
junior justices. Five versions of this model are examined in Table 5.8.
Table 5.8
The Length-of-Tenure Model
First four years versus later
First four years 62% (6/104)
Later tenure 52% (6/278)
Holdover effects
Newly appointed justices 62% (6/104)
Holdover justices 56% (/790)
Replacement effects
Newly appointed justices 62% (6/104)
Retired justices 49% (6/82)
Senior justices
Prior to 20th year 60% (7/443)
20th year or later 54% (7/235)
Senior and nonsenior justices
Senior justices 54% (7/235)
Nonsenior justices 56% (/538)
Early and Late Court Tenure

The first test simply compares a newly appointed justices voting
record during his or her first four years to that same justices own
voting record thereafter. During the Rehnquist Court six new justices
were appointed, two each by Presidents Reagan (Scalia and Kennedy),
Bush (Souter and Thomas), and Clinton (Ginsburg and Breyer). If the
length-of-tenure model is correct, then these six justices should be
most consistent with American public opinion during their early years
on the Court.
This prediction finds some support. Together, these six justices
agreed with public opinion 62% of the time during their first four
years, compared to only 52% of the time during their subsequent years
on the Court. All the justices except Breyer and Ginsburg were more
consistent during their first four years than in later years.27 During the
Rehnquist Court, then, justices were more consistent with public opin-
ion early in their Court tenure than thereafter.28
Holdover Effects
A different comparison is whether newly appointed justices are more
consistent with public opinion than are holdover justicesthose who
continued to serve during a new justices first four years. Table 5.8 com-
pares newly appointed justices voting records with that of holdover jus-
tices. As an example, during his first four terms, Justice Scalia voted
consistently with public opinion 75% of the time. By comparison,
holdover justices (those who were already on the Court at the time of
Justice Scalias appointment and who were not themselves in their own
first four years) voted consistently with public opinion only 57% of the
time. Thus, during his first four years, Justice Scalia was 18% more con-
sistent with public opinion than were the holdover justices.
On the average, the Rehnquist Courts six newly appointed justices
were consistent with public opinion 62% of the time during their first
four yearsversus only 56% for the holdover justices. In other words,
newly appointed justices were 6% more consistent during their first
four years than holdover justices. Except for Justices Souter and
Ginsburg, all newly appointed justices were more consistent than the
holdover justices.29 By appointing president, the average effects were
higher for appointees of President Reagan (+10%) than Bush (+5%) or
Clinton (+1%).
Replacement Effects
During the Rehnquist Court six justices were appointed, each replacing
a retiring justice. During their first four years on the Court, these six
new justices cast 104 votes that could be matched with nationwide
polls, of which 62% were consistent with public opinion, and 38%
were not. How does this compare to the last four terms of the six jus-
tices they replaced?
Together, the six retiring justices cast eighty-two poll-matched
votes during their last four years on the Court, only 49% of which
agreed with public opinion. This difference of 13% (or 62%, versus
49%) is termed the replacement effect. Admittedly, this is an indirect
test of the tenure and aging model, since each pair of justices cast votes
on different decisions. Even so, the result is similar to all the other tests
reported here. New justices are more often consistent with American
public opinion than were the justices they replaced.30
Senior Justices
Historically, about a third of justices served more than two decades on
the Court.31 Long-serving justices are appointees of much-earlier presi-
dents, are much older than the average American, and represent the last
survivors of their cohorts. On the Rehnquist Court seven justices
(Brennan, Marshall, White, Blackmun, Rehnquist, Stevens, and
OConnor) reached or exceeded twenty years of service. Their agree-
ment with public opinion after their twentieth year on the Court was
compared, first, to their own record in earlier years, and then to that of
more recently appointed justices on votes after the justices twentieth
year on the Court.
On the average, the Rehnquist Courts longest-serving justices
became less consistent with public opinion late in their tenure.
Together, these seven justices cast votes that were consistent with
public opinion 54% of the time after their twentieth year of service on
the Court, compared to 60% consistent before that stagea drop of
6% over time.32 Most justices became less consistent with public opin-
ion over time.33 This seniority effect was not observed for justices
prior to the Rehnquist Court.
By the final test, the most senior justices were compared to other
justices votes during the period after which the senior justices reached
their twentieth year of service. The seven Rehnquist Court justices
who reached their twentieth year of service were consistent with public
opinion 54% of the time thereafter. By comparison, during that period
the less senior justices (as well as senior justices prior to their own
twentieth year of service) were consistent with public opinion 56% of
the time.
By all these five tests, just as predicted, senior justices are less
often consistent with public opinion, although the differences vary in
size. Justices are 10% more consistent with the polls in their first four
years, versus thereafter. Newly appointed justices are 6% more consis-
tent than holdover justices. Newly appointed justices are 13% more
consistent in their first four years than were retiring justices in their
last four years. Justices who reach their twentieth year of service are
thereafter 6% less consistent with public opinion than they were in
earlier years. The Rehnquist Courts most senior justices were also 2%
less consistent than less senior justices. Apparently, growing longevity
does decrease the accuracy with which the justices represent public
opinion. This consistent evidence merits inclusion in chapter nines
empirical model of representation.
THE REALIGNMENT MODEL
Robert Dahls thoughtful essay (1957) argued that the Supreme Court
would seldom be out-of-line with a contemporary political majority
except during and shortly after political realignments.34 During
realignments, justices appointed by an earlier president might resist
newly popular political, social, and economic ideas, and, at least tem-
porarily, succeed in overturning laws passed by a newly elected presi-
dent and congressional majority. Eventually, those justices deaths and
retirements would allow a new president to appoint new justices more
in line with the current political majority. Once a new presidents
appointees gained control of the Court, it would again become part of
a national majority.
The Dahl thesis was historical in nature and focused on the
Supreme Courts record of overturning legislation during historical
realignments. Because public opinion polls were not available for
most of American history, Dahls argument could not directly test
whether the Courts decisions during realignment eras actually agreed
with poll majorities. Comparing decisions to poll results is a more
direct test of whether political realignments affect the Courts pattern
of representation.
The evidence for the New Deal period did not support Dahls argu-
ment (Marshall 1989: 118). Republican justices appointed prior to the
1930s were actually slightly more, not less consistent with public opin-
ion than were New Deal appointees. Unfortunately, realignments occur
so seldom (by several accounts, about every thirty years) that there are
few opportunities to reexamine the Dahl thesis. Many justices are con-
firmed, serve on the Court, and retire or die entirely between realign-
ment periods.
To be sure, political scientists do not always agree what time peri-

ods constitute a realignment.35 Some accounts (Aldrich 1995, Edsall
and Edsall 1992, Shafer 1991) argue that the New Deal realignment
ended by 1968, and was followed by years of dealignment, during
which party loyalties weakened, ticket-splitting increased, and elections
became more television- and candidate-centered (Wattenberg 1994).
Some accounts argue that 19801984 or 19801988 was a realignment
period (Adams 1997; Brunell and Grofman 1998; Burnham 1981,
1991; Gunning and Saunders 1998; Ladd 1985; Lublin and Voss 1999;
Rosenof 2003, Schneider 1985; Sundquist 1983; Tuchfarber, Smith,
Rademacher, and Downing 1997). The percentage of Americans identi-
fying as Republicans came close to the percentage identifying as
Democrats, and the GOPs strength increased most greatly among
white males, conservative Christians, and southerners (Black 2004).
Beginning in 1980, Republicans won three presidential elections in a
row, gained control of the U.S. Senate for several years, and began to
fare much better in state elections, particularly in the South. On bal-
ance, then, the Reagan presidency of the 1980s is at least a modest
political test of the Dahl thesis.36
If Dahls argument is correct, the Reagan appointees to the Court
should vote more consistently with public opinion than the appointees
of earlier presidents, or, arguably, than appointees of later presidents.
In fact, Reagan appointees agreed with public opinion 59% of the time
during the Rehnquist Court, as often as the votes of earlier Democratic
presidents. That was followed by Clinton appointees (55% consistent),
and by earlier Republican (Eisenhower, Nixon, and Ford) and George
Bush, Sr. appointees (52% consistent), reported in Table 5.9.
Arguably, this test takes too long a view. If a shorter time window
is considered, such as the early Rehnquist Court from 1986 through
1992, and if pre-Reagan appointees are considered together, more sup-
port appears for the realignment thesis. During this shorter window of
time, the three Reagan appointees (OConnor, Scalia, and Kennedy)
voted consistently with public opinion 66% of the time, compared to
54% for the more seven senior justices.
By a weaker test, the justices votes are simply compared before,
versus after their appointing presidents party lost the White House.
Before partisan shifts at the White House, the justices agreed with the
polls 61% of the time, versus 54% thereafter.
Whether the post-1994 period can be considered as a realignment
period is also open to dispute. In 1994, Republicans gained, and then
kept control of the U.S. House until the 2006 election; regained control
of the U.S. Senate for nearly all the next twelve years; and in 2000 and
2004 elected a Republican president, albeit with neither a popular vote

majority nor even a plurality in 2000. Republicans also fared much
better in southern elections (Bullock, Hoffman, and Gaddie 2005).
Scholars disagree as to whether the 1990s were a realignment period
(Abramowitz and Saunders 1998; Burnham 1996; Meffert, Norpoth,
and Ruhil 2001; Rosenof 2003; and Tuchfarber et al. 1995).
Table 5.9
The Realignment Model
Justices by appointing president
Earlier Democratic appointees 59% (2/63)
Earlier Republican appointees 52% (5/235)
Reagan appointees 59% (3/246)
Bush, Sr., appointees 52% (2/123)
Clinton appointees 55% (2/97)
From 1986 to 1992*
Earlier appointees 54% (7/203)
Reagan appointees 66% (3/103)
By White House shift
Prior to shift 61% (/163)
After shift 54% (/601)
For the 1994 realignment
Republican appointees 52% (7/338)
Democratic appointees 56% (2/94)
* = signif. at .05
If the post-1994 period is counted as a realignment period, then a

practical problem arises. Justice Ginsburg and Breyer, both Democrats,
joined the Court in 1993 and 1994, respectively, just before this
realignment. Then there was not another new appointment through the
rest of the Rehnquist Court, until Justice OConnor retired and Chief
Justice Rehnquist died in 2005. Stretching the realignment thesis a bit,
one might argue that Justices Ginsburg and Breyer, as the last (and
only) Democrats in a newly Republican era, might be expected to have
lower-than-average rates of agreement with public opinion after 1994,
compared to the seven earlier Republican appointees (who were
appointed before the 1990s realignment, but were from the Republican
Party). However, the two Democrats voted consistently with public
opinion 56% of the time, versus only 52% for the seven Republicans.
Admittedly, this is at best an indirect test of the realignment thesis for
1994, but no better test is possible, and the results do not support the
realignment thesis. Unfortunately, there were no new Republican

appointees until Justices Alito and Roberts joined the Court, and that
was over a decade after 1994. Further, it will be some years before
Justices Alito and Roberts compile enough of a voting record to allow
reliable comparisons of how well they represented public opinion, com-
pared to earlier-appointed justices.
Robert Dahls realignment thesis remains an intriguing linkage
between Supreme Court decision-making and American public opinion.
Yet it is a difficult argument to test. Many justices are appointed, serve,
and retire or die between realignments. Other justices do not fit neatly
into the realignment thesis appointment sequence (OBrien 2000:
324349). If the early 1990s were a realignment, then the long delay in
any new justices appointed thereafter makes it impossible adequately to
test the realignment thesis.37 The exact time frame required to ade-
quately test the realignment thesis is also unclear. In short, the evidence
for the realignment thesis is mixed during the Rehnquist Court, and
negative for the New Deal realignment. As a result, the realignment
model is not included in chapter nine.
COMBINING THE PREDICTORS
Only a few variables in this chapter significantly predicted whether a

justices votes agreed with public opinion. In a previous study
(Marshall 1989: 124) from 1935 through 1986, three variables best
predicted how often a justice agreed with public opinion: service as
chief justice, attendance at a prestigious law school, and experience as a
close presidential confidant. Over time, however, the last two of these
three predictors lost their predictive value, since all recent justices
attended top law schools and none were close presidential advisors.
What best explains which Rehnquist Court justices most often
agreed with American public opinion? Given the small number of four-
teen justices on the Rehnquist Court, coupled with the high level of
multicollinearity among predictors, a good model should have strong
explanatory power, contain as few predictors as possible, yield a
straightforward interpretation, and reflect current norms for justices.
A simple, three-variable model best describes the Rehnquist Court
justices agreement with public opinion. This model includes three pre-
Court socialization variables: whether a justice moved between regions,
previously served as a judge, and was described as a moderate when
confirmed.38 At one extreme, a politically moderate justice who moved
between regions, but never served as a judge, is predicted to agree with
public opinion 73% of the time. At the other extreme, a justice who
never moved between regions, who did serve as a state or federal judge,
and who was not politically moderate is predicted to agree with public
opinion only 44% of the time. No other variables were statistically sig-
nificant or added to the models predictive power.
How well does this simple model predict the justices voting
behavior during the Rehnquist Court? Despite the small number of
justices, the models predictions are quite accurate, with only a 3.7%
(absolute) error, as reported in Table 5.10. Most justices agreed with
the polls about as often as predicted. For the eight justices appointed
to the Court before Justice Rehnquist became chief justice, the aver-
age (absolute) error was 3.4%, versus 4.2% for the six justices
appointed thereafter.
Table 5.10
Predicted versus Actual Agreement with Public Opinion, by Justice
% Predicted % Actual
Consistent Consistent Difference
Brennan 44% 35% 9%
White 73% 73% 0
Marshall 44% 38% 6%
Blackmun 44% 41% 3%
Powell 54% 50% 4%
Stevens 44% 45% +1%
OConnor 63% 63% 0
Rehnquist 65% 69% +4%
Scalia 55% 58% +3%
Kennedy 52% 56% +4%
Souter 44% 51% +7%
Thomas 55% 53% 2%
Ginsburg 44% 50% +6%
Breyer 63% 60% 3%
The ties between Supreme Court justices and American public

opinion are clearly time-bound. Contemporary appointment norms also
affect representation. From the 1930s through the mid-1980s, serving
as chief justice, attending a prestigious law school, and serving as a
close presidential confidant were the most important ties. Yet all recent
justices graduated from prestigious law schools, and none served as
close presidential advisors.39 The current norm that a Supreme Court
nominee be a sitting (usually federal appeals court) judge also takes its
toll, costing an estimated 10% on the level of representation. For some
justices, this effect is offset by moving between regions or their politi-

cally moderate leanings, while for others it is not.
Using this simple model also allows a prediction as to how often
future justices will agree with public opinion. For example, Chief
Justice John Roberts is predicted to agree with public opinion about
55% of the time. For Justice Samuel Alito, the prediction is 44% of the
time. Both the Roberts and Alito estimates are well below the career
records for Justices Rehnquist and OConnor, respectively, the justices
they replaced. By contrast, the predicted figure for Harriet Miers, who
withdrew her nomination, was 65%. As a caveat, the accuracy of these
predictions depends on there being no major changes in patterns of rep-
resentation. Since justices now typically serve on the Supreme Court for
two or three decades, these predictions can only be disproved some
years in the future.
CONCLUSION
The Rehnquist Courts fourteen justices varied considerably in terms of

how often their votes agreed with public opinion. Some justices agreed
with the polls more than two-thirds of the time, but other justices did
so less than half the time. In short, there are wide variations in how
often the justices agree with American public opinion on disputes that
reach the Court. Further, most of the justices liberal-conservative
voting patterns varied, depending on the prevailing ideology of
American public opinion.
Only a few predictors strongly predicted which justices would most
frequently agree with the polls. A simple, three-variable model accu-
rately predicted how often the Rehnquist Courts justices agree with the
polls. Not surprisingly, given the recent profile of justices, this model is
different from that of earlier Courts. The ties between public opinion
and justices votes are clearly time-bound.
Current political norms greatly affect how well the justices repre-
sent American public opinion. The modern custom of picking career
federal judges without elective experience, and who then remain on the
Court for two or three decades, clearly reduces the ties between public
opinion and Supreme Court decision-making. At the same time, a pres-
idents choices for the Court affect representation. If presidents were to
return to a now-bygone pattern of picking more politically experienced,
ideologically moderate nominees, and nominees with a broader expo-
sure to American life, future Courts would probably more accurately
represent public opinion.
Admittedly, the prospects that any of these changes will actually

occur seem slight. Over the last twenty years, all recently confirmed jus-
tices have been career federal judges with little exposure to the give-
and-take of electoral politics. None have been close presidential
confidants. The failed nomination of White House counsel, Harriet
Miers (herself a former elected Dallas city council member, a graduate
of a less prestigious law school, a close presidential confidant, but lack-
ing experience as a federal judge) may well give future presidents pause
in making another such nomination any time soon. Nor, despite the
complaints of so many law professors, is there now any serious effort
afoot to impose term limits on the justices. In short, several modern
norms of Supreme Court appointments and tenure negatively affect
representation, but, at the same time, seem quite resistant to change.
CHAPTER SIX
Symbolic Representation and

the Court
T he symbolic representation model is quite different from the other

models in this book. It makes two key assumptions. First, once
confirmed to the Supreme Court, the votes of symbolic nominees
(e.g., Jewish, Catholic, women, or black justices) will better represent
the views of Jewish, Catholic, women, or black Americans, respectively,
than do the votes of other justices. In other words, a justice who sym-
bolically represents a religious, gender, or racial group will actually cast
votes that agree with the views of the group so represented significantly
more often than do other justices.1 Second, a symbolic nominee, once
confirmed, will less well represent the views of the comparison group
than he or she represents the nominees own group. For example,
women justices votes will less often represent the views of American
men than they represent the views of American women. Does this link-
age model have any empirical support?
THE POLITICS OF SYMBOLIC REPRESENTATION
Presidents nominate symbolic nominees to the Supreme Court for many

reasons (Abraham 1999; Friedman and Israel 1969, 1980; Perry 1989,
1991). A symbolic nomination might appeal to an important interest
group or voter group in an upcoming election or redeem a past cam-
paign promise (Silverstein 1994: 107110, 156157). A president
might nominate a symbolic nominee as part of that presidents histori-
cal legacy. A widely respected, well-qualified nominee, or a presidents
personal favorite contender might simply happen to be a symbolic
nominee (Perry 1989, 1991; Silverstein 1994: 1018).2 Indeed, some
symbolic nominees have only marginal ties to their demographic group,
and a president might feel very comfortable with the nominees real
politics (Abraham 1999: 175176; Mason: 1956, 563568; Perry and
107
Abraham 1998). A symbolic nominee might have an easier time being

confirmedespecially if the president is nearing the end of his term,
suffers from low approval ratings, or faces a Senate controlled by the
other political party.3
Historically, presidents often made symbolic nominations to the
Supreme Court. From the earliest times, presidents chose Supreme
Court justices from different geographic regions of the country, and,
until the 1950s, a justices regional background was often important.
President George Washingtons nominees to the Supreme Court were
geographically diverse, with no two justices who served at the same
time hailing from the same state. The establishment of judicial circuits,
the practice of circuit-riding, the diversity of regional issues, and the
existence of regionally based political parties doubtlessly helped to per-
petuate the idea that Supreme Court justices should come from differ-
ent geographic regions.
By the early 1900s, at least one Catholic (Perry 1989, 1991) or
Jewish justice (and sometimes more) typically sat on the Court.
Thurgood Marshall became the first black justice in 1967. President
George Bushs 1991 nomination of Clarence Thomas upon Justice
Marshalls retirement continued that tradition. President Reagans
1981 nomination of Sandra Day OConnor fulfilled a campaign
promise to appoint a woman justice (Abraham 1999: 282). Today, the
common meaning of symbolic representation means a woman, racial
minority, Catholic, or Jewish justice.4
Presidents often make symbolic nominations to the Supreme Court.
Nearly half (twenty-four of fifty-five) of associate or chief justices con-
firmed during the 1900s provided some type of symbolic representa-
tion. Fifteen justices symbolically represented a state or geographic
region. Five justices represented the minority political party. Others
filled the Courts Catholic or Jewish or black or woman
seats. A few justices even provided several types of symbolic representa-
tionsuch as William Brennan (a Catholic Democrat from the
Northeast) or Ruth Bader Ginsburg (a Jewish woman). By 1994, most
of the justices provided symbolic representationeither by gender
(OConnor and Ginsburg), race (Thomas), or religion (Scalia, Kennedy,
and Thomas as Catholic,5 Ginsburg and Breyer as Jewish). By the end
of the Rehnquist Court, only Chief Justice Rehnquist and Justices
Stevens and Souter failed to meet any of the modern standards for sym-
bolic representation.
Many Americans believe that it is important to have a woman or a
black Supreme Court justice. However, support for symbolic nomina-
tions varies greatly depending on the poll questions wording. For
Symbolic Representation and the Court 109
example, in 1991 a small majority (53%) of Americans agreed that it is

important that there always be at least one woman on the Supreme
Court, while 44% of Americans said it was not important.6 In 1993
two-thirds (64%) of Americans agreed that Bill Clinton should
appoint a woman to the current vacancy on the Supreme Court. while
13% of Americans disagreed and 23% said they were not sure.7 In
1993, when asked does the fact that (recent nominee) Ruth Bader
Ginsburg is a woman make you more favorable to her nomination, less
favorable, or does it make no difference to you?, some 17% of
Americans said more favorable, while 6% said less favorable.8 In 2000,
about a quarter (28%) of Americans said it would be better for soci-
ety if most Supreme Court justices were women. Nine percent said it
would be worse for society and 60% said neither better nor
worse.9 In 2003, a quarter (26%) of Americans said it was very
important and another 29% said it was somewhat important that
George W. Bush nominate a woman to the Supreme Court.10 In 2005, a
seventh (14%) of Americans said it was essential that a woman be
nominated to replace retiring Justice Sandra Day OConnor, and 43%
of Americans said it was either not essential, or else a good idea, but
not essential.11
Support for a black Supreme Court justice also varies widely
depending on the poll questions wording. In 1991, half (50%) of
Americans said it was important that there always be at least one
black person on the Supreme Court, while 47% of Americans said it
was not important.12 Yet in another 1991 poll only 20% of Americans
said it was important to have at least one black member of the
Supreme Court, while 77% of Americans said race should never be a
factor in choosing Supreme Court justices.13
Many Americans believe that symbolic representation plays an
important role in Supreme Court nominations. In 1991, nearly half
(48%) of Americans said they believed President George Bush nomi-
nated Clarence Thomas to the Supreme Court because he is black,
while nearly as many (39%) disagreed with that statement.14 Many
Americans also believe symbolic nominations affect a justices behavior
on the Court. In 1991, 40% of Americans said that the fact that
Clarence Thomas is black will have (some impact) on the kinds of deci-
sions he will make if he is confirmed as a Supreme Court justice, while
14% said it would have a great deal of impact and 38% said it
would have not much impact.15
Taken together, polls during the Rehnquist Court suggest three
conclusions. First, many Americans accept symbolic representation as a
legitimate criteria for Supreme Court nominations. Second, many
Americans believe presidents should16 or do search for qualified women

or blacks when making Supreme Court nominations. Third, many
Americans believe that symbolic nominees, once confirmed, vote differ-
ently than other justices.17
Public support for symbolic representation should be placed in per-
spective. Large numbers of Americans also agree that a nominees per-
sonal history, character, or views on current issues should be
considered. In 1989, three-quarters (73%) of Americans agreed that
when the Senate votes on a Supreme Court nominee, (it should) con-
sideralong with the nominees legal qualificationsthat persons per-
sonal history and character.18 Half (50%) of Americans agreed that a
senator should base his vote to confirm or deny (an intellectually well-
qualified Supreme Court nominee) on the candidates views on abor-
tion, school prayer, and other social issues.19 Three-fifths (62%) of
Americans said that senators should attach (a lot of importance) to
the nominees positions on major Constitutional issues.20 Clearly, a
nominees symbolic role is not the publics only criterion.
Several presidential statements attest to the importance of symbolic
representation. President Carter was an outspoken supporter of sym-
bolic representation. At a 1978 press conference, President Carter
announced that if Senate approval were not required, 12% of his judi-
cial appointees would be black, 3% Spanish-speaking, and 40%
women (Abraham 1999, 280; Berkson and Carbon 1980; Goldman
1997; see also Gryski, Zuk, and Barrow 1994). During his 1980 presi-
dential campaign, President Reagan promised that he would appoint a
well-qualified woman to the Court (Abraham 1999: 282). His first
Supreme Court nominee, Sandra Day OConnor, fulfilled that promise.
President George Bushs 1991 nomination of Clarence Thomas contin-
ued the Courts black seat. President Clinton spoke positively of the
merits of appointing more women and minorities as federal judges
(Goldman and Slotnick 1997; Perry and Abraham 1998: 162; Stidham,
Carp, and Songer 1996).
Insider accounts provide anecdotal support for the symbolic repre-
sentation model. For example, the Courts mustering a unanimous pro-
affirmative action decision in Hishon v. King & Spaulding (1984) is
sometimes attributed to Justice Sandra Day OConnors early career
failure to win an attorneys job at a major California law firm, despite
her distinguished academic record at Stanford Law School.21 Justice
Thurgood Marshall was sometimes said to be much more interested in
civil rights cases than in other disputes (Abraham 1999; Lazarus 1999;
Schwartz 1997; Woodward and Armstrong 1979). Justice Marshall
also compiled an exceptionally strong record in supporting civil rights
and liberties (Abraham 1999: 222). Justice Ginsburgs post-inaugura-

tion statement did not argue that symbolic nominees voted differently
than other justices, but she did note that symbolic nominees add a
diversity of background and experience to the Courts delibera-
tions.22 Many Americans take special pride in the Courts symbolic
nomineesfor example, Justice OConnor as the first woman justice,
or Justice Ginsburg and Breyer as Jewish justices (Perry 1999: 121,
132). Some symbolic nominees (such as Justice Ginsburg) espouse their
views on symbolic representation from the Court or in off-the-court
public speeches (Biskupic 1995).
Academic studies report mixed results as to whether symbolic
nominees actually vote differently than do other lower federal and
state court judges (Gryski, Zuk, and Barrow 1994). Men and women
judges in federal or state courts often hand down different decisions,
at least on a few types of cases, such as employment discrimination,
family law, search and seizure cases, or the length of jail terms (Allen
and Wall 1987, 1993; Davis, Haire, and Songer 1993; Songer, Davis,
and Haire 1994). Yet other studies report largely negative findings
(Davis 1986; Gottschall 1983; Gruhl, Spohn, and Welch 1981; Kritzer
and Uhlman 1977; Segal 2000; Walker and Barrow 1985). This pat-
tern of mixed results or modest differences also appears for black,
versus white judges (Uhlman 1978; Walker and Barrow 1985; Welch,
Combs, and Gruhl 1988; cf. Gottschall 1983). The impact of race or
gender seems small compared to voting differences between
Republican and Democrat judges.
The evidence that symbolic representation leads to voting differ-
ences on the Supreme Court is mixed.23 During their years together on
the Rehnquist Court, Justice OConnor and Ginsburg voted together
only at an average rate.24 Because Justice Thomas replaced Justice
Marshall, it is not possible to directly compare the two black justices
voting records. Even so, Justice Marshall compiled a strong liberal
voting record, while Justice Thomas compiled an equally strong conser-
vative voting record.25 The wide variations in voting behavior between
Justices Brennan, Scalia, Kennedy, and Thomas (all Catholics who
served on the Rehnquist Court) show that symbolic nominees do not
necessarily vote alike.26 Justices Ginsburg and Breyer (both Jewish) pro-
vide the strongest evidence that symbolic nominees may vote together
at above-average rates (OConnor and Palmer 2001).27 Yet Ginsburg
and Breyer have many other traits in common; both were Democrats,
former federal appeals judges, Easterners, and Bill Clinton appointees.28
Perry and Abraham (1998) argue that symbolic representation now
rarely explains much about the justices voting records.
Adequately testing the symbolic representation argument requires

actual polling evidence. Two symbolic nominees might often vote
together; both might even agree with the legal arguments of a racial,
gender, or religious groups organized interest groups, as represented in
a lawsuit or an amicus brief. Yet those justices might still disagree with
poll majorities of the group they supposedly represent. In an earlier
study (Marshall 1993), only Justice Thurgood Marshall fully met the
most rigorous criteria for the symbolic representation thesisand then
only on civil rights issues. Justice Marshalls votes on civil rights and
racial disputes perfectly agreed with majorities of black Americans, and
he represented black Americans attitudes at a much higher rate than
did the Courts white justices.
SYMBOLIC REPRESENTATION ON THE REHNQUIST COURT
This chapters question is whether, once confirmed, symbolic nominees

actually represent the policy views of the group supposedly so repre-
sented. The symbolic representation model predicts that, once con-
firmed, a symbolic nominees votes will significantly better represent
the attitudes of their symbolic group than do the other justices votes. A
symbolic nominee should also less well represent the views of the non-
symbolic group than do the other justices.
The poll-matching method makes it possible to determine whether
symbolic representation actually existed on the Rehnquist Court. Poll
responses for the full, written decisions here were broken down by
three demographic groups now usually considered as symbolic repre-
sentation: race (black, white), religion (Protestant, Catholic, Jewish),
and gender (men, women).29 Pollsters routinely ask these categories and
group breakdowns are either reported or can be retrieved from the
polling company or from the Roper Archive.
As a baseline comparison, a breakdown by political party
(Republican, Independent, or Democratic) is also reported.30
Typically, presidents pick Supreme Court nominees from their own
political party, and since Justice Lewis Powells 1971 nomination,
every nominee has been from the presidents political party.31 A mix
of six Democratic and eight Republican justices served during the
Rehnquist Court.
Each justice was classified by his or her symbolic status, if any.
Because there were only a few symbolic nominees in each group (e.g.,
only two women, two blacks, four Catholic, and two Jewish justices),
groups of justices were then combined.32 For example, Justices
OConnor and Ginsburgs votes were combined as womens votes,

while the remaining justices votes were combined as nonsymbolic
(mens) votes. Each demographic groups position was defined as the
response of a majority (or occasionally a plurality) of that group, as
measured in the appropriate public opinion poll. A 5% cutoff was used
to indicate the groups majority or plurality position.33
As an example, consider the 54 Lee v. Weisman (1992) decision,
banning prayers at a high school graduation. Gallup Poll majorities of
all ten demographic groups considered here opposed this ruling, except
for Jews who favored it by a 79-to-21% majority.34 The five-member
majority (Kennedy, Blackmun, Stevens, OConnor, and Souter) were
classified as consistent with Jewish attitudes, but as inconsistent
with the remaining nine groups. The four dissenters (Scalia, Rehnquist,
White, and Thomas) were inconsistent with Jewish attitudes, but
consistent with the remaining nine groups.
As Table 6.1 presents, the Rehnquist Courts decisions about
equally often represented the attitudes of most of these ten groups.
Table 6.1 presents the percentage of decisions that agreed with a major-
ity (or plurality) of each groups attitudes, with results both with and
without denials of certiorari. A clear majority of its decisions agreed
with those of the ten groups considered herewith only a single excep-
tion. The Rehnquist Court significantly less often represented the atti-
tudes of American blacks, compared to American whites attitudes.
Only half (51%) of Rehnquist Court decisions agreed with American
blacks attitudesa figure significantly lower than the figure for
American whites. Except for black Americans, the Rehnquist Court
results are similar to even-handed results reported for the Warren
and Burger Courts (Marshall 1992).
How much do the justices voting patterns reflect symbolic repre-
sentation? Tables 6.2 through 6.5 report the justices votes by gender,
race, religion, and political party. The agreement scores figure indi-
cates how often each group of justices votes agreed with the majority
(or plurality) of the group indicated. For example, in Table 6.2 the two
women justices (OConnor and Ginsburg) agreed with majorities of
men 60% of the time, and with majorities of women 58% of the time.
The twelve male justices agreed with men 56% of the time, and agreed
with women 55% of the time. Tables 6.2 through 6.5 report results for
all full, written decisions for which votes were reported.
Table 6.2 results provide little support for the symbolic representa-
tion model. The two women justices votes slightly more often agreed
with majorities (or pluralities) of American women than did the twelve
male justices votesbut also slightly (4%) better represented the atti-
tudes of American men than did the male justices votes. Individually,
neither OConnor nor Ginsburg was the justice most likely to represent
womens attitudes.35
Table 6.1
Percentage of Rehnquist Court Decisions That Agreed with Group Attitudes
Excluding Including
Cert Denials Cert Denials
By gender
Women 58% 64%
Men 64% 68%
By race
Whites 62% 68%
Blacks 52%* 51%*
By religion
Protestants 62% 68%
Catholics 61% 66%
Jews 55% 57%
By political party
Republicans 62% 66%
Democrats 57% 60%
Independents 59% 64%
Note: Percentages in Table 6.1 are based on a maximum of 88 decisions
excluding denials of certiorari, or 111 decisions including denials of certiorari
(or denials of a stay or denials of a request for an immediate review). Excluded
are instances in which a groups opinion was evenly divided (within the .05-
level margin of error), or in which polls showed inconsistent and conflicting
results for a group. * = significant at .05.
Table 6.2
Agreement Scores, by Gender
Men Women
Women justices (2) 60% 58%
Men justices (12) 56% 55%
Difference 4% 3%
Note: Tables 6.2 through 6.7 are based on a maximum of 781 individual votes
by the justices, and exclude denials of certiorari, where individual justices votes
are not reported. Percentages in these tables indicate the percent of times each
group of justices voted with majorities or pluralities of a demographic group.
Significance levels: * =.05 level; ** = .01 level.
Race is now a top example of symbolic representation. Perhaps

unexpectedly, the Rehnquist Courts twelve white justices votes better
represented attitudes of both black and white Americans than did the
two black justices votes (combined). Indeed, white justices represented

black Americans attitudes significantly better than did the two black
justices. Neither Justice Thomas nor Justice Marshall ranked near the
top (or even among the top half of justices) in representing black
Americans attitudes.36 These results provide no support at all for sym-
bolic representation on race.
Table 6.3
Agreement Scores, by Race
Whites Blacks
Black justices (2) 52% 44%
White justices (12) 57% 54%
Difference 5% 10%*
The symbolic representation thesis was reexamined more narrowly

by comparing symbolic justices by gender or race only against other
justices of their own party. For example, Justice OConnor agreed with
women 62% of the time, compared to a 55% rate for the other seven
Republican justices. Although this finding supports the symbolic repre-
sentation thesis, none of three other comparisons did so. Justice Ginsburg
agreed with women less than did the five other Democratic justices (52%
versus 56%, respectively). Justice Thomas agreed with blacks less often
than the seven other Republican justices (43% versus 53%, respectively).
Even Justice Marshall agreed with blacks less often than did the five
other Democratic justices (48% versus 59%, respectively).
Results in Table 6.4 show only mixed results for the symbolic rep-
resentation model for religion. Both the Rehnquist Courts eight
Protestant justices and six Catholic justices about equally well repre-
sented the attitudes of American Protestants and American Catholics.
Only for American Jews did symbolic representation clearly occur. The
two Jewish justices (Breyer and Ginsburg) significantly better repre-
sented Jewish Americans attitudes and less well represented Catholic
or Protestant Americans attitudes. Even so, the Rehnquist Courts two
Jewish justices only slightly (5%) better represented the attitudes of
American Jews than did Protestant justices. The most striking differ-
ence was that Catholic justices represented the attitudes of Jewish
Americans at such low rates.37
Table 6.5 reports base rate results for political party. The
Rehnquist Courts eight Republican justices significantly (12%) more
often represented the views of Republicans than did the six Democratic
justices. Republican justices votes also represented the views of inde-
pendents slightly (2%) more often than did Democratic justices.
Perhaps surprisingly, the Republican justices votes agreed with

majorities of Democrats 2% more often than did the votes of the
Democratic justices. Again, little support for the symbolic representa-
tion model appears.38
Table 6.4
Agreement Scores, by Religion
Protestant Catholic Jewish
Protestant justices (8) 56% 56% 59%
Catholic justices (4) 57% 56% 46%
Jewish justices (2) 52% 53% 64%
Difference 5% 3% 18%*
Table 6.5
Agreement Scores, by Political Party
Republicans Democrats Independents
Republican justices (8) 54% 56% 56%
Democrat justices (6) 42% 54% 54%
Difference 12%** 2% 2%
** = signif. at .01
Earlier in this chapter, the symbolic representation thesis made two

predictions. First, a symbolic justice should better represent his or her
group than do the other justices. Second, the symbolic justice should
less well represent other groups than he or she represents his or her
own group. How often are these conditions met? For each category of
symbolic representation (women, blacks, Jews, and Catholics), eight
predictions were generated (i.e., two predictions for each group of jus-
tices). However, only three of the eight predicted outcomes actually
occurreda pattern less than expected from random guessing alone. In
short, little support exists for the symbolic representation model.
Symbolic nominees, once confirmed, do not typically represent their
groups better than do the remaining justices. Nor do symbolic justices
typically represent their own group better than they represent other
groups. Only for the two Jewish justices did this model yield strictly
accurate predictions. Justices Ginsburg and Breyer significantly better
represented the views of Jewish Americans than did the non-Jewish jus-
tices, and less well represented the views of Protestants and Catholics
than they did the views of American Jews.
Even political party does not provide a strong example of symbolic
representation. The Rehnquist Courts Republican justices better repre-
sented the views of Republicans than did Democratic justices. Yet

Republican justices also represented the views of Democrats and
Independents about as well as they represented the views of
Republican Americans. This may be surprising, since few studies fail
to note the differences between Republicans and Democrats on the
bench. Yet during the Rehnquist Court, party did not translate into
group representation.
Table 6.6 looks only at instances in which different groups actu-
ally disagreed on Supreme Court decisions. For example, as discussed
earlier in this chapter, a 79-to-21% majority of American Jews sup-
ported the Courts ruling in Lee v. Weisman (1992) banning prayers at
a high school graduation, while majorities of Protestants and Catholics
opposed the ruling. Hence, Lee v. Weisman is an instance in which
majorities of different groups disagreed. Table 6.6 reports only
instances in which at least one symbolic nominee sat on the Court and
in which majorities of blacks and whites disagreed, men and women
disagreed, Jews and non-Jews disagreed, or Republicans and
Democrats disagreed.
Table 6.6
Agreement Scores for Types of Justices, Only for Decisions
in which Groups Disagree on a Ruling
By gender of justice (5 rulings)
Women justices (2) 38% agree with women; 62% agree with men
Men justices (12) 37% agree with women; 63% agree with men
By race of justice (15 rulings)
Black justices (2) 20% agree with blacks; 80% agree with whites
White justices (12) 31% agree with blacks; 69% agree with whites
By religion of justices (9 rulings)
Jewish justices (2) 67% agree with Jews; 33% agree with non-Jews
Non-Jewish justices (12) 40% agree with Jews; 60% agree with non-Jews
By political party of justices
(7 rulings)
Republican justices (8) 58% agree with Republicans; 42% agree with
Democrats
Democrat justices (6) 37% agree with Republicans; 63% agree with
Democrats
Note: Percentages in this table are based on cases in which majorities or pluralities of
social groups disagreed on a decision, in which at least one symbolic justice sat on the
Court, and in which the justices cast recorded votes.
Table 6.6 results offer mixed evidence for symbolic representation,

albeit based on a small number of decisions. No support for symbolic
representation appears by gender. By religion (Jewish versus non-Jewish)
and by political party (Republican versus Democrat), there is stronger

evidence of symbolic representation. For religion and political party,
symbolic nominees indeed better represented their own symbolic
groups than did other justicesat least when groups disagreed. As a
caveat, discussed below, poll majorities of social groups seldom dis-
agree on Supreme Court decisions, and Table 6.6 represents only a
small subset of Rehnquist Court decisions.
Perhaps unexpectedly, the Rehnquist Courts two black justices
agreed with American blacks less often than did white justices on
decisions where majorities of American blacks and whites disagreed.
This occurred because Justice Thomas cast more votes (thirteen) than
Justice Marshall (two votes) in these fifteen disputes. When black
and white Americans disagreed, Justice Marshall both times voted
with blacksthereby continuing his career record of perfect agree-
ment in these disputes with black Americans (Marshall 1993). By
comparison, when black and white Americans disagreed, Justice
Thomas nearly always (twelve of thirteen times) voted with a major-
ity of white Americans.39
Finally, Table 6.7 examines full, written decisions in which a
groups core issue was involved by race, gender, or religion. For
example, black and white justices votes are compared on civil rights
and racial decisions. Again, little support appears for symbolic repre-
sentation. Even on a groups core issues, black and women justices
no more often (indeed, less often) voted with their symbolic group than
do nonsymbolic justices. Only Jewish justices better represented their
symbolic group than did non-Jewish justices, and even here the differ-
ences were not statistically significant.40
Table 6.7
Agreement Scores for Selected Core Cases Only
Agreement Scores Type of Case
By race
Black justices (2) 50% Civil rights, racial cases (6)
White justices (12) 66%
By gender
Women justices (2) 50% Gender rights claims (4)
Men justices (12) 70%
By religion
Jewish justices (2) 70% Religious cases (7)
Non-Jewish justices (12) 58%
Note: Agreement scores are reported for black, women, and Jewish Americans,
respectively. None of these results are significant at the .05-level.
CONCLUSION
At least for the modern Supreme Court, symbolic nominees typically

have not better represented their own groups attitudes than do other
justices. Evidence from the Warren, Burger (Marshall 1993), and
Rehnquist Courts yields no consistent support for the symbolic repre-
sentation model.41 The two women justices votes slightly (3%) more
often agreed with majorities of American women than did men justices
votesbut also slightly (4%) more often agreed with the attitudes of
American men than did men justices votes. White justices better repre-
sented both white and black Americans than did the Rehnquist Courts
two black justices. Support for the symbolic representation model
appears only for the Courts two Jewish justices.
Why was there so little evidence for this model? Several reasons
can be offered. First, large social and demographic groups in America
are themselves quite diverse, and one group may seldom hold different
attitudes than another group. For example, Catholics and non-
Catholics differ only occasionally, even on high-profile issues, such as
abortion, where church officials and church-based interest groups take
clear and different stances (Erikson and Tedin 2001, 2003; Gallup and
Castelli 1987, 1989). On all the decisions considered here, majorities of
American men disagreed with majorities of American women in only
5% of these votes. Likewise, Democrats and Republicans disagreed on
only 12% of these issues.42 Protestants and Catholics disagreed on only
2% of these issues.
When poll majorities of different social groups disagreed, it was
most frequently along racial and religious lines. Majorities of black
Americans disagreed with majorities of white Americans in 21% of
these decisions. Majorities of Jewish Americans disagreed with majori-
ties of Protestant Americans in 23% of these decisions. For Jewish
versus Catholic Americans, the figure was 22%. Generally, however,
majorities of different social groups (such as women and men, or even
blacks and whites) agree much more often than they disagree on issues
that reach the Supreme Court. Over time, a justice may seldom vote
on issues that involve group conflictsas witness Justice White being
the justice who best represented American men, women, blacks,
whites, Republicans, Democrats, Independents, Protestant, Catholics,
and Jews alike.
Second, there is no formal tie between symbolic nominees and their
groups. Even if a justice consistently votes against the positions of the
group that he or she symbolically represents, that group (or its orga-
nized interest groups) cannot discipline or remove that justice from the
Court. Some symbolic nominees are quite aware that they did not rep-
resent their symbolic group, as witness Justice Brennans pro-choice
position on abortion (Leeds 1986: 79).
Third, some symbolic nominees are only very marginal members of
their group. In 1941, President Franklin Roosevelt (a Democrat) nomi-
nated Justice Harlan Fiske Stone (a Republican) to be chief justicea
widely acknowledged symbolic nomination (Abraham 1999: 176;
Mason 1956: 563568). Yet Roosevelt felt reassured that Stone had
voted to uphold New Deal legislation. Justice Clarence Thomas like-
wise symbolically represented black Americans, but had few ties to
organized black and civil rights groups, and many black groups
opposed his nomination (Perry and Abraham 1998). To some
observers, Justice Breyer was less clearly identified as Jewish than was
Ruth Bader Ginsburg (Perry and Abraham 1998).
Fourth, justices come to the Court with group ties, but also with a
commitment to particular methods of deciding cases that are little
related to social groups views on particular issues. For example, Justice
Scalia is often described as opposed to loose readings of congressional
intent and as favoring a theory of original intent (Brisbin 1997; Scalia
1989b). Whether such views will agree with the attitudes of Catholics,
Republicans, whites, or men on a given issue is unclear.
Fifth, the largely negative evidence for symbolic representation on
the Rehnquist, Burger, or Warren Courts does not necessarily mean
that there never was or could be stronger support for this model.
Indeed, symbolic representation may be both justice-specific and
time-specific. Denominational differences were probably more impor-
tant historically than they are today. Symbolic nominations may have
been more closely tied to a justices votes before the New Deal
realignment (Gallup and Castelli, 1987, 1989; Perry 1989, 1991).
Unfortunately, modern polling did not exist before the 1930s, and
that question must remain unanswered. Nor does this evidence deny
that a strong pattern of symbolic representation might occur in the
future - for example, when a disabled, gay or lesbian, Asian
American, Islamic American, or Hispanic American justice joins the
Court (Perry and Abraham 1998).
Finally, polling evidence on symbolic representation is admittedly
limited. For example, Justice Scalias minor fame as a symbolic nomi-
nee was not simply as a Catholic, but more particularly, as an Italian
Catholic. Polls do not reliably yield results so finely graded as Italian
Catholics, nor so fine as between Orthodox and Reform Jews. Many
nuances of symbolic representation cannot be adequately measured by
polling samples of six hundred to one thousand Americans.
Although little support for the symbolic representation model

exists, presidents will probably continue to make symbolic nomina-
tions. Presidents can fulfill past campaign pledges, appeal to voters in
an upcoming election, win a few extra votes in the Senate for a nomi-
nee (Caldeira and Smith 1996; Mansbridge and Tate 1992; Moraski
and Shipan 1999), or attract favorable media attention through a sym-
bolic nomination. Women, disabled Americans, gays and lesbians, or
Hispanics, Asian Americans, blacks, or other ethnic groups will likely
seek to win nominations to the Court (Perry and Abraham 1998). A
symbolic justice might provide emotional or psychological satisfaction
for the group so represented, even if he or she does not represent the
groups policy views any better than do other justices (Mansbridge
1999, 2003). Perhaps symbolic representation even adds to the Courts
reservoir of diffuse support, at least among groups so represented. Even
so, symbolic nominations seldom garner extra votes on Supreme Court
decisions for the groups so represented. Because the evidence for sym-
bolic representation is at best mixed, or even largely negative, this
model is not included in chapter nines model linking public opinion
and Supreme Court decision-making.
CHAPTER SEVEN
Did the Rehnquist Court Influence

Public Opinion?
T his chapter asks whether the Rehnquist Courts decisions them-

selves favorably influenced American public opinion, either in the
short term or over the long term. Several early accounts argued that
Americans hold the Court in high esteem and believe that the justices
have special insights into the Constitution. As a result, many Americans
accept Supreme Court rulings as proper and legitimate.1 Other accounts
argued that Court rulings upholding newly passed legislation gave a
stamp of legitimacy to policy changes during realignment eras (Black
1960; Dahl 1957). In another view, the Courts popularity prevents its
political opponents from overturning key decisions through new legisla-
tion or constitutional amendments (Murphy and Pritchett 1961;
Pritchett 1961). In other accounts, the Courts popularity and the per-
ception that the Court is procedurally fair leads to greater public com-
pliance with even its controversial or unpopular decisions (Breyer
2004:137139; Choper 1980; Tyler and Mitchell 1994; Tyler and
Rasinski 1991).2 In all these accounts, the Supreme Courts prestige is a
strong political resource.
PUBLIC IMPRESSIONS OF THE COURT
Did the Rehnquist Courts rulings, once announced, actually become

more popular? Supportive evidence for this idea would be a consistent
pattern of pre-to-post-decision poll shifts toward the Courts decisions,
or, alternatively, a pattern of post-decision poll shifts toward the
Courts decisions. Such shifts would suggest that public opinion fol-
lowed the Courts decisions.3
If this pattern regularly occurs in the short term, it would mean
that, on balance, some Americans changed their attitudes to favor the
Courts ruling, either directly in response to the decision, or indirectly
123
from the post-decision debate. Positive poll shifts that occurred over
several years would mean that Court decisions find acceptance as they
are incorporated into American life. Positive poll shifts are the most
likely to occur if large numbers of Americans recognize that the Court
made a decision, correctly understand what the Court decided, and are
inclined to follow the Courts leadership.
There are sharp limits to how much Supreme Court decisions can
influence public opinion.4 The media heavily publicize only a handful
of decisions each year. Most decisions receive little or no press cover-
age. Only about a quarter of all Court decisions and very few denials of
certiorari receive any network television coverage (Davis 1994;
Franklin and Kosaki 1995; Larson and Tramont 1993; Slotnick and
Segal 1998; Spill and Oxley 2003). Further, press coverage is not
always favorable toward Court decisions (Clawson and Waltenburg
2003). Some Americans learn about decisions from word-of-mouth,
newspaper editorials, or specialized publications, but this is probably
rare (Franklin and Kosaki 1995; Vermeer 1993).
Factual Knowledge
Only about half of Americans can correctly describe even the Courts
highest-profile decisions. During the Rehnquist Court, pollsters wrote
forty-seven poll questions tapping the publics factual knowledge about
specific rulings, about the Supreme Court itself, or about individual jus-
tices.5 What does a review of these poll questions indicate?
Overall, the level of public knowledge varied widely. On twenty-
one closed-ended poll questions tapping factual knowledge about spe-
cific rulings, the percentage of correct answers ranged from a high of
78% to a low of 27%, with an average of 51% correct. Most of these
twenty-one poll questions tapped high-profile (or landmark) rulings,
such as the Miranda warning, school prayer, school integration, or
flag-burning rulings.6
As a comparison, the percentage of Americans giving correct
answers can be compared to pure random guessing. For instance, on a
closed-ended poll question with only two choices (such as a true-or-
false question), half (50%) of Americans could give a correct answer
simply through random guessing (assuming no one volunteered dont
know). If dont know is offered as an explicit choice, the percent-
age correct through pure random guessing would fall to 33.3%. By
this standard Americans are only modestly well informed about the
Courts landmark decisions. By pure random guessing alone, some
44% of Americans could give a correct answer on these poll ques-
Did the Rehnquist Court Influence Public Opinion? 125
tionsversus 51% actually correct. In other words, the percentage of

Americans correctly explaining landmark rulings is only 7% above
pure random guessing.7
Pollsters also often ask questions tapping the publics knowledge
of the Courts textbook role and procedures, such as the number of
justices. On these closed-ended questions the public averaged 55%
correct answers, compared to a hypothetical average of 33% correct
through pure random guessingan improvement of 22% above pure
random guessing.8
Several poll questions measure factual knowledge of the justices
themselves or of recent nominees. On the average, only 14% of
Americans could correctly identify sitting Rehnquist Court justices. On
open-ended questions, only Justice Sandra Day OConnor was recog-
nized by as many as a fifth of Americans.9 By comparison, an average
of 35% of Americans could correctly identify nominees to the
Rehnquist Court.10 This may not be surprising since the media typically
give much more coverage to nominees than to sitting justices.11
Taking this evidence together, Americans knowledge of the
Rehnquist Court was quite uneven. Only about half of Americans cor-
rectly identified even landmark rulingsa figure barely 7% higher than
pure random guessing. The publics textbook knowledge is much
higher than is public knowledge of major rulings. Nominees are better
recognized than sitting justices. Once confirmed, most justices nearly
fade into anonymity.
The Courts Popularity

Pollsters measure the Courts popularity in several ways. Several series
of poll questions measure the Courts popularity, along with that of
Congress, the president, or top courts in other countries. In these poll
questions, the Supreme Court enjoys greater popularity than do most
American political institutions and most foreign courts. Even so, it is
not clear that the Court enjoys sufficient popularity to move public
opinion greatly on specific decisions.
A well-known measure of the Courts popularity is the diffuse
support index developed by Gibson and Caldeira (1992) during the
Rehnquist Courta six-question index that taps public support for the
Court as an institution (Gibson, Caldeira, and Spence 2003). Most
Americans show fairly positive attitudes toward the Supreme Court on
this index. On a 2001 nationwide survey, the percentage of Americans
expressing support for the Court ranged from a high of 83% (not
doing away with the Supreme Court) to a low of 42% (the Court
favors all groups equally).12 This index is based on approval for the
Court as an institution, not on satisfaction with specific rulings.
Three major pollsters (Harris, NORC, and Gallup) repeatedly mea-
sure public confidence in the Supreme Court.13 These measures are each
based only a single poll question, not on a multiquestion index. Yet
these questions do have the advantage of being frequently asked over
long time periods. The Courts ratings can be also compared to those of
other political institutions.14
On the Harris and NORC/GSS confidence measures, the Rehnquist
Court always ranked higher than Congress or the executive branch, but
not always higher than the White House in the Harris Poll, or satis-
faction with the job a particular president is doing in the Gallup Poll.
The Courts ratings were also as stable or more stable over time than
were other institutions ratings.15 Figure 7.1 depicts the percentage of
Americans indicating giving clearly positive ratings (a great deal of
confidence) toward the Supreme Court since the 1970s, based on the
NORC/GSS ratings, along with ratings for Congress and the executive
branch. In Figure 7.1, the Rehnquist Court enjoyed roughly the same
ratings as did the earlier Burger Courtan average of 33% and 31%
favorable, respectively.16
Figure 7.1
Public confidence in the Supreme Court, the Executive Branch,
and Congress, NORC/GSS positive ratings
The NORC/GSS poll ratings compare the Supreme Court to six

other political or high-profile institutions. The Rehnquist Court aver-
aged better ratings than did most institutions, although it was not the
most popular American institution. On the NORC/GSS question, the
Rehnquist Court averaged a 33% favorable ranking and ranked third
of nine frequently polled institutionsbehind medicine (45%) and the
military (39%), but ahead of major companies (24%), organized reli-
gion (24%), the executive branch (18%), the press (14%), Congress
(13%), and organized labor (10%). Harris and Gallup Polls also indi-
cate that the Rehnquist Court was relatively popular, compared to
other American institutions. In the Harris Poll, the Rehnquist Court
placed second (in a three-way tie) among thirteen frequently polled
institutions.17 In the Gallup Poll, the Rehnquist Court ranked third
among eight frequently polled institutions.18
Although higher than most institutions, the Courts popularity
should not be overestimated. On the average, only a third of Americans
expressed positive attitudes toward the Rehnquist Court in the
NORC/GSS or Harris Polls.19 Only about half of Americans did so in
the Gallup Poll.20 On several other poll questions, about half of
Americans express satisfaction with the Court. For example, Gallup,
Quinnipiac, and Los Angeles Times Polls asked whether the respondent
approved or disapproved of the way the Court is handling its job, with
approval figures ranging from 52% to 62%.21 Gallup, Yankelovich/
Clancy/Schulman, and Quinnipac Polls also queried Americans about
the Supreme Courts political leanings, using the question In general,
do you think the current Supreme Court is too liberal, too conservative,
or just about right? An average of 42% described the Rehnquist Court
as just about right. Several pollsters used different wording varia-
tions, each offering a too liberal option, a too conservative option,
and an option such as generally balanced (ABC/Washington Post) or
just where it should be ideologically (Los Angeles Times). On these
measures, an average of 49% of Americans described the Court as ide-
ologically correct.22 Other poll questions report more negative results.23
An alternative measure taps the visibility and prestige of the indi-
vidual justices. Pollsters asked many different questions about the
Rehnquist Court justices popularity and visibility. For eleven (of
fourteen) justices who sat on the Rehnquist Court, an average of 47%
of Americans offered an opinion on the justice.24 Justices Rehnquist,
OConnor, and Thomas were the most widely recognized justices; an
average of 40% or more Americans offered an opinion on them. At
the other extreme, only about a fifth of Americans offered an opinion
on Justices Stevens, Kennedy, Souter, and Breyer. Most justices
favorable responses outnumbered their unfavorable ratings by a ratio

of 2 or 3 to 1.25
Most Americans would make major changes in the way the Court
operates, according to nationwide polls. The justices lifetime appoint-
ments are especially unpopular. Only a fifth of Americans support life
terms, when asked whether Supreme Court justices should serve for life
or for a specific number of years.26 Two-thirds (68%) of Americans
favor amending the Constitution to require the justices to be reap-
pointed after serving a number of years.27 A majority (59%) favor
terms of ten years or less.28 A majority (54%) favor direct elections of
the justices by the voters.29
Cross-national polls during the mid-1990s compared the U.S.
Supreme Courts popularity to that of top courts in seventeen European
countries (Caldeira and Gibson 1995; Gibson, Caldeira, and Baird
1998) or South Africa (Gibson and Caldeira 2003). Compared with top
courts elsewhere, the U.S. Supreme Court ranked as among the most
visible courts, as above average in public satisfaction with the way the
Court has been working, and as about average in terms of the publics
willingness to protect the Court from basic changes that would
decrease its importance.
Why does the Court remain relatively popular among Americans?
Some scholars point to the Courts secretive deliberations and appear-
ance of not being self-interested as very likely improving its popularity
ratings (Hibbing and Theiss-Morse 2002). Other scholars point to the
Courts image of following legalistic procedures (Baird 2001b), being
procedurally fair in the sense of listening to both sides arguments, and
being equally open to different litigants as helping the Courts popular-
ity, as well as the publics agreement with most Court decisions (Benesh
2006; Thibaut and Walker 1975; Tyler, 1990, 2001; Tyler, Casper,
and Fisher 1989; Tyler and Huo 2002; Tyler and Lind 1988; Tyler and
Mitchell 1994; Tyler and Rasinski 1991).
While the Courts popularity varies a little over time, its poll rat-
ings are much more stable than attitudes toward Congress, the execu-
tive branch, or other high-profile political institutions. Even so, highly
publicized rulings can cause at least temporary shifts in public opinion
toward the Court. Two highly publicized, controversial decisions in
1989 led the Rehnquist Courts approval ratings to drop sharply
(Grosskopf and Mondak 1998).30 Some rulings impact the Courts
approval ratings, perhaps even for the next one or two years (Durr,
Martin, and Wolbrecht 2000; Mondak and Smithey 1997).31 In highly
unusual circumstances, a long string of Court rulings can affect the
Courts approval ratings among specific groups (favorably among black

Americans) for decades (Gibson and Caldeira 1992).
For some controversial decisions, losses in popularity among some
groups may be offset by poll gains among other groupswith no
resulting net change in the Courts approval ratings. During the contro-
versial 2000 Florida presidential vote recount, the Rehnquist Courts
approval ratings dropped sharply among Democrats, but rose among
Republicans (Gibson, Caldeira, and Spence 2003; Jones 2001; Kritzer
2001, 2005; Price and Romantan 2004; Simmons 2001). As a result,
the Courts approval ratings remained quite stable throughout that time
period. Many commentators and scholars questioned whether the
Court would suffer a loss in prestige or legitimacy over that dispute,
but five years later, no negative repercussions were apparent (Kritzer
2005). Whether such offsetting shifts are, or should be, of any great
concern to the justices is unclear.32
Influencing Public Opinion

Admittedly, there are serious practical limits in measuring the Supreme
Courts influence on American public opinion. Pollsters seldom repeat
identically worded questions both before and after Court decisions.
Pollsters often change the question wording, often by including some
variation of the phrase The Supreme Court recently ruled . . . after
the decision is announced. Most polls report only dichotomous
responses (e.g., favor or oppose a ruling), and so do not fully capture
changes the intensity of attitudes. Few polls use split-ballot samples,
counterarguments, or experimental vignettes to measure attitudes.
Overall, there are fewer than three dozen instances of identically
worded pre- or post-decision poll matches on rulings during the
Rehnquist Court.
Because of these problems, some accounts turn to experiments or
special surveys, with mixed results. Some studies (Baas and Thomas
1984; Clawson, Kegler, and Waltenburg 1999; Jaros and Roper 1980)
report that mentioning the Court as the source of a decision does not
significantly increase the percentage of respondents who favor a policy.
In other experiments, mentioning the Court as the policy-maker at least
modestly increases the percentage of respondents who favor a decision
(Gibson 1989; Mondak 1990, 1992, 1994). The Courts ability to
increase public opinion support for specific decisions appears to oper-
ate through its high credibility (Hoekstra 1995, 2003), and may be
greatest among those who are the least interested in, or the least well
informed, about a decision (Hoekstra and Segal 1996; cf. Gibson 1989;
Mondak 1994).33
As a caveat, experimental results differ from real-world results for
three reasons. First, respondents in experiments are fully and recently
informed as to what the Court decided. However, on very few deci-
sions can as many as half of Americans correctly describe what the
Court decided. Most decisions receive so little media coverage that
barely a handful of Americans know of the Courts decision. Second,
most experiments measure opinion changes immediately after partici-
pants are informed about the Courts decision, and do not capture
how quickly the public forgets or changes their views. Third, few
experiments consider other influences on public opinion following a
ruling, such as media coverage or the praise or protests of interest
groups or public officials. Experiments that do consider these effects
report that additional information or criticism of the decision reduces
the Courts impact on public opinion (Mondak 1990, 1994). For all
these reasons, experiments probably exaggerate the Courts actual
influence on public opinion.34
Several studies examine the Courts influence using nationwide or
local surveys. The Courts highest profile rulings (such as its abortion
rulings) may raise public awareness, media attention (Flemming, Bohte,
and Wood 1997), and interest group commentary and criticism so
much that public opinion becomes polarized, rather than simply shift-
ing toward the Courts position (Franklin and Kosaki 1989; Unger
2003).35 The Court may more greatly impact public opinion when it
announces a first landmark ruling, rather than when it announces later
decisions in that same policy area (Johnson and Martin 1998). Some
studies examine polls in communities where a dispute originated.
Because of heightened media attention in these localities, decisions may
much more greatly affect local, rather than nationwide public opinion
(Hoekstra 2003; Hoekstra and Segal 1996).
To summarize: whether the Supreme Court can actually move
public opinion in favor of its decisions is unclear. The Rehnquist Court
was relatively popular, but by no means overwhelmingly so. Most deci-
sions receive little media attention. Certainly, the Rehnquist Court had
the opportunity to move public opinion toward its rulings, but did it, in
fact, do so?
The remainder of this chapter tests two linkage models between
Supreme Court decision-making and American public opinion, and
focuses on specific support for particular rulings, rather than diffuse
support (Easton 1965). Simply put, did the Rehnquist Courts rulings,
once announced, actually become more popular with the public?
THE SHORT-TERM MANIPULATION MODEL
In the short-term manipulation model, once the Court announces a

ruling, public opinion quickly and significantly moves in favor of that
decision. Testing this model involves comparing pre-decision with post-
decision poll support for Court rulings. If poll support for decisions
typically increases, that supports the argument that the Rehnquist
Court manipulated public opinion in its favor.36 By contrast, the long-
term manipulation model predicts that Court rulings, once announced,
grow steadily more popular over long periods of time.
Testing the short-term manipulation model requires identically
worded poll items at least once before and at least once after a deci-
sion is announced. Unfortunately, pollsters seldom repeat identically
worded questions both before and after a decision. Only sixteen such
questions were asked both before and after a decision during the
Rehnquist Court.
As an example, consider the husband notification provision in
Planned Parenthood of Southeastern Pennsylvania v. Casey (1992). The
Rehnquist Court, 54, struck down a Pennsylvania law requiring that
the husband of a woman obtaining an abortion be notified. In the last
pre-decision Gallup Poll, some 31% of Americans favored the ruling. In
the first post-decision Gallup Poll, some 26% of Americans favored the
decisiona poll shift of 5% away from the Courts decision.37
Figure 7.2 depicts the pre- to post-decision poll shifts for these six-
teen rulings. Although there was a wide range of poll shifts, the overall
pattern offers no support for the short-term manipulation model. The
average poll shift was a 2.9%a modest shift away from the Courts
rulings. Examined individually, only four poll shifts were toward the
Courts position; eleven were away from the Courts position; in one
instance the polls did not shift. These results are more negative than
those found during earlier years (Marshall 1989; Page, Shapiro, and
Dempsey 1987).
Did some rulings more greatly influence public opinion than others?
These sixteen decisions differ widely in their circumstances. The first
data column in Table 7.1 reports the average poll shifts broken down by
several possible explanations.38 Briefly, either unanimous or full, written
decisions (each coded 1, otherwise 0) were predicted to influence public
opinion more favorably than non-unanimous decisions, cert denials, or
brief (usually one-paragraph) per curiam opinions.39 Politically liberal
and activist decisions (each coded 1, otherwise 0) were predicted to
influence public opinion more favorably than conservative or restraint
decisions. Decisions in racial cases, criminal justice cases, or sex, gender,
Figure 7.2
Short-term poll shifts during the Rehnquist Court
6% Sternberg v. Carhart (2000)
4% Bd. of Ed., Westside Community Schools v. Mergens (1990)
2% Zelman v. Simmons-Harris (2002)
1% Planned Parenthood v. Casey, 24-hour wait (1992)
0% Planned Parenthood v. Casey, informed consent (1992)
-1% Roemer v. Evans (1996)
-2% Podberesky v. Kirwan, cert denied (1995)
-4% Planned Parenthood v. Casey, one-parent consent (1992)
-4% Webster v. Reproductive Health Services (1989)
-5% Planned Parenthood v. Casey, husband notification (1992)
-5% U.S. Term Limits v. Thornton (1995)
-6% Thomasson v. Perry, cert denied (1996)
-6% U.S. v. Eichman (1990)
-7% Curtis v. School Committee, cert denied (1996)

-10%
Dunn v. White, cert denied (1990)
-10%
Lawrence v. Texas (2003)
-15% -10% -5% 0% 5% 10%
or abortion cases (each coded 1, otherwise 0) were predicted to have a

greater positive impact on the polls than other issues. The greater the
media coverage of the decision, or the longer the pre- to post-decision
time lag, the greater was the predicted (positive) impact on public opin-
ion.40 Finally, the poll percentage initially supporting the decision was
predicted to be negatively related to (positive) poll shifts.
The results are consistent and striking. Under no circumstances did
Rehnquist Court decisions produce (average) positive poll shifts.41 On
the average, its rulings had 55.6% poll support in the last pre-decision
poll, then dropped 2.9% in the first post-decision poll. This is not a
large drop in support, but it is unexpectedly negative.
Table 7.1
Explanations for Short-Term and Long-Term Poll Shifts
Short-Term Long-Term
Average Poll Shift Average Poll Shift
Unanimous decisions 6.3% (4) 2.0% (9)
Non-unanimous decisions 1.8% (12) .8% (9)
Full, written decisions 1.8% (12) .3% (11)
Others 6.3% (4)* 7.0% (1)
Liberal decisions 4.0% (7) .6% (5)
Conservative decisions 2.1% (9) .1% (7)
Activist decisions 2.4% (8) .5% (4)
Restraint decisions 3.5% (8) .8% (8)
Racial decisions 2.0% (1) 7.0% (1)
Nonracial decisions 3.0% (15) 1.0% (11)
Criminal rights decisions 10.0% (1) 7.0% (1)
Noncriminal rights decisions 2.5% (15) 1.0% (11)
Sexual and abortion decisions 3.0% (10) 1.3% (7)
Nonsexual and abortion decisions 2.8% (6) 1.0% (5)
Pearson product-moment
correlations
Year of decision .09 (16) .31 (12)
Media coverage .37 (14) .03 (12)
Time lapse (in months) .24 (16) .24 (12)
Initial poll support .03 (16) .13 (12)
Number of dissents .27 (12) .45 (11)
Note: The four unanimous short-term poll shifts are all cert denials.
Numbers in parentheses are the numbers of decisions involved. * = stat. signif.
at .1.
These negative poll shifts contrast to the mixed pattern of poll

shifts for earlier Courts since the mid-1930s (Marshall 1989). For ear-
lier Courts, an activist ruling (overturning a law or policy) produced a
predicted +7.4% poll shift within three years, and a liberal ruling pro-
duced a predicted +5.7% poll shift within that time frame. Under these
circumstances the Rehnquist Courts predicted poll shift would be a
1.6% or a 2.5% poll shift, respectively.42 In not a single regression
did the Rehnquist Court actually produce positive poll shifts. The pre-
dicted poll shifts range from near zero (1.6% for an activist, full writ-
ten decision, three years later) to strongly negative shifts (6.8% for a
full, written, criminal justice decision, three years later).
THE LONG-TERM MANIPULATION MODEL
So far, no evidence appears that Rehnquist Court decisions favorably

influenced American public opinion. But did the Rehnquist Court influ-
ence public opinion after the decision has been announced? Pollsters
often ask questions only after a decision has been handed down and
public interest aroused, or else they may ask the same pre-decision poll
question several times after a ruling is announced. Examining post-deci-
sion poll shifts can answer the question of whether the Court can shape
public opinion in a decisions aftermath.
As an example, consider the one-parent notification requirement,
upheld 63 in Ohio v. Akron Center for Reproductive Health (1990).
The first post-decision ABC/Washington Post Poll reported that a 79%-
to-20% majority favored the decision. Another poll, two years later,
reported that a 80%-to-18% majority favored the decisiona +1%
post-decision shift toward the Courts position.43
Clearly, there are limitations to the data. Only twelve instances of
an identically worded, post-decision poll question could be identified
during the Rehnquist Court. At the conceptual level, it is unclear
exactly whose influence is being measured by long-term, post-decision
poll shifts. The justices typically do not follow up their decisions with
public statements, news releases, op-ed articles, or talk show appear-
ances. The news media seldom cover a decision for more than a few
days. Interest group leaders, public officials, and attorneys dominate
the post-decision debate. Over long time periods, polls may change due
to social trends or demographic changes, such as cohort replacement or
immigration (Brewer 2003; Erikson and Tedin 2003: 8399, 112141).
Given all these caveats, what do long term, post-decision trends
show? First, Figure 7.3 depicts a wide variety of post-decision poll shifts.
Post-decision poll shifts ranged widely, but on the average, were slightly
negativeaveraging .3%, or a negative three-tenths of 1%. Individually,
six poll shifts were toward the Courts position, five shifts were away
from the Courts position, and in one instance, no poll shift occurred.
As the second column in Table 7.1 indicates, once announced,
public opinion support for Rehnquist Court decisions did not grow
steadily.44 Rather, on the average, post-decision poll support dropped
very slightly. When examining a regression model, nearly all the pre-
dicted post-decision poll shifts were zero or slightly negative. For exam-
ple, a unanimous decision with a three-year lag had a predicted poll
shift of 2.8%. A liberal decision with a three-year lag had a predicted
poll shift of 1.0%.45 For a conservative decision with a three-year lag,
the estimate was a +.3% shift.
Figure 7.3
Long-term poll shifts during the Rehnquist Court
7% McClesky v. Zant (1991)

5% Lawrence v. Texas (2003)
3% Lee v. Weissman (1992)
3% Washington v. Glucksberg (1997)
3% Benten v. Kessler (1992)
1% Ohio v. Akron Center for Reproductive Health (1990)

0% Roemer v. Evans (1996)
-2% Zelman v. Simmons-Harris (2002)
-4% Johnson v. Transportation Agency (1987)
-6% Bd. of Ed., Westside Community Schools v. Mergens (1990)

-7% Thomasson v. Perry, cert denied (1996)
-7% Bd. of Directors, Rotary International v. Rotary Club of

Duarte (1987)
-10% -5% 0% 5% 10%
In short, the Rehnquist Courts decisions, once announced, did not

steadily grow in popularity. Looking at both the short-term and long-
term poll shifts, on the average, poll support typically dropped slightly
once a decision was announced, then fell off very slightly thereafter.
There is no evidence that Rehnquist Court decisions favorably moved
American public opinion.
As a further test, nine very long-term poll trends were also exam-
ined, most of which involve decisions prior to the Rehnquist Court.
Pollsters seldom ask identically worded poll questions over long peri-
ods of time, and only nine very long-term poll trends can be identified.
In these nine instances, identically worded, post-decision poll questions
could be identified over five years or longer, some from the Rehnquist
Court era and others from earlier Courts.46
Over very long periods of time, American public opinion barely
shifts at all after a Supreme Court decision - with only a third of 1%
(annual) poll shift toward the Courts decision over an average of eigh-
teen years. Even for poll trends over two decades, the over-time shift is
less than 1% annually. Only over very long periods of time has the
Court won over public acceptance for its decisionsapparently, as
decisions are incorporated into American life over decades, not simply
over a few years.
Finally, if Supreme Court decisions have little influence on

Americans issue attitudes, do its decisions perhaps affect the Courts
own approval ratings? To test this possibility, the Rehnquist Courts
approval ratings in the NORC/GSS and the Gallup Poll were entered
as a dependent variable. For neither of these two indicators of
approval was the Rehnquist Courts recent level of agreement with
public opinion a statistically significant predictor. Nor were the poll
margins of its decisions; nor was the appointment of new justices
(entered as dummy variables). The strongest predictor of the Courts
approval ratings were the presidencys approval ratings.47 At least
during the Rehnquist Court, then, the Courts own approval ratings
did not closely follow the popularity of its recent decisions, but rather
rose and fell with the presidencys approval ratings. Once again, these
results suggest that there was a disconnect between the Rehnquist
Courts actions and American public opinion.48 The modern Courts
relatively consistent agreement with public opinion may help its
approval ratings over the long term (Chemerinsky 1999b;
Stoutenborough and Haider-Markel 2006). Yet there is no evidence of
a year-to-year correlation between the popularity of the Rehnquist
Courts output and its own approval ratings.
CONCLUSION
The evidence is clear and consistent. The Rehnquist Court did not legit-
imize its rulings in the sense of winning over public opinion. Previous
Courts since the mid-1930s demonstrated little ability to move public
opinion. The Rehnquist Court demonstrated even less ability to move
public opinion behind its decisions. Average poll shifts for the
Rehnquist Courts rulings were neutral to slightly negative, either con-
sidering pre- to post-decision poll shifts or post-decision poll shifts.
These results may be unexpected. The Rehnquist Courts approval
ratings were stable and similar to those of past Courts. There was no
sharp or steady decline in the Rehnquist Courts approval ratings.
Further, special surveys and experiments show that the Supreme Court
has at least the potential to move public opinion toward its rulings. Yet
by the available, albeit limited polling evidence, the Rehnquist Courts
rulings actually lost a slight amount of public support. Why this
occurred is puzzling, and while no definitive answer can be offered, two
suggestions may be put forth.
First, it is unlikely that these results occurred simply because these
poll questions were mostly on criminal rights, social, or racial issues
rather than economic or federalism issues. Rulings on racial, criminal,

or social issues attract the most media attention and most commonly
come to the publics attention. The Courts economic or federalism rul-
ings attract much less media attention and are less likely to be well
understood by the public.49 It seems unlikely that the Rehnquist Court
would lose popular support for high-profile decisions, but gain support
for its low-profile decisions.
Second, the modern Courts decision-making style may contribute
to its waning influence on American attitudes. Rehnquist Court deci-
sions were often characterized by long, fragmented opinions, often
based on specific factual issues, and making only incremental changes
in past rulings. Such rulings seem unlikely to sway public opinion
greatly. The modern norms of picking low-profile, career federal
appeals judges for the Court; the justices increasingly elderly profile;
the growing number of lengthy, finely tuned opinions; and the cacoph-
ony of concurring and dissenting opinions may all contribute to the
Courts inability to influence public opinion (Rosen 2003). More
experimental studies would be welcome.
At a normative level, the Courts supporters may be disheart-
ened, while its critics may take hope from these results. At least since
the mid-1930s, the modern Supreme Court never had much success
in winning public opinion support for its rulings. The Rehnquist
Court actually slipped into negative numbers. Neither the short-term
and long-term manipulation model is included in chapter nines
empirical model.
CHAPTER EIGHT

Test of Time
T he test-of-time model is the final linkage model between American

public opinion and Supreme Court decision-making. This model is
quite different from the other models. In the test-of-time model, the
degree of public opinion support a decision receives is closely linked to
how long that decision endures, and popular decisions will endure
longer than unpopular decisions.
At one time, legal scholars described the Supreme Courts decisions
as a final word in American politics (Paschal 1991), and as a policy-
maker whose rulings were seldom successfully challenged (Agresto
1984; Bork 1990). Justice Jacksons well-turned phrase in Brown v.
Allen (1953, at 540) expressed this view: We are not final because we
are infallible, but we are infallible only because we are final.
Yet as judicial scholars have long recognized (Murphy 1962; Note
1958; Pritchett 1961; Stumpf 1965; Warren 1935), not all Supreme
Court decisions endure indefinitely. Some decisions are overturned
quickly. Other decisions are gradually modified over time, or suddenly
abandoned years later. Supreme Court decisions are not a final word,
but only part of a continuing policy debate (Bickel 1962, 1970;
Klarman 2001; Paschal 1991; Rosen 2006).
Not all Supreme Court decisions withstand the test of time. Perhaps
most easily classified as failing the test of time are those few decisions
overturned by a constitutional amendment.1 Yet constitutional amend-
ments overturning a Supreme Court decision are difficult to enact, and
are the least common way in which decisions fail the test of time.
Much more frequently, the Supreme Court itself overrules an ear-
lier decision, the doctrine of stare decisis notwithstanding. By one com-
pilation (Epstein et al., 2003: 203206; see also Spriggs and Hansford
2001), the Rehnquist Court averaged two to three decisions each term
overturning an earlier decisiona record similar to that of the Warren
and Burger Courts.2
139
Congress and the president also often rewrite a statute to overturn

a Supreme Court decision (Baum 1998a: 246248; Eskridge 1991a;
Hausegger and Baum 1998, 1999; Ignagni, Meernik, and King 1998;
Meernik and Ignagni 1997; Paschal 1991; Solomine and Walker 1992;
Spiller and Tiller 1993).3 Some decisions are quickly overturned, while
others are overturned years or even decades later when Congress
rewrites a section of federal law. Congress sometimes overturns several
decisions in a single bill. The Civil Rights Act of 1991 holds the recent
record for overturning the most (nine) Supreme Court decisions in a
single bill (Baum 1998a:259).4 Most overturnings involve statutory
construction decisions, although at times Congress successfully modi-
fies the Courts constitutional decisions.5
Solomine and Walker (1992: 445) reported fifty congressional
statutory changes overruling a total of fifty-six Supreme Court deci-
sions within a twenty-year periodor about two or three congressional
overrides a year, on the average.6 Meernik and Ignagni (1997) reported
forty-one instances in which Congress overturned a Supreme Court
decision over thirty-six years, although this number only includes
instances in which the Court earlier declared a federal law, state law, or
executive order unconstitutional. By yet another count, Congress annu-
ally overturns or modifies an average of about six Supreme Court deci-
sions (Eskridge 1991a).
Constitutional amendments, explicit overrulings of a precedent,
and congressional overrides are the three most obvious ways in
which Supreme Court decisions fail the test of time, and no longer
have any practical effect. Nonetheless, there are two, and arguably
three other ways in which a decision can fail the test of time. First,
some decisions slowly erode over time until the original decisions
logic no longer has any practical effect. Second, some rulings uphold
a specific executive or agency decision, but the executive or agency
later abandons that decision. Third, some decisions are simply
widely ignored or evaded. Deciding whether a ruling has withstood
the test of time by these three standards is not always simple. While
no published lists of such instances appear to exist, examples of each
type stand out.
Many decisions slowly erode over time, as the justices consider
later disputes and qualify an earlier ruling. The justices themselves fre-
quently disagree as to whether an earlier ruling has ceased to be an
important precedent. In a well-known example, Justices OConnor and
Blackmun and Chief Justice Rehnquist all disagreed as to how fully the
undue burden standard in Casey v. Planned Parenthood of
Southeastern Pennsylvania (1992) overturned the Roe v. Wade (1973)
Public Opinion and the Test of Time 141
trimester standard for abortions. At least in the justices own views, this
situation is not uncommon. During the Rehnquist Court, on the aver-
age, about once a term a justice (usually writing in dissent) explicitly
accused other justices (usually those in the majority) of overruling a
decision sub silentiothat is, abandoning one or more precedents with-
out explicitly saying so. Many law journal articles argue that the
Supreme Court, as well as lower federal or state courts, overrules prece-
dents sub silentio.7 The exact point at which a precedent has become so
qualified that it should be considered as failing the test of time is diffi-
cult to determine, and an eitheror classification may inevitably seem
subjective. However, if the standard for failing the test of time is that a
decision loses its practical effect, then the erosion of precedent and
overrulings sub silentio must be considered.
Some decisions lose their practical effect when a president,
Congress, or an agency abandons a policy previously approved by the
Court (Posner 2005: 42). Typically, these instances involve judicial
restraintin that the Court upheld an executive, congressional, or an
agency decision, but, even so, the executive, Congress, or agency later
changed the policy (for whatever reasons), thereby leaving the Supreme
Court decision without any practical effect. Political scientists have not
offered any counts of these instances, but they may well be numerous.
In the 111 Rehnquist Court decisions examined below, examples
include the so-called gag rule on abortion counseling; the ban on the
abortion pill RU-486; and Congresss decision not to renew the
Independent Counsel Act.
To these five ways in which a decision can fail the test of time
might be added a sixth: the decision is simply widely ignored or evaded.
These instances are the most difficult to quantify.8 The Warren Courts
school prayer rulings were widely ignored. Congresss continued use of
the legislative veto is another example, despite the Courts disapproval
in Immigration and Naturalization Service v. Chadha (1983) (Fisher
1993). Historically, a well-known example was President Andrew
Jacksons open disdain for the Marshall Courts Worcester v. Georgia
(1832) ruling in a dispute between the Cherokee Indians and the state
of Georgia (Abraham 1998:370).9
Many Supreme Court decisions eventually fail the test of time.
Some decisions do so very quickly. For example, in Payne v. Tennessee
(1991), a 63 majority explicitly overturned a 54 decision in Booth v.
Maryland (1987) and a 54 decision in South Carolina v. Gathers
(1989), both barring a victim impact statement at a capital sentencing
hearing. Other decisions fail the test of time only after a much longer
time lag. The longest time lag for the Court overturning one of its own
precedents in Chief Justice Rehnquists list in Payne v. Tennessee

(1991) was 107 years.10
In some instances it is difficult to specify exactly when a decision
failed the test of time. For example, in Benten v. Kessler (1992), the
Court, 72, denied an application to vacate a stay, and upheld the
Food and Drug Administrations 1989 ban on the use or importation
of RU-486 (or mifepristone), a pill used to induce abortions early in a
pregnancy. In 1993, the newly elected Clinton administration approved
clinical trials; in fall 2000 the FDA approved the drug for commercial
use. Clearly, the Clinton administration moved beyond Benten v.
Kessler, although one might view either 1993 or 2000 as the effective
date. Further, Congress often limits the scope of a Supreme Court deci-
sion without openly overruling it. Among these decisions, Sable
Communications of California, Inc. v. Federal Communications
Commission (1989) provides such an example. Congress responded to
the Courts decision overturning a ban on indecent, although not
obscene phone services, by denying access to indecent messages unless a
person specifically subscribed to the service (Congressional Quarterly
Almanac 1989: 382; Paschal 1992:225).11
Do popular Supreme Court decisions endure longer than unpopu-
lar decisions? The evidence in this chapter is that they do. The extent of
public opinion support that a ruling initially receives is significantly
related to whether a ruling survives the test of time. Even within a few
years after a decision is announced, unpopular rulings are more often
and more quickly overturned by the Supreme Court itself, or by presi-
dents, Congress, agencies, or state-level actors, compared to popular
rulings. Over several decades, the differences are dramatic. An unpopu-
lar decision has an expected half-life about six years shorter than a
popular decision. While some predictors of whether a ruling will sur-
vive the test of time are time-bound, both public opinion and the pas-
sage of time itself significantly predict whether a ruling will withstand
the test of time, both before and during the Rehnquist Court.
SURVIVING THE TEST OF TIME
Whether or not Supreme Court decisions prevail over time may depend
on several variablesamong them, the Courts own behavior in
announcing the decision, the type of issue involved, the level of public
opinion support for a decision, and the passage of time itself.12 Each of
these four explanations is briefly reviewed below, and then tested, first
with bivariate data, and then in a statistical model.
Court Decision-Making Patterns
Several explanations for whether Supreme Court decisions prevail over

time are based on how the justices themselves handle the decision. First,
was the decision unanimous? Until the early 1940s, most decisions
were announced as unanimous; even if the justices disagreed in confer-
ence, they rarely wrote dissenting or concurring opinions (Epstein,
Segal, and Spaeth 2001; Haynie 1992; Walker, Epstein, and Dixon
1988).13 In later years, the justices occasionally sought unanimity on
highly controversial decisions, such as landmark desegregation deci-
sions (Brams and Muzzio 1977; Hutchinson 1979; Peterson 1981;
Rohde 1972; Ulmer 1971; Woodward and Armstrong 1979). However,
in recent decades unanimity has not been a Court norm.
Unanimous and near-unanimous decisions are predicted to prevail
more frequently for three reasons (Ignagni and Meernik 1994).14 First,
unanimity gives a stronger signal about the Courts determination to
prevail on an issue and about the decisions legal grounding. Second,
unanimity or near-unanimity on a decision means that one or two new
Supreme Court appointments, or a single justice switching his or her
vote, will not lead to the decision being reversed (Spriggs and Hansford
2001). Third, unanimity may simply mean that a broad consensus
exists on the decision.
Whether a decision prevails over time may also depend on the deci-
sions ideology. From the 1930s to the mid-1980s, liberal decisions pre-
vailed more often than did conservative rulings (Eskridge 1991;
Marshall 1989; Solomine and Walker 1992). That finding may now be
time-bound, since New Deal, Fair Deal, and Great Society-era
appointees (most of them liberals by conventional political standards)
were then replacing older, more conservative justices. As well, congres-
sional liberals often succeeded in blocking challenges to rulings they
favored (Adamany and Grossman 1983; Eskridge 1991; Schmidhauser
and Berg 1972: 170-184). However, during the 1980s, 1990s, and
early 2000s, conservatives and Republicans more often held control of
the White House and Congress.
Another explanation is whether a decision demonstrated judicial
activism or restraint, and whether the dispute was from the federal,
versus the state or local level. The Courts judicial restraint decisions
might seem more likely to prevail than its activist decisions in federal-
level disputes (Meernik and Ignagni 1997). In judicial restraint deci-
sions, the Court only upholds a congressional or executive action,
thereby reducing the chances that critics could successfully put together
a coalition to overturn the decision. In disputes over state and local
laws and policies, however, past research indicates that judicial activist
decisions more often prevail (Marshall 1989).
Another explanation is whether or not the decision is a full, writ-
ten opinion. As several justices argue, by handing down a full, written
opinion, the Court speaks with greater finality than in a denial of cer-
tiorari or a denial of a stay. Full, written and per curiam opinions are
predicted to prevail more often than denials of certiorari or denials of
a stay.
Two other explanations are tested. In past research (Ignagni and
Meernik 1994, Meernik and Ignagni 1995; cf., Lindquist and Yalof
2001; Maveety, Turner, and Way 2004; Meernik and Ignagni 1997;
Solomine and Walker 1992), the extent of interest group interest in a
ruling is linked to the likelihood that Congress will react to or overturn
a decision. The total number of interest group amici filed, and the
number of interest group amici filed for either the winning or the losing
side, and the number of losing minus winning amici briefs filed were
compared to whether or not the decision prevailed. The greater the
interest group involvement, the less likely is a decision predicted to pre-
vail. As before, only per curiam and full, written opinions are included
in this test.
Finally, Supreme Court decisions are predicted to fail the test of
time more frequently if the federal government (represented by either
the solicitor general or a federal agency) lost, either as a direct party or
in an amicus brief (Eskridge 1991a, Meernik and Ignagni 1997). Only
per curiam and full, written opinions are examined for this test.
Types of Issues Involved

Whether a decision prevails over time may also depend on the type of
controversy involved. The Courts economic decisions are predicted to
prevail less often than the averagebecause economic interest groups
are more stable over time, because the legislation involved is more
easily subject to congressional or executive revision, because economic
issues are less symbolic and divisive and less often involve constitu-
tional issues, and because over-time economic changes may erode the
original decisions logic.15
By contrast, fundamental freedoms decisions, involving Bill of
Rights or Fourteenth Amendment claims, are predicted to prevail more
often than the average because these issues typically involve symbolic,
highly divisive constitutional issues that are less easily amenable to
compromises in Congress or the executive branch (Epstein, Knight, and
Martin 2001; Epstein, Segal, and Victor 2002). Rulings that involve
core First Amendment claims (speech, religion, press, and assembly)

are predicted to prevail more often than other rulings largely because
such decisions are often grounded in constitutional law, making them
more difficult to overturn.16
Public Opinion Support

Judicial scholars have long argued that unpopular decisions less often
survive the test of time. As Alexander Bickel (1970: 181) wrote, The
Supreme Courts law could not in our system prevailnot merely in
the very long run, but within the decadeif it ran counter to deeply felt
popular needs or convictions.17 Decisions that initially enjoy wide-
spread public support are predicted to prevail more often than deci-
sions that receive little public support (Marshall 1989; Ignagni and
Meernik, 1994; Meernik and Ignagni 1995, 1997). Elected officials and
interest groups may avoid challenging popular decisions, and, even if
they try, may less often succeed in overturning a popular decision than
an unpopular one. Under ideal circumstances, poll questions would be
available repeatedly on Supreme Court rulings, both before and after
the rulings are announced. Since this is seldom the case, attitudes
toward a ruling are measured by the decisions initial popularity (or
poll margin) when the decision was handed down.18
To be sure, few of the Rehnquist Court rulings considered here
may rise to the level of crisis-times casesthose for which public
opinion is especially one-sided and intensely focused, such as during
World War II or the subsequent Cold War era. However, several rul-
ings tapped the publics most important problems, as measured by
the often-repeated Gallup Poll question, while other rulings were less
visible. No hypothesis is offered as to whether highly visible rulings will
prevail more often than other rulings.
The Passage of Time

Finally, time itself is predicted to take its toll on Supreme Court deci-
sions. Over time, new justices join the Court, creating a different Court
majority, particularly on ruling decided by a narrow 54 or 63 major-
ity. During the Rehnquist Court, both Republicans and Democrats cap-
tured the White House, leading to a mix of new justices by political
party and ideology. Economic, social, or technological changes, over
time, mean that Congress or the justices themselves may see a past deci-
sion as unworkable. At least a year or two may be needed for members
of Congress, agency heads, interest groups, or state legislatures to pass
new legislation, publish and approve a new agency rule, or pass a con-
stitutional amendment overturning a Supreme Court decision.19 Here,
all 111 Rehnquist Court decisions are counted with the time lag (in
years) from the original decision through the end of the Rehnquist
Court (the 2004/2005 term).
Classifying the Rulings

Deciding whether (or when) a Supreme Court decision has been over-
turned is not always simple. The question here is whether the decision
any longer has a practical effect as of the end of the Rehnquist Courts
final (2004/2005) term by the standards outlined earlier in this chapter.
As in an earlier study (Marshall 1989), classifying a decision as either
prevailing or overturned hinged, first, on the specific poll item used,
and, second, on subsequent Court rulings and policy decisions.
If the original poll item mentioned a particular individual, then
the decision was classified as having prevailed if the decision pre-
vailed for that specific individual. For example, a 1998 CBS News
Poll question mentioned convicted murderer Karla Faye Tucker, who
was then appealing a death penalty sentence.20 The Supreme Court
denied all her appeals; Tucker was then executed; the decision was
classified as prevailing.
Most matching poll items, however, were worded more generally;
for these poll-to-decision matches, the Courts decision was classified as
prevailing if the rulings core logic prevailed over time, even if the
policy changed over time. A ruling that was later somewhat modified
somewhat was counted as prevailing. For example, Kansas v.
Hendricks (1997) was classified as prevailing, even though Seling v.
Young (2001, at 1315) offers some possible remedies for sexually
violent predators in state court.
A ruling matched with a generally worded poll item was classified
as failing the test of time only if later Supreme Court rulings, congres-
sional or White House or federal agency decisions, or specific state or
local actions caused the ruling to lose its practical effect, according to
the standards outlined earlier in this chapter. All other rulings are clas-
sified as having prevailed. This classification schema is broader than
several other lists, cited above, that rely only upon Supreme Court
reversals of earlier precedents, only on constitutional amendments, or
only on congressional statutory reversals.
By these standards, six Rehnquist Court decisions failed the test of
time. First, Booth v. Maryland (1989) and South Carolina v. Gathers
(1987), barring victim impact statements during capital sentencing
hearings, were overturned by Payne v. Tennessee (1991). Second,

Benten v. Kessler (1992), upholding FDA refusal of RU-486 (abortion
pills), became irrelevant when the Clinton administration allowed clin-
ical trials in 1993 and approved public use of the pill in 2000. Third,
Rust v. Sullivan (1991), upholding the gag rule on abortion counsel-
ing in federally funded health clinics, became irrelevant when the rule
was rescinded by newly elected President Clinton in 1993. Fourth,
Morrison v. Olson (1988), upholding provisions of the Independent
Counsel Act, became irrelevant when the act expired on June 30,
1999, and was not renewed by Congress. Fifth, Stanford v. Kentucky
and Wilkins v. Missouri (1989), allowing the death penalty for mur-
derers under the age of eighteen, were overturned by Roper v.
Simmons (2005). Sixth, Penry v. Lynaugh (1989), allowing the death
penalty for mentally retarded murderers, was overturned by Atkins v.
Virginia (2002).21
These results suggest two conclusions. First, at least to date, most,
but not all, Rehnquist Court rulings prevailed over time. By the broad
test used here, six decisions have already failed the test of time. Several
other decisions were subsequently modified, albeit not yet sufficiently
so as to be classified as failing the test of time. Second, half the deci-
sions that failed the test of time did so indirectly. That is, the
Rehnquist Court exercised judicial restraint and upheld a federal
agency rule or a congressional statutebut a newly elected president
later rescinded the agency rule, or else Congress and the White House
allowed the act to expire.
Should these instances of judicial restraint be counted as failing the
test of time? Admittedly, the Court itself did not overturn an earlier
precedent. Nor did a constitutional amendment or a congressional act
overturn the decision. Even so, the ruling no longer has any practical
effect, and so is classified as failing the test of time. While political sci-
entists pay little attention to these instances, and no list of such deci-
sions is available, they are quite common (Eskridge 1991; Paschal
1991: 206-208).
DATA ANALYSIS
Table 8.1 lists the percent of Rehnquist Court rulings that prevailed
(through its final 2004/2005 term), broken down by the hypotheses
above. Although few results in Table 8.1 reach statistical significance,
most results are in the expected direction. Unanimous decisions, core
First Amendment decisions, recent decisions, and decisions consistent
with the polls all prevailed more often than decisions that were not.
Several early Rehnquist Court decisions have already failed the test of
time, while more recent decisions, as yet, show no such pattern.
Table 8.1
Percentage of Rehnquist Court Decisions Classified as Prevailing
% Prevailed Number of Decisions
Overall results 95% 111
By unanimity
Unanimous decisions 100% 17
Nonunanimous decisions 91% 67
By ideology
Liberal decisions 95% 42
Unclear ideology 67% 3
Conservative decisions 96% 66
By federal-level judicial activism
Activist decisions 100% 13
Restraint decisions 88% 25
By state-level judicial activism
Activist decisions 96% 28
Restraint decisions 96% 46
By outcome, per federal
government position
Federal government won 98% 40
Federal government not involved 87% 31
Federal government lost 94% 17
By type of decision
Full, written and per curiam
decisions 93% 88
All others 100% 23
By economic cases
Economic cases 100% 17
Non-economic cases 94% 94
By fundamental freedoms cases
Fundamental freedoms cases 95% 79
Non-fundamental freedoms cases 94% 32
By First Amendment cases
Core First Amendment claim raised 100% 30
No such claim raised 93% 81
By public opinion*
Consistent with public opinion 99% 68
Unclear decisions 100% 4
Inconsistent with public opinion 87% 39
Table 8.1 (Continued)

Percentage of Rehnquist Court Decisions Classified as Prevailing
% Prevailed Number of Decisions
By public opinion, combined*
Consistent and unclear decisions,
combined 99% 72
Inconsistent with public opinion 87% 39
By most important problem
Top-ranked issue 100% 21
Not a top-ranked issue 93% 90
By time of decision**
1986-1993 85% 41
1994 or later 100% 70
Point-biserial correlations
Number of dissenting votes .12
Date (year) of decision** .29
Initial poll margin .12
Total number of interest group amici .08
Number of interest group amici,
losing side .04
Number of interest group amici,
winning side .09
Net number of interest group amici,
winning minus losing side .10
Note: For unanimity, decisions exclude denials of certiorari and denials of
stays. For federal government and interest group variables, the analysis is based
only on full, written and per curiam decisions. ** = stat. signif. at .01; * = stat.
signif. at .05.
For several variables, no clear relationship appeared, for example,

the total number of interest group amici involved, the number of amici
on either the winning or the losing side, or the net number of winning
versus losing amici. Unexpectedly, full, written and per curiam deci-
sions more often failed the test of time.
In part, the Rehnquist Courts record differed from that of earlier
Courts since the 1930s. The Rehnquist Courts ideologically conserva-
tive decisions as often prevailed as its liberal decisionsa record differ-
ent than that of previous Courts (Marshall 1989; cf. Eskridge 1991a).
Perhaps unexpectedly, federal-level activist decisions and economic
decisions prevailed at above-average rates.
As these results suggest, the types of decisions that survive the test
of time are probably time-bound and vary from era to era. Dramatic
events; shifts in long-term public opinion; the extent of fragmented gov-

ernment; and the balance between conservatives and liberals on the
Court, at the White House, in the agencies, and in Congress all affect
which Supreme Court decisions survive the test of time.
Supreme Court decisions do not endure indefinitely. For the
Rehnquist Court, at least to date, only two variables significantly pre-
dicted which Rehnquist Court decisions survived the test of time: first,
whether the decision itself was originally inconsistent with American
public opinion, and, second, the time elapsed since the ruling. Older,
inconsistent decisions are the most likely to fail the test of time. Other
predictors that were significant for earlier Courts (such as a decisions
ideology, whether the decision was a crisis-times case, or whether the
Court was unanimous) are now time-bound, and no longer signifi-
cantly predict whether or not a decision will prevail.
Most Rehnquist Court decisions that failed the test of time did so
relatively quicklywithin an average of six years of the decision being
announced.22 This six-year average time lag between the Courts origi-
nal decision and the overturning action is longer than the two-year
average time lag for congressional overturnings alone (Eskridge 1991a;
Hettinger and Zorn 2005),23 but shorter than the twenty-four-year
average historical time lag for the Supreme Courts overturning of its
own precedents (Blaustein and Field 1958; Payne v. Tennessee (1991),
at 828).
Supreme Court decisions fail the test of time at a steadily higher
rate as the years pass. Unpopular decisions fail more quickly than pop-
ular decisions. Consider the expected half-life of a decisionthat is, the
time at which exactly half of the Supreme Courts decisions would be
predicted to fail the test of time. For an unpopular decision during the
Rehnquist Court, the expected half-life is seventeen years. That is, after
seventeen years, half of the Courts unpopular decisions are still pre-
dicted to survive the test of time. After a lag of twenty years only 25%
of unpopular decisions would survive the test of time.24
By contrast, the expected half-life of a popular decision during the
Rehnquist Court is twenty-four years, about seven years longer than for
unpopular decisions. After a lag of twenty-seven years, only 25% of
popular decisions would still survive the test of time. Stated otherwise,
when the Rehnquist Court handed down a decision consistent with
public opinion, it added about seven years to that decisions expected
half-life. Figure 8.1 depicts the expected half-life of popular, versus
unpopular decisions for the Rehnquist Court.
This simple two-variable model was also run for 146 poll-matched
Supreme Court decisions from 1935 to 1986 (Marshall 1989). Both
Figure 8.1
Probability that a Rehnquist Cort decision will survive, by whether
the decision is consistent with public opinion or not
Years
variables again strongly predicted how well the Courts decisions

would meet the test of time (through the 1986 cutoff date), although
the predicted half-life for both popular and unpopular decisions was
much longer than during the Rehnquist Court alone. From 1935 to
1986, both predictors are strongly significant, and 90% of the Courts
decisions can be correctly predicted (compared to a base rate of 82%).
For popular decisions, the predicted half-life is fifty-four years, and for
unpopular decisions, the predicted half-life is only twenty-five yearsa
difference of twenty-nine years.25
CONCLUSION
American public opinion affects the longevity of Supreme Court deci-

sions. True, the Court is a very successful policy-maker. Most of its
decisions prevail for many years. Yet, over time, a steadily growing
number of decisions fail the test of time. Constitutional amendments,
congressional overrides, Court decisions that abandon precedent, and
executive or agency actions all take their toll. Over very long periods of
time, most decisions fail the test of time. Of chief interest here, unpopu-
lar decisions fail at a much faster rate than do popular decisions.
The test-of-time model should be included in the linkage model
between public opinion and Supreme Court policy-making. Only two
variables significantly predict whether a decision will meet the test of
time, both for the 19351986 era and for the Rehnquist Court: the pas-
sage of time itself, and whether or not a decision is consistent with
public opinion. Other explanations for how long Supreme Court deci-
sions survive are time-bound. The justices unanimity, a decisions lib-
eral-conservative ideology, or a decisions visibility were strongly
related to a decisions longevity prior to, but not during the Rehnquist
Court. Regrettably, since modern polls are not available prior to the
1930s, the impact of public opinion cannot be measured before the
1930s. As another caveat, the half-life estimates for poll-matched deci-
sions may differ from decisions for which no polls are available.26
In short, American public opinion has a second chance at
Supreme Court decisions. Unpopular decisions last several fewer years
than do popular decisions. Accordingly, the test-of-time model is
included in chapter nines model summarizing the ties between
American public opinion and the U.S. Supreme Court.
CHAPTER NINE
An Empirical Model of Representation
THE MAJORITARIAN COURT
T he modern Supreme Court has compiled a remarkably consistent

record of representing American public opinion. Since the origin of
modern polling in the 1930s, about three-fifths of Supreme Court deci-
sions represented majority public opinion, as measured by nationwide
public opinion polls among American adults. In this sense, the Supreme
Court is a majoritarian institution. Since the 1930s, all six Courts, by
chief justice, compiled a mostly majoritarian record. By agreeing with
public opinion in 64% of its decisions, the Rehnquist Court continued
this long tradition, now spanning seven decades.
Further, the Supreme Courts record is remarkably consistent across
even short time periods. By decade, the modern Courts decisions were
always mostly consistent with American public opinion, with the per-
centage of consistent decisions ranging from 56% consistent during the
1980s, to 69% consistent during the 1940s, 1950s, and 1990s. Even
across much shorter three-year periods, most Supreme Court decisions
were consistent with public opinion in 81% of all three-year periods,
evenly divided 14% of the time, and mostly inconsistent only 4% of the
time. When a mostly inconsistent or countermajoritarian three-year
period occurred, the Court quickly returned to its majoritarian habits
always doing so within only one year. By this standard, the Courts last
inconsistent or countermajoritarian three-year period was the
19791981 period. No American who is younger than middle-aged
could possibly remember a time during which the Courts decisions were
mostly out-of-line with public opinion for three years in a row.
The Rehnquist Courts relationship to American public opinion
was quite typical of earlier Courts. Sixty-four percent of its poll-
matched decisions were consistent with majority public opiniona
figure nearly identical to the 61% figure for the Warren Court, or the
62% figure for the Burger Court. For every single three-year period,
153
most Rehnquist Court decisions agreed with majority public opinion.

At least half of the Rehnquist Courts decisions agreed with the polls in
every major policy area, with the sole exception of speech cases.
Not only was the modern Courts record a majoritarian record
overall, its record was also majoritarian for all major demographic
groups. The Rehnquist Court least well represented American blacks,
but, even so, a slight majority of the Rehnquist Courts decisions agreed
with African Americans attitudes. At the other extreme, the Rehnquist
Court best represented the attitudes of males, whites, and Protestants.
Sixty-eight percent of Rehnquist Court decisions, including cert denials,
agreed with majority attitudes among these groups.
Overall, the Rehnquist Courts record of representing American
public opinion places it squarely within the modern Courts majoritar-
ian tradition. The modern Courts consistent record over seven decades
of representing public opinion is remarkably broad and deep, given the
wide range of disputes before the Court, the diversity of justices since
the mid-1930s, and the Courts changing norms and practices. The
Rehnquist Court maintained the modern Courts record as a majoritar-
ian branch.
If it is not quite the most majoritarian branch, then the modern
Courts record places it not far behind Congress and the president, and
sometimes a little ahead and sometimes a little behind the states and
localities. By the evidence here, federal laws and policies represented
poll majorities 72% of the time prior to the Rehnquist Court, and 73%
of the time during the Rehnquist Court. For state and local govern-
ments, the figures are 58% and 70%, respectively. For the Supreme
Court, the figures are 63% and 64%, respectively. The modern
Supreme Court has consistently been more majoritarian than the lower
state and federal courts. In short, the modern Supreme Court fares well,
perhaps surprisingly well, as a majoritarian policy-maker.
RECONSIDERING THE EVIDENCE
If the Rehnquist Court was, like the five earlier Courts, a majoritarian
Court, what best explains this pattern? Chapter one introduced and
outlined fifteen linkage models to explain the ties between American
public opinion and Supreme Court decision-making. Individually, each
model might seem plausible, but the real question is, does it hold up
empirically? Chapters two through eight tested each model. Some
proved to be good explanatory models, while others failed to win any
support at all. A few linkages that worked well in earlier years lost their
An Empirical Model of Representation 155
explanatory value during the Rehnquist Court. Occasionally, the

reverse was true.
The strongest empirical model linking public opinion and
Supreme Court decision-making should have several characteristics.
First, it should include only those linkages that strongly and consis-
tently link public opinion and Court decision-making. Statistically
significant results, or at least a clear and consistent pattern of results,
provides such evidence. Ideally, a good model would find support
during both the Rehnquist Court and earlier Courts. However, this
may not occur if the linkages are time-bound. Sometimes, a signifi-
cant linkage occurred in earlier years, but not during the Rehnquist
Court, and sometimes, the reverse occurred. When a linkage is time-
bound, a preference should be given to a linkage that works well
during the Rehnquist Court, even if not in earlier years. Second, a
linkage should remain a significant predictor, even when appropriate
controls are made for other explanatory variables. Third, ideally, a
linkage should be theoretically meaningful, and offer a coherent, per-
suasive explanation of the ties between public opinion and Supreme
Court decision-making.
By these standards, several linkage models described in chapter one
clearly failed. The federal policy model is now time-bound, and less
important than in earlier years. True, a large majority of federal laws
and policies reflected American public opinion during the Rehnquist
Court, and a large part of the Rehnquist Courts docket involved fed-
eral laws and policies. Yet the Rehnquist Court upheld fewer consistent
federal policies than did earlier Courts since the 1930s. As a result,
Rehnquist Court decisions in these cases were less often consistent with
public opinion than the average. The Rehnquist Courts views of judi-
cial activism and restraint eroded a linkage that was previously impor-
tant in connecting American public opinion and Supreme Court
policy-making. Whether this continues to be true during the Roberts
and subsequent Courts, or whether those Courts will be more support-
ive of federal laws and policies remains to be seen.
The state/local policy model is also time-bound and can no longer
be included in the empirical linkage model, but for different reasons.
This model contends that state and local laws and policies are much
less often consistent with nationwide public opinion than federal laws
and policies. The model also assumes that the Court less often upholds
state and local laws and policies, compared to federal laws and policies.
Both of these descriptions were indeed true from the 1930s until the
Rehnquist Court, but during the Rehnquist Court, these differences
largely disappeared. As a result, the Rehnquist Courts decisions were
no more consistent with American public opinion in federal-level than

in state or local-level disputes.
There is also no evidence that interest groups have mediated the
linkage between public opinion and Supreme Court decision-making at
any time since the 1930s. While interest groups often file lawsuits and
amicus briefs, no type of interest group outweighs the impact of public
opinion on the Court. Indeed, interest groups that disagree with public
opinion find that their winloss rates drop off significantly.
Likewise, there is no evidence that the justices political party ties
mediate the relationship between public opinion and Supreme Court
decisions. Knowing whether a justice is a Republican or a Democrat, or
appointed by a Republican or a Democratic president, does not help to
explain how often a justices votes will agree with public opinion. Nor
has political party apparently ever been closely tied to representation
since the 1930s.
The appointments process model also found no empirical support.
Whether a nominee was a presidents first choice, or not, did not pre-
dict how often a justices votes would agree with public opinion.
Perhaps surprisingly, justices who had the most difficulty being con-
firmed in the Senate slightly more often agreed with American public
opinion, compared to justices who were confirmed unanimously, or
nearly so. The evidence on crossover nominees (justices who were not
from the presidents own party) is mixed since the 1930s, and no such
nominations have occurred in over three decades. In short, nothing
about the nomination or confirmation process explains how often a
justice will agree with public opinion.
Neither was there much support for the symbolic representation
model. Presidents often (indeed, now usually) nominate symbolic
nominees, such as women, blacks, Catholics, or Jews. Yet there is little
evidence that most symbolic nominees, once confirmed, better represent
their gender, racial, or religious groups than do other justices. Only by
political party and for Jewish justices was there any support at all for
this model, and, even here, the evidence was not especially strong. This
model fails to hold up for several reasons, among them the diversity of
demographic and religious groups, the lack of formal ties between sym-
bolic justices and the groups they supposedly represent, and the impor-
tance of the justices legal ideas.
Nor is there any support for the short-term or long-term manipula-
tion linkage models. At one time, liberal, activist decisions routinely
gained poll support, as measured by pre- to post-decision poll shifts.
During the Rehnquist Court, however, that effect disappeared, and no
set of decisions routinely gained popularity. Indeed, pre- to post-deci-
sion polls show that Rehnquist Court decisions, once announced, usu-
ally lost popularity. Post-decision polls point to the same conclusion.
Prior to the Rehnquist Court, Supreme Court decisions showed at least
modest gains in popularity after the decision was announced. During
the Rehnquist Court, however, those modest positive poll gains
dropped into negative numbers. Arguably, only over very long time
periods do Supreme Court decisions now actually gain in public
approval. The modern Court never had much ability to legitimize its
decisions, in the sense of winning public approval. During the last two
decades, the Court lost whatever persuasive ability that it previously
had with the public. Exactly why this occurred is somewhat puzzling,
since the Rehnquist Court was not an unpopular Court, and its deci-
sions agreed with public opinion as often as did earlier Courts. Even so,
the weight of the evidence is that the Supreme Court no longer can win
over public opinion by announcing decisions.
Robert Dahls realignment model does little to explain the ties
between public opinion and Supreme Court decision-making. In part,
this is because realignment cycles are often hard to identify, and many
justices are confirmed, serve, and then retire or die between realignment
periods. In part, this is because Dahls argument is more about the rela-
tionship between Congress, the White House, and the Court than
about the relationship between public opinion and the Court. In part,
lags in judicial appointments make it difficult to compare a group of
pre-realignment, versus post-realignment justices. Where comparisons
are available, the evidence is, at best, mixed. Republican justices
appointed prior to the 1930s realignment actually voted slightly more,
not less often consistently with public opinion than did the justices
appointed thereafter, or than the twelve justices appointed by
Presidents Roosevelt and Truman. If the mid-1980s are considered to
be a realignment period, then the evidence for the realignment thesis is
mixed. If the mid-1990s are taken as a realignment period, then no evi-
dence so far appears for the realignment thesis. While Dahls realign-
ment thesis remains interesting, it does not explain the relationship
between Supreme Court decision-making and American public opinion.
AN EMPIRICAL LINKAGE MODEL
Although the evidence for some linkage models was mixed or negative,
more support appears for other models. In some cases, part of a linkage
model showed support, while in a few cases, the entire linkage model
showed support.
Part of the state of public opinion model should be included in the

empirical linkage model. Two such linkages were important, both in
the half century prior to, and also during the Rehnquist Court. First,
when public opinion is closely focused on an issue in a Supreme Court
case, the Court more often reflects public opinion. Second, when nearly
all Americans express an opinion on a controversy, the Court more
often agrees with public opinion. In these instances, as Chief Justice
Rehnquist once wrote, the justices can most easily sense and share pre-
vailing public opinion. By contrast, one-sided public opinion, per se,
was not strongly linked to representation, either before or during the
Rehnquist Court.
Unlike interest groups, the U.S. solicitor generals position is closely
tied to whether a Supreme Court decision will agree with public opin-
ion, or not. At one extreme, if the solicitor generals position agrees
with public opinion, then the Courts decisions reflect public opinion
over four-fifths of the time. At the other extreme, if the solicitor gen-
erals position disagrees with public opinion, then the Courts decisions
reflect public opinion barely a quarter of the time. Fortunately for
American public opinion, the solicitor general takes a position agreeing
with public opinion much more often than not.
Two judicial roles are tied to representation. Most notably, the
chief justice usually ranks among the justices most likely to represent
public opinion. Chief Justice Rehnquist continued this pattern. His
votes agreed with majority public opinion 69% of the time during his
years as chief justice, ranking him near the top of the Rehnquist Courts
fourteen justices. Rehnquists rate of agreeing with public opinion also
jumped from 60% to 69% after being elevated from associate to chief
justice. It may well be that the chief justices role as the Courts
spokesperson, organizational defender, and lobbyist sensitizes a chief
justice to public opinion.
The Courts intellectual leaders significantly more often voted con-
sistently with public opinion. To a lesser degree, the Courts task lead-
ers, its tie-breaking justices, and the most frequent authors of important
opinions a little more often reflected public opinion in their votes. By
contrast, justices with the strongest (or weakest) reputations, and jus-
tices who are especially aggressive during oral arguments did not stand
out from the other justices in terms of agreeing with public opinion. In
short, the chief justices role and that of an intellectual leader clearly
deserve to be included in an empirical linkage model between Supreme
Court decision-making and American public opinion.
The justices ideology should also be considered. Justices thought
to be ideologically moderate at the time of their nomination more often
cast votes agreeing with American public opinion than do justices

described as either very liberal or very conservative. To be sure, presi-
dents seldom any longer nominate ideological moderates to the Court,
and perhaps only politically weak presidents, or those in their last year
or two in office and facing a hostile Senate, will do so.
The tenure and aging model also merits inclusion in an empirical
linkage model. Justices now typically remain on the Court for over two
decades, until retiring in old age and ill health, or dying. Long service
on the Court is not unprecedented. Since the 1930s, several justices
served at least twenty years, and six justices (Black, Douglas, Brennan,
White, Rehnquist, and Stevens) served into their third decade. If presi-
dents continue to choose relatively young justices who have no off-the-
Court ambitions, and if the average lifespan continues to increase, more
justices will serve a quarter of a century or more.
Does lengthy tenure affect representation? The evidence is that it
does. Justices are the most consistent with American public opinion in
their first four years on the Court. At this time, a new justice is also
typically more consistent with public opinion than are the justices who
continue to sit on the Court, and more consistent than was the outgo-
ing justice in that justices last four years on the Court. Later on, a jus-
tice becomes less consistent with public opinion after his or her
twentieth year on the Court, either compared to a justices own record
in earlier years, or to less senior justices. Individually, some of these dif-
ferences are not dramatic, but they consistently point in the same direc-
tion. Simply put, long tenure does not just limit a presidents ability to
appoint new justices. Long tenure also reduces the ties between
Supreme Court decision-making and American public opinion.
Political scientists often study the freshman effectthat is, how
newly appointed justices adjust to the Court during their first few
terms. However, so far, little attention has been given to the senior
effect. Arguably, however, senior effects are more important than
freshman effects, since there are typically fewer justices in their first
term or two than there are justices in their twentieth term or beyond.
Very likely, this will continue to be true.
Exactly why justices vote less consistently with public opinion, over
their time on the Court, is not entirely clear, but the reason may partly
lie in changing issues. Over two decades or more, the issues coming
before the Court change considerably. Historically, there were few
racial, free speech, or dissent cases before the Court in the 1930s, but
from the mid-1940s through the 1960s, the Court heard many more
such cases. Sometimes public attitudes change dramatically over a
decade or two. If a justice was in line with majority public opinion
when confirmed, that justices values may no longer be in the main-

stream two decades later.
By this evidence, term limits at two decades, or so, would modestly
increase the number of Supreme Court decisions that agree with public
opinion. Admittedly, it seems unlikely that such limits will be imposed
any time soon. The likely result: an increasingly senior Court, whose
justices slowly grow less consistent with public opinion. Stated other-
wise, the justices increasingly long tenure reduces the ties between
American public opinion and the Supreme Court.
In the socialization model, the justices come to the Court with a
very wide variety of backgrounds, a few of which are tied to represent-
ing American attitudes. Among the strongest experiences now linked
to representing public opinion are moving between regions, holding a
top post in Washington, DC, prior candidate or campaign experience,
serving as a presidential confidant, and not previously having been a
judge. Underlying all these experiences is a broad, diverse, and politi-
cized experience.
It may seem ironic, then, that modern norms of picking Supreme
Court nominees effectively weaken the ties between public opinion and
Supreme Court decision-making. Since the 1960s, fewer and fewer
nominees have brought any electoral experience to the Court. As of
Sandra Day OConnors retirement, not a single justice had previously
run for an elective post. Indeed, after her nomination, a quarter-century
ago, no new justice had such experience.
The custom of sending White House confidants to the Supreme
Court also seems to have passed into history, partly a result of crony-
ism charges leveled against Lyndon Johnson nearly half a century ago.
George W. Bushs failed nomination of White House counsel, Harriet
Miers, can only reinforce that perception. All the justices of the early
Roberts Court were federal appeals judges at the time of their appoint-
mentor, as this norm is sometimes described, from inside the
monastery. This now-common practice reduces the ties between
public opinion and Supreme Court decision-making.
Some practices, once common, that mediated the ties between public
opinion and Supreme Court decisions have now become time-bound and
irrelevant. At one time, several justices came from less prestigious law
schools, or were self-taught. This practice, which produced justices
whose votes less often reflected majority public opinion, has passed into
history. Justice Thurgood Marshall was the last such nominee, now over
four decades ago. Again, Harriet Miers would have been another such
nominee, but her nomination failed, and her lower-status law school
experience was probably a modest negative mark against her.
Judicial scholars too seldom consider how the Courts routine

norms and customs affect representation. Yet the Courts routine prac-
tices do, indeed, matter. Typically, the Court disproportionately uses its
grants of certiorari to pick out lower court decisions that disagree with
public opinion, and then to reverse many inconsistent lower court deci-
sions. If certiorari is granted on an appeal, an inconsistent lower court
decision is much more often reversed than is a consistent lower court
decision. This pattern of poll correction is common to higher courts,
and brings judicial decision-making back into line with American
public opinion. The Rehnquist Courts record of poll correction against
lower state and federal courts was very similar to that of earlier Courts
since the 1930s.
The modern Court, however, has not engaged in poll correction
against nonjudicial policy-makers. Overturning federal, state, and local
laws and policies effectively reduces the degree to which public policy
reflects American public opinion. If the Supreme Court more highly
valued judicial restraint, and upheld more federal, state, and local laws
and policies, its decisions would more closely reflect public opinion. If
presidents more often nominated, and the Senate more often confirmed,
justices who favored judicial restraint, then Supreme Court decisions
would more often agree with public opinion.
Among the Rehnquist Courts most important norms was cutting
the number of full, written decisions, per term, by about half, com-
pared to earlier Courts. The shrunken docket norm also reduced the
number of decisions consistent with majority public opinion, arguably
by half a dozen consistent decisions, per term. Since Supreme Court
decisions are precedents for lower state and federal courts, fewer full,
written decisions probably translates into fewer lower court decisions
consistent with public opinion, as well.
That the justices typically vote their long-standing, liberal-versus-
conservative ideological preferences is now well-recognized. Indeed, it
is difficult to foresee a future Court in which this is not the case. This
norm is an important one for representation, since a one- or two-vote
ideological shift would sometimes dramatically change the percentage
of Supreme Court decisions that reflect majority public opinion.
Taken together, the Courts routine norms and practices dramati-
cally affect the extent to which its decisions reflect American public
opinion. A future Court that significantly increased its caseload, contin-
ued to reverse inconsistent lower court decisions, and was deferential to
elected office-holders, would almost certainly be a Court more of
whose decisions agreed with American public opinion. The Courts
routine norms and practices can change dramatically over time;
whether any these changes will actually occur in the future remains to
be seen.
Finally, the test-of-time linkage was a very strong one, both before
and during the Rehnquist Court. Decisions that were unpopular, when
announced, do not endure as long as those with greater public sup-
port. Exactly how decisions fail the test of time can vary. Some deci-
sions are overturned by the Supreme Court itself, others by Congress
or the president or the agencies, or, in very rare instances, by a consti-
tutional amendment. Since the 1930s, popular Supreme Court deci-
sions have prevailed for several years longer than have unpopular
decisions. During the five decades from the mid-1930s until the
Rehnquist Court, the difference was about twenty-nine years. During
the Rehnquist Courts nearly two decades, the difference was about
seven years. Public opinion clearly gets a second chance at Supreme
Court decisions.
To summarize the evidence: the Rehnquist Courts relationship to
American public opinion was complex, but heavily rested on three ties:
the context of a controversy, the justices themselves, and the Courts cur-
rent norms. First, when public opinion was closely focused on an issue,
when nearly all Americans expressed an opinion on an issue, and when
the solicitor generals position agreed with public opinion, Rehnquist
Court decisions typically reflected public opinion. Rehnquist Court deci-
sions that received widespread public support also lasted longer than
unpopular decisions. Second, justices who moved between regions before
coming to the Court; justices with elective experience or experience in
top, nonjudicial positions; justices without prior judicial experience; less
senior justices; the chief justice; and the Courts intellectual leaders most
often cast votes that agreed with majority public opinion. Third, the
norm of judicial deference to elected policy-makers, and the norm of poll
correction against the lower courts both fostered representation. The
norm of a smaller docket of full, written decisions did not.
At least since the 1930s, most Supreme Court decisions agreed with
majority public opinion. Whether the Supreme Court will continue to
be an essentially majoritarian institution in the future seems likely, but
by no means guaranteed. Some trends, such as longer tenure for the jus-
tices, or the now-complete ascendancy of federal appeals court judges
as justices, suggest that future Courts may be much less majoritarian
than were past Courts since the 1930s. Whether these trends are offset
by prevailing public opinion, by other judicial experiences and roles,
and by Court norms that foster representation will decide whether the
future Supreme Court continues to reflect American public opinion
more often than not.
APPENDIX ONE
Poll-to-Ruling Matches
(1) Affirmative action in hiring and promotionJohnson v.

Transportation Agency of Santa Clara County (March 25, 1987)
(2) Victim impact statements at sentencing stageBooth v. Maryland
(June 15, 1987) and South Carolina v. Gathers (June 12, 1989)
(3) Clubs right to discriminate among members by sexBoard of
Directors of Rotary International v. Rotary Club of Duarte (May
4, 1987) and Roberts v. United States Jaycees (July 3, 1984)
(4) Teaching creation scienceEdwards v. Aguillard (June 19, 1987)
(5) Federal funding cutoff for states with below-twenty-one drinking
age limitsSouth Dakota v. Dole (June 23, 1987)
(6) Principals censoring high school newspapersHazelwood School
District v. Kuhlmeier (January 13, 1988)
(7) Independent Counsel ActMorrison v. Olson (June 29, 1988)
(8) Drug testing for safety employeesSkinner v. Railway Labor
Executives Association (March 21, 1989)
(9) Phone sexSable Communications v. Federal Communications
Commission (June 23, 1989)
(10) Death penalty for mentally retarded murderersPenry v.
Lynaugh (June 26, 1989)
(11) Flag-burning protestsTexas v. Gregory Lee Johnson (June 21,
1989) and U.S. v. Eichman (June 11, 1990)
(12) Death penalty for teenage murderersStanford v. Kentucky and
Wilkins v. Missouri (June 26, 1989)
(13) Punitive damagesBrowning Ferris Industries of Vermont v.
Kelco Disposal (June 26, 1989)
163
(14) Public hospital restrictions on abortionWebster v. Reproductive

Health Services (July 3, 1989)
(15) Fetal viability testsWebster v. Reproductive Health Services
(July 3, 1989)
(16) AIDS testing of health care workersEastern Nebraska
Community Office of Retardation v. Glover, cert denied (October
30, 1989)
(17) AIDs testing of prison inmatesDunn v. White, cert denied
(January 22, 1990)
(18) Student religious groups use of school facilitiesBoard of
Education of Westside Community Schools v. Mergens (June 4,
1990)
(19) Roadblocks for drunk driversMichigan Department of State
Police v. Stitz (June 14, 1990)
(20) Two-parent notification for minors abortionsHodgson v.
Minnesota (June 25, 1990)
(21) One-parent notification for minors abortionsOhio v. Akron
Center for Reproductive Health (June 25, 1990)
(22) Withdrawal of life supportCruzan v. Missouri (June 25, 1990)
(23) Gag rule for abortion providersRust v. Sullivan (May 23, 1991)
(24) Luggage searches for drugs on interstate busesFlorida v. Bostick
(June 20, 1991)
(25) Nude dancingBarnes v. Glen Theatre (June 21, 1991)
(26) Victim testimony at sentencing stagePayne v. Tennessee (June
27, 1991)
(27) Life sentences for first offenders, drug casesHarmelin v.
Michigan (June 27, 1991)
(28) Racial disparities in death penalty sentencesMcCleskey v.
Bowers, application for stay of execution denied (September 24,
1991) and prior cases.
(29) AIDS testing of hospital patientsLee v. Baptist Medical Center
of Oklahoma, cert denied (June 8, 1992)
(30) Prayers at high school graduationsLee v. Weisman (June 24,
1992)
(31) Tobacco litigationCipollone v. Liggett Group (June 24, 1992)

(32) Husband notificationPlanned Parenthood of Southeastern
Pennsylvania v. Casey (June 29, 1992)
(33) Informed consentPlanned Parenthood of Southeastern
Pennsylvania v. Casey (June 29, 1992)
(34) Twenty-four-hour waiting rulePlanned Parenthood of South-
eastern Pennsylvania v. Casey (June 29, 1992)
(35) One-parent consent for minors abortionsPlanned Parenthood
of Southeastern Pennsylvania v. Casey (June 29, 1992)
(36) RU-486 abortion pillBenten v. Kessler, denial of application to
vacate a stay (July 17, 1992)
(37) Operation RescueBray v. Alexandria Womens Health Clinic
(January 13, 1993)
(38) Limit on death penalty appealsGraham v. Collins, Herrera v.
Collins, and Lockhart v. Fretwell (January 25, 1993).
(39) Hate crimesWisconsin v. Mitchell (June 11, 1993)
(40) Deporting HaitiansSale v. Haitian Centers Council (June 21,
1993)
(41) Affirmative action in federal contractsAdarand v. Pena (June
12, 1995)
(42) Racial redistrictingShaw v. Reno (June 28, 1993) and Miller v.
Johnson (June 29, 1995)
(43) Baby JessicaDeBoer v. Schmidt, application for a stay denied
(July 26, 1993)
(44) Abortion clinic accessMadsden v. Womens Health Center (June
30, 1994)
(45) Term limitsU.S. Term Limits v. Thornton (May 22, 1995)
(46) One-race college scholarshipsPodberesky v. Kirwan and Greene
v. Podberesky, cert denied (May 22, 1995)
(47) Drug testing for high school studentsVernonia School District
47J v. Acton (June 26, 1995)
(48) Condom distribution in public schoolsCurtis v. School
Committee, cert denied (January 8, 1996)
(49) Motor voter lawWilson v. Voting Rights Coalition, cert denied

(January 22, 1996)
(50) AIDS education in public schoolsBrown v. Hot, Sexy and Safer
Productions, Inc., cert denied (March 4, 1996)
(51) 1990 Census adjustmentWisconsin v. City of New York and
companion cases (March 20, 1996)
(52) Gay rights Colorado ballot issueRoemer v. Evans (May 20,
1996)
(53) Dont ask, dont tellThomasson v. Perry, cert denied (October
21, 1996)
(54) Paula Jones lawsuitClinton v. Jones (May 27, 1997)
(55) Constitutional right to assisted suicideWashington v.
Glucksberg and Vacco v. Quill (June 26, 1997)
(56) Background checks in the Brady billMack v. U.S. and Printz v.
U.S. (June 27, 1997)
(57) Civil commitment of violent sexual predatorsKansas v.
Hendricks (June 23, 1997)
(58) Internet pornography to minorsReno v. ACLU (June 26, 1997)
(59) Assisted suicide lawLee v. Harcleroad, cert denied (October 14,
1997)
(60) California civil rights initiativeCoalition for Economic Equity v.
Wilson, cert denied (November 3, 1997)
(61) Karla Faye Tucker death penalty caseTucker v. Texas, four denials
of cert or denial of a writ of habeas corpus (February 3, 1998)
(62) Megans lawDoe v. Pataki and Verniero v. W.P., cert denied
(February 23, 1998)
(63) Mandatory prison term for using a gun in a crimeMuscarello v.
U.S. (June 8, 1998)
(64) Line-item vetoClinton v. City of New York (June 25, 1998)
(65) Three-strike lawMonge v. California (June 26, 1998)
(66) Executive privilege claimsOffice of the President v. Office of
Independent Counsel, cert denied (November 9, 1998) and Rubin
v. U.S., cert denied (November 9, 1998)
(67) Suspicionless car searches during a routine traffic stopKnowles

v. Iowa (December 8, 1998)
(68) Oklahoma City bombing death penalty appealTimothy Jones
McVeigh v. U.S. (March 9, 1999)
(69) Additional federal death penalty offensesJones v. U.S. (June 21,
1999)
(70) Drivers license data basesReno v. Condon (January 12, 2000)
(71) PAC limit of $1,000Nixon v. Shrink Missouri Government
PAC (January 24, 2000)
(72) Elian GonzalezIn re Hirschfeld, cert denied (February 22, 2000)
(73) FDA regulation of tobacco as a drugFDA v. Brown &
Williamson Tobacco Corp. (March 21, 2000)
(74) Welfare cutoff for legal immigrantsCity of Chicago v. Shalala,
cert denied (March 27, 2000)
(75) Grandparent visitationTroxel v. Granville (June 5, 2000)
(76) HMO lawsuits for denial of carePegram v. Herdich (June 12,
2000)
(77) Student prayers at high school football gamesSanta Fe
Independent School District v. Doe (June 19, 2000)
(78) Teaching creationismTangipahoa Parish Bd. of Education v.
Freiler, cert denied (June 19, 2000)
(79) Miranda warningsDickerson v. U.S. (June 26, 2000)
(80) Judge versus jury sentencingApprendi v. New Jersey (June 26,
2000)
(81) Blanket primariesCalifornia Democratic Party v. Jones (June
26, 2000)
(82) Computers to parochial schoolsMitchell v. Helms (June 28,
2000)
(83) Partial birth abortionsSternberg v. Carhart (June 28, 2000)
(84) Gay boy scoutsBoy Scouts of America v. Dale (June 28, 2000)
(85) Specific zone for abortion clinic protestsHill v. Colorado (June
28, 2000)
(86) Florida presidential election recountBush v. Gore (December

12, 2000)
(87) Cocaine testing of pregnant womenFerguson v. City of
Charleston (March 21, 2001)
(88) Disability status of golfersPGA Tour v. Martin (May 29,
2001)
(89) Religious groups in public schoolsGood News Club v. Milford
Central School (June 11, 2001)
(90) Moment of silenceBrown v. Gilmore, cert denied (September
29, 2001)
(91) Death penalty for mentally retarded killersAtkins v. Virginia
(June 20, 2002)
(92) Independent HMO reviewRush Prudential HMO v. Moran
(June 20, 2002)
(93) Sentencing in death penalty casesRing v. Arizona (June 24,
2002)
(94) VouchersZelman v. Simmons-Harris (June 27, 2002)
(95) Any willing provider HMO planKentucky Association of
Health Plans v. Miller (April 2, 2003)
(96) Legal status of gay sexLawrence v. Texas (June 26, 2003)
(97) Affirmative action in college admissionsGratz v. Bollinger and
Grutter v. Bollinger (June 23, 2003)
(98) Internet filtering in librariesU.S. v. American Library
Association (June 23, 2003)
(99) Judge suspension in Ten Commandments caseIn re Moore,
petition for writ of mandamus denied (November 23, 2003)
(100) Soft moneyMcConnell v. Federal Election Commission
(December 10, 2003)
(101) Candidate identification in campaign adsMcConnell v. Federal
Election Commission (December 10, 2003)
(102) Pledge of allegianceElk Grove Unified School District v.
Newdow (June 14, 2004)
(103) Confidentiality of vice presidential proceedingsCheney v. U.S.

District Court (June 23, 2004)
(104) Terrorist detention and trial delaysRasul v. Bush and Al Odah
v. United States (June 28, 2004)
(105) American-born terroristsPadilla v. Rumsfield (June 28, 2004)
(106) Do-not-call listMainstream Marketing Services v. Federal
Trade Commission, cert denied (October 4, 2004)
(107) Gay adoptionsLofton v. Secretary of Department of Children
and Family Services, cert denied (January 10, 2005)
(108) Death penalty for juvenile murderersRoper v. Simmons
(March 1, 2005)
(109) Medical marijuana suppliersGonzales v. Raich (June 6, 2005)
and earlier cases
(110) Terri Schiavo caseSchindler v. Schiavo (March 17, 2005),
application for a stay, denied, and related cases
(111) Ten Commandments on Texas Capitol groundsVan Orden v.
Perry (June 27, 2005)
APPENDIX TWO
List of Cases
Abrams v. Johnson, 521 U.S. 74 (1996)

Adams v. Florida, 437 U.S. 947 (1978)
Adarand v. Pena, 515 U.S. 200 (1995)
Al Odah v. United States, 541 U.S. 933 (2004)
Apprendi v. New Jersey, 530 U.S. 466 (2000)
Atkins v. Virginia, 536 U.S. 304 (2002)
Banks v. Dretke, 540 U.S. 668 (2004)
Barber v. Tennessee, 513 U.S. 1185 (1995)
Barnes v. Glen Theatre, 501 U.S. 560 (1991)
Bartnicki v. Vopper, 532 U.S. 514 (2001)
Bates v. Little Rock, 361 U.S. 516 (1960)
Benten v. Kessler, Commissioner of FDA, 505 U.S. 1084 (1992)
Beuharnais v. Illinois, 343 U.S. 250 (1952)
BMW of North America, Inc. v. Gore, 517 U.S. 559 (1996)
Board of Directors of Rotary International v. Rotary Club of Duarte,
481 U.S. 537 (1987)
Board of Education of Westside Community Schools v. Mergens, 496
U.S. 226 (1990)
Board of Trustees v. Garrett, 531 U.S. 356 (2001)
Booth v. Maryland, 482 U.S. 496 (1987)
Borgner v. Florida Board of Dentistry, 537 U.S. 1080 (2002)
Bowen v. American Hospital Association, 476 U.S. 610 (1986)
171
Boy Scouts of America v. Dale, 530 U.S. 640 (2000)

Bray v. Alexandria Womens Health Clinic, 506 U.S. 263 (1993)
Brown v. Allen, 344 U.S. 443 (1953)
Brown v. Board of Education of Topeka, Kansas, 347 U.S. 483 (1954)
and 349 U.S. 294 (1955)
Brown v. Gilmore, 533 U.S. 1301 (2001)
Brown v. Hot, Sexy and Safer Productions, Inc., 516 U.S. 1159 (1996)
Brown v. Texas, 522 U.S. 940 (1997)
Browning Ferris Industries of Vermont v. Kelco Disposal, 492 U.S. 257
(1989)
Buckley v. American Constitutional Law Foundation, 525 U.S. 182
(1999)
Buckley v. Valeo, 424 U.S. 1 (1976)
Burlington Northern Railroad Company v. Brotherhood of Mainte-
nance of Way Employees, 481 U.S. 429 (1987)
Burnet v. Coronado Oil and Gas Co., 285 U.S. 393 (1932)
Bush v. Gore, 531 U.S. 98 (2000)
California v. Brown, 479 U.S. 538 (1987)
California Democratic Party v. Jones, 530 U.S.567 (2000)
Campbell v. Wood, 511 U.S. 1119 (1994)
Cheney v. U.S. District Court, 124 S. Ct. 2576 (2004)
Cipollone v. Liggett Group, 505 U.S. 504 (1992)
Ciraolo v. City of New York, cert denied, 2000 WL 1376675 (2000)
Cleveland v. U.S., 531 U.S. 12 (2000)
Clinton v. City of New York, 524 U.S. 417 (1998)
Clinton v. Jones, 520 U.S. 681 (1997)
Coalition for Economic Equity v. Wilson, 521 U.S. 1141 (1997)
Coker v. Georgia, 433 U.S. 584 (1977)
Coleman v. Balkcom, 451 U.S. 949 (1981)
Collins v. Youngblood, 497 U.S. 37 (1990)

Cooper Industries, Inc. v. Leatherman Toll Group, Inc., 532 U.S. 424
(2001)
Cruzan v. Director, Missouri Department of Health, 497 U.S. 261
(1990)
Curtis v. School Committee of Falmouth, 516 U.S. 1067 (1996)
DeBoer v. Schmidt, 509 U.S. 938 (1993)
Dial Information Services Corporation of New York v. Barr, cert
denied, 502 U.S. 1072 (1992)
Dial Information Services Corporation of New York v. Thornburgh,
938 F.2d 1535 (1991)
Dickerson v. U.S., 530 U.S. 428 (2000)
Doe v. Pataki, 522 U.S. 1122 (1998)
Dretke v. Haley, 541 U.S. 386 (2004)
Dunn v. White, 493 U.S. 1059 (1990)
Eastern Nebraska Community Office of Retardation v. Glover, 493
U.S. 932 (1989)
Edenfield v. Fane, 507 U.S. 761 (1993)
Edwards v. Aguillard, 482 U.S. 578 (1987)
Elk Grove Unified School District v. Newdow, 542 U.S. 1(2004)
Elledge v. Florida, 119 S.Ct. 366 (1998)
Federal Election Commission v. Colorado Republican Federal
Campaign Committee, 533 U.S. 431 (2001)
Ferguson v. City of Charleston, 532 U.S. 67 (2001)
Florida v. Bostick, 501 U.S. 429 (1991)
Florida Bar v. Went For It, Inc., 515 U.S. 618 (1995)
Florida v. Riley, 488 U.S. 445 (1989)
Food and Drug Administration v. Brown & Williamson Tobacco
Corp., 529 U.S. 120 (2000)
44 Liquormart, Inc. v. Rhode Island, 517 U.S. 484 (1996)
Frank v. Magnum, 237 U.S. 309 (1915)

Furman v. Georgia, 408 U.S. 238 (1972)
Gentile v. State Bar of Nevada, 501 U.S. 1030 (1991)
Gomez v. U.S. District Court, 503 U.S. 653 (1992)
Gonzales v. Raich, 125 S.Ct. 2195 (2005)
Good News Club v. Milford Central School, 533 U.S. 93 (2001)
Graham v. Collins, 506 U.S. 461 (1993)
Gratz v. Bollinger, 539 U.S. 244 (2003)
Greater New Orleans Broadcasting Association v. U.S., 517 U.S. 173
(1990)
Greene v. Podberesky, 514 U.S. 1128 (1995)
Gregg v. Georgia, 428 U.S. 153 (1976)
Grutter v. Bollinger, 539 U.S. 306 (2003)
Harmelin v. Michigan, 501 U.S. 957 (1991)
Harris v. Alabama, 513 U.S. 504 (1995)
Hazelwood School District v. Kuhlmeier, 484 U.S. 260 (1988)
Herrera v. Collins, 506 U.S. 390 (1993)
Hill v. Colorado, 530 U.S. 703 (2000)
Hodgson v. Minnesota, 497 U.S. 417 (1990)
Hunt v. Cromartie, 532 U.S. 234 (2001)
Illinois v. Wardlow, 528 U.S. 119 (2000)
Immigration and Naturalization Service v. Chadha, 462 U.S. 919
(1983)
In re Hirschfeld, 528 U.S. 1152 (2000)
In re Moore, 124 S. Ct. 467 (2003)
In re Stanford, 537 U.S. 968 (2002)
Johnson v. Transportation Agency of Santa Clara County, 480 U.S.
616 (1987)
Jones v. U.S., 527 U.S. 373 (1999)
Kansas v. Hendricks, 521 U.S. 346 (1997)

Kentucky Association of Health Plans v. Miller, 538 U.S. 329 (2003)
Knight v. Florida, 120 S. Ct. 459 (1999)
Knowles v. Iowa, 525 U.S. 113 (1998)
Kolstad v. American Dental Association, 527 U.S. 526 (1999)
Kring v. Missouri, 107 U.S. 221 (1883)
Lackey v. Texas, 514 U.S. 1045 (1995)
Lambert v. City and County of San Francisco, 529 U.S. 1045 (2000)
Lawrence v. Texas, 539 U.S. 558 (2003)
Lawyer v. Department of Justice, 521 U.S. 567 (1997)
Legal Services Corporation v. Velazquez, 531 U.S. 533 (2001)
Lee v. Baptist Medical Center of Oklahoma, 504 U.S. 973 (1992)
Lee v. Harcleroad, 522 U.S. 927 (1997)
Lee v. Weisman, 505 U.S. 577 (1992)
Lochner v. New York, 198 U.S. 45 (1905)
Lockhart v. Fretwell, 506 U.S. 364 (1993)
Lofton v. Secretary of the Department of Children and Family Services,
125 S.Ct. 869 (2005)
Lorillard Tobacco Company v. Reilly, 533 U.S. 525 (2001)
Mack v. U.S., 521 U.S. 898 (1997)
Madsden v. Womens Health Center, 512 U.S. 753 (1994)
Mainstream Marketing Services v. Federal Trade Commission, 125
S.Ct. 47 (2004)
Maryland v. Baltimore Radio Show, 338 U.S. 912 (1950)
McCarver v. North Carolina, 521 U.S. 941 (2001) and 533 U.S. 975
(2001)
McCleskey v. Bowers, 501 U.S. 1281 (1991)
McCleskey v. Kemp, 481 U.S. 279 (1987)
McCleskey v. Zant, 499 U.S. 467 (1991)
McConnell v. Federal Election Commission, 124 S. Ct. 619 (2003)

McCray v. New York, 461 U.S. 961 (1983)
McIntyre v. Ohio Elections Commission, 514 U.S. 334 (1995)
McVeigh v. United States, 526 U.S.1007 (1999)
Michigan Department of State Police v. Stitz, 496 U.S. 444 (1990)
Mickens v.Taylor, 535 U.S. 162 (2002)
Milkovich v. Lorain Journal Co., 497 U.S. 1 (1990)
Miller v. Johnson, 515 U.S. 900 (1995)
Minor v. Happersett, 21 Wall. 162 (1875)
Mitchell v. Helms, 530 U.S. 793 (2000)
Monge v. California, 524 U.S. 721 (1998)
Moore v. Dempsey, 261 U.S. 86 (1923)
Muscarello v. U.S., 524 U.S. 125 (1998)
NAACP v. Alabama, 357 U.S. 499 (1958)
National Archives and Records Administration v. Favish, 541 U.S. 157
(2004)
National Endowment for the Arts v. Finley, 524 U.S. 569 (1998)
Neder v. U.S., 517 U.S. 1 (1999)
Nixon v. Shrink Missouri Government PAC, 528 U.S. 377 (2000)
Northern Securities Company v. U.S., 193 U.S. 197 (1904)
Office of the President v. Office of Independent Counsel, 525 U.S. 990
(1998)
Ohio v. Akron Center for Reproductive Health, 497 U.S. 502 (1990)
ONeil v. Vermont, 144 U.S. 323 (1892)
Oregon v. Mitchell, 400 U.S. 112 (1970)
Pacific Mutual Life Insurance v. Haslip, 499 U.S. 1 (1991)
Padilla v. Rumsfield, 124 U.S. 2711 (2004)
Patterson v. Texas 536 U.S. 984 (2002)
Payne v. Tennessee, 501 U.S. 808 (1991)
Pegram v. Herdich, 530 U.S. 211 (2000)

Penry v. Lynaugh, 492 U.S. 302 (1989)
Peters v. Hobby, 349 U.S. 331 (1955)
PGA Tour v. Casey Martin, 532 U.S. 661 (2001)
Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S.
833 (1992)
Plessy v. Ferguson, 163 U.S. 587 (1896)
Podberesky v. Kirwan, 514 U.S. 1128 (1995)
Powell v. Alabama, 287 U.S. 45 (1932)
Printz v. U.S., 521 U.S. 898 (1997)
Ramdass v. Angelone, 530 U.S. 156 (2000)
Rasul v. Bush, 124 S.Ct. 2686 (2004)
Reno v. American Civil Liberties Union, 521 U.S. 844 (1997)
Reno v. Condon, 121 S.Ct. 666 (2000)
Republican Party of Minnesota v. White, 536 U.S. 765 (2002)
Ring v. Arizona, 536 U.S. 584 (2002)
Roberts v. Louisiana, 428 U.S. 325 (1976) and 431 U.S. 633 (1977)
Roberts v. United States Jaycees, 468 U.S. 609 (1984)
Roe v. Wade, 410 U.S. 113 (1973)
Roemer v. Evans, 517 U.S. 620 (1996)
Roper v. Simmons, 125 S.Ct. 1183 (2005)
Rubin v. U.S., 525 U.S. 990 (1998)
Rush Prudential HMO v. Moran, 536 U.S. 355 (2002)
Rust v. Sullivan, 500 U.S. 173 (1991)
Sable Communications v. Federal Communications Commission, 492
U.S. 115 (1989)
Sale v. Haitian Centers Council, 509 U.S. 155 (1993)
Santa Fe Independent School District v. Doe, 530 U.S. 290 (2000)
Schindler v. Schiavo, 544 U.S. 915 (2005)
Seling v. Young, 531 U.S. 250 (2001)

Shaw v. Reno, 509 U.S. 630 (1993)
Simmons v. South Carolina, 512 U.S. 154 (1994)
Singleton v. Commissioner of Internal Revenue, 439 U.S. 940 (1978)
Skinner v. Railway Labor Executives Association, 489 U.S. 602 (1989)
Smith and Botelho v. Doe, 538 U.S. 84 (2003)
South Carolina v. Gathers, 490 U.S. 805 (1989)
South Dakota v. Dole, 483 U.S. 203 (1987)
Stack v. Boyle, 342 U.S. 1 (1951)
Stanford v. Kentucky, 492 U.S. 361 (1989)
Sternberg v. Carhart, 530 U.S. 914 (2000)
Sweezy v. New Hampshire, 354 U.S. 234 (1957)
Tangipahoa Parish Board of Education v. Freiler, 530 U.S. 1251
(2000)
Texas v. Hopwood, 518 U.S. 1033 (1996)
Texas v. Johnson, 491 U.S. 397 (1990)
Thomasson v. Perry, 519 U.S. 948 (1996)
Thompson v. Oklahoma, 487 U.S. 815 (1988)
Thompson v. Western States Medical Center, 535 U.S. 357 (2002)
Timothy Jones McVeigh v. U.S., 526 U.S. 1007 (1999)
Trop v. Dulles, 356 U.S. 86 (1957)
Troxel v. Granville, 530 U.S. 57 (2000)
Tucker v. Texas, and companion cases, 522 U.S. 1100 (1998)
TXO Production Corp. v. Alliance Resources Corp., 509 U.S. 443
(1993)
U.S. v. American Library Association, 539 U.S. 194 (2003)
U.S. v. Atkinson, 297 U.S. 157 (1936)
U.S. v. Eichman, 496 U.S. 310 (1990)
U.S. v. Morrison, 529 U.S. 598 (2000)
U.S. v. Olano, 507 U.S. 725 (1993)

U.S. v. Playboy Entertainment Group, Inc., 529 U.S. 803 (2000)
U.S. Term Limits v. Thornton, 514 U.S. 779 (1995)
Utah v. Evans, 536 U.S. 452 (2002)
Vacco v. Quill, 521 U.S. 793 (1997)
Van Orden v. Perry, 123 S.Ct. 2854 (2005)
Vernonia School District v. Acton, 515 U.S. 646 (1995)
Wal-Mart Stores, Inc. v. Samara Bros., 529 U.S. 205 (2000)
Washington v. Glucksberg, 521 U.S. 640 (1997)
Watchtower Bible and Tract Society of New York v. Village of
Stratton, 536 U.S. 150 (2002)
W.P. v. Verniero and Verniero v. W.P., 522 U.S. 1109, 1110 (1998)
Webster v. Reproductive Health Services, 492 U.S. 490 (1989)
Weems v. U.S., 217 U.S. 349 (1910)
Weiman v. Updegraff, 344 U.S. 183 (1952)
West Coast Hotel v. Parrish, 300 U.S. 379 (1937)
Wilkins v. Missouri, 492 U.S. 361 (1989)
Wilson v. Voting Rights Coalition, 516 U.S. 1093 (1996)
Wisconsin v. City of New York and companion cases, 517 U.S. 1
(1996)
Wisconsin v. Mitchell, 508 U.S. 476 (1993)
Worcester v. Georgia, 6 Peters 515 (1832)
Zelman v. Simmons-Harris, 536 U.S. 639 (2002)
Notes
CHAPTER ONE
1. Madisons essay, Federalist 10, also reflects a negative view of majority

public opinion (Wright 1961). Madison wrote that society was inevitably
divided into competing groups, that it was natural for groups to vex and
oppress each other, and that, when possible, majorities would carry into
effect schemes of oppression. Until the late 1800s, writers usually used the
term public opinion in a negative sense to describe possible or actual abuses
by majorities against minorities (Friedman 2003, Neumann 1977, Palmer 1936,
Sheehan 2004, Spier 1950).
2. . . . do you generally favor or oppose. . . . Making public school facili-
ties available after school hours for use by student religious groups? Favor
72%, oppose26%, no opinion2%. Gallup, 2/911/2001.
3. Do you favor or oppose the death penalty for mentally retarded indi-
viduals convicted of serious crimes, such as murder? Favor27%, oppose
61%, no opinion12%. Yankelovich Clancy Shulman, 6/28/1989.
4. Questions on individual cases are the most difficult to identify, since
only about half actually use the term Supreme Court in the question word-
ing. Key word searches on issues in cases identified appropriate questions that
did not include the term Supreme Court, thereby identifying over three hun-
dred additional questions. Admittedly, a number of questions may have gone
uncounted. The Roper Archive does not include privately commissioned polls
that are never released, such as those for presidential candidates, interest
groups, or litigants. Nor does it include many of the polls conducted by acade-
mics and published in academic journals, or specialized types of public opinion
research, such as focus groups.
5. For most poll-to-decision matches, pollsters only write one or two poll
questions. Unfortunately, pollsters rarely track long-term opinion changes on
Supreme Court decisions, as chapter eight later describes.
6. Supreme Court opinions sometimes use phrases or expressions that less
clearly refer to the American publics attitudes or beliefs, or that only do in an
abstract or argumentative sensefor example, the resident could not reason-
ably have expected . . . (in Florida v. Riley 1989), or so as to lower him in
the estimation of the community (in Milkovich v. Lorain Journal Co. 1990),
181
182 Notes to Chapter 1
or the public greeted it with a collective yawn (in U.S. v. Playboy

Entertainment Group 2000). These indirect mentions are not counted in
Table 1.1, and appear to be much less frequent during the Rehnquist Court
than during earlier years (Marshall 1989: 3132). Decisions that refer to
polls in the sense of elections, or to polling a jury are not counted as direct
mentions. Nor are opinions that refer to surveys of the law, historical events,
documents, or judicial decisions. Nor are references to a public forum.
7. For historical comparisons, see Marshall (1989: 3167). The number
of written decisions, per term, dropped by nearly half during the Rehnquist
Court, perhaps making the rising number of references to public opinion even
more dramatic.
8. These figures only include the justice who authored the majority, con-
curring, or dissenting opinion, and not justices who joined an opinion. The
range of direct mentions, per term, was surprisingly narrowranging only
from 1.8 for Justice Stevens to .5 for Justice Souter. The seven more liberal jus-
tices averaged 1.1 direct mentions, per term, compared to .95 for the seven
more conservative justices.
9. Many trademark decisions rely upon the personal impressions of
judges and juries, or on other tests unrelated to public opinion, such as property
interests, obscenity, or parody, rather than on specific surveys (Bartow 2004).
10. The American Association for Mental Retardations brief includes
twenty-seven nationwide and statewide poll questions from recent years.
11. For the vast majority of Supreme Court decisions, no nationwide polls
are available, according to this authors search of the Roper Poll archives, and
the justices could not cite poll results even if they were so inclined. For an inter-
esting exception, see Greenhouse (2005): 91.
12. Justice Rehnquist also wrote that surely even the suggestion (of rely-
ing on public opinion to decide controversial cases) is totally at war with the
idea of legitimacy in whose name it is invoked. The Judicial Branch derives its
legitimacy, not from following public opinion, but from deciding by its best
lights whether legislative enactments of the popular branches of Government
comport with the Constitution. The doctrine of stare decisis is an adjunct of
this duty, and should be no more subject to the vagaries of public opinion than
is the basic judicial task, at 962.
13. For a similar view, see Justice OConnors opinion in Penry v.
Lynaugh (1989, at 335), another death penalty case: The public sentiment
expressed in these and other polls and resolutions may ultimately find expres-
sion in legislation, which is an objective indicator of contemporary values upon
which we can rely. In Thompson v. Oklahoma (1988), Justices Stevens,
OConnor, and Scalia all debated contemporary public opinion toward a death
penalty sentence for a convicted murderer who was fifteen years old at the time
of the crime by reviewing state and federal laws, jury decisions, and (for Justice
Notes to Chapter 1 183
Stevens) statements of professional associations or international organizations,

but not American public opinion polls. Roper v. Simmons (2005), on the death
penalty for juveniles convicted of murder, offers a similar example.
14. For examples from the extensive law journal literature on the use of
surveys in capital punishment cases, see Bowers, Vandiver, and Dugan (1994);
Eisenberg and Wells (1993); Hughes (1993); Note: The Meaning of Life for
Virginia Jurors (1989); and Paduano and Smith (1987). For a discussion on
the standards used in surveys, see Becker (1991).
15. Ramdass v. Angelone (2000), at 3034. In the same case, at 7079,
Justice Stevens favorably cited the same polls in his dissent, as did Justice
Blackmuns opinion in Simmons v. South Carolina (1994), at 158169. Justice
Kennedy criticized low-quality polling in Florida Bar v. Went For It, Inc
(1995), dissenting, at 640641. Justice Rehnquist criticized poll question word-
ing and unexplained survey methodology in Atkins v. Virginia (2002), as did
Justice Scalia in Atkins and Roper v. Simmons (2005); see also Robinson
(2004), Becker (1991), and Sorensen and Sorensen (1953). For a similar, but
earlier critique of polls in death penalty cases, see Vidmar and Ellsworth
(1974). For a discussion of sampling theory and census adjustments, see Utah
v. Evans (2002).
16. At 628, citing Edenfield v. Fane (1993).
17. For a historical review, see Wilson (1993).
18. Nor is there any evidence that the justices either commission polls on
disputes before them or contact pollsters for poll results, a practice long
common among members of Congress and by presidents dating back to
Franklin D. Roosevelt (Eisenger 2003).
19. To Justice Kennedy, this theory allowed Internet publication of names
of convicted sex offenders in Smith and Botelho v. Doe (2003).
20. See, e.g., Austin v. Michigan State Chamber of Commerce (1990), at
694695, and Lehnert v. Ferris Faculty Association (1991), at 533, both in dis-
sent. See also Justice Scalias defense of judicial candidates discussing policy
views, in Republican Party of Minnesota v. White (2002).
21. See, e.g., Watchtower Bible and Tract Society of New York v. Village
of Stratton (2002); as well as Meese v. Keene (1987), at 485493; and NEA v.
Finley (1998), at 603, both in dissent.
22. Bartnicki v. Vopper (2001), allowing journalists to air and publish
illegally intercepted cell phone conversations, if the journalists themselves got
the information from an unknown source and did not directly participate in its
interception.
23. McConnell v. FEC (2003) used this theory to defend disclosure
requirements and requirements that broadcasters keep records of campaign
commercials.
24. For an early instance of this argument, see ONeil v. Vermont (1892),
in which three justices, in dissent, argued that a heavy fine or a fifty-five-year
hard-labor sentence for illegal liquor sales was cruel and unusual punishment.
25. This theory might seem to be the most likely to elicit mentions of spe-
cific poll questions. However, a large majority of such decisions do not report
specific poll questions, but rather rely on indirect evidence of public opinion,
such as state law, jury decisions, or even the positions of professional associa-
tions or religious groups (Robinson 2004). In part, this may be because polling
evidence is seldom available or presented in litigants briefs or in amicus briefs.
For an exception, see the survey of poll questions on the death penalty for men-
tally retarded murderers, presented in Atkins v. Virginia (2002), originally from
a brief submitted in another death penalty case, McCarver v. North Carolina
(2001). In Roper v. Simmons (2005) a 54 majority barred the death penalty
for convicted murderers who were under eighteen years old when the crime was
committed. In Roper the majority relied on objective indicators of attitudes,
particularly state laws, the trend in state laws, and the behavior of juries; no
polling evidence was reported in the majority, concurring, or dissenting opin-
ions. This third theorys growing popularity may inspire litigants to report sup-
portive poll questions, or even to commission polls, but this does not now
appear to be common.
26. A few rulings cite international public opinion, or at least opinion in
Anglo-European nations that share common cultural ties to the United States.
Opinions such as Atkins v. Virginia (2002) or Roper v. Simmons (2005) come
close to arguing that public opinion abroad is relevant to the standard of evolv-
ing public opinion. To date, such views rely more upon statutes and courtroom
practices than upon foreign polls, however. Dissenting justices vigorously criti-
cized the view that overseas practices and attitudes should affect Supreme
Court decisions; see Justice Scalias dissent in Atkins and Roper. See also the
exchange between justices in Knight v. Florida (1999). For a review of this
argument, see Comment (2005); Jackson (2005); and Young (2005).
27. Weems v. U.S. (1910) held that the Eighth Amendments ban on cruel
and unusual punishment is not fastened to the obsolete, but may acquire
meaning as public opinion becomes enlightened by a humane justice.
28. Trop v. Dulles (1957) held that expatriation for military desertion was
inconsistent with the evolving standards of decency that mark the progress of
a maturing society.
29. Well-known examples include death penalty cases, e.g., Furman v.
Georgia (1972), Gregg v. Georgia (1976), Roberts v. Louisiana (1976, 1977),
and Coker v. Georgia (1977).
30. Dickerson v. U.S. (2000), at 443.
31. At 656, in dissent, arguing against a death penalty sentence to be car-
ried out by using cyanide gas.
32. At 331335, discounting poll evidence that the death penalty for men-
tally retarded convicted murderers is unpopular, and arguing that juries and
statutes are better indicators of contemporary public opinion.
33. Thompson v. Oklahoma (1988), at 822, 849858, struck down a
death penalty sentence for a convicted murderer who was fifteen years old at
the time of the crime, partly on grounds that such a sentence did, or at least
might offend evolving standards of decency. See also Justice Stevenss dissents
in Patterson v. Texas (2003) and In re Stanford (2002) concerning the death
penalty for convicted murderers who were under age eighteen at the time of
the crime.
34. At 2126, in dissent, arguing that hanging no longer finds public accep-
tance.
35. OConnor also argued that Supreme Court justices, unlike elected
politicians, could not gain back legitimacy by winning at the polls, at 868869.
As a result, popular support, or legitimacy, once lost, would be very difficult to
recover.
36. U.S. v Olano (1993), at 732; see also Mickens v. Taylor (2002). This
rule dates back to U.S. v. Atkinson (1936), at 160: In exceptional circum-
stances, especially in criminal cases, appellate courts, in the public interest,
may, of their own motion, notice errors to which no exception has been taken,
if the errors are obvious or if they otherwise seriously affect the fairness,
integrity or public reputation of judicial proceedings. In Dretke v. Haley
(2004), Justice Kennedy, dissenting, made a similar argument regarding execu-
tive clemency for convicted felons who were actually innocent. For another dis-
cussion, see Stewart (2003).
37. Jones v. U.S. (1999), at 388389, concerning the Federal Death
Penalty Act; and Neder v. U.S. (1999), at 9, 18, 34. See also the dissenting
opinion in the stop-and-frisk ruling in Illinois v. Wardlow (2000), at 127, or
Justice Ginsburgs opinion in Banks v. Dretke (2004) concerning disclosure of
exculpatory evidence in a murder case.
38. Buckley v. American Constitutional Law Foundation (1999), at 647,
upholding a restriction on campaign spending to avoid the appearance of cor-
ruption; and Nixon v. Shrink Missouri Government PAC (2000), at 383,
390391, allowing a restriction on PAC donations to a candidate to avoid
corruption or the perception of corruption. See also McConnell v. Federal
Election Commission (2003) and Federal Election Commission v. Colorado
Republican Federal Campaign Committee (2001). This argument dates back to
the landmark decision on campaign finance, Buckley v Valeo (1976), at 2526.
39. Cleveland v. U.S. (2000), at 371.
40. Crandon v. U.S. (1989), and U.S. v. National Treasury Employees
Union (1995), at 501, in dissent.
41. U.S. v. Winstar Corporation (1996); U.S. v. Gaubert (1991); FDIC v.

Mallen (1988).
42. Legal Services Corp. v. Velazquez (2001), at 1051.
43. West Coast Hotel v. Parrish (1937), at 404.
44. Frank v. Magnum (1915), at 335, 347; Moore v. Dempsey (1923), at
86; Powell v. Alabama (1932), at 5158.
45. Weiman v. Updegraff (1952), at 191; Sweezy v. New Hampshire
(1957), at 248; Stack v. Boyle (1951), at 10; or Peters v. Hobby (1955), at 345.
46. Beuharnais v. Illinois (1952), at 261; NAACP v. Alabama (1958), at
462463; or Bates v. Little Rock (1960), at 524.
47. See also Justices Stevens and Ginsburgs argument in Smith and
Botelho v. Doe (2003) that Internet posting of sex offenders names poses a
threat to these offenders.
48. Lambert v. City and County of San Francisco (2000), at 15511552,
Justice Scalia, dissenting.
49. Lee v. Weisman (1992), at 623; McIntyre v. Ohio Elections
Commission (1995), at 357.
50. Olmstead v. L.C. (1999), at 588, 608609.
51. Lawrence v. Texas (2003).
52. National Archives and Records Administration v. Favish (2004).
53. Ironically, these are also the cases on which public opinion polls are
most often available before the decision is announced.
54. Since 1960, the rate at which state and local laws have been over-
turned averages about fifteen such laws a year. Further, although only a single
state or local law may be overturned in a decision, many other similar laws
may also fall as a result (Baum 1998a: 20204).
CHAPTER TWO
1. The last serious discussion of impeaching a sitting Supreme Court jus-

tice was during the 19681970 period for Justices Douglas and Fortas (Baum
1998a: 76). For an argument that Justice Clarence Thomas merits impeach-
ment, see Smith and Baugh (2000).
2. In addition to Chief Justice Rehnquist, Justices Scalia, Breyer,
Kennedy, and OConnor authored books on legal issues or legal history
(Abraham 1999:388410).
3. For a review of the reaction of Congress, the president, lower courts,

state legislatures, or governors to Supreme Court decisions, see Baum (1998a:
246261).
4. Historically, justices were also much more often male, and less often
an ethnic or religious minority (Perry 1991); see also chapters five and six.
5. By the end of the Rehnquist Court in 2005, only Justice OConnor
had ever run for (and won) an elective office, as a state senator in Arizona.
6. During the Rehnquist Court, six justices (Blackmun, Powell, Marshall,
White, Brennan, and OConnor) retired from the Rehnquist Court after an
average of twenty-five years. The first five of these justices averaged eighty-one
years old at the time of their retirement, and then lived another six years, on
the average, before dying. Chief Justice Rehnquist died on September 3, 2005,
after serving thirty-three years on the Court.
7. A February 1974 National Opinion Research Center (NORC) Poll
reported that a 66-to-31% majority disapproved of the ruling that no state or
local government may require the reading of the Lords Prayer or Bible verses
in public schools. More recently, in a NORC survey conducted between
August 2004 and January 2005, a 62-to-35% majority opposed the ruling.
8. A June 1989 Yankelovich Clancy Shulman poll reported that a 67-to-
29% majority favored passing a constitutional amendment which would make
it illegal to burn the American flag. A June 2005 Gallup Poll reported a 55-to-
42% majority favored a constitutional amendment that would allow Congress
and state governments to make it illegal to burn the American flag.
9. For a review of findings from lower federal courts and state courts, see
Caldeira (1991: 316319).
10. The public mood index also inspired attempts to create a general
index of public support for the Supreme Court (Durr, Martin, and Wolbrecht
2000; Mondak and Smithey 1997; for an earlier report, see Caldeira 1986).
11. Robert Dahl (1957) first popularized the argument that the Supreme
Court responds to public opinion through presidential elections and new
appointments. Dahls argument was later developed by Funston (1975) and
criticized by Casper (1976).
12. Flemming and Wood (1997) report that most justices, not just moder-
ate justices, respond to public mood changes, and that large changes in the
public mood have a modest direct impact on Supreme Court decision-making
for most, but not all types of cases, a result also reported by Link (1995).
Unlike Mischler and Sheehan (1993, 1994, 1996), Flemming and Wood (1997)
and Stimson, MacKuen, and Erikson (1995) report that the public mood typi-
cally affects the Court within a short time lag of about a year or two, rather
than over five to seven years. Hurwitz, Mischler, and Sheehan (2004) argue
that public opinion had a direct impact on Supreme Court decisions from the
early 1950s to about 1980, roughly at a lag of two to four years, but found no
evidence that the Courts rulings themselves affected the public mood.
13. Two-thirds of changes in the public mood are felt within 2.3 years at
the Supreme Court, and virtually all the impact of changes in the public mood
are felt within about seven years, according to Stimson, MacKuen, and Erikson
(1995: 558559).
14. For examples in Germany or Canada, see Brettschneider (1996) or
Petry (1999). For a study of the U.S. Supreme Courts abortion policies, see
Vitiello (1999).
15. During the Rehnquist Court this represents about 5% of all per
curiam and full, written opinions.
16. Because the pairwise method does not require that poll items be
repeated, it also allows an analysis going back to the origin of modern polling
during the mid-1930s. Unfortunately, the practice of repeating identically
worded poll items was uncommon until the 1950s. The public mood index
(Stimson 1999) begins with data in the 1950s, as do other trend studies (Devine
1970). For studies tracking repeat poll items before the 1950s, see Page and
Shapiro (1992) and Davis and Smith (1980).
17. When the phrase Supreme Court appears in a poll question describ-
ing a specific decision, it is typically used to describe a recently decided case, for
example, (T)he Supreme Court recently decided. . . .
18. In computing the poll margins, the poll question closest in time to the
decision was usedexcept in the case of inconsistent poll results, for which all
available polls were averaged to compute the poll margin.
19. Sources other than U.S. Reports or Supreme Court Reporter or slip
opinions were especially useful in identifying matches involving denials of cer-
tiorari, denials of applications for a stay, or denials of a writ of habeas
corpusall of which can seldom be identified from reading U.S. Reports, the
Supreme Court Reporter, or slip opinions.
20. Princeton Survey Research Associates, 6/2930/2000.
21. Gallup Poll, 8/1213/1997.
22. Princeton Survey Research Associates, 6/2930/2000.
23. Bush v. Gore (2000) is the Rehnquist Courts most controversial,
highly publicized unclear decision, due to multiple polls with closely divided
and inconsistent results.
24. Thus, Seling v. Young (2000) was not counted as an additional match,
since Kansas v. Hendricks (1997) was previously counted on the issue of civil
incarceration of sex offenders.
25. In this case, the carrying and using offense was invoked for transport-
ing the firearm in a locked glove compartment or the trunk of a car during a
drug trafficking offense.
26. However, the Rehnquist Court never so ruled; see Seling v. Young
(2001).
27. Justice Stevens wrote most often on the meaning of cert denials; see
Singleton v. Commissioner of Internal Revenue, 439 U.S. 940 (1978). For a
similar, earlier argument, see Justice Frankfurters views in Maryland v.
Baltimore Radio Show, 338 U.S. 912 (1950).
28. In DeBoer v. Schmidt (1993), for example, Justices Blackmun and
OConnor dissented from the denial of an application for stay, but did not
clearly indicate how they would have voted on the merits of the case.
29. The death penalty appeal of Texas inmate Karla Faye Tucker came to
the Court as an application for a stay of execution, temporary restraining
order, and preliminary injunction, in In Re Tucker (1998), Tucker v. Texas
(1998), and Tucker v. Texas Board of Pardons and Paroles (1998). Ms.
Tuckers appeals were denied as a petition for writ of habeas corpus or a writ
of certiorari, and she was then executed.
30. The two justices often dissented from denials of certiorari in death
penalty cases, writing that Adhering to our views that the death penalty is in
all circumstances cruel and unusual punishment prohibited by the Eighth and
Fourteenth Amendments . . . we would grant certiorari and vacate the death sen-
tences in these cases. See, e.g., Adams v. Florida (1978).
31. Justices Breyer and Stevens used dissents from denials of certiorari to
describe their views of lengthy delays in death penalty appeals; see Lackey v.
Texas (1995); Elledge v. Florida (1998); or the debate among Justices Stevens,
Thomas, and Breyer in Knight v. Florida (1999). See also Justice Stevenss com-
ments in Barber v. Tennessee (1995), or the debate between justices in Coleman
v. Balkcom (1981) or McCray v. New York (1983), all cert denials. Justice
Scalia, Thomas, and Rehnquists dissent from denial of certiorari in
Tangipahoa Parish Board of Education v. Freiler (2000) was considered suffi-
ciently clear to indicate how the three justices would have voted.
32. These eighty-eight full, written decisions were adjusted to the figures
reported in Epstein et al. (2003: 8085) for five major (combined) issues: crim-
inal cases, economic cases, federalism and intergovernmental relations cases,
civil liberties cases, and miscellaneous cases.
33. If these figures are reweighted against the number of full, written deci-
sions reported in the annual reports of the Harvard Law Review, the percent-
age of consistent, inconsistent, and unclear decisions would be 64%, 32%, and
4%, respectively.
34. Gallup Poll, 8/1213/1997.
35. For the original, uncollapsed, but absolute poll margins, the point-bis-
erial correlation is an insignificant .07.
36. Justice Black and Douglass dissent in Dennis v. U.S. (1951), at 580,
describes the crisis-times thesis: Public opinion being what it is now, few will
protest the conviction of these Communist petitioners. There is hope, however,
that in calmer times, when present pressures, passions, and fears subside, this
or some later Court will restore the First Amendment liberties to the high pre-
ferred place where they belong in a free society.
37. If any of the five top concerns in the Gallup Polls most important
problem question were used, the results would also be in the predicted direc-
tion, but statistically insignificant. The Rehnquist Court was consistent with
public opinion in 69% of decisions involving one of the top five national con-
cerns, versus 59% of cases in which the issue raised was not.
38. Chapter three discusses at greater length the Rehnquist Courts chang-
ing theories of federalism and intergovernmental relations.
39. For the history of this doctrine and a description of the Carolene
Products lawsuit, see Abraham (1982), Miller (1987), Rogers (1999), and
Yackle (1994).
40. This argument also assumes that local, state, or federal laws chal-
lenged at the Supreme Court restrict, not support a particular groups claims,
and that nationwide public opinion also opposes the groups claim.
41. When public opinion was politically conservative on these claims and,
hence, opposed the claim, the Rehnquist Court agreed with public opinion
68% of the time, versus supported the claim only 32% of the time. By contrast,
when public opinion was politically liberal, and, hence, supported the claim,
the Rehnquist Court agreed with public opinion 56% (as well as the fundamen-
tal freedoms claims) of the time. Overall, public opinion was conservative
(hence, opposed to the claim) about twice as often (69% versus 31% of the
time) as it was liberal (or supportive of the claim) during the Rehnquist Court.
42. Chapter seven reviews these studies at length.
43. Testing another possible polling artifact, the average percentage of
poll respondents who agreed with the Courts decision was also higher (55%,
versus 47%) when the matching poll did not mention the Supreme Court, com-
pared to when it did.
44. The terms liberal and conservative are defined according to the
meaning in Rathjen and Spaeth (1979).
45. The Supreme Courts docket is now nearly completely discretionary,
since Congress in 1988 virtually eliminated the Courts mandatory docket
(Baum 1998a).
46. Friedman (2004) and Merrill (2003) put the cutting point slightly
later, in 2004/2005 term.
CHAPTER THREE
1. Of the fourteen justices serving during the Rehnquist Court, Justices

Marshall Brennan, and Blackmun most often voted in a liberal direction
doing so 96%, 94%, and 94% of the time, respectively, in the votes counted
here. Justices Scalia, Rehnquist, and Thomas least often voted in a liberal direc-
tiondoing so only 18%, 20%, or 18% of the time, respectively. For a review
of the justices voting records by appointing president, see Carp and Stidham
(1998: 93100).
2. Princeton Survey Research Associates, 6/247/8/2003.
3. Roper v. Simmons (2005) illustrates both the impact of vote switching
and new appointments. In overturning two 1989 decisions that allowed the
execution of convicted killers under age eighteen, Justice Kennedy switched his
vote, thereby providing the decisive vote in Roper. The remaining four justices
involved in both decisions (Justices Rehnquist, Scalia, OConnor, and Stevens)
voted in the same way as they had sixteen years earlier. Justices Souter,
Thomas, Ginsburg, and Breyer split their votes in the same mix as the four jus-
tices who voted in 1989, but had since left the Court.
4. These percentages differ from those in chapter two because only full,
written opinions with a clear ideological content and a clear poll majority (or
plurality) are included here. Votes that reach 90 are still counted, although
they have reached the maximum coalition size.
5. Figures 3.1 and 3.2 are asymmetric because decisions occurred over
multiple terms, and because the number of liberal-versus-conservative votes dif-
fered over time. These calculations necessarily disregard the possibility that one
or more additional conservative (or, alternatively, liberal) votes would result in
certiorari being granted to different appeals (Caldeira, Wright, and Zorn 1999).
6. Figure 3.2 includes all full, written opinions in which vote totals can
be computed, including four decisions not included in Marshall (1989).
Percentages in Figure 3.2 are reweighted by caseload for the Hughes, Stone,
Vinson, Warren, and Burger Courts.
7. Civil confinement allows a prisoner convicted of certain crimes
(here, violent sexual offenses) to be held for lengthy, even indefinite periods
even after he or she would otherwise be released.
8. Most (67%) of these cases were from a federal appeals court, while
28% were from a state court. The remaining cases came directly from a federal
district court, or else were appeals combined from different types of lower
courts.
9. A decision to affirm was counted as upholding a lower court decision.
Decisions to reverse, reverse and remand, vacate, or vacate and remand were
counted as overturning a lower court ruling. Figures reported in this section
exclude a few decisions of original jurisdiction (and in which there was therefore
no lower court decision) or in which different lower courts gave conflicting

decisions.
10. Rates for individual justices are computed only if justices had at least
twenty votes coded here; accordingly, Justice Powell is excluded.
11. The Supreme Court now rarely accepts an appeal from a district court
decision without an appeals court hearing, but at one time this practice was
fairly common.
12. These figures are unweighted. If the figures, per Court, were
reweighted for caseload, the +7% figure would become +9% (or 61% consis-
tent for the Hughes, Stone, Vinson, Warren, and Burger Courts, versus 52%
consistent for lower courts). By Court, the poll correction rates are +58% for
the Hughes Court, +8% for the Stone Court, 9% for the Vinson Court,
37% for the Warren Court, and +10% for the Burger Court. As a caveat,
these rates are based on a small number of decisions, per Court, prior to the
Burger Court.
13. Well-known examples include Will v. Michigan Department of State
Police (1991), Gregory v. Ashcroft (1989), Idaho v. Couer dAlene Tribe of
Idaho (1997), and Alden v. Maine (1999).
14. Well-known examples include U.S. v. Lopez (1995), City of Boerne v.
Flores (1997), New York v. U.S. (1992), Printz v. U.S. (1997), Seminole Tribe
of Florida v. Florida (1996), and Jones v. U.S. (2000).
15. Well-known examples include College Savings Bank v. Florida
Prepaid Postsecondary Education Expense Board (1999), U.S. v. Morrison
(2000), Vermont Agency of Natural Resources v. U.S. ex rel. Stevens (2000),
and Kimel v. Florida Board of Regents (2000).
16. Well-known decisions that favored the federal government in federal-
ism cases include Pennsylvania Department of Corrections v. Yeskey (1998),
Reno v. Condon (2000), AT&T Corp. v. Iowa Utilities Board (1991), and
Tennessee v. Lane (2004).
17. See, for example, Justice Stevenss dissent in Kimel v. Florida Board of
Regents (2000), at 653654.
18. This figure slightly exceeds 100% because a few cases involve a con-
flict between a federal and a state statutefor example, Reno v. Condon
(2000), involving a conflict between the federal Drivers Privacy Protection Act
(which forbade the sale of drivers license personal information without the
drivers consent), and South Carolinas policy of selling the data bases.
19. These figures exclude a few unclear cases of inconsistent or evenly
divided polls, and a few cases in which federal laws and policies disagree with
state/local laws and policies.
20. The five previous Courts since the mid-1930s agreed with public opin-
ion 63% of the time, versus 62% of the time under a pure judicial restraint
model, but only 38% of the time under a pure judicial activism model. For
these five Courts, this pattern would also result if only considering federal laws
and policies (70% actually consistent, versus 68% under pure judicial restraint,
but only 32% under pure judicial activism), or only state and local laws and
policies (59% actually consistent, versus 56% under pure judicial restraint, but
only 41% under pure judicial activism).
21. By type of decision, some 62% percent of the Rehnquist Courts fun-
damental freedoms decisions agreed with public opinion, a figure that would
have increased to 73% under pure judicial restraint. For economic cases the
actual figure was 53% consistent, versus 65% under pure judicial restraint.
22. The Harvard Law Review figures include a few per curiam opinions
that have sufficiently lengthy legal reasoning to be counted as a full, written
opinion. The Supreme Court Compendium (Epstein 2003) puts the number of
signed opinions at 145 during the 1986/1987 term, falling steadily to 74 in the
1999 term and 76 in the 2001 term.
23. According to the Harvard Law Review, there were 6,590 cases denied,
dismissed, or withdrawn during the final (2004/2005) term of the Rehnquist
Court.
CHAPTER FOUR
1. Ironically, filing an amicus brief against certiorari increases the

chances that the Court will grant certiorari. Groups that do not want the Court
to accept a case may forego filing an amicus at the cert stage. Filing amicus
briefs is also less frequent at the cert stage than at the full merits stage (Caldeira
and Wright 1988; McGuire 1994; Stern, Gressman, and Shapiro 1986).
2. Seventy-eight amicus briefs were filed in Webster v. Reproductive
Health Services (1989), and fifty-seven were filed in Regents of University of
California v. Bakke (1978) (Epstein 1993: 646). Eighty-three briefs supported
the University of Michigans affirmative action policies in Grutter v. Bollinger
and Grantz v. Bollinger (2003) (Taylor 2004).
3. The justices often cite amicus briefs in their opinions. On the average,
the Rehnquist Courts ten most senior justices did so in 62% of their opinions,
versus only 29% for the four least senior justices (Souter, Thomas, Ginsburg,
and Breyer) (Epstein 2003: 690).
4. Interest groups, particularly economic-based groups, occasionally limit
the number of their appeals when the group perceives the Court as hostile,
thereby limiting the number of unfavorable precedents until new appointments
lead to a more favorable Court (Baird 2001a).
5. In a few case studies, adverse publicity, negative public opinion, or the
fear of financial retaliation dissuade groups from filing amicus briefs (ONeill
1985: 151156), but some groups pursue a lawsuit despite potentially negative
consequences (Lawrence 1990).
6. Because many individual groups took only a few positions, this coding
is for types of groups, not individual groups. The twelve types of groups (and
better-known examples in each category) include: (1) civil liberties, free speech,
and media (ACLU, People for the American Way); (2) states and local govern-
ments (typically state attorneys general or individual cities); (3) pro-civil rights
(NAACP, MALDEF); (4) education (NEA, National School Board Association,
AAUP); (5) corporations, businesses, and trade associations (Chamber of
Commerce, individual corporations); (6) bar associations (ABA); (7) labor
unions (AFL-CIO); (8) feminist, pro-gender equality, and privacy (Planned
Parenthood, NOW); (9) liberal religious (BNai BRith, National Council of
Churches, American Jewish Congress); (10) Catholic (U.S. Catholic
Conference); (11) conservative (NRA, Concerned Women for America, Family
Defense Council, Eagle Forum LDEF); (12) pro-law enforcement (Americans
for Effective Law Enforcement, Fraternal Order of Police).
A group was counted as taking a position either by being directly involved
as a party in a case, or (much more often) by filing an amicus brief. If different
groups in the same category took opposing positions (e.g., when state govern-
ments, typically represented by their respective attorneys general, took different
positions in amicus briefs), no position was coded for the group. Different
groups in the same category seldom took different positions, and intragroup
conflict occurred in less than three percent of all cases.
As a caveat, state, counties, and cities are not usually viewed as interest
groups, although their associations might be. Because units of state and local
governments and their associations often participated, any of these were
counted in the category of state and local governments. States, counties, and
cities participated as a direct party or as an amicus much more often than did
associations of these governments.
Two types of interest groups (states/counties/cities and legal groups) filed
enough lawsuits or amicus briefs to be included here, but not in a previous
study (Marshall 1989). Conservative groups and law enforcement groups are
broken out separately here, since both have become much more active since the
early 1980s.
7. From the 1950s through the mid-1980s, the solicitor general won
75% of the time when filing an amicus brief (Segal 1988). This rates held up
even when controlling for the lower courts decision and case facts (Segal and
Reedy 1988), new appointments to the Supreme Court (Segal 1990), or the
opposing attorneys experience (McGuire 1993, 1995). During the early
Rehnquist Court, the solicitor general averaged 66% success (through the 1996
term) and the federal government averaged 61% success on all cases as a direct
party (through the 2001 term). By contrast, states averaged 56% success
through the 2001 term (Epstein et al. 2003). By comparison, the solicitor gen-
eral won 69% of decisions in this book, versus 58% for state and local govern-
mentsfigures very close to those in Epstein (2003).
8. These figures exclude the few cases in which public opinion itself was
unclearthat is, cases of inconsistent or closely divided (at the .05 level) poll
results.
9. Catholic and pro-life groups alone agreed with American public opin-
ion more often than did the federal government, doing so in 86% of their
twenty-two positions. Before the Rehnquist Court, however, Catholic and pro-
life groups only agreed with public opinion 50% of the time (Marshall 1989:
96).
10. Stated otherwise, American public opinion won 67% of the time, on
the average, when it agreed with a groups position. Public opinion won 59%
of the time, on the average, when it disagreed with a groups position.
11. These twelve types of groups took 104 positions agreeing with, but
112 positions disagreeing with the solicitor general. In the remaining instances
either the group or the solicitor general (or both) took no position.
12. From 1935 until 1986, the federal government was also very success-
ful, whether or not its position agreed with public opinion. When the federal
government was involved as a party or as an amicus, it won 76% of the time if
its position agreed with public opinion, versus 63% of the time when its posi-
tion disagreed with public opinion. For states, counties, and cities, the percent-
age of wins were 58% and 46%, respectively.
13. These groups positions were ideologically liberal over 96% of the
time; the very rare exceptions include the ACLUs conservative position in
Hill v. Colorado (2000) and McConnell v. Federal Election Commission
(2003).
14. The relatively small sample of full, written and per curiam decisions
and the large number of interest groups (twelve, plus the solicitor general) dic-
tated that group positions be entered one by one, rather than jointly.
15. The equation is .122 + .758 (solicitor generals position, with a stan-
dard error of .327, significant at .02). 2LLR = 108.7329; Cox & Snell R-2 =
.06; Nagelkerke R-2 = .09. The number of correctly predicted decisions rises
from 60% to 66%, or an improvement over the base rate of 10%.
16. This figure excludes the few cases with inconsistent or evenly divided
polls.
17. As a comparison, this is roughly the impact of the solicitor generals
filing an amicus on behalf of a respondent; see Kearney and Merrill (2000:
804810) who also report that the ACLU, the AFL-CIO, and the states enjoy
modest success in their amicus filings.
CHAPTER 5
1. Gallup Poll for Phi Delta Kappa, 5/216/9/1993.

2. This chapter is based on justices votes in full, written decisions (when
individual votes are reported), and per curiam votes, if reported. In denials
from certiorari or denials of habeas corpus, only in rare cases does a justice dis-
sent with an opinion sufficiently clear to indicate how he or she would have
voted. For multiple authored opinions, each justices opinion was examined
separately; the justices often wrote separate opinions (Maveety 2005), even
more often than they had previously done as federal appeals judges (Gerber and
Park 1997).
3. For the five earlier Courts since the mid-1930s, some 59% of individ-
ual votes were consistent with poll majorities or pluralities, while 41% were
not.
4. Information in this chapter comes from a variety of sources, particu-
larly Abraham (1999), Baum (1998a), Biskupic and Witt (1996), Epstein et al.
(1996, 2003), Gottlieb (2000), Lamb and Halpern (1991), and Schwartz
(1997).
5. By 2003, the Gallup Poll reported that the percentages of Americans
who identified as Republicans or as Democrats was tied, at 45% apiece.
6. From the 1930s to the mid-1960s, the figure was 59% consistent for
both appointees of Republican presidents or Democratic presidents.
7. Justice Rehnquist, initially appointed by President Nixon, was
included in the early Republican category.
8. In Marshall (1989: 109) only small differences appeared for five
groups of presidents, by era and political party; the percentage of consistent
votes by group of justices ranged only between 57% and 65%.
9. The three groups include: three liberals (Brennan, Marshall,
Ginsburg), five moderates (White, OConnor, Kennedy, Souter, and Breyer),
and six conservatives (Thomas, Powell, Blackmun, Scalia, Rehnquist, and
Stevens). The original scale ranged from +1 (most liberal) to 1 (most conserv-
ative); the cutoffs in Table 5.4 were: liberal (+.35 or higher), moderates (+.34 to
.34), and conservative (.35 or lower).
10. The justices were divided according to the groups in Table 5.3, with
the two moderate categories combined.
11. In the common pattern, justices moved to Washington, DC, prior to
their nomination. The justices were classified as follows: East (Brennan,
Marshall, Rehnquist, Scalia, Thomas, Souter, Breyer, Ginsburg, and White);
Midwest (Blackmun and Stevens); South (Powell); and West (OConnor and
Kennedy).
12. Justices White, Rehnquist, OConnor, Scalia, Thomas, and Breyer

were counted as moving between regions.
13. For the nine justices, the midpoint of assets, per justice, was a break-
point of $815,000. Justices Ginsburg, Breyer, OConnor, Stevens, and Souter
ranked as above-average; Justices Scalia, Rehnquist, Kennedy, and Thomas
ranked as below-average (Epstein et al. 2003). As a caveat, some assets, such as
houses or retirement accounts, are not included in this tally.
14. Justices Brennan, Powell, Stevens, Rehnquist, and White all served in
the military during World War II. Justice OConnor served as a civilian attor-
ney for the military. Justices Breyer and Kennedy served in the army reserve or
national guard.
15. Justices Blackmun and Ginsburg clerked for federal judges. If they are
combined with justices who clerked for earlier justices (White, Stevens,
Rehnquist, and Breyer), justices with a federal clerkship experience voted con-
sistently with public opinion 55% of the time, an identical rate as for the other
eight justices.
16. Only Justice Powell was in private practice at the time of his appoint-
ment to the Court.
17. If Justice Rehnquist were so considered (Baum 1998a: 4647), then
presidential confidants were consistent with the polls 70% of the time, versus
52% for other justices.
18. The literature on Senate confirmation votes for Supreme Court nomi-
nees is extensive. For examples, see Caldeira and Wright (1998); Cameron,
Cover, and Segal (1990); Epstein et al. (2006); Guliuzza, Reagan, and Barrett
(1994); Maltese (1995); Melone (1991); Melone, Morris, and Pufong (1992);
and Reynolds (1991). For a discussion of whether increased Senate scrutiny of
candidates weakens the quality of justices, see Comiskey (2004).
19. These figures include the twenty-six justices who served from the mid-
1930s, but not on the Rehnquist Court.
20. The bivariate difference was 4% for chief justices, but in a regression
equation, the difference rose to 11% for Chief Justices Hughes, Stone, Vinson,
Warren, and Burger.
21. Biskupic and Witt (1996:841) cite Chief Justice Rehnquists role in
persuading Congress to grant a 1990 pay raise for federal judges. Eskridge and
Frickey (1994, at note 196) cite Rehnquists opposition to the growing number
of federal crimes. For a recent review of the chief justices role as leader of the
judicial branch, see Wheeler (2005). Other current or past justices may also
seek to influence policy-makers; for an example concerning restructuring the
Ninth Circuit Court of Appeals, see White (1999: 421424). For a discussion
of Chief Justice Tafts earlier success in getting Congress to pass the Judiciary
Act of 1925, increasing the Supreme Courts certiorari power, reorganizing the
federal courts, and appropriating funds for a new Supreme Court building, see
White (1976: 200206) and Baum (1998a: 1415).
22. Prior to the Rehnquist Court, Justice Brennan ranked about the
middle of justices in agreeing with public opinion (Marshall 1989: 106). His
earlier 55% rating fell to 35% during the Rehnquist Court.
23. A composite ranking of the justices (Epstein et al. 1992) yielded no
significant results, at least since the mid-1930s. The pearson product-moment
correlation between a composite ranking and the percentage of consistent votes
was insignificant. If justices rated as failures are compared to the remaining
justices, then low-ranked justices are only a little less likely to vote consistently
with public opinion, compared to other justices. The figures are 55% or 56%
consistent for failures, compared to 60% for other justices, broken down by
the Blaustein and Mersky ranking or the Schwartz ranking, respectively
(Abraham 1999: 369372; Marshall 1989: 106). As a caveat, none of the
Rehnquist Courts fourteen justices are rated as failures.
24. Baum (1998a:135) lists Justices Scalia, Souter, Ginsburg and Breyer as
the most assertive on the current Court at asking questions. If these four jus-
tices are compared to the remaining five justices at that time (Rehnquist,
Stevens, OConnor, Kennedy, and Thomas) the percent consistent with public
opinion would be 55% versus 58%, respectively.
25. By another test of intellectual leadership, some justices publish much
more in law journals or in book form than do other justices. Four of the
Rehnquist Courts fourteen justices published two or more books (through-
out their entire lives, through fall 2005), and can be considered as publishing
leaders. These four justices (Rehnquist, Scalia, OConnor, and Breyer) agreed
with majority public opinion 61% of the time (over their entire time on the
Court), versus 52% for the remaining ten justices, a difference significant at
the .01 level.
26. This variable is computed by averaging the number of significant
opinions (Epstein et al. 2003). Those with an average of 1.0 or more significant
opinion, per term, were scored as above average; those with fewer were scored
as below average, both for the justices entire careers (through 1996) or solely
for the Rehnquist Court.
27. Justice Scalia was 75% consistent during his first four years, versus
51% in later yearsan early term effect of +24%. For Justices Kennedy,
Souter, Thomas, Ginsburg, and Breyer, the early term effects would be +3%,
+6%, +21, 4%, and 1%, respectively.
28. No such effect appears prior to the Rehnquist Court, however
(Marshall 1989: 117). Since new justices are typically the most precedent-ori-
ented (Hurwitz and Stefko 2004), these results are unexpected. For a review of
freshman effects, see Hagle (1993) and Wood et al. (1998).
29. For Scalia, Kennedy, Thomas, and Breyer the figures were +18%,
+1%, +12% and +5%, and for Justices Souter and Ginsburg, 1% and 4%,
respectively.
30. Replacement effects occurred for five of the six newly appointed jus-
tices, all except Ginsburg. The size of these effects varies considerably; by pair-
ings, replacement effects are +14% (Scalia for Burger), +12% (Kennedy for
Powell), +30% (Souter for Brennan), +36% (Thomas for Marshall), 27%
(Ginsburg for White), and +4% (Breyer for Blackmun).
31. Excluding the justices serving on the Rehnquist Court in 2005, thirty-
one of ninety-eight justices served twenty years or more. The average length of
service varied from ten years (those appointed during the 1700s) to twenty
years (those appointed from 1800 to 1850).
32. These figures also include a justices service before the Rehnquist
Court.
33. Justices Brennan, Marshall, Blackmun, and OConnor dropped in the
frequency with which their votes agreed with public opinion (20%, 18%,
32%, and 4%, respectively); Justices Whites level of consistency rose (by
4%); Justices Rehnquist and Stevens figures did not change.
34. V. O. Keys three seminal articles (1952, 1955, 1959) describe realign-
ments as critical historical events that destabilize an existing party system and
create a new balance between the political parties. The Great Depression and
the New Deal period between 1929 and 1940 are usually cited as a dramatic
(or critical) realignment that made the Democratic Party the majority party
in presidential and congressional elections for the next three decades. Many
political scientists consider realignments to have occurred about 1828 to 1830
(or, alternatively, about 1836 or 1838) (Nardulli 1995; Silbey 1991), 1856 to
1860, 1896, and 1928 to 1932. Gates (1992) argues that the Court behaves
differently during and immediately before realignments. For recent critiques,
see Aldrich (1995), Carmines and Stimson (1989), Edsall and Edsall (1992),
Mayhew (2002), Nardulli (1995, Paulson (2000), Rosenof (2003), and Shafer
(1991).
35. Some state party systems can survive national realignments for long
time periods (Gimpel 1996). Shafer (1991: 3784) argues that the idea of polit-
ical realignments is less useful than electoral orders, and that political scien-
tists seldom agree upon a contemporary realignment.
36. To Burnham (1989, 1991), the Reagan era was only an incomplete
realignment.
37. For example, Justices Ginsburg and Breyer, both Democratic
appointees of a Democrat president, were the next justices after the Reagan-
Bush presidencies; the next two justices (Roberts and Alito, both Republicans)
joined the Court after a lag of eleven terms.
38. The linear regression model is: .536 (constant) + .112 (if moved
between regions) .099 (if previously served as a judge + .081 (if described as a
political moderate at confirmation time). For this model, the R-2 is .81, with an
adjusted R-2 of .75. Region is significant at .01; prior judicial experience and
ideology are significant at .05.
39. Arguably, a case can be made for chief justice status. After becoming
chief justice, Justice Rehnquists rate of agreement with public opinion jumped
from 60% to 69%. Statistically, however, the chief justice variable fell victim to
the small number of chief justices (only one per Court) and to multicollinearity
(Chief Justice Rehnquist also moved between regions and was never a judge
before his appointment to the Supreme Court).
CHAPTER SIX
1. For the distinction between symbolic and substantive representative,

see Pitkin (1967) and Edelman (1967, 1971); for somewhat recent discussions,
see Mansbridge (1991, 2003) and Dovi (2002).
2. Admittedly, since only a fourth of adult Americans are now white,
male, and Protestant, the term symbolic representation itself may seem
increasingly outmoded.
3. Research hints, but does not clearly establish, that symbolic nominees
fare slightly better than other nominees; see Caldeira and Smith (1996);
Caldeira, Hojnacki, and Wright (2000); Overby and Brown (1995); Overby,
Henschen, Walsh, and Strauss (1992); Ruckman (1993). President George W.
Bushs failed nominee, Harriet Miers, shows that even a symbolic nominee may
not be successful, however. President George W. Bushs minority nominees to
the federal bench may have been aimed at reducing hostility among Senate
Democrats (Harwood and Greenberger 2001). For an overview, see Goldman
(1997) and Yalof (1999).
4. Because the justices are so much older, more affluent, and better-edu-
cated than average Americans, few accounts consider symbolic representation
in terms of age, income, or education (but see Marshall 1993).
5. After his 1991 confirmation, Justice Thomas rejoined the Catholic
Church in 1996, thereby giving the Court its first non-Protestant majority
(Perry and Abraham 1998).
6. CBS News/New York Times Poll, 9/35/1991.
7. Yankelovich Partners Inc., 5/2627/1993.
8. When Ginsburg was described as a Jewish woman, the percentage of
Americans reporting more favorable attitudes dropped from 17% to 6%, while
those reporting less favorable attitudes dropped from 6% to 3% (Tarrance
Group & Mellman, Lazarus & Lake, 6/1516/1993, registered voters only).
9. Hart and Teeter, 6/1418/2000.

10. Quinnipiac University Polling Institute, 2/263/3/2003. For a black,
the percentages were 24% and 31%, respectively, and for a Hispanic, 23% and
29%, respectively.
11. Gallup Poll, 10/2830/2005. For a black justice, the percentages were
5% and 26%, respectively, and for a Hispanic justice, 3% and 26%, respec-
tively.
12. CBS News/New York Times, 9/35/1991.
13. Yankelovich Clancy Shulman, 7/1718/1991.
14. ABC News/Washington Post, 7/1/1991.
15. CBS News/New York Times, 9/35/1991.
16. In 1991, some 18% of Americans said that President Bush should
make a special effort to find a qualified woman, and 30% of Americans said
that President Bush should make a special effort to find a qualified minority
candidate to replace Justice Thurgood Marshall. Los Angeles Times Poll,
6/2830/1991.
17. For two discussions, see Dovi (2002) and Manbridge (1999).
18. CBS/New York Times, 9/35/1991.
19. Gallup Poll, 9/23/1987.
20. CBS/New York Times, 7/2122/1987. In September 2005, a CBS/New
York Times Poll reported that 36% of Americans said the Senate should con-
sider only a nominees legal qualifications and backgrounds, but 54% said the
Senate should also consider how the nominee might vote on major issues the
Supreme Court decides.
21. The Rehnquist Courts most prominent symbolic justice was Sandra
Day OConnor. For studies reaching mixed conclusions on the impact of
gender on her decision-making, see Biskupic (2005), Maveety (1996), Van
Sickel (1998), and Zelnick (2004).
22. Justice Ginsburgs comments are cited in Martin (1993, 126): I have
no doubt that women, like persons of different racial groups and ethnic origins,
contribute (to) . . . a distinctive medley of views. . . . A system of justice will be
the richer for diversity of background and experience. It will be the poorer, in
terms of appreciating what is at stake and the impact of its judgments, if all of
its members are cast from the same mold; see also Davis (1989).
23. Some scholars argue that gender differences appear primarily in a
judges legal reasoning; see Allen and Wall (1993); Gilligan (1982) and cf.
Davis (1992, 1993) and Van Sickel (1998).
24. Justices Ginsburg and OConnor averaged 64% voting agreement on
all full opinions, according to the Harvard Law Review. By comparison,
Justices Ginsburg and Breyer, both Jewish, had a much higher (79%) average
voting agreement score during their joint tenure on the Rehnquist Court, and
voted with each other at above-average rates. In the decisions sampled in this
book, Justices Ginsburg and OConnor had a 61% voting agreement rate,
while Justices Ginsburg and Breyer had an 85% voting agreement rate.
25. During his last two terms, Justice Marshall voted with Chief Justice
Rehnquist only 43% of the time, and voted with Justice Scalia only 41% of the
timein both instances Justice Marshalls lowest voting agreement rates with
any justice. By comparison, during his first two terms, Justice Thomas voted at
above-average rates with both Justice Rehnquist (averaging 81% agreement)
and Justice Scalia (averaging 86% agreement); see also Epstein, Hoekstra,
Segal, and Spaeth (1998).
26. During his last two terms, Justice Brennans average voting agreement
with Justices Scalia and Kennedy was only 46% and 49%, respectivelyboth
well below average. During those same two terms, Justice Scalia and Kennedy
voted together at well above average rates (85%), according to the Harvard
Law Review.
27. Even so, the two Jewish justices agreement was not complete, even on
questions of religious observances in public schools. Justices Breyer voted to
allow after-hours religious clubs in public schools, while Justice Ginsburg voted
against the practice in The Good News Club v. Milford Central Schools (2001).
28. For an account of federal judges backgrounds by gender or race, see
Slotnick (1984).
29. No comparisons for Hispanics or for those without religious prefer-
ences are presented, since no such justices served on the Rehnquist Court, and,
until recently, pollsters did not often break down responses by these categories.
30. No Independents sat on the Rehnquist Court; this category is provided
only as a comparison.
31. For an analysis of whether presidential appointees to the Supreme
Court vote together, see Lindquist, Yalof, and Clark (2000).
32. Justice Thomas, a non-Catholic at the time of his Senate confirmation,
was classified as Catholic, due to his Catholic upbringing and his rejoining the
Catholic Church in 1996. Justices Brennan and Powell were classified as
Democrats, since that was their party affiliation, although not that of their
appointing president.
33. The 5% standard for judging a demographic groups position on a
poll question roughly represents the 95% confidence level for men or women in
polls of 600 to 1,000 respondents. This standard is more problematic for
smaller groups (such as blacks or Jews) who may number only about one hun-
dred or twenty-five respondents, respectively, in a poll of one thousand adults
nationwide. Results for small groups should be treated more cautiously than
results for larger groups, such as men or women.
34. Gallup Poll, 5/216/9/1993.

35. Justice White and Rehnquist (at 72% and 67%, respectively) were the
justices most likely to represent womens attitudes; Justice OConnor ranked as
the third most likely justice to reflect womens attitudes (at 62%). Justice
Ginsburg tied for eighth of fourteen justices in representing womens attitudes
(at 52%). Justice Marshall ranked last (at 40%) in representing womens atti-
tudes.
36. Justice White and Breyer (at 67% and 62%, respectively) most often
represented black Americans attitudes. Justice Marshall ranked as tenth of
fourteen justices (at 48%) and Justice Thomas ranked as thirteenth of fourteen
(at 43%). During the limited time period in which Justice Thomas served on
the Rehnquist Court, he ranked as the tenth (of eleven) most likely justice to
represent black attitudes (at 45% ahead only of Justice Scalia); Justice Breyer
again ranked first (at 62%).
37. Justices Breyer and Ginsburg ranked second and third (at 65% and
63%, respectively) among all fourteen justices in representing Jewish
Americans attitudes, with Justice White in first place (at 69%).
38. Justices White, Rehnquist, OConnor, and Scalia best represented
Republicans (at 76%, 72%, 65%, and 64%, respectively). Justices Blackmun,
Marshall, and Brennan least often represented Republicans (at 38%, 35%, and
30%, respectively). Justices White, Rehnquist, Breyer, and OConnor best rep-
resented Democrats (at 68%, 66%, 63%, and 60%, respectively). Justices
Blackmun, Marshall, and Brennan least often represented Democrats (at only
40%, 38%, and 37%, respectively). Justices White, Rehnquist, and OConnor
best represented Independents (at 69%, 67%, and 62%, respectively). Justices
Blackmun, Marshall and Powell least often represented Independents (at only
44%, 44%, and 42%, respectively). By this evidence, Justice Rehnquist was a
Republican presidents most successful appointment, and Justice Brennan the
least successful. Justice White was a Democratic presidents most successful
appointment, and Justice Marshall the least successful, at least during the
Rehnquist Court. For earlier evidence, see Marshall (1993).
39. When majorities of men and women disagreed, Justice Ginsburg
always (three times) voted with majorities of women; Justice OConnor always
(five times) voted with majorities of men. By religion, the results are more
mixed. When majorities of Jews and non-Jews disagreed on a decision, Justice
Breyer voted four times with Jews and four times with non-Jews; Justice
Ginsburg voted six times with Jews and two times with non-Jews, perhaps con-
firming the view that Justice Ginsburg is a stronger example of symbolic repre-
sentation than Justice Breyer (Perry and Abraham 1998).
40. On these decisions, majorities of men and women never disagreed
(four decisions); majorities of blacks and whites disagreed 67% of the time (six
decisions); poll majorities of Jews and non-Jews disagreed 86% of the time (six
of seven decisions).
41. For a similar finding based on litigant claims, see Segal (2000).
42. Republicans and Independents disagreed on only 3% of these issues;
Democrats and Independents disagreed on only 8% of these issues.
CHAPTER SEVEN
1. For descriptions of mythic views, see Corwin (1936), Frank (1930),

and Lerner (1937). For survey research on these views, see Casey (1974) and
Scheb and Lyons (2000, 2001).
2. The justices most explicit recent debate over the importance of public
opinion, the Courts prestige, and compliance with its decisions came in
Planned Parenthood of Southeastern Pennsylvania v. Casey (1992); see chapter
one.
3. Whether poll shifts favor the Courts decision is a different question
than whether Americans would actually comply with an unpopular decision.
For studies that examine compliance, see Gibson (1989), Thibaut and Walker
(1975), Tyler (1990), and Tyler and Rasinski (1991). The results here do not
address that debate.
4. Arguably, the Rehnquist Courts influence on public opinion also
declined as the number of decisions, per term, fell by half.
5. These include all the identifiable poll questions in the Roper Archive.
All polls in this chapter are based on nationwide samples of adult Americans.
6. In another study, the national awareness of six Supreme Court deci-
sions ranged from a low of 0% to a high of 40% (Franklin and Kosaki 1995;
Franklin, Kosaki, and Kritzer 1993).
7. On the three most commonly asked issues, the percent of correct
answers above pure random guessing did not vary greatly+9.2% (abortion
questions), +4.3% (school prayer and bible reading questions), or +4.7% (flag-
burning questions). The percentage of correct answers on open-ended poll
questions cannot be compared to pure random guessing, and pollsters do not
commonly ask open-ended poll questions. The percentage correct on open-
ended (factual) questions averaged 43%, as reported by pollsters themselves.
8. The percentage correct through pure random guessing is lower on
these questions, since several poll items specified . . . or dont you know as a
choice.
9. Washington Post, June 1989 survey.
10. Because nearly all these questions are open-ended, no comparison to
pure random guessing is offered.
11. Presidents now typically pick Supreme Court nominees from the fed-
eral appeals courts, and describe their nominees in nonpartisan and non-ideo-
logical termsperhaps thereby boosting the Courts nonpolitical or textbook

image during high publicity periods (Caliendo and Medvic 1999). The Courts
popularity may also help presidents who describe their nominees in nonpolitical
terms (Gibson and Caldeira 2006). Political elites also seldom criticize individ-
ual justices, even when criticizing specific decisions (Caliendo 2000).
12. Gibson, Caldeira, and Spence (2003) argue that the confidence mea-
sure taps both long-term institutional commitment and short-term policy satis-
faction toward the Court.
13. Kritzer (2005) and Hibbing and Theiss-Morse (1995) review these
and other poll ratings.
14. The National Election Studys thermometer scale asks respondents to
report their attitudes toward the Supreme Court (and other institutions) on a
scale from 100 degrees (most favorable) to 0 degrees (least favorable). This
measure is only available for national election years and is not reported here;
see Kritzer (2005).
15. NORC/GSS uses the wording, I am going to name some institutions
in this country. As far as the people in charge of running . . . the U.S. Supreme
Court . . . are concerned, would you say you have a great deal of confidence,
only some confidence, or hardly any confidence at all in them? Harris uses all
but the first sentence of the NORC/GSS question. Gallup uses a different word-
ing with clearly two positive (a great deal and quite a lot), but only one
clearly negative (very little) response: Now I am going to read you a list of
institutions in American society. Please tell me how much confidence you, your-
self have in each onea great deal, quite a lot, some, or very little? . . . Supreme
Court. Harris and Gallup use nationwide telephone samples, while
NORC/GSS uses in-person interviews. Because the question wording, response
categories, method of data collection, and polling dates vary, the three series
cannot be combined into a single time series.
16. NORC/GSSs favorable ratings for Congress and the executive branch
during the Rehnquist and Burger Courts, respectively, averaged 12%, versus
15% (for Congress) and 19%, versus 17% (for the executive branch). For
another survey with similar results, see Scheb and Lyons (1998).
17. In the Harris Poll, the military (averaging 46% favorable reviews)
ranked first, followed by the Supreme Court, medicine, and colleges and uni-
versities (each averaging 31%), then organized religion (24%), television news
and the White House (both at 23%), major companies (18%), the executive
branch (17%), the press and Congress (15%), and organized labor and major
law firms (both at 13%).
18. In the Gallup Poll for institutions polled over the entire Rehnquist
Court, the Supreme Courts combined positive (a great deal and quite a lot
combined) ratings (averaging 49%) ranked behind the military (at 62%) and
the church/organized religion (53%), but ahead of public schools (42%), banks
(41%), Congress (27%), and big business and organized labor (both at 26%).
For the partial time periods in which both were asked, the Supreme Court
ranked behind the police and the presidency, but ahead of HMOs, the medical
system, television news, and electrical power utilities.
19. In a much earlier account (Murphy and Tanenhaus 1969), some 37%
of Americans held positive views toward the Court and 22% held negative
views, with the remaining 41% as indifferent or uninformed. More recently,
Tyler and Rasinsky (1991: 627) reported that the percent of Americans
expressing positive views of the Court ranged from 45% (gives all interested
citizens a chance to express their views before making decisions) to 72% (only
makes decisions after assembling relevant information).
20. Even those expressing hardly any confidence in the Supreme Court
may not support extreme measures, such as doing away with the Court or dis-
obeying the Court when they disagree with its decisions (Gibson, Caldeira, and
Spence 2003, at 362).
21. Do you approve or disapprove of the way the United States Supreme
Court is handling its job? reported for various dates between June 2000 and
March 2003.
22. Comparing the positive responses to negative responses, the Rehnquist
Court averaged a 3.4 to 1 favorable ratio on the Gallup Poll, a 2.3 to 1 favor-
able ratio on the NORC/GSS Poll, and a 2.1 to 1 favorable ratio on the Harris
Poll.
23. For example, a 1991 Los Angeles Times Poll reported a 45-to-43%
plurality disagreeing that the Supreme Court represents the opinions and
beliefs of people like you. A 1989 Harris Poll reported a 55%-to-41% major-
ity saying that the Supreme Court is not really in touch with what is going on
in the country.
24. On each of these twenty-two poll questions, the respondent was told
the justices name and then asked for a favorable or unfavorable rating. This
type of poll question results in many more respondents offering an opinion
than if the respondent were forced to volunteer the justices name. No ques-
tions were identified for Justices Brennan, Blackmun, Powell, or White.
25. Justices OConnor, Marshall, Stevens, and Rehnquist scored the best
positive-to-negative ratings (at 5 to 1, 4.2 to 1, 4 to 1, and 3.1 to 1, respec-
tively). Justice Thomas scored the worst rating (at 1.5 to 1).
26. CBS/New York Times, 5/1114/1987 and Los Angeles Times, 6/28
30/1991.
27. Research and Forecasts, 10/2011/2/1986.
28. CBS/New York Times, 5/1114/1987.
29. Los Angeles Times, 6/2830/1991: As you may know, the President
appoints justices of the Supreme Court. Would you favor or oppose the direct
election of Supreme Court justices by the voters?54% favor, 41% oppose,

5% not sure.
30. The two rulings were Webster v. Reproductive Health Services (1989),
on abortions in public facilities, and Texas v. Johnson (1989), on flag burning.
The percentage of Americans reporting a great deal of confidence in the
Court dropped from 30% prior to the decisions to 18% after the decisions. For
an earlier account tying drops in the Courts popularity to its criminal justice
rulings, see Caldeira (1986). For similar findings in four localities affected by
Rehnquist Court rulings, see Hoekstra (2000, 2003).
31. To Scheb and Lyons (2000, 2001), a greater reliance on original intent
and less on the justices ideology and partisanship would probably increase the
Courts popularity (cf. Mondak 1991).
32. After the Florida recount, several critics such as Ackerman (2002),
Balkin (2002), Dershowitz (2001), Radin (2002), Tribe (2002), and Tushnet
(2002), suggested that the Courts decisions might erode its legitimacy; others
(Gillman 2001: 197206; Greene 2001; Posner 2001) suggested that no real
consequences were likely. That the media treated the Florida recount decision
as atypical and soon returned to portraying the Court in nonpartisan text-
book terms may have reduced any long-term fallout (Caliendo 2001). In the
NORC/GSS, Harris, and Gallup Polls, the Rehnquist Courts favorable ratings
rose slightly (+3%, +1%, and +2%, respectively) from the last poll prior to the
Florida recount dispute to the first poll thereafter.
33. In Hoekstra (1995, 2003), Court rulings on moral and social issues
have more influence than rulings on property rights. In other experiments, a
steady stream of unpopular decisions lowers the Courts popularity, thereby
limiting the Courts ability to produce future positive poll shifts for its decisions
(Mondak 1990, 1992, 1994; Peters 1995; Segal 1999).
34. With few exceptions (Caldeira and Gibson 1992; Gibson, Caldeira,
and Spence 2003), most experiments rely either on small, local adult samples or
else larger samples of undergraduates, typically political science students
(Kinder and Palfrey 1993; Segal 1994).
35. For a similar account of the effects of media coverage and interest
group activity during the Clarence Thomas nomination or the 2000 Florida
presidential vote recount, see Hutchings (2001) and Kritzer (2005).
36. The term manipulate simply means that public opinion follows
Court decisions, not that the justices try to win public support for their deci-
sions through the techniques commonly used by other political figures, such as
talk show appearances, press releases, op-ed pieces, speeches, rallies, or media
interviews. Justices occasionally discuss specific decisions in off-the-bench
speeches, articles, and books, but not very often.
37. Do you favor or oppose . . . a law requiring that the husband of a
married woman be notified if she decides to have an abortion? Source: Gallup
Poll, 3/304/5/1992, 67% favor, 31% oppose, 2% dont know/refused;

7/2628/1996, 71% favor, 26% oppose, 4% dont know/refused. In the short-
term model, the last pre-decision poll results were matched with the first post-
decision poll results. In the long-term model, the first post-decision poll results
were matched with the last post-decision poll results during the Rehnquist
Court.
38. For a review of these hypotheses, see Marshall (1989). The publicity
variable is coded as 3 (if the decision was covered in each of the New York
Times, Washington Post, and Los Angeles Times), 2 (covered in two papers), 1
(covered in only one paper), or 0 (no coverage in any of these three papers).
39. None of these sixteen rulings was unanimous. Six rulings were split
54; six rulings had larger, but non-unanimous majorities; four rulings were
cert denials. The average poll shift was 2.05% for 54 decisions, 1.6% for
decisions with six or seven or eight votes in the majority, and 6.3% for cert
denials. In a previous account (Marshall 1989: 152), positive poll gains
occurred for unanimous rulings (+1.5%, versus .1% for non-unanimous deci-
sions).
40. The average pre- to post-decision time lag is thirty months with a
standard deviation of thirty-six months; results should not be generalized
beyond that time frame.
41. When the rulings were divided into earlier, versus later rulings (1992
or earlier, versus later than 1992), both average pre- to post-ruling poll shifts
were negative (3.0% and 2.9%, respectively). For the post-post ruling poll
shifts divided at the same point, both average shifts were also negative (.4%
and 2.0%, respectively).
42. Not only did the Rehnquist Court fail to produce positive average poll
shifts, the correlations with a variety of explanatory variables were also much
weaker. For example, the explained variance (adjusted R-2) for activist or lib-
eral rulings was only .09 and .07, respectively, versus .42 and .41 for earlier
Courts (Marshall 1989: 153).
43. The Akron decision was announced June 25, 1990. ABC/Washington
Post Poll, 7/1923/1990 and 3/2731/92: Do you think that a pregnant
teenager under the age of 18 should be required by law to notify a parent
before she could obtain a legal abortion?
44. The time lag for these twelve rulings averaged thirty-five months;
results should not be generalized beyond that time frame. Support for these rul-
ings averaged 50.5% in the first post-decision poll.
45. Only the predictions for race, crime, and death yielded predicted positive
poll shifts, and these are based on only a single decision, McClesky v. Zant (1991).
None of the regression estimates was statistically significant at the .05 level.
46. The nine poll questions identified include a patients right to die
(19982005), Bible reading and prayers in public schools (19742004), abor-
tion rights (1974-2006), birth control information (19741983), gender dis-

crimination and affirmative action (19851997), racial discrimination and
affirmative action (19851997), school busing (19711980), racial intermar-
riage (19632002), and overturning Roe v. Wade (19882002). The average
poll shift toward the Courts ruling was +.34% annually. For poll series that
went on twenty years or longer, the average poll shift toward the Courts ruling
was +.48% annually. Only for the racial intermarriage poll series did the
annual rate of change (toward the Courts ruling) exceed 1% annually. In the
average initial post-decision poll, some 42% of Americans supported the
Courts decision, compared to an average of 49% in the last post-decision poll.
47. The NORC/GSS approval ratings are measured less frequently than
the Gallup approval ratings: thirteen times, versus twenty-one times, respec-
tively, during the Rehnquist Court. The strongest predictor of the Rehnquist
Courts (positive) approval ratings (in the Gallup Poll) were the presidencys
approval ratings, with an equation of 33.5% (constant) + .272 (presidency rat-
ings, significant at .01), with an adjusted R-2 of .49. No other variable was sta-
tistically significant when entered into this equation, including congressional
approval ratings, high-profile decision popularity, low-profile decision popular-
ity, combined (all) decision popularity, or individual justice confirmation events
(each entered as 0,1).
48. In part, these largely negative results may occur because there was rel-
atively little variation in the percentage of Rehnquist Court decisions consistent
with public opinion. No significant relationship appeared when the Courts
approval ratings were compared to the last years decisions, or for cases up to
three years previously.
49. Hoekstras (2003) four local surveys show no poll shifts in favor of
economic rulings, and only minimal shifts in favor of non-economic rulings.
CHAPTER EIGHT
1. At least six constitutional amendments overturned a Supreme Court

decision (the eleventh, thirteenth, fourteenth, fifteenth, sixteenth, and twenty-
sixth), and arguable, three others (the seventeenth, nineteenth, and twenty-
fourth) did so (Abraham 1998). Epstein et al. (2003) (five amendments),
Romano (1992) (a different list of five amendments), Spiller and Spitzer (1992)
(five amendments), and Paschal (1991) (seven amendments) offer different
counts.
2. In Payne v. Tennessee (1991), at 828, Justice Rehnquist listed deci-
sions during the past twenty terms overruling thirty-three precedents, in whole
or in part, an average of one to two precedents overruled, per term. For an ear-
lier account, see Blaustein and Field (1958).
3. Since some decisions openly invite Congress and the president to rewrite
a statute, not all congressional overturnings can be considered as hostile
(Hausegger and Baum 1998, 1999: 166; see also Barnes 2004 and Spiller and
Tiller 1993). On the average, about eight majority opinions, per term, openly
invited a congressional override for the 1986 through 1990 terms; see Eskridge
(1991). Congress more often overturns decisions that include an invitation to
review (Ignagni and Meernik 1994). Congress may also use its control over
judicial budgets to signal disagreement with the Courts decisions (Toma 1991).
For a historical review, see Whittington (2001).
4. Paschal (1991) lists the 1991 Civil Rights Act as overturning ten deci-
sions.
5. Under the dormant commerce theory, Congress can agree to state leg-
islative restrictions that would otherwise likely fail on constitutional grounds,
restrict state laws previously accepted by the Supreme Court by legislating in
the area, or extend constitutional protections in an area where the Court previ-
ously allowed restrictions on rights claims (Paschal 1991: 206208).
6. Because some decisions are overridden many years after the original
decision, these numbers doubtlessly undercount the number of Congressional
overrides (Eskridge 1991a; Solomine and Walker 1992).
7. For law journal arguments debating whether lower federal or state
courts overruled precedents sub silentio, see Fowler (2000); Kash (1997);
Leathers (1998/1999); Ledewitz (1999); Nagareda (1998); or Tilleman and
Swindle (1999). For law journal articles debating whether Supreme Court rul-
ings overturned precedents sub silentio, see Cosby (1995); c.f. Esbeck (1998);
Macleod (1997; Ellington, Higashi, Kim, and Murakami (1998); Nice (1999);
Olivas 1997; Ray (2000); and Umhofer (2001).
8. For historical examples, see NoteBecker and Feeley (1973);
Evasion of Supreme Court Mandates (1954); Murphy (1959); Sheldon (1974:
197); and Wasby, dAmato, and Metrailer (1977).
9. President Jackson declined to enforce the Courts Worcester v.
Georgia (1832) decision; Georgia had imprisoned two white missionaries for
living on Cherokee territory, despite the Courts decision that the Georgia law
unconstitutionally regulated commerce with Indian tribes. The two missionaries
remained in jail from 1831 until 1833, until eventually pardoned by Georgias
governor. President Jackson finally did react to South Carolinas nullification
declaration on tariffs by requesting expanded federal judicial power; Congress
approved his request in 1833 (Biskupic and Witt 1997:15). No Rehnquist
Court decisions considered here appear to fall into this sixth category, with the
possible exception of the Courts medical marijuana rulings.
10. At 828, note 1, Collins v. Youngblood (1990), overruling Kring v.
Missouri (1883). The lag for precedents overruled by the Court itself ranged
from 3 years to 107 years, averaging 40 years, with a median of 23 years.
Blaustein and Field (1958) reported a 24-year average lag before the Court
overturned a precedent, compared to only a 2.4 year time lag for decisions
overturned by Congress. For constitutional amendments overturning a Supreme

Court decision, lags vary from one year [the Twenty-sixth Amendment over-
turning Oregon v. Mitchell (1970] to 45 years [(the Nineteenth Amendment
overturning Minor v. Happersett (1875] (Marshall 1989: 170172). For earlier
listings, see Douglas (1949) and Justice Brandeiss dissent in Burnet v.
Coronado Oil and Gas Co. (1932), 405407, at note 2.
11. Congress response to Sable included writing a ban on obscene dial-a-
porn into the 1990 deficit-reduction reconciliation bill, an action later upheld
in Dial Information Services Corp. v. Thornburgh, 938 F.2d 1535 (1991) and
in the denial of certiorari in Dial Information Services Corp. v. Barr, 502 U.S.
1072 (1992).
12. Eskridge (1991a) and Meernik and Ignagni (1995) also cite a reliance
on plain text, or state or federal amicus or interest group activity as related to
congressional overturnings.
13. Most accounts attribute the rising number of dissenting and concur-
ring opinions to Chief Justice Stones leadership during the 1940s, rather than
to institutional changes, such as the largely discretionary docket and reduced
caseload under the 1925 Judiciary Act, or to the younger justices appointed
during the 1930s and 1940s (Epstein et al. 2003; see also Ryan 1998). For evi-
dence on unanimity during the 1800s, see Caldeira and Zorn (1998) and
Epstein, Segal, and Spaeth (2001). For consensus on the Rehnquist Court, see
Gerber and Park (1997).
14. Unanimous decisions also more often survive congressional overriding
efforts (Solomine and Walker 1992:447), cf. Eskridge (1991a).
15. See Congressional Reversals of Supreme Court Decisions, 194557
(1958); Eskridge (1991); Henschen (1983); Meernik and Ignagni (1995);
Murphy (1962); and Pritchett (1961).
16. Eskridge (1991b: 652) and Spiller and Spitzer (1992) argue that the
justices ground decisions in constitutional law if they fear that Congress might
override a decision based on statutory grounds. For a similar argument based
on agencies and lower courts, see Tiller and Spiller (1999).
17. Bickel (1970: 181) added or even if it was opposed by a determined
and substantial minority and received with indifference by the rest of the coun-
try. This, in the end, is how and why judicial review is consistent with the
theory and practice of political democracy.
18. The poll margin is explained in chapter two. Because chapter seven
and earlier research (Marshall 1989) show that Supreme Court decisions little
affect American public opinion, poll support at the time of the original decision
is an unbiased indicator of later poll support.
19. About half of congressional overrides occur within two years of the
Supreme Court decision overridden, two-thirds within five years, and three-
fourths within ten years (Eskridge 1991; Hettinger and Zorn 2001; Solomine
and Walker 1992). Lindquist and Yalof (2001) put the average time for
Congress to overturn an appeals court decision at five years.
20. Karla Faye Tucker is on death row in Texas for having brutally mur-
dered two people in 1983. Some people who believe she has had a religious
conversion and has been rehabilitated want the State of Texas to reduce her
sentence to life in prison without parole. From what youve heard or read, do
you think her sentence should be reduced to life in prison without parole, or do
you think she should be executed as scheduled? Results: 37%reduced to life
imprisonment; 54%executed; 9%dont know (results for the 74% of
respondents who had heard of Tucker), CBS News Poll, 2/1/1998. For a survey
among national registered voters with similar results, see Fox News,
1/78/1998.
21. In addition to Sable, discussed earlier, Shaw v. Reno (1993) and
Miller v. Johnson (1995), making it more difficult, but not impossible or unnec-
essary, to draw minority-dominated congressional districts, were classified as
prevailing, notwithstanding Hunt v. Cromartie (2001), Lawyer v. Department
of Justice (1997), or Abrams v. Johnson (1996). Since Sable, Shaw, and Miller
were all inconsistent with the polls, the statistical relationships noted would be
stronger had these decisions been classified as failing the test of time or had an
ordinal dependent variable (such as prevailed, substantially modified, or failed)
been used.
22. This figure was computed by taking the minimum lag for overturn-
ings.
23. Congress overturns Supreme Court decisions much more quickly than
the Court itself overrides its own precedents, or than constitutional amend-
ments can be passed. Historically, for the 19351986 period, the hazard rate
(Box-Steffensmeier and Jones 1997) for Supreme Court decisions failing was
the highest in the first or second year after the ruling: 5.5% in the first year and
2.2% in the second year (Marshall 1989).
24. Because all arguable cases were classified as prevailing, these estimates
are a lower bound estimate of the half-life of Court decisions. Had decisions
such as Sable (1989), Browning Ferris (1989), Shaw v. Reno (1993), and Miller
v. Johnson (1995) been classified as failing, the expected half-life would be
shorter.
25. Both analyses are complicated by the limited time frame usedfor the
19351986 period (fifty-one years) or the 19862005 period (nineteen years).
Data for longer time periods may well yield longer half-life estimates. The
19351986 logit equation is: 2.16 (constant) 1.15 (if inconsistent with public
opinion) .04 (per year of the decision prior to 1986). The constant and both
predictors are significant at the .001 level, and the equation correctly predicts
90% of the decisions, compared to a base rate of 82%, or an improvement of
44% over the base rate. The 19862005 logit equation is 6.883 .394 (per lag
year) + 2.456 (if consistent with public opinion). No other variables were sig-
nificant predictors for either time period, either alone or in combination with
years and public opinion. Results using rare events logit were virtually identical
to these results (King and Zeng 2001).
26. For most Supreme Court decisions no polling evidence is available,
and the importance of public opinion cannot be estimated, although other
explanations for a decisions longevity can be. For a discussion of congressional
overrides, see Hettinger and Zorn (2005).
Bibliography
Abraham, Henry J. 1982. Freedom and the Court. New York: Oxford
University Press.
Abraham, Henry J. 1998. The Judicial Process. New York: Oxford
University Press.
Abraham, Henry J. 1999. Justices, Presidents, and Senators. Lanham,
MD: Rowman & Littlefield.
Abramowitz, Alan, and Kyle Saunders. 1998. Ideological Realignment
in the U.S. Electorate. Journal of Politics 60: 634652.
Ackerman, Bruce. 2002. Off Balance. In Bruce Ackerman, ed., Bush
v. Gore. New Haven: Yale University Press: 192209.
Adams, Greg. 1997. Abortion: Evidence of an Issue Evolution.
American Journal of Political Science 41: 718737.
Agresto, John. 1984. The Supreme Court and Constitutional
Democracy. Ithaca: Cornell University Press.
Aldrich, John. 1995. Why Parties? The Origin and Transformation of
Party Politics in America. Chicago: University of Chicago Press.
Allen, David W., and Diane E. Wall. 1987. The Behavior of Women
State Supreme Court Justices: Are They Tokens or Outsiders?
The Justice System Journal 12: 232244.
Allen, David W., and Diane E. Wall. 1993. Role Orientations and
Women State Supreme Court Justices. Judicature 77: 156165.
Amar, Akhil Reed. 1998. The Bill of Rights: Creation and
Reconstruction. New Haven: Yale University Press.
Atkinson, David N. 1999. Leaving the Bench: Supreme Court Justices
at the End. Lawrence: University Press of Kansas.
Baas, Larry, and Dan Thomas. 1984. The Supreme Court and Policy
Legitimation: Experimental Tests. American Politics Quarterly
12: 335360.
215
216 Bibliography
Bailey, Michael A., Brian Karmoie, and Forrest Maltzman. 2005.

Signals from the Tenth Justice: The Political Role of the Solicitor
General in Supreme Court Decision Making. American Journal
of Political Science 49: 7285.
Baird, Vanessa A. 2001a. Profit, Political Context and Judicial Power:
Why the Supreme Court Cannot Make Liberal Economic Policy.
Paper delivered at the 2001 annual meeting of the Midwest
Political Science Association, Chicago.
Baird, Vanessa A. 2001b. Building Institutional Legitimacy: The Role
of Procedural Justice. Political Research Quarterly 54: 333354.
Baird, Vanessa A. 2004. The Effect of Politically Salient Decisions on
the U.S. Supreme Courts Agenda. Journal of Politics 66:
755772.
Balkin, Jack M. 2002. Legitimacy and the 2000 Election. In Bruce
Ackerman, ed., Bush v. Gore. New Haven: Yale University Press:
192209.
Banks, William. 1995. At the Halfway PointLight Docket Makes it
Hard to Read Trends in Supreme Court Decisions. ABA Journal
April: 5052.
Barnes Jeb. 2004. Overruled? Legislative Overrides, Pluralism, and
Contemporary Court-Congress Relations. Stanford: Stanford
University Press.
Barnum, David G. 1985. The Supreme Court and Public Opinion:
Judicial Decision Making in the Post-New Deal Period. Journal
of Politics 47: 652666.
Bartow, Ann. 2004. Likelihood of Confusion. San Diego Law
Review 41: 721817.
Baum, Lawrence. 1997. The Puzzle of Judicial Behavior. Ann Arbor:
University of Michigan Press.
Baum, Lawrence. 1998a. The Supreme Court. Washington, DC:
Congressional Quarterly.
Baum, Lawrence. 1998b. American Courts: Process and Policy. Boston:
Houghton Mifflin.
Becker, Susan. 1991. Public Opinion Polls and Surveys as Evidence:
Suggestions for Resolving Confusing and Conflicting Standards
Governing Weight and Admissibility. Oregon Law Review 70:
463522.
Bibliography 217
Becker, Theodore. 1970. Comparative Judicial Politics. Chicago: Rand

McNally.
Becker, Theodore, and Malcolm Feeley. 1973. The Impact of Superior
Court Decisions. New York: Oxford University Press.
Behuniak-Long, Susan. 1991. Friendly Fire: Amici Curiae and
Webster v. Reproductive Health Services. Judicature 74:
261270.
Belsky, Martin H. 2000. Privacy: The Rehnquist Courts
Unmentionable Right. Tulsa Law Journal 36: 4358.
Belsky, Martin H. ed. 2002. The Rehnquist Court. New York: Oxford
University Press.
Benesh, Sara C. 2006. Understanding Public Confidence in American
Courts. Journal of Politics 68: 697707.
Bentley, Arthur. 1908. The Process of Government. Evanston, IL:
Principia Press.
Berkson, Larry, and Susan B. Carbon. 1980. The United States Circuit
Judge Nominating Commission: Its Mmembers, Procedures, and
Candidates. Chicago: American Judicature Society.
Best, Samuel. 1999. The Sampling Problem in Measuring Policy
Mood: An Alternative Solution. Journal of Politics 61: 721740.
Bickel, Alexander. 1962. The Least Dangerous Branch. New Haven:
Yale University Press.
Bickel, Alexander. 1970. The Supreme Court and the Idea of Progress.
New Haven, CN: Yale University Press.
Biskupic, Joan. 1995. High Court Justice with a Cause. Washington
Post, April 17: A1.
Biskupic, Joan. 1996. The Shrinking Docket: Attorneys Try to Make
an Issue Out of the Dramatic Decline in High Court Rulings.
Washington Post, March 18: A15.
Biskupic, Joan. 2005. Sandra Day OConnor. New York:
HarperCollins.
Biskupic, Joan, and Elder Witt. 1997. Guide to the U.S. Supreme
Court. Third Edition. Washington, DC: Congressional Quarterly
Press.
218 Bibliography
Black, Charles L., Jr. 1960. The People and the Court. New York:
Macmillan.
Black, Merle. 2004. The Transformation of the Southern Democratic
Party. Journal of Politics 66: 10011017.
Blaustein, Albert, and Andrew Field. 1958. Overruling Opinions in
the Supreme Court. Michigan Law Review 57: 151194.
Boles, Donald E. 1987. Mr. Justice Rehnquist, Judicial Activist: The
Early Years. Ames: Iowa State University Press.
Bork, Robert. 1990. The Tempting of America. New York: Free Press.
Boucher, Robert L. Jr., and Jeffrey A. Segal. 1995. Supreme Court
Justices as Strategic Decision Makers: Aggressive Grants and
Defensive Denials on the Vinson Court. Journal of Politics 57:
824837.
Bowers, William, Margaret Vandiver, and Patricia Dugan. 1994. A
New Look at Public Opinion on Capital Punishment: What
Citizens and Legislators Prefer. American Journal of Criminal
Law 22: 77150.
Box-Steffensmeier, Janet M., and Bradford Jones. 1997. Time Is of the
Essence: Event History Models in Political Science. American
Journal of Political Science 41: 14141461.
Brams, Steven J., and Douglas Muzzio. 1977. Unanimity in the
Supreme Court: A Game-Theoretic Explanation of the Decision in
the White House Tapes Cases. Public Choice 32: 6783.
Brettschneider, Frank. 1996. Public Opinion and Parliamentary
Action: Responsiveness of the German Bundestag in Comparative
Perspective. International Journal of Public Opinion Research 8:
292311.
Brewer, Paul. 2003. The Shifting Foundations of Public Opinion
about Gay Rights. Journal of Politics 65: 12061220.
Breyer, Stephen G. 1982. Regulation and Its Reform. Cambridge:
Harvard University Press.
Breyer, Stephen G. 1992. On the Uses of Legislative History in
Interpreting Statutes. Southern California Law Review 65:
845874.
Breyer, Stephen G. 1993. Breaking the Vicious Circle: Toward Effective
Risk Management. Cambridge: Harvard University Press.
Bibliography 219
Breyer, Stephen G. 2004. Judicial Checks upon the Electoral Process.

In Robert Badinter and Stephen Breyer, eds., Judges in
Contemporary Democracy. New York: New York University
Press: 117174.
Breyer, Stephen G. 2005. Active Liberty. New York: Alfred A. Knopf.
Breyer, Stephen G., and Paul W. Macavoy. 1974. Energy Regulation by
the Federal Power Commission. Washington, DC: Brookings.
Brisbin, Richard A., Jr. 1997. Justice Antonin Scalia and the
Conservative Revival. Baltimore: Johns Hopkins University Press.
Bronner, Ethan. 1989. Battle for Justice: How the Bork Nomination
Shook America. New York: Norton.
Brunell, Thomas, and Bernard Grofman. 1998. Explaining Divided
U.S. Senate Delegations, 17881996: A Realignment Approach.
American Political Science Review 92: 391399.
Bullock, Charles, III, Donna Hoffman, and Ronald Gaddie. 2005. The
Consolidation of the White Southern Congressional Vote.
Political Research Quarterly 58: 231244.
Burnham, Walter Dean. 1981. The 1980 Earthquake: Realignment,
Reaction, or What? In Thomas Ferguson and Joel Rogers, eds.,
The Hidden Election. New York: Pantheon Books: 98140.
Burnham, Walter Dean. 1989. The Reagan Heritage. In Gerald M.
Pomper, ed., The Election of 1988. Chatham, NJ: Chatham
House: 132.
Burnham, Walter Dean. 1991. Critical Realignment: Dead or Alive.
In Byron E. Shafer, ed., The End of Realignment? Madison:
University of Wisconsin Press: 101140.
Burnham, Walter Dean. 1996. Realignment Lives: The 1994
Earthquake and Its Implications. In Colin Campbell and Bert A.
Rockman, eds. The Clinton Presidency: First Appraisals.
Chatham, NJ: Chatham House.
Calabresi, Steven G., and James Lindgren. 2006. Term Limits for the
Supreme Court: Life Tenure Reconsidered. Harvard Journal of
Law & Public Policy 29: 769877.
Caldeira, Gregory A. 1986. Neither the Purse Nor the Sword:
Dynamics of Public Confidence in the Supreme Court. American
Political Science Review 80: 12091226.
220 Bibliography
Caldeira, Gregory A. 1991. Courts and Public Opinion. In John B.

Gates and Charles A. Johnson, eds., The American Courts.
Washington, DC: Congressional Quarterly.
Caldeira, Gregory A., and James L. Gibson. 1992. The Etiology of
Public Support for the Supreme Court. American Journal of
Political Science 36: 635664.
Caldeira, Gregory A., and James L. Gibson. 1995. The Legitimacy of
the Court of Justice in the European Union: Models of
Institutional Support. American Political Science Review 89:
356376.
Caldeira, Gregory A. and Charles E. Smith Jr. 1996. Campaigning for
the Supreme Court: The Dynamics of Public Opinion on the
Thomas Nomination. Journal of Politics 58: 655681.
Caldeira, Gregory A., and John Wright. 1988. Organized Interests
and Agenda Setting in the U.S. Supreme Court. American
Caldeira, Gregory A., and John Wright. 1989. Why Organized
Interests Participate as Amicus Curiae in the U.S. Supreme
Court. Paper presented at the annual meeting of the Midwest
Caldeira, Gregory A., and John Wright. 1990. Amicus Curiae Before
the Supreme Court: Who Participates, When, and How Much?
Journal of Politics 52: 782806.
Caldeira, Gregory A., and John Wright. 1998. Lobbying for Justice:
Organized Interests, Supreme Court Nominations, and the United
States Senate. American Journal of Political Science 42:
499523.
Caldeira, Gregory A. and Christopher Zorn. 1998. Of Time and
Consensual Norms in the Supreme Court. American Journal of
Caldeira, Gregory A., Marie Hojnacki, and John R. Wright. 1996.
The Informational Roles of Organized Interests in the Politics of
Federal Judicial Nominations. Paper presented at the annual
meeting of the Midwest Political Science Association, Chicago.
Caldeira, Gregory A., Marie Hojnacki, and John R. Wright. 2000.
The Lobbying Activities of Organized Interests in Federal
Judicial Nominations. Journal of Politics 62: 5169.
Bibliography 221
Caldeira, Gregory, John Wright, and Christopher Zorn. 1999.

Sophisticated Voting and Gate-Keeping in the Supreme Court.
The Journal of Law, Economics & Organization 15: 549571.
Caliendo, Stephen M. 2000. For Public Consumption: Elite Rhetorical
Framing of Salient Supreme Court Decisions. Paper presented at
the annual meeting of the Southern Political Science Association,
Atlanta.
Caliendo, Stephen M. 2001. The Great Legitimacy Gamble: Media
Coverage and Potential Effects of the U.S. Supreme Courts
Involvement in the 2000 Presidential Election. Paper presented
at the annual meeting of the Midwest Political Science
Association, Chicago.
Caliendo, Stephen M., and Stephen K. Medvic. 1999. The Politics of
Supreme Court Nominations: An Analysis of Presidential
Rhetoric from Truman to Clinton. Paper presented at the annual
meeting of the American Political Science Association, Atlanta.
Cameron, Charles, Albert Cover, and Jeffrey A. Seal. 1990. Senate
Voting on Supreme Court Nominees: A Neoinstitutional Model.
Canon, Bradley C. 1982. A Framework for the Analysis of Judicial
Activism. In Stephen C. Halpern and Charles M. Lamb, eds.,
Supreme Court Activism and Restraint. Lexington, MA:
Lexington Books.
Caplan, Lincoln. 1987. The Tenth Justice. New York: Vintage Books.
Carmines, Edward, and James Stimson. 1989. Issue Evolution: Race
and the Transformation of American Politics. Princeton, NJ:
Princeton University Press.
Carp, Robert A., and Ronald Stidham. 1998. The Federal Courts.
Casey, Gregory. 1974. The Supreme Court and Myth: An Empirical
Investigation. Law & Society Review 8: 385419.
Casper, Jonathan. 1976. The Supreme Court and National Policy
Making. American Political Science Review 70: 5063.
Chemerinsky, Erwin. 1996. The Shrinking Docket. Trial May:
7071.
222 Bibliography
Chemerinsky, Erwin. 1999a. The Rehnquist Court and Justice: An

Oxymoron? Washington University Journal of Law & Policy 37:
3652.
Chemerinsky, Erwin. 1999b. The Supreme Court, Public Opinion,
and the Role of the Academic Commentator. South Texas Law
Review 40: 943955.
Chemerinsky, Erwin. 2001. Why the Rehnquist Court Is Wrong
About the Establishment Clause. Loyola University Chicago
Law Journal 33: 221235.
Choi, Stephen J., and G. Mitu Gulati. 2004. Choosing the Next
Supreme Court Justice: An Empirical Ranking of Judge
Performance. Southern California Law Review 78: 78115.
Choper, Jesse. 1967. On the Warren Court and Judicial Review.
Catholic University Law Review 17: 243.
Choper, Jesse. 1974. The Supreme Court and the Political Branches:
Democratic Theory and Practice. University of Pennsylvania
Law Review 122: 811855.
Choper, Jesse. 1980. Judicial Review and the National Political
Process. Chicago: University of Chicago Press.
Clark, John A., and Kevin T. McGuire. 1996. Congress, the Supreme
Court, and the Flag. Political Research Quarterly 49: 771781.
Clawson, Rosalee A., Elizabeth R. Kegler, and Eric N. Waltenburg.
1999. The Legitimacy-Conferring Authority of the U.S. Supreme
Court: An Experimental Design. Paper presented at the annual
meeting of the Southern Political Science Association, Savannah.
Clawson, Rosalee A., and Eric N. Waltenburg. 2003. Support for a
Supreme Court Affirmative Action Decision. American Politics
Research 31: 251279.
Clayton, Cornell. 1999. Law, Politics, and the Rehnquist Court:
Structural Influences on Supreme Court Decision Making. In
Howard Gillman and Cornell Clayton, eds., The Supreme Court
in American Politics: New Institutionalist Interpretations.
Lawrence: University Press of Kansas: 151177.
Closen, Michael L. 1998. The Decade of Supreme Court Avoidance of
AIDS: Denial of Certiorari in HIV-AIDs Cases and its Adverse
Effects on Human Rights. Albany Law Review 61: 897987.
Bibliography 223
Collins, Paul M., Jr. 2004. Friends of the Court: Examining the
Influence of Amicus Curiae Participation in U.S. Supreme Court
Litigation. Law & Society Review 38: 807832.
Comiskey, Michael. 2004. Seeking JusticesThe Judging of Supreme
Court Nominees. Lawrence: University of Kansas Press.
Comment. 2005. The Debate Over Foreign Law in Roper v.
Simmons. Harvard Law Review 119: 103108.
Congressional Quarterly Almanac. 1989. Dial-A-Porn Restrictions.
Washington, DC: Congressional Quarterly: 382.
Converse, Jean M. 1987. Survey Research in the United States: Roots
and Emergence 18901960. Berkeley, CA: University of Cali-
fornia Press.
Cortner, Richard C. 1968. Strategies and Tactics of Litigants in
Constitutional Cases. Journal of Public Law 17: 287307.
Cortner, Richard C. 1970a. The Jones & Laughlin Case. New York:
Alfred A. Knopf.
Cortner, Richard C. 1970b. The Apportionment Cases. Knoxville:
University of Tennessee Press.
Cortner, Richard C. 1988. A Mob Intent on Death: The NAACP and
the Arkansas Riot Cases. Middletown, CT: Wesleyan University
Press.
Cortner, Richard C. 1993. The Iron Horse and the Constitution.
Westport, CT: Greenwood Press.
Corwin, Edward. 1936. The Constitution as Instrument and as
Symbol. American Political Science Review 30: 10711085.
Cosby, Tricia 1995. Note: Strictly Speaking: Viewing J.E.B. v.
Alabama Ex Rel. T.B. As Sub Silentio Application of Strict
Scrutiny to Gender-Based Classifications. Houston Law Review
32: 869893.
Cramton, Roger C., and Paul D. Carrington, eds. 2006. Reforming the
Court: Term Limits for Supreme Court Justices. Durham, NC:
Carolina Academic Press.
Cushman, Barry. 2002. Mr. Dooley and Mr. Gallup: Public Opinion
and Constitutional Change in the 1930s. Buffalo Law Review
50: 777.
224 Bibliography
Dahl, Robert. 1957. Decision-making in a Democracy: The Supreme

Court as a National Policy-Maker. Journal of Public Law 6:
279295.
Danielski, David. 1972. The Influence of the Chief Justice in the
Decisional Process. In Walter F. Murphy and C. Herman
Pritchett, eds., Courts, Judges, and Politics. New York: Random
House.
Danielski, David. 1986. Causes and Consequences of Conflict and Its
Resolution in the Supreme Court. In Sheldon Goldman and
Charles Lamb, eds., Judicial Conflict and Consensus. Lexington:
University of Kentucky Press.
Davis, Abraham. 1989. Blacks in the Federal Judiciary. Bristol, IN:
Wyndham Hall.
Davis, James A., and Tom W. Smith. 1980. Conservative Weather in a
Liberalizing Climate. Social Forces 58: 11291156.
Davis, Richard. 1994. Decisions and Images: The Supreme Court and
the Press. Englewood Cliffs, NJ: Prentice Hall.
Davis, Sue. 1986. The Impact of President Carters Judicial Selection
Reforms: A Voting Analysis of the United States Courts of
Appeals. American Politics Quarterly 14: 320344.
Davis, Sue. 1989. Justice Rehnquist and the Constitution. Princeton,
NJ: Princeton University Press.
Davis, Sue. 1992. Do Women Speak in a Different Voice?
Wisconsin Womens Law Journal 8: 143173.
Davis, Sue. 1993. The Voice of Sandra Day OConnor. Judicature
77: 134139.
Davis, Sue, Susan Haire, and Donald R. Songer. 1993. Voting
Behavior and Gender on the U.S. Courts of Appeals. Judicature
77: 129133.
Davis, Sue, Susan Haire, and Donald R. Songer. 1999. The Chief
Justice and Judicial Decision-Making: The Institutional Basis for
Leadership on the Supreme Court. In Cornell W. Clayton and
Howard Gillman, eds., Supreme Court Decision-Making: New
Institutionalist Approaches. Chicago: University of Chicago Press:
135174.
Bibliography 225
Days, Drew S. III. 1998. The Rehnquist Years: A Supreme Court

Retrospective. Nova Law Review 22: 679691.
Deen, Rebecca E., Joseph Ignagni, and James Meernik. 2003. The
Solicitor General as Amicus 19532000: How Influential?
Judicature 87: 6071.
DeGregorio, Christine, and Jack F. Rossotti. 1994. Resources,
Attitudes and Strategies: Interest Group Participation in the Bork
Confirmation Process. American Review of Politics 15: 119.
DeGregorio, Christine, and Jack F. Rossotti. 1995. Campaigning for
the Court: Interest Group Participation in the Bork and Thomas
Nomination Processes. In Interest Group Politics, 4th ed., ed.
Allan Cigler and Burdett Loomis. Washington, DC: Con-
gressional Quarterly: 215218.
Dershowitz, Alan M. 2001. Supreme Injustice. New York: Oxford
University Press.
Devine, Donald. 1970. The Attentive Public: Polyarchial Democracy.
Chicago: Rand McNally.
Devins, Neal. 2004. The Majoritarian Rehnquist Court? Law &
Contemporary Problems 67: 6381.
Dinan, John J. 1999. The Rehnquist Courts Federalism Decisions in
Perspective. Journal of Law & Politics 15: 127194.
Dorf, Michael C. 1998. The Limits of Socratic Deliberation. Harvard
Law Review 112: 483.
Douglas, William C. 1949. Stare Decisis. Columbia Law Review 49:
735758.
Dovi, Suzanne. 2002. Preferable Descriptive Representation: Will Just
Any Woman, Black, or Latino Do? American Political Science
Review 96: 729743.
Durr, Robert, Andrew Martin, and Christine Wolbrecht. 2000.
Ideological Divergence and Public Support for the Supreme
Court. American Journal of Political Science 44: 768776.
Easton, David. 1965. A Systems Analysis of Political Life. Chicago:
Wiley.
Edelman, Murray. 1967. The Symbolic Uses of Politics. Urbana:
University of Illinois Press.
226 Bibliography
Edelman, Murray. 1971. Politics as Symbolic Action. New York:

Harcourt Brace Jovanovich.
Edelman, Paul H., and Jim Chen. 2001. The Most Dangerous Justice
Rides Again: Revisiting the Power Pageant of the Justices.
Minnesota Law Review 86: 131222.
Edsall, Thomas, and Mary Edsall. 1992. Chain Reaction: The Impact
of Race, Rights, and Taxes on American Politics. New York:
Norton.
Eisenberg, Theodore, and Martin T. Wells. 1993. Deadly Confusion:
Juror Instructions in Capital Cases. Cornell Law Review 79:
117.
Eisinger, Robert M. 2003. The Evolution of Presidential Polling. New
York: Cambridge University Press.
Ellington, Toni J., Sylvia K. Higashi, Jayna K. Kim, Mark M.
Murakami. 1998. Justice Ruth Bader Ginsburg and Gender
Discrimination. University of Hawaii Law Review 20: 699796.
Epstein, Lee. 1985. Conservatives in Court. Knoxville: University of
Tennessee Press.
Epstein, Lee. 1993. Interest Group Litigation during the Rehnquist
Court Era. Journal of Law & Politics 9: 639717.
Epstein, Lee, and Jack Knight. 1998. The Choices Justices Make.
Epstein, Lee, and Jack Knight. 1999. Mapping Out the Strategic
Terrain: The Informational Role of Amici Curiae. In Cornell W.
Clayton and Howard Gillman, eds., Supreme Court Decision-
Making: New Institutionalist Approaches. Chicago: University of
Chicago Press: 215236.
Epstein, Lee, and Jack Knight. 2000. Toward a Strategic Revolution
in Judicial Politics: A Look Back, A Look Ahead. Political
Research Quarterly 53: 625661.
Epstein, Lee, and Joseph F. Kobylka. 1992. The Supreme Court and
Legal Change: Abortion and the Death Penalty. Chapel Hill:
University of North Carolina Press.
Epstein, Lee, and Carol Mershon. 1996. Measuring Political
Preferences. American Journal of Political Science 40: 261294.
Bibliography 227
Epstein, Lee, and Claude K. Rowland. 1991. Debunking the Myth of

Interest Group Invincibility in the Courts. American Political
Science Review 85: 205217.
Epstein, Lee, and Thomas Walker. 1995. The Role of the Supreme
Court in American Society: Playing the Reconstruction Game. In
Lee Epstein, ed., Contemplating Courts. Washington, DC:
Congressional Quarterly: 315346.
Epstein, Lee, Tracey George, Michael Giles, and Thomas Walker.
1992. Rating the Justices: Lessons from Another Court. Paper
presented at the annual meeting of the Midwest Political Science
Epstein, Lee, Valerie Hoekstra, Jeffrey A. Segal, and Harold J. Spaeth.
1998. Do Political Preferences Change? A Longitudinal Study of
U.S. Supreme Court Justices. The Journal of Politics 60:
801818.
Epstein, Lee, Jack Knight, and Andrew D. Martin. 2001. The Supreme
Court as a Strategic National Policy-maker. Emory Law Journal
50: 583611.
Epstein, Lee, Rene Lindstadt, Jeffrey A. Segal, and Chad Westerland.
2006. The Changing Dynamics of Senate Voting on Supreme
Court Nominees. Journal of Politics 68: 296307.
Epstein, Lee, Jeffrey A. Segal, and Harold J. Spaeth. 2001. The Norm
of Consensus on the U.S. Supreme Court. American Journal of
Epstein, Lee, Jeffrey A. Segal, and Jennifer Nicoll Victor. 2002.
Dynamic Agenda-Setting on the United States Supreme Court:
An Empirical Assessment. Harvard Journal on Legislation 39:
395433.
Epstein, Lee, Jeffrey A. Segal, Harold J. Spaeth, and Thomas G.
Walker. 1996, 2003. The Supreme Court Compendium.
Erikson, Robert S., and Kent L. Tedin. 2001, 2003. American Public
Opinion. New York: Longman.
Esbeck, Carl H. 1998. The Establishment Clause as a Structural
Restraint on Governmental Power. Iowa Law Review 84:
1113.
228 Bibliography
Eskridge, William N., Jr. 1989. Public Values in Statutory

Interpretation. University of Pennsylvania Law Review 137:
10071094.
Eskridge, William N., Jr. 1991a. Overriding Supreme Court Statutory
Interpretation Decisions. Yale Law Journal 101: 331455.
Eskridge, William N., Jr. 1991b. Reneging on History? Playing the
Court/Congress/President Civil Rights Game. California Law
Review 79: 613684.
Eskridge, William N., Jr., and Philip Frickey. 1994. The Supreme
Court 1993 Term. Forward: Law as Equilibrium. Harvard Law
Review 108: 26108.
Fallon, Richard H., Jr. 1997. Implementing the Constitution.
Harvard Law Review 111: 56152.
Fallon, Richard H., Jr. 2002. The Conservative Paths of the
Rehnquist Courts Federalism Decisions. University of Chicago
Farnsworth, Ward. 2005. The Regulation of Turnover on the
Supreme Court. University of Illinois Law Review 2005:
407453.
Farnsworth, Ward. 2006. The Ideological Stakes of Eliminating Life
Tenure. Harvard Journal of Law & Public Policy 29: 879889.
Feeley, Malcolm M., and Edward L. Rubin. 1999. Judicial Policy
Making and the Modern State: How the Courts Reformed
Americas Prisons. New York: Cambridge University Press.
Fisher, Louis. 1990. Separation of Powers: Interpretation Outside the
Courts. Pepperdine Law Review 18: 5793.
Fisher, Louis. 1993. The Legislative Veto: Invalidated, it Survives.
Law and Contemporary Problems 56: 273292.
Flemming, Roy B., and B. Dan Wood. 1997. The Public and the
Supreme Court: Individual Justice Responsiveness to American
Policy Moods. American Journal of Political Science 41:
468498.
Flemming, Roy B., John Bohte, and B. Dan Wood. 1997. One Voice
Among Many: The Supreme Courts Influence on Attentiveness to
Issues in the United States, 194792. American Journal of
Bibliography 229
Foley, Michael A. 2003. Arbitrary and Capricious: The Supreme Court,

the Constitution, and the Death Penalty. Westport, CT: Praeger.
Fowler, Thomas L. 2000. Of Moons, Thongs, Holdings and Dicta:
State v. Fly and the Rules of Law. Campbell Law Review 22:
253315.
Frank, Jerome. 1930. Law and the Modern Mind. New York:
Brentano.
Franklin, Charles H., and Liane C. Kosaki. 1989. Republican
Schoolmaster: The U.S. Supreme Court, Public Opinion, and
Abortion. American Political Science Review 83: 751771.
Franklin, Charles H., and Liane C. Kosaki. 1995. Media, Knowledge,
and Public Evaluations of the Supreme Court. In Lee Epstein,
Contemplating Courts. Washington, DC: Congressional
Quarterly: 352375.
Franklin, Charles H., Liane C. Kosaki, and Herbert Kritzer. 1993.
The Salience of United States Supreme Court Decisions. Paper
presented at the annual meeting of the American Political Science
Association, Washington, DC.
Friedelbaum, Stanley H. 1994. The Rehnquist Court: In Pursuit of
Judicial Conservativism. Westport, CT: Greenwood.
Friedman, Barry. 2002. The Birth of an Academic Obsession: The

History of the Countermajoritarian Difficulty. Yale Law Journal
112: 153259.
Friedman, Barry. 2003. Mediated Popular Constitutionalism.

Michigan Law Review 101: 25962636.
Friedman, Barry. 2004. The Importance of Being Positive: The Nature
and Function of Judicial Review. University of Cincinnati Law
Review 72: 12571303.
Friedman, Barry. 2005. The Politics of Judicial Review. Texas Law
Review 84: 257337.
Friedman, Leon, and Fred Israel. 1969. The Justices of the United
States Supreme Court, 17891969. New York: Chelsea House.
Friedman, Leon, and Fred Israel. 1980. The Justices of the United
States Supreme Court, 17891978. New York: Chelsea House.
230 Bibliography
Funston, Richard. 1975. The Supreme Court and Critical Elections.

Gallup, George, Jr., and Jim Castelli. 1987. The American Catholic
People: Their Beliefs, Practices, and Values. Garden City, NY:
Doubleday.
Gallup, George, Jr., and Jim Castelli. 1989. The Peoples Religion. New
York: Macmillan.
Garrow, David J. 2000. Mental Decrepitude on the U.S. Supreme
Court: The Historical Case for a 28th Amendment. University of
Chicago Law Review 67: 9951087.
Gates, John B. 1992. The Supreme Court and Partisan Realignment.
Boulder, CO: Westview.
George, Tracy E., and Lee Epstein. 1992. On the Nature of Supreme
Court Decision Making. American Political Science Review 86:
323337.
Gerber, Scott D., and Keeok Park. 1997. The Quixotic Search for
Consensus on the U.S. Supreme Court: A Cross-Judicial Empirical
Analysis of the Rehnquist Court Justices. American Political
Gerber, Scott D., and Keeok Park. 1999. First Principles: The
Jurisprudence of Clarence Thomas. New York: New York
University Press.
Gibson, James L. 1989. Understandings of Justice: Institutional
Legitimacy, Procedural Justice, and Political Tolerance. Law &
Society Review 23: 469496.
Gibson, James L., and Gregory A. Caldeira. 1992. Blacks and the
Supreme Court: Models of Diffuse Support. Journal of Politics
52: 11201145.
Gibson, James L., and Gregory A. Caldeira. 2003. Defenders of
Democracy? Legitimacy, Popular Acceptance, and the South
African Constitutional Court. Journal of Politics 65: 130.
Gibson, James L., and Gregory A. Caldeira. 2006. Confirmation
Politics and the Legitimacy of the U.S. Supreme Court.
Unpublished paper.
Gibson, James L., Gregory A. Caldeira, and Vanessa Baird. 1998. On
the Legitimacy of National High Courts. American Political
Bibliography 231
Gibson, James L., Gregory A. Caldeira, and Lester Kenyatta Spence.

2003. The Supreme Court and the U.S. Presidential Election of
2000: Wounds, Self-Inflicted or Otherwise? British Journal of
Political Science 33:535536.
Gibson, James L., Gregory A. Caldeira, and Lester Spence. 2003.
Measuring Attitudes Toward the United States Supreme Court,
American Journal of Political Science 47: 354367.
Gillman, Howard. 2001. The Votes That Counted. Chicago: University
of Chicago Press.
Gilligan, Carol. 1982. In a Different Voice: Psychological Theory and
Womens Development. Cambridge: Harvard University Press.
Ginsburg, Ruth Bader. 1985. Some Thoughts on Autonomy and
Equality in Relation to Roe v. Wade. North Carolina Law
Review 63: 375386.
Ginsburg, Ruth Bader. 1988. Informing the Public about the U.S.
Supreme Courts Work. Loyola University Chicago Law Journal
29: 275288.
Ginsburg, Ruth Bader. 1990. Remarks on Writing Separately.
Washington Law Review 65: 133150.
Ginsburg, Ruth Bader. 1992. Speaking in a Judicial Voice. New
York University Law Review 67: 11851209.
Gimpel, James. 1996. National Elections and the Autonomy of
American State Parties. Pittsburgh: University of Pittsburgh Press.
Goldman, Sheldon. 1997. Picking Federal Judges: Lower Court
Selection from Roosevelt through Reagan. New Haven: Yale
University Press.
Goldman, Sheldon, and Elliot Slotnick. 1997. Clintons First Term
Judiciary: Many Bridges to Cross. Judicature 80: 254273.
Gottlieb, Stephen E. 2000. Morality Imposed: The Rehnquist Court
and Liberty in America. New York: New York University Press.
Gottschall, Jon. 1983. Carters Judicial Appointments: The Influence
of Affirmative Action and Merit Selection in Voting on the US
Courts of Appeals. Judicature 67: 165173.
Greene, Abner. 2001. Understanding the 2000 Election. New York:
New York University Press.
232 Bibliography
Greenhouse, Linda. 2005. Becoming Justice Blackmun. New York:

Henry Holt.
Grosskopf, Anke, and Jeffery Mondak. 1998. Do Attitudes Toward
Specific Supreme Court Decisions Matter? The Impact of Webster
and Texas v. Johnson on Public Confidence in the Supreme
Court. Political Research Quarterly 51: 633654.
Gruhl, John. 1997. The Impact of Term Limits for Supreme Court
Justices. Judicature 81: 6672.
Gruhl, John, Cassia Spohn, and Susan Welch. 1981. Women as Policy
Makers: The Case of Trial Judges. American Journal of Political
Science 25: 308322.
Gryski, Gerard S., Gary Zuk, Deborah Barrow. 1994. A Bench That
Looks Like America? Representation of African Americans and
Latinos on the Federal Courts. Journal of Politics 56:
10761086.
Guliuzza, Frank III, Daniel Reagan, and David Barrett. 1994. The
Senate Judiciary Committee and Supreme Court Nominees:
Measuring the Dynamics of Confirmation Criteria. Journal of
Politics 56: 773787.
Gunning, Matthew, and Kyle Saunders. 1998. The Politics of
Realignment: Wattenberg Revisited. Paper presented at the annual
meeting of the Southern Political Science Association, Atlanta.
Hagle, Timothy M. 1993. Freshmen Effects for Supreme Court
Justices. American Journal of Political Science 37: 11421157.
Halpern, Stephen C., and Charles M. Lamb, eds. 1982. Supreme Court
Activism and Restraint. Lexington, MA: Lexington Books.
Hansford, Thomas G. 2004a. Information Provision, Organizational
Constraints, and the Decision to Submit an Amicus Curiae Brief
in a U.S. Supreme Court Case. Political Research Quarterly 57:
219230.
Hansford, Thomas G. 2004b. Lobbying Strategies, Venue Selection,
and Organized Interest Involvement at the U.S. Supreme Court.
American Politics Research 32: 170197.
Harwood, John, and Robert S. Greenberger. 2001. Bushs Judicial
Picks Signal the Beginning of Battle for Courts. Wall Street
Journal. May 10: A1.
Bibliography 233
Hausegger, Lori, and Lawrence Baum. 1998. Behind the Scenes: The
Supreme Court and Congress in Statutory Interpretation. In
Herbert F. Weisberg and Samuel C. Patterson, eds., Great
Theater: The American Congress in Action. New York:
Cambridge University Press: 224247.
Hausegger, Lori, and Lawrence Baum. 1999. Inviting Congressional
Action: A Study of Supreme Court Motivations in Statutory
Interpretation. American Journal of Political Science 43:
162185.
Haynie, Stacia L. 1992. Leadership and Consensus on the U.S.
Supreme Court. Journal of Politics 54: 11581169.
Hellman, Arthur D. 1996. The Shrunken Docket of the Rehnquist
Court. Supreme Court Review 1996: 403438.
Henschen, Beth. 1983. Statutory Interpretation of the Supreme Court:
Congressional Response. American Politics Quarterly 11:
441458.
Hettinger, Virginia, and Christopher Zorn. 2005. Explaining the
Incidence and Timing of Congressional Responses to the U.S.
Supreme Court. Legislative Studies Quarterly 30: 528.
Hibbing, John R., and Elizabeth Theiss-Morse. 1995. Congress as
Public Enemy. Cambridge: Cambridge University Press.
Hibbing, John R., and Elizabeth Theiss-Morse. 2001. What Is It About
Government That Americans Dislike? New York: Cambridge
University Press.
Hibbing, John R., and Elizabeth Theiss-Morse. 2002. Stealth
Democracy. New York: Cambridge University Press.
Hoekstra, Valerie. 1995. The Supreme Court and Opinion Change.
American Politics Quarterly 23: 109129.
Hoekstra, Valerie. 2000. The Supreme Court and Local Public
Opinion. American Political Science Review 94: 89100.
Hoekstra, Valerie. 2003. Public Reaction to Supreme Court Decisions.
New York: Cambridge University Press.
Hoekstra, Valerie, and Jeffrey A. Segal. 1996. The Shepherding of
Local Public Opinion: The Supreme Court and Lambs Chapel.
234 Bibliography
Holmes, Oliver Wendell. 1881. The Common Law. Boston: Little,

Brown.
Hughes, James M. 1993. Informing South Carolina Capital Juries
About Parole. South Carolina Law Review 44: 383410.
Hurwitz, Mark, S., William Mishler, and Reginald S. Sheehan. 2004.
The Influence of Public Opinion on Supreme Court Decision
Making, Revisited. Paper presented at the annual meeting of the
American Political Science Association, Chicago.
Hurwitz, Mark S., and Joseph V. Stefko. 2004. Acclimation and
Attitudes: Newcomer Justices and Precedent Conformance on
the Supreme Court. Political Research Quarterly 57: 121129.
Hutchings, Vincent L. 2001. Political Context, Issue Salience, and
Selective Attentiveness: Constituent Knowledge of the Clarence
Thomas Confirmation Vote. Journal of Politics 63: 846868.
Hutchinson, Dennis. 1979. Unanimity and Desegregation: Decision-
making in the Supreme Court, 19481958. Georgetown Law
Journal 68: 196.
Ignagni, Joseph, and James Meernik. 1994. Explaining Congressional
Attempts to Reverse Supreme Court Decisions. Political
Ignagni, Joseph, James Meernik, and Kimi King. 1998. Statutory
Construction and Congressional Response. American Politics
Quarterly 26: 459484.
Ivers, Gregg. 1992. Religious Organizations as Constitutional
Litigants. Polity 25: 243266.
Ivers, Gregg. 1995. To Build a Wall: American Jews and Separation of
Church and State. Charlottesville: University of Virginia Press.
Ivers, Gregg, and Karen OConnor. 1987. Friends as Foes: The
Amicus Curiae Participation and Effectiveness of the American
Civil Liberties Union and the Americans for Effective Law
Enforcement in Criminal Cases, 19691982. Law and Policy 9:
161178.
Jackson, Vicki. 2005. Constitutional Comparisons: Convergence,
Resistance, Engagement. Harvard Law Review 119: 109128.
Jageman, Jason F. 2001. Assessing the Effectiveness of Coalitional
Activity in Affirmative Action Cases Argued Before the Supreme
Bibliography 235
Court. Paper presented at the annual meeting of the Midwest

Jaros, Dean, and Robert Roper. 1980. The U.S. Supreme Court:
Myth, Diffuse Support, Specific Support, and Legitimacy.
American Politics Quarterly 8: 85105.
Johnson, Scott. 1999. The Influence of Case Complexity on the

Opinion Writing of the Rehnquist Court. Ohio Northern Law
Review 25: 4562.
Johnson, Timothy R. 2003. The Supreme Court, the Solicitor General,

and the Separation of Powers. American Politics Research 31:
426451.
Johnson, Timothy R. 2004. Oral Arguments and Decision Making on

the United States Supreme Court. Albany: State University of
New York Press.
Johnson, Timothy R., and Jason M. Roberts. 2004. Presidential

Capital and the Supreme Court Confirmation Process. Journal
of Politics 66: 663683.
Johnson, Timothy R., Paul Wahlbeck, and James Spriggs, II. 2006.
The Influence of Oral Arguments on the U.S. Supreme Court.
Johnson, Timothy R., Paul Wahlbeck, James Spriggs II, and Andrew D.
Martin. 1998. The Publics Conditional Response to Supreme
Court Decisions. American Political Science Review 92:
299309.
Jones, Jeffrey M. January 3, 2001. Opinion of U.S. Supreme Court

Has Become More Politicized. Gallup Opinion Poll:
http://www.gallup.com/poll/releases/pr010103b.asp.
Kahn, Ronald. 1999. Institutional Norms and Supreme Court

Decision-Making: The Rehnquist Court on Privacy and
Religion. In Cornell W. Clayton and Howard Gillman, eds.,
Supreme Court Decision-Making: New Institutionalist
Approaches. Chicago: University of Chicago Press: 175198.
Kash, Douglas A. 1997. Legal Development: Prewarrant Thermal

Imaging as a Fourth Amendment Violation: A Supreme Court
Question in the Making. Albany Law Review 60: 12951313.
236 Bibliography
Kearney, Joseph D., and Thomas W. Merrill. 2000. The Influence of

Amicus Curiae Briefs on the Supreme Court. University of
Pennsylvania Law Review 148: 743855.
Kennedy, Anthony M. 1990. The Constitution and the Spirit of
Freedom. Washington, DC: National Legal Center.
Key, V. O., Jr. 1952. The Future of the Democratic Party. Virginia
Quarterly Review 28: 161175.
Key, V. O., Jr. 1955. A Theory of Critical Elections. Journal of
Politics 17: 318.
Key, V. O., Jr. 1959. Secular Realignment and the Party System.
Kinder, Donald R., and Thomas R. Palfrey, eds. 1993. Experimental
Foundations of Political Science. Ann Arbor: University of
Michigan Press.
King, Gary, and Langche Zeng. 2001. Explaining Rare Events in
International Relations. International Organization 55:
693715.
Klarman, Michael J. 2001. How Great Were the Great Marshall
Court Decisions? Virginia Law Review 87: 11111183.
Kluger, Richard. 1976. Simple Justice. New York: Alfred A. Knopf.
Kobylka, Joseph F. 1991. The Politics of Obscenity. Westport, CT:
Greenwood Press.
Krislov, Samuel. 1963. The Amicus Curiae Brief: From Friendship to
Advocacy. Yale Law Journal 72: 694721.
Kritzer, Herbert M. 2001. The Impact of Bush v. Gore on Public
Perceptions and Knowledge of the Supreme Court. Judicature
85:3338.
Kritzer, Herbert M. 2005. The American Publics Assessment of the
Rehnquist Court. Judicature 89: 168176.
Kritzer, Herbert M., and Thomas M. Uhlman. 1977. Sisterhood in the
Courtroom: Sex of Judge and Defendant in Criminal Case
Disposition. Social Science Quarterly 14: 7788.
Kurland, Philip. 1970. Politics, the Constitution, and the Warren
Court. Chicago: University of Chicago Press.
Bibliography 237
Ladd, Everett Carll. 1985. On Mandates, Realignments, and the 1984

Presidential Election. Political Science Quarterly 100: 125.
Lamb, Charles M., and Stephen C. Halpern, eds. 1991. The Burger
Court. Urbana: University of Illinois Press.
Larson, Stephanie G., and Bryan Tramont. 1993. The Supreme Court
and Television: Predicting Case Coverage. Paper presented at the
annual meeting of the Midwest Political Science Association,
Chicago.
Lawrence, George. 1966. Tran., Alexis de Tocqueville, Democracy in
America. New York: Harper & Row.
Lawrence, Susan. 1990. The Poor in Court: The Legal Services
Program and Supreme Court Decision Making. Princeton, NJ:
Lazarus, Edward. 1999. Closed Chambers. New York: Penguin Books.
Leathers, John R. 1998/1999. Choice of Law in Kentucky. Kentucky
Law Journal 87: 583678.
Ledewitz, Bruce. 1999. When Federal Law Is Also State Law.
Temple Law Review 72: 561596.
Leeds, Jeffrey T. 1986. A Life on the Court. New York Times
Magazine, October 5: 79.
Lerner, Max. 1937. Constitution and Court as Symbols. Yale Law
Journal 46: 12911319.
Lichtman, Judith. 1990. Public Interest Groups and the Bork
Nomination. Northwestern University Law Review 84:
978979.
Lindquist, Stefanie, and David Yalof. 2001. Congressional Responses
to Federal Circuit Court Decisions. Judicature 85: 6168.
Lindquist, Stefanie, David Yalof, and John Clark. 2000. The Impact
of Presidential Appointments to the U.S. Supreme Court:
Cohesive and Divisive Voting within Presidential Blocs. Political
Link, Michael W. 1995. Tracking Public Mood in the Supreme Court:
Cross-Time Analysis of Criminal Procedure and Civil Rights
Cases. Political Research Quarterly 48: 6178.
238 Bibliography
Lublin, David, and D. Stephen Voss. 1999. Racial Redistricting and

Realignment in Southern State legislatures. Paper presented at
the annual meeting of the Southern Political Science Association,
Savannah.
Mace, George. 1972. The Antidemocratic Character of Judicial
Review. California Law Review 60: 11401149.
Macleod, Peter R. 1997. Latin in Legal Writing: An Inquiry into the
Use of Latin in the Modern Legal World. Boston College Law
Review 39: 235251.
Maltese, John Anthony. 1995. The Selling of Supreme Court
Nominees. Baltimore: Johns Hopkins University Press.
Maltz, Earl M. 1994. No Rules in a Knife Fight: Chief Justice
Rehnquist and the Doctrine of Stare Decisis. Rutgers Law
Journal 25: 669677.
Maltzman, Forrest, and Paul J. Wahlbeck. 1996. May It Please the
Chief? Opinion Assignments in the Rehnquist Court. American
Maltzman, Forrest, James F. Sprigs II, and Paul J. Wahlbeck. 2001.
Crafting Law on the Supreme Court: The Collegial Game. New
York: Cambridge University Press.
Mansbridge, Jane. 1999. Should Blacks Represent Blacks and Women
Represent Women? A Contingent Yes. Journal of Politics 61:
628657.
Mansbridge, Jane. 2003. Rethinking Representation. American
Mansbridge, Jane, and Katherine Tate. 1992. Race Trumps Gender:
The Thomas Nomination in the Black Community. PS: Political
Science and Politics 25: 488492.
Manwaring, David R. 1962. Render Unto Caesar: The Flag Salute
Controversy. Chicago: University of Chicago Press.
Marshall, Thomas R. 1989. Public Opinion and the Supreme Court.
Winchester, MA: Unwin Hyman.
Marshall, Thomas R. 1992. The Supreme Court and the Grass Roots:
Whom Does the Court Represent Best? Judicature 76: 2228.
Bibliography 239
Marshall, Thomas R. 1993. Symbolic versus Policy Representation on

the U.S. Supreme Court. Journal of Politics 55: 140150.
Marshall, Thomas R. 2003. Representing Public Opinion: American
Courts and the Appeals Process. Politics & Policy 31: 726747.
Martin, Andrew, and Kevin Quinn. 2002. Dynamic Ideal Point
Estimation via Markov Chain Monte Carlo for the U.S. Supreme
Court, 19531999. Political Analysis 10: 134153.
Martin, Elaine. 1993. Women on the Bench: A Different Voice?
Mason, Alpheus T. 1956. Harlan Fiske Stone: Pillar of the Law. New
York: Viking.
Mason, Alpheus T. 1958. The Supreme Court from Taft to Warren.
Baton Rouge: Louisiana State University Press.
Massaro, John. 1990. Supremely Political. Albany: State University of
New York Press.
Massey, Calvin. 2002. Federalism and the Rehnquist Court.
Hastings Law Journal 53: 431477.
Mauro, Tony. 2004. Glasnost at the Supreme Court. In Neal Devins
and Davison M. Douglas, eds., A Year at the Supreme Court.
Durham, NC: Duke University Press: 191208.
Maveety, Nancy. 1996. Justice Sandra Day OConnor: Strategist on the
Supreme Court. Lanham, MD: Rowan & Littlefield.
Maveety, Nancy. 2005. The End of the Choral Court. Judicature 89:
138145.
Maveety, Nancy, Charles Turner, and Lori Beth Way. 2004. Changes
in the Use of Concurrence: The Burger and Rehnquist Courts
Compared. Paper presented at the annual meeting of the
American Political Science Association, Chicago.
Mayhew, David. 2002. Electoral Realignments: A Critique of an
American Genre. New Haven: Yale University Press.
McCloskey, Robert G. 1960. The American Supreme Court. Chicago:
University of Chicago Press.
McGuire, Kevin T. 1993. Lawyers and the U.S. Supreme Court: The
Washington Community and Legal Elites. American Journal of
240 Bibliography
McGuire, Kevin T. 1994. Amici Curiae and Strategies for Gaining

Access to the Supreme Court. Political Research Quarterly 47:
821837.
McGuire, Kevin T. 1995. Repeat Players in the Supreme Court: The
Role of Experienced Lawyers in Litigation Success. Journal of
McGuire, Kevin T. 1998. Explaining Executive Success in the U.S.
Supreme Court. Political Research Quarterly 51: 505526.
McGuire, Kevin T., and Gregory A. Caldeira. 1993. Lawyers,
Organized Interests, and the Law of Obscenity: Agenda Setting in
the Supreme Court. American Political Science Review 87:
717728.
McGuire, Kevin T., and James A. Stimson. 2004. The Least
Dangerous Branch Revisited: New Evidence on Supreme Court
Responsiveness to Public Preferences. Journal of Politics 66:
10181035.
McIntosh, Wayne V. 1985. Litigating Scientific Creationism or Scopes
I, II, III. Law and Policy 7: 375385.
Medalie, Richard C. 1968. From Escobedo to Miranda: The Anatomy
of a Supreme Court Decision. Washington, DC: Lerner Book Co.
Meernik, James, and Joseph Ignagni. 1995. Congressional Attacks on
Supreme Court Rulings Involving Unconstitutional State Laws.
Political Research Quarterly 48: 4359.
Meernik, James, and Joseph Ignagni. 1997. Judicial Review and
Coordinate Construction of the Constitution. American Journal
of Political Science 41: 447467.
Meffert, Michael F., Helmut Norpoth, and Anirudh V.S. Ruhil. 2001.
Realignment and Macropartisanship. American Political
Melone, Albert. 1991. The Senates Confirmation Role in Supreme
Court Nominations and the Politics of Ideology versus
Impartiality. Judicature 75: 6879.
Melone, Albert, Alan Morris, and Marc-George Pufong. 1992. Too
Little Advice, Senatorial Responsibility, and Confirmation
Politics. Judicature 75: 187192.
Bibliography 241
Merrill, Thomas. 2003. The Making of the Second Rehnquist Court:

A Preliminary Analysis. St. Louis Law Journal 47: 569653.
Mezey, Susan Gluck. 1996. Children in Court. Albany: State University
of New York Press.
Miller, Geoffrey. 1987. The True Story of Carolene Products. The
Supreme Court Review. Chicago: University of Chicago Press.
Mischler, William, and Reginald S. Sheehan. 1993. The Supreme
Court as a Countermajoritarian Institution? The Impact of Public
Opinion on Supreme Court Decisions. American Political
Mischler, William, and Reginald S. Sheehan. 1994. Response: Popular
Influence on Supreme Court Decisions. American Political
Mischler, William, and Reginald S. Sheehan. 1996. Public Opinion,
the Attitudinal Model, and Supreme Court Decision Making: A
Micro-Analytic Perspective. Journal of Politics 58: 169200.
Mondak, Jeffery J. 1990. Perceived Legitimacy of Supreme Court
Decisions: Three Functions of Source Credibility. Political
Behavior 12: 363384.
Mondak, Jeffery J. 1991. Substantive and Procedural Aspects of
Supreme Court Decisions as Determinants of Institutional
Approval. American Politics Quarterly 19: 174188.
Mondak, Jeffery J. 1992. Institutional Legitimacy, Policy Legitimacy,
and the Supreme Court. American Politics Quarterly 20:
457477.
Mondak, Jeffery J. 1994. Policy Legitimacy and the Supreme Court:
The Sources and Contexts of Legitimation. Political Research
Mondak, Jeffery, and Shannon Smithey. 1997. The Dynamics of
Public Support for the Supreme Court. Journal of Politics 59:
111442.
Monroe, Alan D. 1979. Consistency between Policy Preferences and
National Policy Decisions. American Politics Quarterly 7: 318.
Monroe, Alan D. 1998. Public Opinion and Public Policy,
19801993. Public Opinion Quarterly 62: 628.
242 Bibliography
Moore, Michael. 2001. Justifying the Natural Law Theory of

Constitutional Interpretation. Fordham Law Review 69:
20872117.
Moraski, Bryon J., and Charles R. Shipan. 1999. The Politics of
Supreme Court Nominations: A Theory of Institutional
Constraints and Choices. American Journal of Political Science
43: 10691095.
Murphy, Walter. 1959. Lower Court Checks Upon Supreme Court
Power. American Political Science Review 53: 10171036.
Murphy, Walter. 1962. Congress and the Court. Chicago: University of
Chicago Press.
Murphy, Walter. 1964. Elements of Judicial Strategy. Chicago:
University of Chicago Press.
Murphy, Walter, and C. Herman Pritchett. 1961. Courts, Judges, and
Politics: An Introduction to the Judicial Process. New York:
Random House.
Murphy, Walter, and John Tanenhaus. 1969. Public Opinion and the
United States Supreme Court: A Preliminary Mapping of Some
Prerequisites for Court Legitimation of Regime Changes. In Joel
Grossman and John Tanenhaus, eds. Frontiers in Judicial
Research. New York: Wiley: 273303.
Nagareda, Richard A. 1998. Outrageous Fortune and the
Criminalization of Mass Torts. Michigan Law Review 96:
11211198.
Nardulli, Peter. 1995. The Concept of a Critical Realignment:
Electoral Behavior, and Political Change. American Political
Neumann, Elizabeth Noelle. 1977. Public Opinion and the Classical
Tradition: A Reevaluation. Public Opinion Quarterly 43:
143156.
Nice, Julie A. 1999. The Emerging Third Strand in Equal Protection
Jurisprudence. University of Illinois Law Review 1999:
12091271.
Norpoth, Helmut, and Jeffrey A. Segal. 1994. Comment: Popular
Influence on Supreme Court Decisions. American Political
Bibliography 243
Note. 1954. Evasion of Supreme Court Mandates in Cases Relating to

State Courts since 1941. Harvard Law Review 67: 12511259.
Note. 1958. Congressional Reversals of Supreme Court Decisions
19451957. Harvard Law Review 71: 13241337.
Note. 1989. The Meaning of Life for Virginia Jurors and Its Effect
on Reliability in Capital Sentencing. Virginia Law Review 75:
16051638.
Note. 1976. Opinion Polls and the Law of Evidence. Virginia Law
Review 62: 11011133.
OBrien, David M. 1997. The Rehnquist Courts Shrinking Plenary
Docket. Judicature 81: 5865.
OBrien, David M. 1999. Institutional Norms and the Supreme Court
Opinions: On Reconsidering the Rise of Individual Opinions. In
Cornell W. Clayton and Howard Gillman, eds., Supreme Court
Decision-Making: New Institutionalist Approaches. Chicago:
University of Chicago Press: 91114.
OBrien, David M. 2000. Storm Center: The Supreme Court in
American Politics. New York: Norton.
OBrien, David M. 2005. A Diminished Plenary Docket: A Legacy of
the Rehnquist Court. Judicature 89: 134137.
OConnor, Karen. 1980. Womens Organizations Use of the Courts.
Lexington, MA: Lexington Books.
OConnor, Karen. 1996. No Neutral Grounds: Abortion Politics in an
Age of Absolutes. Boulder, CO: Westview.
OConnor, Karen, and Lee Epstein. 1984. The Role of Interest Groups
in Supreme Court Policymaking. In Robert Eyestone, ed., Public
Policy Formation, Greenwich, CT: JAI Press: 6381.
OConnor, Karen, and Bryan Scott McFall. 1992. Conservative
Interest Group Litigation in the Reagan Era and Beyond. In
Mark Petracca, ed., The Politics of Interests: Interest Groups
Transformed. Boulder, CO: Westview: 263281.
OConnor, Karen, and Barbara Palmer. 2001. The Clinton Clones:
Ginsburg, Breyer, and the Clinton Legacy. Judicature 84:
262273.
244 Bibliography
Olivas, Michael A. 1997. Affirmative Action: Diversity of Opinions:

Constitutional Criteria: The Social Science and Common Law of
Admissions Decisions in Higher Education. Colorado Law
Review 68: 10651121.
Olson, Mancur. 1984. Clients and Lawyers: Securing the Rights of
Disabled Persons. Westport, CT: Greenwood Press.
ONeill, Timothy J. 1985. Bakke & the Politics of Equality.
Middletown, CT: Wesleyan University Press.
Overby, L. Marvin, Beth M. Henschen, Michael Walsh, and Julie
Strauss. 1992. Courting Constituents? An Analysis of the Senate
Confirmation Vote on Justice Clarence Thomas. American
Overby, L. Marvin, and Robert Brown. 1995. Electoral Coalitions
and the Politics of Supreme Court Confirmation Votes. Paper
Owens, Ryan J., and Lee Epstein. 2005. Amici Curiae During the
Rehnquist Years. Judicature 89: 127133.
Pacelle, Richard, Jr. 1991. The Transformation of the Supreme Courts
Agenda. Boulder, CO: Westview.
Pacelle, Richard, Jr. 2003. Between Law & Politics: The Solicitor
General and the Structuring of Race, Gender, and Reproductive
Rights Litigation. College Station: Texas A&M University Press.
Paduano, Anthony, and Clive A. Stafford Smith. 1987. Deathly
Errors: Juror Misperceptions Concerning Parole in the Imposition
of the Death Penalty. Columbia Human Rights Law Review 18:
211257.
Page, Benjamin I., and Robert Y. Shapiro. 1983. Effects of Public
Opinion on Policy. American Political Science Review 77:
175190.
Page, Benjamin I., and Robert Y. Shapiro. 1992. The Rational Public.
Chicago: University of Chicago Press.
Page, Benjamin I., Robert Y. Shapiro, and Glenn R. Dempsey. 1987.
What Moves Public Opinion? American Political Science
Review 81: 2343.
Bibliography 245
Palmer, Paul A. 1936. The Concept of Public Opinion in Political

Theory. In Essays in History and Political Theory in Honor of
Charles H. McIlwain. Cambridge: Harvard University Press.
Paschal, Richard. 1991. The Continuing Colloquy: Congress and the
Finality of the Supreme Court. Journal of Law and Politics 8:
142226.
Paulson, Arthur. 2000. Realignment and Party Revival. Westport, CT:
Praeger.
Pederson, William D., and Norman W. Provizer. 2003. Leaders of the
Pack: Polls & Case Studies of Great Supreme Court Justices. New
York: Peter Lang.
Peretti, Terri Jennings. 1999. In Defense of a Political Court. Princeton,
NJ: Princeton University Press.
Perine, Keith. 2005. Fiercest Fight in Partisan War May Be Over
Supreme Court. Congressional Quarterly Weekly Review:
January 7: 5864.
Perrin, L. Timothy. 2001. Pricking Boils, Preserving Error: On the
Horns of a Dilemma After Ohler v. United States. U.C. Davis
Perry, Barbara A. 1989. The Life and Death of the Catholic Seat on
the United States Supreme Court. Journal of Law & Politics 55:
5592.
Perry, Barbara A. 1991. A Representative Supreme Court? The
Impact of Race, Religion, and Gender on Appointments.
Westport, CT: Greenwood Press.
Perry, Barbara A. 1999. The Priestly TribeThe Supreme Courts
Image in the American Mind. Westport, CT: Praeger.
Perry, Barbara A., and Henry J. Abraham. 1998. A Representative
Supreme Court? The Thomas, Ginsburg, and Breyer
Appointments. Judicature 81: 158165.
Peters, C. Scott. 1995. Public Support for the Supreme Court: Values,
Awareness and Contextual Information. Paper presented at the
Chicago.
Peterson, Steven A. 1981. Dissent in American Courts. Journal of
246 Bibliography
Petry, Francois. 1999. The Opinion-Policy Relationship in Canada.

Phillips, Carter G. 2004. Was Affirmative Action Saved by Its
Friends? In Neal Devins and Davison M. Douglas, eds., A Year
at the Supreme Court. Durham, NC: Duke University Press:
113130.
Pitkin, Hanna. 1967. The Concept of Representation. Berkeley:
University of California Press.
Ponnuru, Ramesh. 2004. The Courts Faux Federalism. In Neal
Devins and Davison M. Douglas, eds. A Year at the Supreme
Court. Durham, NC: Duke University Press: 131150.
Posner, Richard A. 1996. The Federal Courts. Cambridge: Harvard
University Press.
Posner, Richard A. 2001. Breaking the Deadlock. Princeton, NJ:
Posner, Richard A. 2003. Law, Pragmatism, and Democracy.
Cambridge: Harvard University Press.
Posner, Richard A. 2005. Foreword: A Political Court. Harvard Law
Review 119: 32102.
Price, Vincent, and Anca Romantan. 2004. Confidence in Institutions
Before, During, and After Indecision 2000. Journal of Politics
66: 939956.
Pritchett, C. Herman. 1961. Congress versus the Supreme Court.
Minneapolis: University of Minnesota Press.
Radin, Margaret Jane. 2002. Can the Rule of Law Survive Bush v.
Gore? In Bruce Ackerman, ed., Bush v. Gore. New Haven: Yale
University Press: 110125.
Rathjen, Gregory J., and Harold Spaeth. 1979. Access to the Federal
Courts: An Analysis of Burger Court Policy Making. American
Ray, Laura Krugman. 2000. The Road to Bush v. Gore. Nebraska
Rehnquist, William H. 1986. Constitutional Law and Public
Opinion. Suffolk University Law Review 20: 751769.
Bibliography 247
Rehnquist, William H. 1987. The Supreme Court: How It Was, How It

Is. New York: Morrow.
Rehnquist, William H. 1998. All The Laws But One. New York:
Alfred A. Knopf.
Reynolds, William Bradford. 1991. The Confirmation Process: Too
Much Advice and Too Little Consent. Judicature 75: 8082.
Richards, Mark J., and Herbert M. Kritzer. 2002. Jurisprudential
Regimes in Supreme Court Decision Making. American Political
Ring, Kevin A., ed. 2004. Scalia Dissents. Washington, DC: Regnery.
Robinson, Claude. 1932. Straw PollsA Study of Political Prediction.
New York: Columbia University Press.
Robinson, Tracy. 2004. By Popular Demand? George Mason
University Civil Rights Law Journal 14: 107144.
Roesch, Benjamin. 2006. Crowd Control: The Majoritarian Court
and the Reflection of Public Opinion in Doctrine. Suffolk
University Law Review 39: 379423.
Rogers, James R. 1999. Legislative Incentives and Two-Tiered Judicial
Review: A Game Theoretic Reading of Carolene Products
Footnote Four. American Journal of Political Science 43:
10961121.
Rohde, David W. 1972. Policy Goals and Opinion Coalitions in the
Supreme Court. Midwest Journal of Political Science 16:
208224.
Romano, Roberta. 1992. Comment on Spiller and Spitzers Judicial
Choice of Legal Doctrines. Journal of Law, Economics, and
Organization 8: 4758.
Rosen, Jeffrey. 2003. How to Reignite the Culture Wars. New York
Times Magazine, September 7: 4750.
Rosen, Jeffrey. 2006. The Most Democratic Branch: How the Courts
Serve America. New York: Oxford University Press.
Rosenberg, Gerald N. 1993. The Hollow Hope: Can Courts Bring
About Social Change? Chicago: University of Chicago Press.
248 Bibliography
Rosenof, Theodore. 2003. Realignment: The Theory That Changed the

Way We Think about American Politics. Lanham, MD: Rowman
& Littlefield.
Ruckman, P. S., Jr. 1993. The Supreme Court, Critical Nominations,
and the Senate Confirmation Process. Journal of Politics 55:
793805.
Ryan, John Francis. 1998. Is There Leadership on this Court? A
Cross-Time Examination of Supreme Court Disunity. Paper pre-
sented at the annual meeting of the Southern Political Science
Association meeting, Atlanta.
Salokar, Rebecca Mae. 1992. The Solicitor General: The Politics of
Law. Philadelphia: Temple University Press.
Savage, David G. 1992. Turning Right: The Making of the Rehnquist
Supreme Court. New York: Wiley.
Savage, David G. 1995. Docket Reflects Ideological ShiftsShrinking
Caseload, Cert Denials Suggest an Unfolding Agenda. ABA
Journal, December: 4042.
Scalia, Antonin. 1989a. The Rule of Law as a Law of Rules.
University of Chicago Law Review 56: 11751188.
Scalia, Antonin. 1989b. Originalism: The Lesser Evil. University of
Cincinnati Law Review 57: 849865.
Scalia, Antonin. 1997. A Matter of Interpretation. Princeton, NJ:
Scheb, John M. II, and William Lyons. 1998. Public Perception of the
Supreme Court in the 1990s. Judicature 82: 6669.
Scheb, John M. II, and William Lyons. 2000. The Myth of Legality
and Public Evaluation of the Supreme Court. Social Science
Scheb, John M. II, and William Lyons. 2001. Judicial Behavior and Public
Opinion: Popular Expectations Regarding the Factors that Influence
Supreme Court Decisions. Political Behavior 23: 181194.
Schlozman, Kay, and John Tierney. 1986. Organized Interests and
American Democracy. New York: Harper & Row.
Schmidhauser, John, and Larry Berg. 1972. The Supreme Court and
Congress: Conflict and Interaction, 19451968. New York: Free
Press.
Bibliography 249
Schneider, William. 1985. The November 6 Vote for President: What

Did It Mean. In Austin Ranney, ed., The American Elections of
1984. Washington, DC: American Enterprise Institute: 203244.
Schoen, Ryan, and Paul J. Wahlbeck. 2006. The Discuss List and
Agenda-Setting on the Supreme Court. Paper presented at the
annual meeting of the Southern Political Science Association,
Atlanta.
Schultz, David A., and Christopher E. Smith. 1996. The Jurisprudential
Vision of Justice Antonin Scalia. Lanham, MD: Rowman &
Littlefield.
Schwartz, Bernard. 1993. A History of the Supreme Court. New York:
Oxford University Press.
Schwartz, Bernard. 1997. A Book of Legal Lists. New York: Oxford
University Press.
Segal, Jeffrey A. 1988. Amicus Curiae Briefs by the Solicitor General
during the Warren and Burger Courts. Western Political
Segal, Jeffrey A. 1990. Supreme Court Support for the Solicitor
General: The Effects of Presidential Appointments. Western
Political Quarterly 43: 137152.
Segal, Jeffrey A., and Albert D. Cover. 1989. Ideological Values and
the Votes of U.S. Supreme Court Justices. American Political
Segal, Jeffrey A., and Cheryl D. Reedy. 1988. The Supreme Court and
Sex Discrimination: The Role of the Solicitor General. Western
Political Quarterly 41: 553568.
Segal, Jeffrey A., and Harold J. Spaeth. 1993. The Supreme Court and
the Attitudinal Model. New York: Cambridge University Press.
Segal, Jeffrey A., and Harold J. Spaeth. 2002. The Supreme Court and
the Attitudinal Model Revisited. New York: Cambridge
University Press.
Segal, Jeffrey A., Lee Epstein, Charles M. Cameron, and Harold J.
Spaeth. 1995. Ideological Values and the Votes of Justices
Revisited. Journal of Politics 57: 812823.
Segal, Jeffrey A., Richard Timpone, and Robert Howard. 2000. Buyer
Beware? Presidential Success through Supreme Court
Appointments. Political Research Quarterly 53: 557595.
250 Bibliography
Segal, Jennifer. 1999. Diffuse Support for the U.S. Supreme Court:
Reliable Reservoir or Fickle Foundation? American Review of
Politics 20: 124.
Segal, Jennifer. 2000. Representative Decision Making on the Federal
Bench: Clintons District Court Appointees. Political Research
Shafer, Byron, ed. 1991. The End of Realignment? Madison: University
of Wisconsin Press.
Shapiro, David. 1976. Mr. Justice Rehnquist: A Preliminary View.
Harvard Law Review 90: 293357.
Shapiro, Martin. 1986. APA: Past, Present, Future. Virginia Law
Review 72: 447490.
Sheehan, Colleen A. 2004. Madison v. Hamilton: The Battle Over
Republicanism and the Role of Public Opinion. American
Sheehan, Reginald S., William Mishler, and Donald R. Songer. 1992.
Ideology, Status, and the Differential Success of Direct Parties
Before the Supreme Court. American Political Science Review
86: 464471.
Sheldon, Charles. 1974. The American Judicial Process. New York:
Dodd, Mead.
Silbey, Joel H. 1991. Beyond Realignment and Realignment Theory:
American Political Eras, 17891989. In Byron E. Shafer, ed.,
The End of Realignment? Madison: University of Wisconsin
Press: 323.
Silverstein, Mark. 1994. Judicious Choices. New York: W.W. Norton.
Simmons, Wendy M. January 16, 2001. Election Controversy
Apparently Drove Partisan Wedge Into Attitudes Toward
Supreme Court. Gallup Opinion Poll: http://www.gallup.com/
poll/releases/pr010116.asp.
Simon, James F. 1996. The Center Holds: The Power Struggle Inside
the Rehnquist Court. New York: Simon & Schuster.
Slotnick, Elliot. 1984. The Paths to the Federal Bench: Gender, Race,
and Judicial Recruitment Variation. Judicature 67: 370388.
Bibliography 251
Slotnick, Elliot, and Jennifer Segal. 1998. Television News and the
Supreme Court: All the News Thats Fit To Air? Cambridge:
Cambridge University Press.
Smith, Christopher E. 1993. Critical Judicial Nominations and Political
Change: The Impact of Clarence Thomas. Westport, CT: Praeger.
Smith, Christopher E. 1997. Courts, Politics, and the Judicial Process.
Chicago: Nelson-Hall.
Smith, Christopher E., and Joyce A. Baugh. 2000. The Real Clarence
Thomas: Confirmation Veracity Meets Performance Reality. New
York: Peter Lang.
Smith, Tom W. 1990a. Liberal and Conservative Trends in the United
States since World War II. Public Opinion Quarterly 54:
479507.
Smith, Tom W. 1990b. The First Straw? A Study of the Origins of
Election Polls. Public Opinion Quarterly 54: 2136.
Solomine, Michael E., and James L. Walker. 1992. The Next Word:
Congressional Response to Supreme Court Statutory Decisions.
Temple Law Review 65: 425458.
Songer, Donald R., Sue Davis, and Susan Haire. 1994. A Reappraisal
of Diversification in the Federal Courts: Gender Effects in the
Courts of Appeals. The Journal of Politics 56: 425439.
Songer, Donald R., Ashlyn Kiersten, and Erin Kaheny. 2000. Why the
Haves Dont Always Come Out Ahead: Repeat Players Meet
Amici Curiae for the Disadvantaged. Political Research
Sorauf, Frank J. 1976. The Wall of Separation: Constitutional Politics
of Church and State. Princeton, NJ: Princeton University Press.
Sorensen, Robert, and Theodore Sorensen. 1953. The Admissibility
and Use of Opinion Research Evidence. New York University
Law Review 28: 12131261.
Spaeth, Harold, and Jeffrey Segal. 1999. Majority Rule or Minority
Will: Adherence to Precedent on the U.S. Supreme Court.
Cambridge: Cambridge University Press.
Spier, Hans. 1950. Historical Development of Public Opinion.
American Journal of Sociology 55: 378388.
252 Bibliography
Spill, Rorie, and Zoe Oxley. 2003. Philosopher Kings or Political

Actors? How Do Reporters Cover the Supreme Court?
Spiller, Pablo T., and Matthew Spitzer. 1992. Judicial Choice of Legal
Doctrines. Journal of Law, Economics and Organization 8:
846.
Spiller, Pablo T., and Emerson H. Tiller. 1993. Invitations to
Override: Congressional Reversals of Supreme Court Decisions.
Typescript. University of California-Berkeley.
Spohn, Cassie, John Gruhl, and Susan Welch. 1981. The Effects of
Race on Sentencing. Law & Society Review 16: 7188.
Spriggs, James F., II, and Paul J. Wahlbeck. 1997. Amicus Curiae and
the Role of Information at the Supreme Court. Political
Spriggs, James F., II and Thomas G. Hansford. 2001. Explaining the
Overruling of U.S. Supreme Court Precedent. Journal of Politics
63: 10911111.
Spriggs, James F. II, Forrest Maltzman, and Paul Wahlbeck. 1999.
Bargaining on the U.S. Supreme Court: Justices Responses to
Majority Opinion Drafts. Journal of Politics 61: 485506.
Stephenson, Donald Grier, Jr. 1999. Campaigns and the Court: The
U.S. Supreme Court in Presidential Elections. New York:
Columbia University Press.
Stern, Robert H., Eugene Gressman, and Stephen M. Shapiro. 1986.
Supreme Court Practice, 6th ed. Washington, DC: Bureau of
National Affairs.
Stewart, Carl. 2003. A Reflection on Contemporary Ethical Issues
Facing Judges. University of St. Thomas Law Journal 1:
464479.
Stewart, David O. 1996. The Uncertainty of CertPredicting Courts
Choices of Cases Can Be Anyones Guess. ABA Journal, April:
5051.
Stidham, Ronald, Robert A. Carp, and Donald R. Songer. 1996. The
Voting Behavior of President Clintons Judicial Appointees.
Bibliography 253
Stimson, James A. 1991, 1999. Public Opinion in AmericaMoods,

Cycles, and Swings. Boulder, CO: Westview.
Stimson, James A., Michael B. MacKuen, and Robert S. Erikson. 1995.
Dynamic Representation. American Political Science Review
89: 543565.
Storing, Herbert J. 1985. The Anti-Federalist. Chicago: University of
Chicago Press.
Stoutenborough, James W., and Donald Haider-Markel. 2006. Public
Confidence in the U.S. Supreme Court: A New Look at the
Impact of Court Decisions. Paper presented at the annual meet-
ing of the Midwest Political Science Association, Chicago.
Strauss, David A. 1998. The Solicitor General and the Interests of the
United States. Law and Contemporary Problems 61: 165.
Stumpf, Harry. 1965. Congressional Response to Supreme Court
Rulings: The Interaction of Law and Politics. Journal of Public
Law 14: 377395.
Sundquist, James L. 1983. Dynamics of the Party System. Washington,
DC: Brookings.
Sungaila, Mary-Christine. 1999. Effective Amicus Practice before the
United States Supreme Court: A Case Study. Southern California
Review of Law and Womens Studies 8: 187196.
Sunstein, Cass R. 1996. Leaving Things Undecided. Harvard Law
Review 110: 4101.
Szmer, John, and Donald R. Songer. 2005. The Effects of Information
on the Accuracy of Presidential Assessments of Supreme Court
Nominee Preferences. Political Research Quarterly 58: 151160.
Taylor, Stuart, Jr. 2004. The Affirmative Action Decisions. In Neal
Devins and Davison M. Douglas, A Year at the Supreme Court.
Durham, NC: Duke University Press: 87112.
Thibaut, John, and Laurens Walker. 1975. Procedural Justice. New
York: Wiley.
Thomas, Clarence. 1987. Toward a Plain Reading of the
ConstitutionThe Declaration of Independence in Constitutional
Interpretation. Howard Law Journal 30: 983995.
254 Bibliography
Thomas, Clarence. 1989. The Higher Law Background of the

Privileges or Immunities Clause of the Fourteenth Amendment.
Harvard Journal of Law and Public Policy 12: 6370.
Tilleman, Karl, and Shane Swindle. 1999. Closing the Book on
CERCLA Section 107 Joint and Several Claims by Liable Private
Parties. Virginia Environmental Law Journal 18: 159182.
Tiller, Emerson, and Pablo T. Spiller. 1999. Strategic Instruments:
Legal Structure and Policy Games in Administrative Law.
Journal of Law, Economics & Organization 15: 34977.
Todd, James S. 2004. Overturning Acts of Congress on the Rehnquist
Court: Will the Real Judicial Activist Please Stand Up? Paper
presented at the annual meeting of the American Political Science
Association, Washington, DC.
Toma, Eugenia. 1991. Congressional Influence and the Supreme
Court: The Budget as a Signaling Device. Journal of Legal
Studies 20: 131147.
Toner, Robin. 2001. Interest Groups Set for Battle on a Supreme
Court Vacancy. New York Times, April 21: 1A.
Tribe, Lawrence. 2002. Through the Looking Glass. In Bruce
Ackerman, ed., Bush v. Gore. New Haven: Yale University Press.
Tuchfarber, Alfred J., Stephen Bennett, Andrew Smith, and Eric
Rademacher. 1995. The Republican Tidal Wave of 1994:
Testing Hypotheses About Realignment, Restructuring, and
Rebellion. PS: Political Science & Politics: 689696.
Tuchfarber, Alfred J., Andrew Smith, Eric Rademacher, and Kimberly
Downing. 1997. Realignment: Evidence from the 1996 Election
and a Look Ahead to 1998. Paper presented at the annual meet-
ing of the American Political Science Association, Washington,
DC.
Tushnet, Mark. 1987. The NAACPs Legal Strategy against Segregated
Education, 19251950. Chapel Hill: University of North
Carolina Press.
Tushnet, Mark. 1999. The New Constitutional Order and the

Chastening of Constitutional Aspiration. Harvard Law Review
113: 29109.
Bibliography 255
Tushnet, Mark. 2002. The Conservativism in Bush v. Gore. In Bruce

Ackerman, ed., Bush v. Gore. New Haven: Yale University Press:
110121.
Tyler, Tom R. 1990. Why People Obey the Law. New Haven: Yale
University Press.
Tyler, Tom R. 2001. The Psychology of Public Dissatisfaction with
Government. In John R. Hibbing and Elizabeth Theiss-Morse,
eds. What Is It About Government That Americans Dislike? New
York: Cambridge University Press: 227242.
Tyler, Tom R., and Yuen J. Huo. 2002. Trust in the Law. New York:
Russell Sage.
Tyler, Tom R., and E. Allan Lind. 1988. The Social Psychology of
Procedural Justice. New York: Plenum Press.
Tyler, Tom R., and Gregory Mitchell. 1994. Legitimacy and the
Empowerment of Discretionary Legal Authority, The United
States Supreme Court, and Abortion Rights. Duke Law Journal
42: 703802.
Tyler, Tom R., and Kenneth Rasinski. 1991. Procedural Justice,
Institutional Legitimacy, and the Acceptance of Unpopular U.S.
Supreme Court Decisions: A Reply to Gibson. Law and Society
Review 25: 621630.
Tyler, Tom R., Jonathan D. Casper, and Bonnie Fisher. 1989.
Maintaining Allegiance toward Political Authorities: The Role of
Prior Attitudes and the Use of Fair Procedures. American
Umhofer, Matthew D. 2001. Confusing Pursuits: Sacramento v.
Lewis and the Future of Substantive Due Process in the Executive
Setting. Santa Clara Law Review 41: 437508.
Uhlman, Thomas. 1978. Black Elite Decision Making: The Case of
Trial Judges. American Journal of Political Science 22: 884895.
Ulmer, S. Sidney. 1971. Earl Warren and the Brown Decision.
Unger, Michael. 2003. Changing Minds? The U.S. Supreme Courts
Impact on Abortion and Death Penalty Attitudes in 1989. Paper
256 Bibliography
Van Sickel, Robert. 1998. Not a Particularly Different Voice: The

Jurisprudence of Sandra Day OConnor. 1998. New York: Peter
Lang.
Vermeer, Jan P. 1993. Supreme Court Cases as Guides to State and
Local Policy Making: An Examination of Local Newspaper
Editorials. Paper presented at the annual meeting of the Midwest
Vidmar, Neil, and Phoebe Ellsworth. 1974. Public Opinion and the
Death Penalty. Stanford Law Review 26: 12451270.
Vitiello, Michael. 1999. How Imperial Is the Supreme Court? An
Analysis of Supreme Court Abortion Doctrine and Popular Will.
University of San Francisco Law Review 34: 4998.
Vose, Clement E. 1957. The National Consumers League and the
Brandeis Brief. Midwest Journal of Political Science 1: 178190.
Vose, Clement E. 1959. Caucasians Only. Berkeley: University of
California Press.
Wahlbeck, Paul J., James F. Spriggs II, and Forrest Maltzman. 1998.
Marshalling the Court: Bargaining and Accommodation on the
United States Supreme Court. American Journal of Political
Science 42: 294315.
Walker, Thomas, and Deborah Barrow. 1985. The Diversification of
the Federal Bench: Policy and Process Ramifications. Journal of
Walker, Thomas, Lee Epstein, and William Dixon. 1988. On the
Mysterious Demise of Consensual Norms in the United States
Supreme Court. The Journal of Politics 50: 361389.
Ward, Artemus. 2003. Deciding to LeaveThe Politics of Retirement
from the United States Supreme Court. Albany: State University
of New York Press.
Warren, Charles. 1935. Congress, The Constitution, and the Supreme
Court. Boston: Little, Brown.
Wasby, Stephen L., Anthony dAmato, and Rosemary Metrailer. 1977.
Desegregation From Brown to Alexander. Carbondale:
University of Illinois Press.
Wasby, Stephen L. 1993. The Supreme Court. Chicago: Nelson-Hall.
Bibliography 257
Wasby, Stephen L. 1995. Race Relations Litigation in an Age of

Complexity. Charlottesville: University Press of Virginia.
Wattenberg, Martin. 1994. The Decline of American Political Parties
19521992. Cambridge: Harvard University Press.
Way, Frank, and Barbara J. Burt. 1983. Religious Marginality and the
Free Exercise Clause. American Political Science Review 77:
652665.
Weissberg, Robert. 1976. Public Opinion and Popular Government.
Englewood Cliffs, NJ: Prentice-Hall.
Weizer, Paul. 2004. The Opinions of Justice Antonin Scalia. New
York: Peter Lang.
Welch, Susan, Michael Combs, and John Gruhl. 1988. Do Black
Judges Make a Difference? American Journal of Political Science
32: 126136.
Wellington, Harry. 1982. The Nature of Judicial Review. Yale Law
Journal 91: 486520.
Wheeler, Russell R. 2005. Chief Justice Rehnquist as third branch
leader. Judicature 89: 116120.
White, Byron R. 1999. ForwardShould the Ninth Circuit Be
Saved? Journal of Law & Politics 15: 379382.
White, G. Edward. 1976. The American Judicial Tradition. New York:
Oxford University Press.
Whittington, Keith E. 2001. Presidential Challenges to Judicial
Supremacy and the Politics of Constitutional Meaning. Polity
33: 365396.
Wilcox, Clyde. 1992. Gods Warriors: The Christian Right in
Twentieth-Century America. Baltimore: Johns Hopkins University
Press.
Wilcox, Clyde. 1996. Onward Christian Soldiers: The Christian Right
in Politics. Boulder, CO: Westview.
Wilson, James G. 1993. The Role of Public Opinion in Constitutional
Interpretation. Brigham Young University Law Review 1993:
10371138.
Wood, Sandra, Linda Camp Keith, Drew Noble Lanier, and Ayo
Ogundele. 1998. Acclimation Effects for Supreme Court
258 Bibliography
Justices: A Cross-Validation, 19881940. American Journal of

Woodward, Bob, and Scott Armstrong. 1979. The Brethren: Inside the
Supreme Court. New York: Simon and Schuster.
Wright, Benjamin F., ed. 1961. The Federalist. Cambridge: Belknap
Press.
Yackle, Larry. 1994. Reclaiming the Federal Courts. Cambridge:
Harvard University Press.
Yalof, David. 1999. Pursuit of Justices: Presidential Politics and the
Selection of Supreme Court Nominees. Chicago: University of
Chicago Press.
Yarbrough, Tinsley E. 2000. The Rehnquist Court and the
Constitution. New York: Oxford University Press.
Yates, Jeff. 2002. Popular Justice: Presidential Prestige and Executive
Success in the Supreme Court. Albany: State University of New
York Press.
Yee, Albert S. 1997. Thick Rationality and the Missing Brute Fact:
The Limits of Rationalist Incorporations of Norms and Ideas.
Young, Earnest. 2005. The Supreme Court, 2004 Term: Foreign Law
and the Denominator Problem. Harvard Law Review 119:
148167.
Young, Gregory D. 2001. Caught in the Middle: Are Moderate
Supreme Court Justices Guided by the Public, the Congress, and
the President More Than the Extremists? Paper presented at the
Chicago.
Zelnick, Robert. 2004. Swing Dance: Justice OConnor and the
Michigan Muddle. Stanford: Hoover Institution Press.
Index
abortion, 47, 141142, 147 Barnes, Jeb, 210n3

Abraham, Henry, 91, 94, 107108, Barnum, David G., 25
110, 111, 120, 121, 131 186n2, Barrett, David, 197n18
190n39, 196n4, 198n23, 200n5, Barrow, Deborah, 111
203n39, 209n1 Bartow, Ann, 182n9
Abramowitz, Alan, 101 Baugh, Joyce A., 186n1
Ackerman, Bruce, 207n32 Baum, Lawrence, 16, 34, 52, 58, 63,
active liberty, theory of, 65 81, 140, 186n54, 186n1, 187n3,
Adams, Greg, 100 190n45, 196n4, 197n17,
aging, and representation, 85, 88, 197198n21, 198n21, 210n3
9599, 159 Becker, Susan, 183nn1415
Agresto, John, 139 Becker, Theodore, 39, 210n8
Aldrich, John, 100, 199n34 Behuniak-Long, Susan, 67
Alito, Samuel, 102, 104, 199n37 Belsky, Martin H., 61
Allen, David W., 111, 201n23 Benesh, Sara C., 128
Amar, Akhil Reed, 65 Bentley, Arthur, 68
amicus briefs, 6768, 182n10, Berg, Larry, 143
193194nn17 Best, Samuel, 28
Antifederalists, 1 Bickel, Alexander, 56, 139, 145,
Atkins v. Virginia (2002), 6, 184n25 211n16
Appointment Process linkage model, Biskupic, Joan, 39, 63, 111, 196n4,
18, 9092, 104105, 156 197n21, 210n21, 210n9
Armstrong, Scott, 101, 143 Black, Charles L., Jr., 123
Atkinson, David N., 96 Black, Hugo, 190n36
attorneys, quality of, 66 Black, Merle, 100
awareness of Supreme Court Blackmun, Harry, 46, 52, 58, 7779,
decisions, 124125, 129130, 87, 98, 103, 115, 140, 183n15,
204n6 189n28, 191n1, 196197nn915,
199nn3033, 203n38, 206n24
Baas, Larry, 129 Blaustein, Albert, 198n23, 209n2
Bailey, Michael A., 70 Bohte, John, 130
Baird, Vanessa, 64, 128, 193n4 Bork, Robert, 34, 139
balancing tests, 65 Boucher, Robert L., Jr., 66
Balkin, Jack M., 207n32 Bowers, William, 183n14
Banks, William, 62 Box-Steffensmeier, Janet M., 212n23
bargaining on the Court, 66 Brams, Stephen J., 143
259
260 Index
Brandeis, Louis, 11, 77, 210211n10 Canon, Bradley C., 59

Brennan, William, 34, 46, 52, 7879, Caplan, Lincoln, 70
81, 82, 91, 9495, 98, 103, 108, career experiences of justices, 8586,
111, 187n6, 187n6, 191n1, 8890, 160
196197nn914, 198n22, 199n30, Carmines, Edward, 199n34
202n25, 202n32, 203n38, 206n24 Carp, Robert A., 81, 94, 110, 191n1
Brettschneider, Frank, 188n14 Carrington, Paul D., 96
Brewer, Paul, 134 Carter, Jimmy, 110
Breyer, Stephen, 12, 60, 65, 7879, caseload, 6264, 161, 204n4
82, 89, 97, 101, 103, 108, 111, Casey, Gregory, 204n1
115, 120, 123, 127, 186n2,
Casper, Jonathan D., 128, 187n11
189n31, 191n3, 193n3,
Castelli, Jim, 119, 120
196197nn919, 198nn2427,
CBS/New York Times Poll, 30,
199nn2933, 199n37, 202n24,
202n7, 203nn3638 200n6, 201n12, 201n15, 201n18,
Brisbin, Richard A., Jr., 53, 120 201n20, 206n26, 206n28
Bronner, Ethan, 67 certiorari, grants and denials of,
Brown, Robert, 200n3 3335, 46, 48, 189nn2731,
Brunnell, Thomas, 100 196n2; longevity of decisions, 144,
Bullock, Charles, III, 101 148
Burger, Warren, 197n20, 199n30 Chemerinsky, Erwin, 53, 59, 64, 136
Burger Court, 27, 3537, 48, 52, 56, Chen, Jim, 95
59, 64, 7273, 77, 113, 199, 120, Chief justice, and representation,
139, 153, 191n6, 192n12, 196n3, 9293, 158, 200n39
205n16 Choi, Stephen J., 94
Burnham, Walter Dean, 100, 101, Choper, Jesse, 56, 123
199n34 Clark, John A., 202n31
Burt, Barbara J., 68 Clawson, Rosalee A., 124, 129
Bush v. Gore (2000), 129, 188n23, Clayton, Cornell, 53
207n32 Closen, Michael L., 64
Bush, George H.W., 55, 82, 97, 108, clear statement rules, 65
100101, 109110, 201n16 clear error standard, 65
Bush, George W., 2, 109, 160, 200n3
Clinton, Bill, 142; and appointees of,
Business as Normal linkage model,
82, 97, 100, 109111; and realign-
17, 28, 5166, 161162
ments, 101102
coding rules for poll matches, 3033,
Calabresi, Steven G., 96 188n23
Caldeira, Gregory A., 63, 6768, 70, Collins, Paul M., Jr., 67
121, 125, 128, 129, 187n9, Combs, Michael, 111
187n10, 191n5, 193n1, 197n18,
Comiskey, Michael, 197n18
200n3, 206n20, 207n34, 211n13
compliance with Supreme Court deci-
Cameron, Charles, 45, 197n18
sions, 141, 204n3
capital punishment, 34, 146147,
183n14, 184185nn2534, confidence in the Rehnquist Court,
189nn2931 125129
Caliendo, Stephen M., 205n11, confirmation votes, U.S. Senate,
207n32 9192, 200n3
Index 261
congressional overrides of Supreme diffuse support, 125126

Court decisions, 140143, direct mentions of public opinion,
209210nn36, 210211nn1016, 410
211212n19, 212n23 Dixon, William, 92, 143
consensus on the Supreme Court, Dorf, Michael C., 64
143, 211n13 dormant commerce theory, 65, 210n5
Constitutional amendments, Douglas, William O., 186n1,
139140, 209n1, 210211n10 190n36, 211n10
contemporary public opinion, theory Dovi, Suzanne, 200n1
of, 10, 1213 Downing, Kimberly, 100
contextualizing, 65 Dugan, Patricia, 183n14
Converse, Jean M., 1, 2 Durr, Robert, 128, 187n10
Cortner, Richard C., 68, 69
Corwin, Edward, 204n1 early term effects, 9697, 198n2728
Cosby, Tricia, 210n7 Easton, David, 130
Cover, Albert, 45, 53, 83, 197n18 economic issues, 4345, 47, 148
Cramton, Roger C., 96 Edelman, Murray, 200n1
criminal cases, 8, 47, 133 Edelman, Paul H., 95
crisis times, 39, 145149, Edsall, Mary, 100, 199n34
190nn3637 Edsall, Thomas, 100, 199n34
education issues, 47
Crossley, Archibald, 2
effective risk management, 65
crossover appointments, 91
Eisenberg, Theodore, 183n14
cruel and unusual punishment, 12,
Eisenger, Robert M., 183n1
184nn2431
Eisenhower, Dwight, 82, 100101
Cushman, Barry, 2
election issues, 47
Ellington, Toni J., 210n7
Dahl, Robert, 55, 99104, 123, 157, Ellsworth, Phoebe, 183n15
187n11 Epstein, Lee, 16, 23, 45, 53, 55, 58,
dAmato, Anthony, 210n8 6768, 70, 82, 84, 87, 92, 94, 95,
Danielski, David, 92 139, 143, 144, 189n32, 193nn23,
data sources. See poll matching 194195n7, 196n4, 197n18,
method 198n23, 198n26, 202n25, 209n1,
Davis, Abraham, 210n22 211n13
Davis, James A., 188n16 Erikson, Robert S., 26, 2829, 119,
Davis, Richard, 124 134, 187n12, 188n13
Davis, Sue, 60, 92, 111 Esbeck, Carl H., 210n7
Days, Drew S., III, 201n23 Eskridge, Willian N., Jr., 65, 140,
death penalty. See capital 143, 144, 147, 149, 150, 197n21,
punishment. 210n3, 210n6, 211n12,
Deen, Rebecca E., 70 211nn1416, 212n19
DeGregorio, Christine, 67 ex post facto clause, 65
Dempsey, Glenn R., 130
Dershowitz, Alan M., 207n32 Fallon, Richard H., Jr., 59
deTocqueville, Alexis, 2, 27 Farnsworth, Ward, 96
Devine, Donald, 25, 188n16 Federal Policy linkage model, 16,
Devins, Neal, 46 4041, 44, 48, 155, 161
262 Index
federalism decisions, 47, 5962 Gerber, Scott D., 196n2, 211n13

Feeley, Malcolm, 210n8 Gibson, James L., 125, 128, 129,
Field, Andrew, 209n2 204n3, 205n12, 206n20, 207n34
first amendment issues, 44, 4748, Giles, Michael, 94
148 Gillman, Howard, 207n32
Fisher, Bonnie, 128 Gilligan, Carol, 201n23
Fisher, Louis, 141 Gimpel, James, 199n35
Flemming, Roy B., 26, 130, 187n12 Ginsburg, Ruth Bader, 5, 58, 60, 63,
Florida recount of 2000. See Bush v. 7879, 91, 94, 97, 101, 103,
Gore (2000) 108109, 111115, 120, 185n37,
Ford, Gerald, 82, 100101 186n47, 191n3, 193n3,
Fortas, Abe, 90, 186n1 196197nn915, 198n24,
Fowler, Thomas L., 210n7 199nn2930, 199n37, 200n8,
Frank, Jerome, 204n1 200n24, 202n27, 203nn3539
Frankfurter, Felix, 11 Goldman, Sheldon, 81, 110, 200n3
Franklin, Charles H., 124, 130, Gottlieb, Stephen E., 45, 46, 52, 53,
204n6 59, 196n4
freshman effect, 159, 198n28. See Gottschall, Jon, 111
also Length of tenure model Greenberger, Robert S., 200n3
Frickey, Philip, 65, 197n21 Greene, Abner, 207n32
Friedelbaum, Stanley H., 45 Greenhouse, Linda, 182n11
Friedman, Barry, 1, 25, 50, 181n1, Gressman, Eugene, 193n1
190n46 Grofman, Bernard, 100
Friedman, Leon, 107 Grosskopf, Anke, 128
fundamental freedoms, 4243, 48; Gruhl, John, 96, 111
longevity of such decisions, Gryski, Gerard S., 111
144145, 148 Gulati, G. Mitu, 94
Funston, Richard, 187n11 Guliuzza, Frank III, 197n18
Gunning, Matthew, 100
Gaddie, Ronald, 101
Gallup, George, 2 Hadley, Charles, vii
Gallup, George, Jr., 119, 120 Hagle, Timothy M., 198n28
Gallup Poll, 27, 29, 31, 131, 189n34, Haider-Markel, Donald, 136
190n37, 196n1, 201n11, 201n19, Haire, Susan, 111
203n34, 205n15, 205n18, 206n22, half life of Supreme Court decisions,
207n32, 207208n37, 209n47 142, 150152, 212n24
Garrow, David J., 96 Halpern, Stephen C., 59, 196n4
Gates, John B., 199n34 Hamilton, Alexander, 1, 24
George, Tracey, 94 Hansford, Thomas G., 68, 139, 143
gender, and representation, 8588, Harris Poll, 30, 205n15, 205n17,
107121, 201202nn2124, 206nn2223, 207n32
203n35, 203nn3940. See also Hart and Teeter Poll, 201n9
Symbolic Representation model Harwood, John, 200n3
geography, and representation, Hausegger, Lori, 140, 210n3
8487, 108. See also Symbolic Haynie, Stacia L., 92, 143
Representation model hazard rate, 212n23
Index 263
health care issues, 47 Jackson, Andrew, 141, 210n9

Hellman, Arthur D., 6263 Jackson, Robert, 139
Henschen, Beth, 200n3, 211n15 Jackson, Vicki, 184n26
Hettinger, Virginia, 150, 212n19, Jaros, Dean, 129
213n26 Johnson, Lyndon, 90, 160
Hibbing, John R., 128, 205n13 Johnson, Scott, 5, 6566
Higashi, Sylvia, 210n7 Johnson, Timothy R., 70
Hoekstra, Valerie, 129, 130, 202n25, Jones, Bradford, 212n23
207n30, 207n33, 209n49 Jones, Jeffrey M., 129
Hoffman, Donna, 101 judicial activism, 1112, 5962,
Hojnacki, Marie, 67, 200n3 143144, 148, 186n54
holdover effects, and representation, judicial restraint. See judicial activism
9697, 199n29 Judicial Roles linkage model, 18,
Holmes, Oliver Wendell, 11 9295, 104105, 158
Howard, Robert, 84 justices, mentions of public opinion
Hughes Court, 3637, 48, 56, 59, and polls, 414, 181182n6; theo-
7273, 77, 191n6, 192n12, 196n3, ries of public opinion, 814; agree-
197n20 ment with public opinion, 77,
Hughes, James M., 183n14 103105, 196n3. See also names of
Hurwitz, Mark S., 37, 187n12, individual justices
198n28
Hutchings, Vincent L., 207n35 Kaheny, Erin, 67
Kahn, Ronald, 53
Hutchinson, Dennis, 143
Karmoie, Brian, 70
Huo, Yuen J., 128
Kash, Douglas, 210n7
Kearney, Joseph D., 67, 68, 70,
ideology, 4546, 48, 5155, 190n41;
195n17
meaning of, 190n44; justices ideol-
Kegler, Elizabeth R., 129
ogy and agreement with public
Kennedy, Anthony, 5, 8, 53, 58, 60,
opinion, 7884, 158159, 161;
63, 7779, 91, 95, 97, 103, 108,
decisions and longevity, 143,
111, 115, 127, 183n15, 183n36,
148149
186n2, 191n2, 196197nn914,
Ignagni, Joseph, 70, 140, 143, 144,
198nn2427, 199nn2930,
145, 210n3, 211n12, 211n15 202nn2526
Interest Groups linkage model, 17, Kennedy, John F., 90
6775, 144, 156 Key, V. O., Jr., 199n34
issues, types of and representation, Kim, Jayna, 210n7
4647 Kinder, Donald R., 207n34
impeachment, 23, 186n1 King, Gary, 213n25
importance of decisions, 39, 44 King, Kimi, 140
informed public opinion, theory of, Klarman, Michael J., 201
1011 Kluger, Richard, 68
intertextualism, 65 Knight, Jack, 67, 144
Israel, Fred, 107 knowledge of Supreme Court
issue clarity, 3940, 44 decisions, 124125, 127,
Ivers, Gregg, 67, 68 204nn710
264 Index
Kobylka, Joseph F., 67, 68, 69 Maltese, John Anthony, 197n18

Kosaki, Liane C., 124, 130, 204n6 Maltz, Earl M., 53
Krislov, Samuel, 67 Maltzman, Forrest, 66, 70, 92
Kritzer, Herbert, 64, 111, 129, Mansbridge, Jane, 123, 200n1
204n6, 205nn1314, 207n35 Manwaring, David R., 68, 69
Kurland, Philip, 56 Marshall, Thomas R., 2, 2729,
3033, 35, 38, 40, 43, 45, 46, 49,
Ladd, Everett Carll, 100 56, 58, 69, 70, 72, 77, 81, 82, 88,
Lamb, Charles M., 58, 196n4 89, 90, 93, 99, 102, 111, 113, 118,
Larson, Stephanie G., 124 119, 131, 133, 143, 144, 145, 146,
Lawrence, George, 2 149, 150, 182nn67, 191n6,
Lawrence, Susan, 68, 69, 194n5 194n6, 195n9, 196n4,
Lazarus, Edward, 63, 110 198nn2223, 198n28, 200n4,
leadership on the Court, 9395, 158 203n38, 208n39, 208n42, 211n10,
Leathers, John R., 210n7 211n18, 212n23
Ledewitz, Bruce, 210n7 Marshall, Thurgood, 34, 46, 52, 55,
Leeds, Jeffrey T., 120 7879, 82, 8788, 95, 98, 103,
legitimacy, 1213, 129, 185nn3536 108, 110112, 118, 160, 191n1,
Length of Tenure linkage model, 19, 196nn911, 199n20, 201n16,
9599, 104105, 159, 187n6 202n25, 203nn3638, 206n25
Lerner, Max, 204n1 Martin, Andrew, 82, 128, 133,
Lichtman, Judith, 67 187n10
Lind, E. Allan, 128 Martin, Elaine, 200n22
Lindgren, James, 96 Mason, Alpheus T., 92, 107, 120
Lindquist, Stefanie, 144, 202n31, Massey, Calvin, 59
212n19 Mauro, Tony, 23
Link, Michael W., 26, 187n12 Maveety, Nancy, 65, 144, 196n2,
linkage models, 1421, 4247, 200n21
153162. See also names of Mayhew, David, 199n34
individual linkage models McCloskey, Robert G., 16
logit analysis, 49, 74, 195n15, McFall, Bryan Scott, 68
212213n25 McGuire, Kevin T., 26, 66, 6768,
Long Term Manipulation linkage 70, 89, 194n7
model, 20, 133137, McIntosh, Wayne V., 67
208209nn4446 Medalie, Richard C., 68
Los Angeles Times, 208n38 media coverage, 124, 208n38
Los Angeles Times Poll, 30, 206n23, Medvic, Stephen K., 205n11
206n29 Meernik, James, 70, 140, 143, 144,
Lublin, David, 100 145, 210n3, 211n12, 211n15
Lyons, William, 204n1, 205n16, Meffert, Michael F., 101
207n31 Melone, Albert, 197n18
Merrill, Thomas W., 67, 68, 70,
Mace, George, 56 190n46, 195n17
MacKuen, Michael B., 26, 2829, Metrailer, Rosemary, 210n8
187n12, 188n13 Mezey, Susan Gluck, 68
Macleod, Peter R., 210n7 Miers, Harriet, 104, 105, 160, 200n3
Index 265
Miller, Geoffrey, 65, 190n39 ONeill, Timothy J., 67, 68,

Mishler, William, 26, 2829, 37, 70, 193194n5
187n12 opinion authorship, 95
Mitchell, Gregory, 123, 128 oral arguments, 65
Mondak, Jeffery, 128, 129, 130, original intent, 65
187n10, 207n31, 207n33 Overby, L. Marvin, 200n3
Monroe, Alan D., 2627, 37 Owens, Ryan J., 67
Moore, Michael, 65 Oxley, Zoe, 124
Moraski, Bryan J., 121
Morris, Alan, 197n18 Pacelle, Richard J., 35
most dangerous justice, 95 Paduano, Anthony, 183n14
most important problem, 39, 44, 48, Page, Benjamin, 25, 130, 188n16
149 pairwise method, 2629, 51, 53
Murakami, Mark, 210n7 Palfrey, Thomas R., 207n34
Murphy, Walter, 92, 123, 139, Palmer, Barbara, 111
206n19, 210n8, 211n15 Palmer, Paul A., 181n1
Muzzio, Douglas, 143 Park, Keeok, 196n2, 211n13
Pashal, Richard, 139, 140, 142, 147,
Nagredo, Richard A., 210n7 209n1, 210nn34
Nardulli, Peter, 199n34 Paulson, Arthur, 199n34
National Election Study poll, 205n14 Pederson, William D., 94
National Opinion Research Center penalty for disagreeing with public
Poll, 205nn1516, 206n22, opinion, 7075. See also interest
groups.
207n32, 209n47
Peretti, Terri Jennings, 50
natural law, 65
Perine, L. Timothy, 55
NBC/ Wall Street Journal Poll, 30
Perry, Barbara, 107108, 110, 111,
Neumann, Elizabeth Noelle, 181n1
120, 121, 187n4, 200n5, 203n39
New York Times, 208n38
Peters, C. Scott, 207n33
Nice, Julie A., 210n7
Peterson, Steven A., 143
Nixon, Richard, 82, 100101, 196n7 Petry, Francois, 188n14
Norpoth, Helmut, 26, 101 Phillips, Carter G., 67
null model, 17, 69, 72. See also inter- Pitkin, Hanna, 200n1
est groups plain errors, 13
Planned Parenthood of Southeastern
OBrien, David M., 5, 63, 102 Pennsylvania v. Casey (1992), 78,
OConnor, Karen, 67, 68, 69, 111 3233
OConnor, Sandra Day, 5, 7, 8, Political Parties and Ideology linkage
1314, 53, 58, 60, 77, 95, 98, 101, model, 1718, 8084, 104105,
103, 108115, 127, 140, 160, 112121, 156, 161, 203n38
182n13, 185n35, 186n2, Political Socialization linkage model,
187nn67, 189n28, 191n3, 18, 8490, 104105, 156, 162
196197nn914, 198nn2425, poll correction, 5559, 64
201202nn2224, 203nn3538, polls and surveys, 12, 8, 120,
206n25 183n15; sources of poll matches,
Olivas, Michael A., 210n7 2930; questions on Supreme
Olson, Mancur, 68 Court, 35
266 Index
poll margins, 3839, 188n18, Rasinski, Kenneth, 123, 128, 204n3,

211n18 206n19
poll matching method, 23, 2635, Rathjen, Gregory J., 190n44
181n4, 188nn1417, 188n19 Ray, Laura Krugman, 210n7
Ponnuru, Ramesh, 60 Reagan, Daniel, 197n18
popularity, of Supreme Court, Reagan, Ronald, 82, 97, 99102,
125129, 204206nn1119; of jus- 108, 110
tices, 127128, 206nn2425; of Realignment linkage model, 19,
rulings, and longevity, 132135, 99102, 157, 199nn3437
211n18 Reedy, Cheryl D., 194n7
Posner, Richard A., 63, 65, 141, Rehnquist, William, 5, 6, 7, 15, 23,
207n32 24, 38, 53, 58, 60, 7779, 89,
Powell, Lewis, 7879, 81, 8791, 94, 9295, 98, 101, 103, 108, 115,
103, 187n6, 196197nn916, 127, 140, 142, 158, 182n12,
199n30, 202n32, 203n38, 206n24 183n15, 186n2, 187n6, 189n31,
pragmatism, 65 191n1, 191n3, 196197nn727,
preferred position, 4243, 65, 197n21, 198nn2425, 199n33,
190n39 200n39, 202n25, 203nn3538,
presidential appointments, 8183, 206n25
9091 religion, and representation, 47,
presidential confidants, 86, 90 8588, 107121, 201212n24,
Price, Vincent, 129 202nn2627, 203n37, 203nn3940
Princeton Survey Research Associates region, and representation, 8487
Poll, 2930, 191n2 regression analysis, 103104,
Pritchett, C. Herman, 123, 139, 200nn3839
211n15 replacement effects, 9698, 199n30
procedural fairness, 123, 128 representation, on the Rehnquist
Provizer, Norman W., 94 Court, 3538, 153162; cf.
public opinion, classical theory of, 1, Congress and the executive branch,
181n1; Supreme Courts agreement 6162, 154; cf. states and
with, 2, 3538, 4750, 5556, 59, localities, 154; cf. lower courts,
7274, 148152; international 5559, 154; cf. prior Courts since
public opinion, 184n26 the 1930s, 3537, 4749, 59,
public mood, 2526, 187n10, 153154; by justice, 7880; by
187188nn1213 issue, 4647, 57
Pufong, Marc-George, 197n18 Research and Forecasts Poll, 206n27
reweighting, 38, 189nn3233,
Quinn, Kevin, 82 192n12
Quinnipiac University polling Reynolds, William Bradford, 197n18
Institute, 201n10 Richards, Mark J., 64
Ring, Kevin A., 65
race, and representation, 8588, Roberts, John, 4, 102, 104, 155,
107121, 202n25, 203n36, 199n37
203n40 Roberts Court, 155, 160
Rademacher, Eric, 100 Robison, Claude, 1
Radin, Margaret Jane, 207n32 Robinson, Tracy, 8, 183n15, 184n25
Index 267
Roesch, Benjamin, 8 Shapiro, David, 60

Rogers, James R., 65, 190n39 Shapiro, Martin, 65
Rohde, David W., 143 Shapiro, Robert Y., 25, 130, 188n16
Romano, Roberta, 209n1 Shapiro, Stephen M., 192n1
Romantan, Anca, 129 Sheehan, Colleen A., 181n1
Roper, Elmo, 2 Sheehan, Reginald S., 26, 2829, 37,
Roper Poll, 30 70, 187n12
Roper Archive, 3, 181n4, 181n4, Sheldon, Charles, 210n8
182n11, 204n5 Shipan, Charles R., 121
Roper, Robert, 129 Short Term Manipulation linkage
Roosevelt, Franklin, 120, 157, model, 20, 123, 129133,
183n18 136137, 156157, 208nn3942
Rosen, Jeffrey, 62, 137, 139 Silbey, Joel H., 199n34
Rosenof, Theodore, 100, 101, Silverstein, Mark, 107
199n34 Simmons, Wendy M., 129
Rossotti, Jack F., 67 Slotnick, Elliot, 81, 110, 124, 202n28
Rowland, Claude K., 67 Smith, Andrew, 100
Ruckman, P.S., Jr., 200n3 Smith, Charles E., Jr., 121, 200n3
Ruhil, Anirudh V.S., 101 Smith, Christopher E., 46, 186n1
Ryan, John Francis, 211n13 Smith, Clive A. Stafford, 183n14
Smith, Tom W., 1, 188n16
Smithey, Shannon, 128, 188n10
Salokar, Rebecca Mae, 70
Solicitor General, 63, 7072, 144,
Saunders, Kyle, 100, 101
148, 158, 195nn1112,
Savage, David G., 45, 52, 63
195nn1417
Scalia, Antonin, 5, 7, 14, 53, 60, 63,
Solomine, Michael E., 140, 143, 144,
65, 7779, 82, 89, 94, 97, 103,
210n6, 211n14, 212n19
108, 111, 115, 120, 182n13,
Songer, Donald R., 67, 70, 81, 84,
183n15, 183n20, 184n26, 186n2, 110, 111
189n31, 191n1, 191n3, Sorauf, Frank, J., 68, 69
196197nn913, 198nn2425, Sorensen, Robert, 183n14
198n27, 199nn2930, Sorensen, Theodore, 183n14
202nn2526, 203n38 Souter, David, 12, 58, 60, 63, 7779,
Scheb, John M. II, 204n1, 205n16, 103, 108, 115, 127, 182n8, 191n3,
207n31 193n3, 196197nn913, 198n24
Schlozman, Kay, 67 Spaeth, Harold J., 23, 45, 53, 65,
Schmidhauser, John, 143 143, 190n44, 202n25, 211n13
Schneider, William, 100 Spence, Lester Kenyatta, 125, 129,
Schoen, Ryan, 66 206n20, 207n34
Schwartz, Bernard, 94, 110, 196n4, Spier, Hans, 181n1
198n23 Spill, Rorie, 124
Segal, Jeffrey A., 23, 26, 45, 53, 65, Spiller, Pablo T., 140, 209n1, 210n3,
66, 83, 84, 130, 143, 144, 194n7, 211n16
197n18, 204n41, 205n25, 211n13 Spitzer, Matthew, 209n1, 211n16
Segal, Jennifer, 124, 207nn3334 Spohn, Cassia, 111
seniority, 9899, 159, 199nn3133 Spriggs, James, II, 6566, 67, 139,
Shafer, Bryan, 100, 199n34 143
268 Index
stare decisis, 65, 139 Taylor, Stuart, Jr., 193n2

State/local linkage model, 1617, Tedin, Kent L., 119, 134
4142, 44, 48, 155156, 161 term limits, 105, 160
State of Public Opinion linkage Test of Time linkage model, 2021,
model, 1516, 3840, 44, 48, 158 139152, 162, 209210nn17,
statutory construction, 65 210211nn910, 211212n19,
Stefko, Joseph V., 198n28 212213nn2126
Stern, Robert H., 193n1 textualism, 65
Stevens, John Paul, 5, 6, 58, 60, Theiss-Morse, Elizabeth, 128,
7779, 87, 89, 95, 98, 103, 108, 205n13
115, 127, 182n8, 182n13, 183n15, thermometer scales, 205n14
185n33, 186n47, 189n27, 189n31, Thibaut, John, 128, 204n3
191n3, 192n17, 196197nn915, Thomas, Clarence, 34, 46, 53, 55,
198n24, 199n33, 206n25 60, 65, 7779, 82, 8789, 92, 97,
Stewart, David O., 185n36 103, 108111, 115, 118, 120, 127,
Stewart, Potter, 26 186n1, 189n31, 191n1, 191n3,
Stidham, Ronald, 81, 94, 110, 191n1 193n3, 196197nn915,
Stimson, James A., 2526, 2829, 198nn2427, 199nn2930,
198n12, 188n13, 188n16, 199n34 199n3033, 200n5, 202n25,
Stone, Harlan Fiske, 11, 120, 202n32, 203n36, 206n25
197n20, 211n13 Thomas, Dan, 129
Stone Court, 3637, 48, 56, 59, Thornberry, Homer, 90
7273, 77, 191n6, 192n13, 196n3 Tierney, John, 67
Storing, Herbert J., 1 Tilleman, Karl, 210n7
Stoutenborough, James W., 136 Tiller, Emerson H., 140, 210n3,
strategic voting, 66 211n16
Strauss, David A., 70 Timpone, Richard, 84
Strauss, Julie, 200n3 Todd, James S., 59
straw polls, 1 Toma, Eugenia, 210n3
Stumpf, Harry, 139 Toner, Robin, 67
sub silentio overrulings, 140141, Tramont, Bryan, 124
210n7 trend method, 2526
substantive due process, 65 Tribe, Lawrence, 207n32
Sundquist, James L., 100 Truman, Harry, 157
Sungalia, Mary-Christine, 67 Tuchfarber, Alfred J., 100, 101
Sunstein, Cass R., 64 Tucker, Karla Faye, 146, 212n20
Swindle, Shane, 210n7 Turner, Charles, 144
Symbolic Representation linkage Tushnet, Mark, 63, 207n32
model, 1920, 8788, 107121, Tyler, Tom R., 123, 128, 204n3,
156, 200n2, 202n33 206n19
Szmer, John, 84
Uhlman, Thomas, 111
Tanenhaus, John, 206n19 Ulmer, S. Sidney, 143
Tarrance Group Poll, 200n8 Umhofer, Matthew D., 210n7
Taft, William Howard, 197n21 unanimity, of Supreme Court
Tate, Katherine, 121 decisions, 143, 148, 208n39
Index 269
undue burden standard, 140141 Weissberg, Robert, 25

Unger, Michael, 130 Weizer, Paul, 65
Welch, Susan, 111
Vandiver, Margaret, 183n14 Wellington, Harry, 56
Van Sickel, Robert, 65, 200n21, Wells, Martin T., 183n14
210n23 Wheeler, Russell R., 197n21
Vermeer, Jan P., 124 White, Byron, 26, 60, 7779, 8790,
Victor, Jennifer Nicoll, 144 98, 103, 115, 119, 197n21,
Vidmar, Neil, 183n15 203nn3538, 206n24
Vinson, Fred, 77, 197n20 White, G. Edward, 92, 197198n21
Vinson Court, 3637, 48, 56, 59, Whittington, Keith E., 210n3
7273, 77, 191n6, 192n12, 196n3 Wilcox, Clyde, 68
Vitiello, Michael, 188n14 Wilson, James G., 183n17
Vose, Clement E., 68, 69 Witt, Elder, 39, 196n4, 197n21,
Voss, D. Stephen, 100 210n9
Wolbrecht, Christine, 128, 187n10
Wahlbeck, Paul, 6566, 67, 92 Wood, B. Dan, 26, 130, 187n12
Walker, James, 140, 143, 144 Wood, Sandra, 198n28
Walker, Laurens, 128, 204n3, 210n6, Woodward, Bob, 110, 143
211n14, 212n19 Wright, Benjamin F., 181n1
Walker, Thomas, 23, 92, 94, 111, Wright, John, 63, 67, 68, 70, 191n5,
143 193n1, 197n18, 200n3
Wall, Diane, 111, 201n23
Walsh, Michael, 200n3 Yackle, Larry, 190n39
Waltenburg, Eric N., 124, 129 Yalof, David, 144, 200n3, 202n31,
Ward, Artemus, 95 212n19
Warren, Charles, 139 Yankelovich Clancy Shulman Poll,
Warren, Earl, 197n20 187n8, 200n7, 201n13
Warren Court, 3537, 48, 52, 56, 59, Yarborough, Tinsley E., 45, 46, 52,
64, 7273, 77, 113, 119, 120, 139, 59, 61
141, 153, 191n6, 192n12, 196n3 Yates, Jeff, 70
Wasby, Stephen L., vii, 68, 69, 210n8 Young, Earnest, 184n26
Washington, George, 108 Young, Gregory D., 82
Washington Post, 208n38
Washington Post/ABC Poll, 30, 134, Zelnick, Robert, 200n21
201n14, 204n9, 208n43 Zeng, Langche, 213n25
Wattenberg, Martin, 100 Zorn, Christopher, 63, 150, 191n5,
Way, Frank, 144 211n13, 212n19, 213n26
Way, Lori Beth, 68 Zuk, Gary, 111
POLITICAL SCIENCE
PUBLIC OPINION AND THE

REHNQUIST COURT
Thomas R. Marshall
Public Opinion and the Rehnquist Court offers the most thorough evidence yet in favor of the
U.S. Supreme Court representing public opinion. Thomas R. Marshall analyzes more than two
thousand nationwide public opinion polls during the Rehnquist Court era and argues that a clear
majority of Supreme Court decisions agree with public opinion. He explains that the Court
represents American attitudes when public opinion is well informed on a dispute and when the
U.S. Solicitor General takes a position agreeing with poll majorities. He also finds that certain
justices best represent public opinion and that the Court uses its review powers over the state and
federal courts to bring judicial decision making back in line with public opinion. Finally, Marshall
observes that unpopular Supreme Court decisions simply do not endure as long as do popular
decisions.
This is an excellent, well-written, and important book. Not only is it comprehensive in its
examination of opinions rendered by the Supreme Court, but it is also comprehensive with respect
to the relevant literature produced by Supreme Court scholars and with a clear understanding of
said literature.
Charles D. Hadley, coauthor of Womens PACs: Abortion and Elections
Marshall constructs a well-developed set of models, which are effectively used for organizing
previous studies and as the framework for his own research. This is an important addition to the
fields of law and courts and public opinion.
Stephen L. Wasby, author of The Supreme Court in the Federal Judicial System
THOMAS R. MARSHALL is Professor of Political Science at the University of Texas at

Arlington. He is the author of Presidential Nominations in a Reform Age and Public Opinion and the
Supreme Court.
State University of New York Press

www.sunypress.edu

Public Opinion and The Rehnquist Court - Marshall, Thomas R. (Author)

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Public Opinion and The Rehnquist Court - Marshall, Thomas R. (Author)

Enviado por

Direitos autorais:

Formatos disponíveis

Public Opinion

and the Rehnquist Court

2008 State University of New York

All rights reserved

Printed in the United States of America

No part of this book may be used or reproduced in any manner whatsoever

For information, contact State University of New York Press, Albany, NY

Production by Kelli W. LeRoux

Library of Congress Cataloging-in-Publication Data

Marshall, Thomas R., 1949

Chapter Three: Judicial Norms, Representation,

Chapter Five: Justices and Representation 77

Chapter Six: Symbolic Representation and the Court 107

Chapter Seven: Did the Rehnquist Court Influence

Chapter Eight: Public Opinion and the Test of Time 139

Chapter Nine: An Empirical Model of Representation 153

Appendix One: Poll-to-Ruling Matches 163

Appendix Two: List of Cases 171

Figure 1.1 Number of poll questions, per year, during

Table 1.1 Frequency of Direct Mentions of Public Opinion

Table 5.3 Rehnquist Court Justices Voting Patterns, by

Public Opinion and Supreme Court

S hould U.S. Supreme Court decisions reflect American public opin-

conducted on a nationwide basis. Further, until the 1930s, most polls

Supreme Court decisions were more often overturned by constitutional

HOW POLLSTERS VIEW THE COURT

American pollsters often write questions about Supreme Court rulings,

Thomas and Robert Bork. A small number of questions tapped

HOW THE JUSTICES VIEW PUBLIC OPINION

At least 123 Rehnquist Court opinions directly mentioned public opin-

As Table 1.1 indicates, over two centuries the number of direct

Although the justices often discussed public opinion, they rarely

polling data as too inadequately explained and poorly presented to be

In part, the inattention to specific polls may occur because several

Parenthood of Southeastern Pennsylvania v. Casey (1992) is a well-

JUDICIAL THEORIES OF PUBLIC OPINION

Long before modern public opinion polling developed during the

As a caveat, most direct mentions of public opinion are not based

An Informed Public Opinion

union actions (such as picketing), or commercial speech (such as adver-

during the Rehnquist Court. Examples include Justice Breyers argu-

Contemporary Public Opinion

Public Opinion as a Threat

LINKING PUBLIC OPINION AND THE REHNQUIST COURT

By this books count, three-fifths to two-thirds of Rehnquist Court

other models focus on public reactions to Court decisions, or on key

The State of Public Opinion Model

American attitudes. As one Court-watcher wrote: (T)he justices have

The Federal Policy Model

The State/Local Policy Model

constitute a very large part of the Courts annual docketin recent

The Business-as-Normal Model

The Interest Groups Model

The Political Parties and Ideology Model

Democratic presidents typically pick liberal Democratic justices; con-

The Political Socialization Model

The Appointment Process Model

The Judicial Roles Model

The Length of Tenure Model

The Realignment Model

The Symbolic Representation Model

The Short-Term Manipulation Model

The Long-Term Manipulation Model