The Psychology of Judicial Decision Making

The Psychology of Judicial Decision Making
American Psychology-Law Society Series
Series Editor
Ronald Roesch
Editorial Board
Gail S. Goodman
Thomas Grisso
Craig Haney
Kirk Heilbrun
John Monahan
Marlene Moretti
Edward P. Mulvey
J. Don Read
N. Dickon Reppucci
Gary L. Wells
Lawrence S. Wrightsman
Patricia A. Zapf
Books in the Series

Trial Consulting
Amy J. Posey and Lawrence S. Wrightsman
Death by Design
Craig Haney
Psychological Injuries
William J. Koch, Kevin S. Douglas, Tonia L. Nicholls, and Melanie L. ONeill
Emergency Department Treatment of the Psychiatric Patient
Susan Stefan
The Psychology of the Supreme Court
Proving the Unprovable
Christopher Slobogin
Adolescents, Media, and the Law
Roger J. R. Levesque
Oral Arguments Before the Supreme Court
God in the Courtroom
Brian H. Bornstein and Monica K. Miller
Expert Testimony on the Psychology of Eyewitness Identification
Edited by Brian L. Cutler
The Psychology of Judicial Decision Making
David Klein and Gregory Mitchell
The Psychology of Judicial
Decision Making
Edited by
David Klein
Gregory Mitchell
1
2010
1
Oxford University Press, Inc., publishes works that further
Oxford Universitys objective of excellence
in research, scholarship, and education.
Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Copyright 2010 by Oxford University Press.
Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016
www.oup.com
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data

The psychology of judicial decision making / edited by David Klein, Gregory Mitchell.
p. cm. (American Psychology-Law Society series)
Includes bibliographical references.
ISBN 978-0-19-536758-4
1. Judicial processUnited StatesPsychological aspects. I. Klein, David E., 1970
II. Mitchell, Gregory, J.D.
KF8775.P78 2010
347.730 14019dc22
2009023896
9 8 7 6 5 4 3 2 1
Printed in the United States of America

on acid-free paper
Series Foreword
This book series is sponsored by the American Psychology-Law Society

(APLS). APLS is an interdisciplinary organization devoted to scholarship,
practice, and public service in psychology and law. Its goals include advancing
the contributions of psychology to the understanding of law and legal institu-
tions through basic and applied research; promoting the education of
psychologists in matters of law and the education of legal personnel in matters
of psychology; and informing the psychological and legal communities and
the general public of current research, educational, and service activities in the
field of psychology and law. APLS membership includes psychologists from
the academic research and clinical practice communities as well as members
of the legal community. Research and practice is represented in both the civil
and criminal legal arenas. APLS has chosen Oxford University Press as a
strategic partner because of its commitment to scholarship, quality, and the
international dissemination of ideas. These strengths will help APLS reach
our goal of educating the psychology and legal professions and the general
public about important developments in psychology and law. The focus of the
book series reflects the diversity of the field of psychology and law as we will
publish books on a broad range of topics.
David Klein and Gregory Mitchell are the editors of the latest book in the
series, The Psychology of Judicial Decision Making. The book is a perfect fit for
this series, as its intended audience is students and scholars from psychology,
law, and political science who are engagedor may be encouraged to
engagein exploring the intersection of psychology and judicial behavior.
v
vi Series Foreword
While the book is grounded in psychological theory and research, the editors
recognize that the study of the behavior of judges is complex, so both theory
and research would be enhanced through debate and discussion by contri-
butors from many professional backgrounds. To accomplish their goal, Klein
and Mitchell assembled an impressive interdisciplinary group representing
law, political science, and, of course, psychology. This group first came
together at a conference in Virginia, where the participants had an opportu-
nity to share and critique each others ideas. Klein and Mitchell had a
forward-looking perspective, as they wanted the book to provide an agenda
for future research rather than a review of prior studies of judicial decision
making. The contributors were asked to identify theories, concepts, or find-
ings from psychology that could usefully be incorporated into thinking about
how judges make decisions, and describe new research questions and the
accompanying methodology to test hypotheses generated from this process.
Having worked in an interdisciplinary faculty for a few years early in my
career, I appreciate that bringing together an interdisciplinary group does not
easily result in increased collaborations. Each discipline has its own traditions
and approaches to scholarship, and the interdisciplinary boundaries often
seem insurmountable. As the editors note in their introduction, examples of
other disciplines drawing on psychology to inform the study of judicial
decision making are rare. That the participants in the Klein and Mitchell
project were able to overcome these interdisciplinary barriers is an impressive
achievement. Klein and Mitchell wanted a book that would encourage stu-
dents of judicial behavior to incorporate psychology into their work and also
persuade psychologists and other students of decision making to pay more
attention to the decision-making process used by judges. This book serves this
purpose well.
Ronald Roesch
Series Editor
Contents
Series Foreword v
Contributors ix
Introduction xi
Part I. Judges and Human Behavior
1 Motivation and Judicial Behavior: Expanding the Scope of Inquiry 3

Lawrence Baum
2 Multiple Constraint Satisfaction in Judging 27

Jennifer K. Robbennolt, Robert J. MacCoun, and John M. Darley
3 Top-Down and Bottom-Up Models of Judicial Reasoning 41

Brandon L. Bartels
4 Persuasion in the Decision Making of U.S. Supreme Court Justices 57

5 Judges as Members of Small Groups 73

Wendy L. Martinek
6 The Supreme Court, Social Psychology, and Group Formation 85

Neal Devins and Will Federspiel
vii
viii Contents
Part II. Judging as Specialized Activity
7 Is There a Psychology of Judging? 103

Frederick Schauer
8 Features of Judicial Reasoning 121

Emily Sherwin
9 In Praise of Pedantic Eclecticism: Pitfalls and Opportunities in the

Psychology of Judging 131
Dan Simon
10 Judges, Expertise, and Analogy 149

Barbara A. Spellman
11 Thresholds for Action in Judicial Decisions 165

Len Dalgleish, James Shanteau, and April Park
12 Every Jury Trial Is a Bench Trial: Judicial Engineering of Jury

Disputes 183
C. K. Rowland, Tina Traficanti, and Erin Vernon
13 Searching for Constraint in Legal Decision Making 203

Eileen Braman
Part III. Evaluating and Improving Judging
14 Evaluating Judges 221

Gregory Mitchell
15 Defining Good Judging 249

Andrew J. Wistrich
16 Expertise of Court Judges 269

James Shanteau and Len Dalgleish
17 Cognitive Style and Judging 279

Gregory Mitchell and Philip E. Tetlock
18 Building a Better Judiciary 285

Daniel A. Farber and Suzanna Sherry
References 297
Index 335
Contributors
Brandon L. Bartels David Klein

Assistant Professor of Political Associate Professor of Politics
Science University of Virginia
George Washington University
Robert J. MacCoun
Lawrence Baum Professor of Law and Public Policy,
Professor of Political Science UC-Berkeley
The Ohio State University
Wendy L. Martinek
Eileen Braman Associate Professor of Political
Assistant Professor of Political Science
Science Binghamton University
Indiana University
Gregory Mitchell
Len Dalgleish Daniel Caplin Professor of Law
Professor of Decision Making University of Virginia
University of Stirling, and
head of the Decision Making April Park
program of HealthQwest Ph.D. Candidate in Psychology
Kansas State University
John M. Darley
Warren Professor of Psychology Jennifer K. Robbennolt
Princeton University Professor of Law and Psychology
Neal Devins University of Illinois
Goodrich Professor of Law and C. K. Rowland
Professor of Government Professor Emeritus of Political
College of William and Mary Science at Kansas University and
Daniel A. Farber Founder
Sho Sato Professor of Law Litigation Insights
Boalt Hall Frederick Schauer
UC-Berkeley
David and Mary Harrison
Will Federspiel Distinguished Professor of Law
Associate at McGuireWoods University of Virginia
ix
x Contributors
James Shanteau Philip E. Tetlock

University Distinguished Professor Professor of Organizational
of Psychology Behavior and Mitchell Chair in
Kansas State University Leadership
UC-Berkeley
Suzanna Sherry
Herman O. Loewenstein Professor Tina Traficanti
of Law Attorney at Litigation
Vanderbilt University Insights
Emily Sherwin Erin Vernon
Professor of Law Law student at Duke
Cornell Law School University
Dan Simon Andrew J. Wistrich
Professor of Law and United States Magistrate
Psychology Judge
University of Southern California Central District of California
Barbara A. Spellman Lawrence S. Wrightsman
Professor of Law and Professor of Professor Emeritus of Social
Psychology Psychology
University of Virginia University of Kansas
Introduction
David Klein
Over the years, psychologists have devoted uncountable hours to learning

how human beings make judgments and decisions. Legal scholars and
political scientists have expended immeasurable intellectual energy trying
to understand why those particular human beings who sit on courts act as
they do in presiding over and deciding cases. It might seem obvious that
fertile intellectual ground lies at the intersection of these disciplines, and
certainly some scholars have seen it this way. As far back as 1930, Jerome
Frank drew on contemporary psychology to explain judging in his Law and
the Modern Mind. And yet, nearly eighty years on, the area under active
cultivation is quite small. To be sure, psychological concepts crop up in
studies of judicial behavior from time to time, but it would be difficult to
name a score of published studies that have relied extensively on current
ideas and evidence in psychology to generate major theoretical propositions
about judging. This is partly because students of judicial behavior tradi-
tionally have not engaged deeply with scholarship in psychology, but only
partly; it is also the case that psychologists have tended not to focus on the
kinds of questions that would be most helpful for understanding what
professional judges do. This volume of essays grows from a belief that
students of both judges and psychology would benefit from a dramatic
expansion of research into the psychology of judicial decision making and
closely related behavior.
xi
xii Introduction
The study of judicial decision making has indisputably made great strides
in recent years, through the labors of hundreds of scholars from political
science, law, economics, and other disciplines. Nevertheless, one could argue
that there remains a lack of both depth and breadth to our understanding of
what judges do. Even where scholars can make consensual and successful
predictions of a judges behaviorfor example, that Justice J will vote for the
conservative position in case Cthey will often disagree sharply about exactly
what happens in the judges mind to generate the predicted result. (Does
Justice J vote conservatively in a conscious effort to further his policy pre-
ferences, in an unconscious effort to do so despite a sincere desire to be guided
by legal texts, or as a result of a method of interpretation that is independent
of his ideology?) And as soon as we move beyond ideology, we enter areas
where good predictions are much harder to come by. How will a judges
decision on a motion, verdict, or appeal be affected by precedents, the
presence of an amicus curiae brief from the federal government, the plaintiffs
race, a particularly eloquent brief or oral argument by the defendants
attorney, the preferences and arguments of other panelists on a collegial
court, the opinions of the local bar, the presentations of expert witnesses,
other demands on the judges time? Why will it be affected that way? Some of
these questions have been the subject of excellent scholarly analysis, but none
have received definitive answers.
Naturally, various methodological difficulties unrelated to psychology
have hindered attempts to study judging, and as scholars devise creative new
ways to measure previously intractable concepts, observe hidden behaviors
and influences, and design studies so as to control for more confounding
factors, our understanding of judging will continue to improve. Still, anyone
who has ever tried to choose fairly between serious competing legal argu-
ments must have been struck by the depth, complexity, and mysteriousness of
the mental processes involved in the evaluation. It is hard to see how we can
hope to achieve a profound understanding of the far more complex and
difficult undertaking we call judicial decision making without a close analysis
of these underlying mental processes.
Thinking about the intersection of psychology and judicial decision
making can do more than help us answer questions that have long troubled
scholars; it can also point us toward equally exciting but less explored ques-
tions. To give just a few examples: What does it mean to judge well? Are some
circumstances, personalities, or cognitive styles more conducive to good
judging than others? Do most judges possess special reasoning skills that
other people lack? Do judges care what other people think about them, and,
if so, how does this affect their decision making? When different motivations
come into play at the same time, which have the most influence on judges
behavior, and why?
While students of judging may be the primary beneficiaries of an engage-
ment with psychology, the topics covered in this book should also interest
academics doing basic research in the psychology of expertise, analogical
Introduction xiii
reasoning, judgment and decision-making, and the psychology of small

group behavior, with applications to the real-world behavior of professional
decision makers rather than ordinary people providing opportunities to test
the limits of basic theories and experimental studies. Do professional judges
think the same way as ordinary people? Is their behavior affected by the same
forces that affect the behavior of other people? If not, in what ways do their
thinking and behavior differ, and why? Confronting questions like these can
only strengthen research in psychology.
The authors of the essays in this book do not always agree about exactly
how we should blend the study of psychology and judging or what we can
expect to learn from doing so. But all agree that more rigorous thinking about
the nature of the mental processes involved in judges work will lead to deeper
understandings of professional judging and psychology generally. Their essays
can best be understood as invitations to other scholars to join in this enterprise,
offering suggestions for research and surveying the theoretical and methodo-
logical promise and problems of different approaches. The authors occasionally
present original empirical evidence, but more often their emphases are theore-
tical. In fact, the authors were encouraged to engage in free speculation, with
the intention that the book raise more questions than it answers.
The book is divided into three sections. Essays in the first two sections are
concerned with the empirical investigation of decision making. The third part
of the book raises questions about whether and how we can evaluate judicial
performance, with implications for the possibility of improving judging
through the selection and training of judges and structuring of judicial
institutions.
What chiefly distinguishes the first set of essays from the second is the
perspective from which each set approaches the incorporation of psychology
into the study of judging. Essays in the first section take as their starting point
the fact that judges are human beings. From this perspective, one asks,
Knowing what we do about people generally, what should we expect of
people put in the positions that judges are and asked to do what they do?
For example, people generally tend to engage in certain suboptimal reasoning
processes at times (see Guthrie, Rachlinski, & Wistrich, 2001). How often and
under what circumstances does judicial reasoning fall short in these ways?
Among the general population, people vary in their characteristic ways of
thinking (Stanovich, 1999). What kind of variation in cognitive styles might
we find among judges, and with what impact on their behavior and the
outputs of the judicial system? Or, to take the most common theme in the
first sections essays, human beings act on a wide range of motivations. Which
of those motivations influence the work of judges, and how?
An alternative approach to the psychology of judicial decision making
eschews this focus on what judges have in common with other people and
instead begins with what makes, or is supposed to make, judges different from
other people or the mental processes judges employ different from the thinking
other people do. The second set of essays begins with Frederick Schauers
xiv Introduction
argument for this approach. The other essays either engage this argument
directly or direct our attention to specific tasks judges are required to perform
or particular modes of reasoning in which they are expected to engage.
The essays in the third part of the book turn from what we know and can
learn about how judges make decisions to reflections on the assessment of
judicial performance. Understanding judicial behavior is not, after all, simply
an academic exercise. Judges wield substantial power, including the power to
make policy, and we naturally want our judges to exercise that power as fairly,
competently, and appropriately as possible. The ability to draw on research to
improve judging, say by predicting which of two candidates was more likely
to excel as a judge or how court practices could be changed to help sitting
judges perform better, would provide important practical benefits to society.
In the final essay of the book, Daniel Farber and Suzanna Sherry make the best
of current knowledge to offer some prescriptions. But the dominant focus of
the essays in this section, beginning with Gregory Mitchells, is on the
challenges posed by this project, on the theoretical side as we attempt to
determine precisely what it means for judges to perform well, and on the
methodological side as we seek to assess judges performance.
The practical and technical difficulties involved in studying judges are
daunting, to say the least. Psychological research most often entails conducting
experiments with the subjects of interest, but this method can take us only so far
in studying judges. Judges are far from the most accessible subjects and typically
are considerably less willing than undergraduates to participate in experiments.
Even when researchers can persuade judges to participate (e.g., Guthrie,
Rachlinski, & Wistrich, 2001), the experiments must necessarily lack something
of the complexity and unpredictability of real-world judging situations; more
importantly, they lack the stakes. The challenges involved in analyzing judges
thinking occupy the attention of a number of the authors here.
Methodological challenges is just one of several themes running through
the book. Two others are particularly important. One of these is skepticism
about the theoretical approaches to judicial decision making that dominate
the political science and legal literatures. Viewed through the lens of psycho-
logical research, these approaches (fully described in Lawrence Baums
chapter in this volume) can appear both overly simplistic and unrealistically
demanding. The essays point to a number of ways in which the dominant
theories seem psychologically implausible and in which we can improve our
understanding of judging by going beyond them.
The other major theme is the importance of differences across indivi-
duals, tasks, and situations. To be sure, students of judging have not simply
ignored such differences in the past. But it seems fair to say thatperhaps
because of excessive attention to the U.S. Supreme Courtwe have often
given less attention than we should to variation in types of cases judges hear,
the environments in which they operate, and the exact nature of the reasoning
tasks they are asked to perform. Psychology teaches us that we should not
expect the mental processes judges engage in to remain invariant across very
Introduction xv
different conditions. It also teaches us to expect variation across individual

judges in what they want to accomplish and how they think, but this kind of
variation in particular has been the subject of precious little analysis.
As noted, the essays in this volume raise more questions than they
answer. Furthermore, they are naturally not comprehensive in their coverage,
and there may be some topics discussed only briefly here that should be part
of a full-fledged psychology of judging. For example, psychologists in recent
years have devoted considerable attention to the roleboth positive and
negativethat emotion plays in peoples thinking (Forgas, 2000; Thagard,
2006). This could be an important area of inquiry for students of judges (see
Posner, 2008, pp. 105107). But the aim of this volume is not to lay out a
complete framework for the study of psychology in judicial decision making;
rather it is meant to encourage more scholars to engage in that study and
provide suggestions for where to begin. To the extent it convinces readers that
doing so can be intellectually exciting and practically important, it will have
succeeded in its task.
This volume grew out of a workshop held in Charlottesville, Virginia, in
March, 2007. The editors and contributors are grateful to the National
Science Foundation for support of the workshop and book.
This page intentionally left blank
Part I
JUDGES AND HUMAN BEHAVIOR
1
Motivation and Judicial Behavior: Expanding
the Scope of Inquiry
Lawrence Baum
Students of judicial behavior have taken only limited steps to incorporate

psychological theory into research on judicial decision making. In my view,
this represents a missed opportunity. It is true that judges and judging differ
in important ways from the people and activities that psychologists generally
study, so it is uncertain how much of what they have learned can be applied to
judges choices as decision makers.1 Yet even if we approach psychological
research with some caution, we can gain considerable insight on judicial
behavior from that research. In this essay I discuss the value of psychological
theory for an understanding of judicial behavior, both in broad terms and in
relation to one key issue in judicial behavior.
The first section of the essay provides a preface to this Part of the book by
discussing the dominant theoretical models and perspectives in the study of
judicial behavior. I describe the state of theory about judging and evaluate
that body of work from a psychological perspective. This perspective, I argue,
highlights some important limitations to the ways that scholars generally
think about judicial behavior.
The second section considers the motivations on which judges act. The
study of judicial behavior implicitly centers on motivation, in that legal
scholars and political scientists explain judges choices in terms of what
they seek to accomplish with their decisions. I discuss ways that psycho-
logical theory can help scholars to dig more deeply into judges motives and
thus to gain a richer understanding of the motivational bases for judicial
decisions.
3
4 Judges and Human Behavior
The Study of Judicial Behavior

Explanation of judges choices is an important issue in legal scholarship and
the central concern of political scientists who study the courts. Of course,
judicial behavior takes many forms and occurs in many settings. The aspect of
judicial behavior on which theory and research concentrate is the decisions
that appellate courts (and especially the Supreme Court) reach on the merits
of the cases they consider.
Models of Judicial Behavior

Students of judicial behavior in law and political science do not always make
their theoretical premises explicit. For this reason, it can be difficult to sort
out how particular pieces of empirical scholarship relate to theoretical
models. However, taken as a whole, the work that is currently being done in
political science is considerably more explicit in this respect than the work of
past eras.2
Legal and SemiLegal Models

A good deal of legal scholarship rests on the normative premise that judges
should devote themselves to interpreting the law correctly by applying appro-
priate rules of interpretation such as adherence to the plain meaning of
statutory language and to relevant precedents.3 In some legal scholarship of
the past, this normative ideal was translated into an explanation: in deciding
cases, judges try only to make good law in this sense. This explanation was
effectively debunked by the legal realism movement of the early twentieth
century (Fisher, Horwitz, & Reed, 1993). Although something like a pure legal
model is reflected in some law-school teaching and occasionally appears in
legal scholarship (see Cross & Nelson, 2001, pp. 14391443), it has essentially
been discarded.
For some legal scholars and political scientists, the lesson of legal realism
is that legal considerations have no impact on judges choices. For others, the
lesson is more complex: the law does not determine what judges do, but
judges efforts to interpret the law well do much to influence their choices.
This complex version of legal realism, though usually implicit, is probably the
majority position in legal scholarship. It also has supporters in political
science, though the most prominent theoretical conceptions of judicial beha-
vior ascribe little or no impact to legal considerations. Indeed, Jeffrey Segal
and Harold Spaeth (2002), the leading proponents of what is called the
attitudinal model of decision making, treat the traditional legal model and
what might be called semilegal models as a foil that they seek to refute.
The view that judges give some weight to legal considerations is incor-
porated into a theoretical approach to the study of politics that has been
Motivation and Judicial Behavior 5
labeled historical institutionalism. Adherents to this school, such as Howard

Gillman (2001), Mark Graber (2006), and Keith Whittington (2000), argue
that judges care about making good law as well as good policy. They articulate
that position primarily through qualitative analysis of the course of judicial
doctrine.
There is also a growing body of quantitative research aimed at deter-
mining whether the law in its various forms affects judges choices. Two books
(Spaeth & Segal, 1999; Hansford & Spriggs, 2006) have probed the Supreme
Courts treatment of its own precedents. Mark Richards and Herbert Kritzer
(2002; Kritzer & Richards, 2003, 2005) have done a series of studies in which
they conclude that new Supreme Court doctrines reshape the justices
approaches to the cases to which those doctrines apply. In a book that
combined quantitative and qualitative analysis, David Klein (2002) argued
that judges on the federal courts of appeals respond to the decisions of other
courts largely on the basis of a commitment to interpret the law well.
Attitudinal Models
Legal realism was reflected in the quantitative studies of Supreme Court

decision making that political scientists began to undertake in the 1940s.
The premise of these studies was that the justices acted primarily on the basis
of their conceptions of good public policytheir policy preferences. Efforts
to follow relevant rules of legal interpretation might also influence their
choices, but on the whole policy considerations outweighed legal considera-
tions. These scholars chose as their main dependent variable a judges vote on
the outcome of a case (in other words, which party wins).
C. Herman Pritchett (1954), the leading early scholar, was largely implicit
in sketching out a theory of judicial behavior. Later, Harold Spaeth (1979)
and Glendon Schubert (1965) adopted more explicit theories that were based
partly on attitude theory in psychology. In their theoretical and empirical
work, Spaeth and Schubert took the position that attitudes toward alternative
policy positions accounted for all or nearly all of what Supreme Court justices
do. This conception is reflected in the label of the attitudinal model. Spaeth
and his collaborators (Rohde & Spaeth, 1976; Segal & Spaeth, 1993, 2002)
have applied the attitudinal model specifically to the Supreme Court, based in
part on the Courts institutional attributes. But other scholars extended the
model to lower courts, especially the federal courts of appeals and state
supreme courts.
From the 1960s to the 1990s, most of the political science scholarship on
judicial behavior followed the attitudinal model. One body of work probed
the dimensionality of votes in judicial decisions, a dimensionality that was
assumed to reflect the structure of judges attitudes toward policy questions
(e.g., Schubert, 1965). Another body of work analyzed the determinants of
judges votes under the assumption that the primary basis for those votes was
personal policy preferences (e.g., Tate, 1981).
In its original form, the attitudinal model was not linked to judges
motivations: the linkage between attitudes and votes was treated as more or
less reflexive. By the 1960s, however, both scholars who embraced the attitu-
dinal model and other students of judicial behavior were doing research that
assumed a conscious judicial goal of achieving good policy (e.g., Rohde,
1972). In stages of the decision process such as selection of cases to hear on
the merits (Schubert, 1962), judges were depicted as making calculations
based on their desire to advance the policies they favored. In other words,
they were behaving strategically. The reflexive and conscious conceptions of
the linkage between attitudes and behavior have continued to coexist unea-
sily. But in pure strategic models, considered next, the conscious conception
is dominant.
Strategic Models
Students of judicial behavior use the term strategic in multiple ways. The
most common usage relates to intent. In this usage, strategic judges seek to
achieve a desirable outcome for their actions by taking into account the
responses of other people to those actions. In the original form of the
attitudinal model, judges are not strategic when they cast votes on case
outcomes. Instead, they take the position that best reflects their policy pre-
ferences regardless of how others might react to what they do. In strategic
models, in contrast, judges might deviate from their most preferred positions
if doing so would achieve a better result. To take the most prosaic example, an
appellate judge might take a less liberal doctrinal position in a case than she
would prefer in order to secure a majority for a relatively liberal position. As
this description indicates, strategic models shift the focus from votes on
dichotomous case outcomes as the dependent variable to doctrinal positions
on an ideological spectrum.
Interest in judicial strategy goes back a long time, and in 1964 Walter
Murphy published an influential book in which he explored an array of
strategies that policy-oriented Supreme Court justices might employ. But in
the 1990s scholars began giving more attention to strategic behavior, pri-
marily because of the influence of rational choice models imported from
economics. Pulling together these developments, Lee Epstein and Jack Knight
(1998) wrote a book advocating a strategic model of Supreme Court decision
making. In depictions by them and by other scholars, strategic considerations
might lead the justices to take into account the prospective reactions of their
colleagues as well as a variety of groups outside the Court, including the other
branches of government and the general public.
Within political science, strategic models are now quite popular. Scholars
routinely consider the possibility of strategic behavior, and much of the
research on judicial behavior posits or assumes that judges are strategic.
One example is the book on Supreme Court decision making by Forrest
Maltzman, James Spriggs, and Paul Wahlbeck (2000), which uses
information in the justices papers to analyze strategy at several stages of

decision making (see also Hammond, Bonneau, & Sheehan, 2005). Like that
book, a good deal of empirical analysis by other scholars is aimed at doc-
umenting strategic action by judges or at ascertaining the extent to which
such action occurs. Just as the attitudinal model represented the standard way
of thinking about judicial behavior for a long time, strategic models now
occupy that position (Epstein & Knight, 2000).
Some scholars move back and forth between attitudinal and strategic
models without reconciling the two. Segal and Spaeth (2002) have sketched
out a reconciliation. They suggest that Supreme Court justices are strategic,
and they argue that strategic considerations affect what the justices do in most
stages of decision making. But in voting on the outcome of cases, they argue,
justices have no strategic reasons to depart from their most preferred policy
positions. In this view, the justices can be said to think strategically and act
attitudinally when they vote to affirm or reverse.
The difference between intent and result comes into play in a different
way in Epstein and Knights strategic model. The justices in that model care
only about making good policy. But because the justices are concerned with
acceptance of their decisions as legitimate, they sometimes act on legal
considerationsmost important, adhering to precedents that are inconsis-
tent with their policy preferences (Epstein & Knight, 1998, pp. 163177). In
this view, the justices are policy-oriented but act as if they gave some weight to
making good law.
Thus far, I have described strategy as a means to advance judges policy
goals. Judges could act strategically in the service of other goals as well. A
legally oriented judge could use strategy to advance the judges conception of
good law, and strategy plays a role of a sort even in pure legal models. If judges
conceive of good law as encompassing coherent legal rules, then judges on an
appellate court may compromise with each other as a means to achieve clear,
consensual decisions (Edwards, 2003).
Judges might also act strategically to advance multiple goals, sometimes
balancing them against each other: good law and good policy (Spiller & Tiller,
1996), good policy and continued tenure in office (M. Hall, 1992). The work
of some historical institutionalists treats Supreme Court justices as people
who use strategy on behalf of both legal and policy goals (e.g., Gillman, 1997).
But by and large, analyses of judicial strategy have focused on policy-oriented
strategy.
Probing the Models

As the summaries of their tenets make clear, the dominant models of judicial
behavior differ in important respects. Those differences are reflected in
debates among scholars about the relative importance of legal and policy
considerations for judges and about the extent and impact of strategic
behavior. But in any field of scholarship, disagreements typically occur
within a limited range, and competing camps share basic assumptions that are
largely unnoticed. That is certainly true of the several models of judicial
behavior.
It is useful on occasion to step back to identify and consider the assump-
tions that serve as foundations for work in a field. From a psychological
perspective, several key assumptions seem problematical in certain respects.
First, each of the models of judicial behavior implicitly centers on motivation,
in that judges actions are treated as driven solely by their goals: legal and
policy preferences turn directly into choices in cases. Scholars who utilize
these models seldom mention cognition explicitly (but see Segal, 1986), but
the implicit assumption is that the cognitive processes involved in judicial
decision making are straightforward and unproblematic.
This assumption is questionable on its face. The processing of informa-
tion, analysis of alternatives, and selection among those alternatives that
culminate in judges choices are hardly straightforward. Even if we conceive
of judicial decision making primarily in motivational terms, cognitive pro-
cesses surely intervene between goals and choices. In reality, motivation and
cognition are closely intertwined (Kruglanski, 1996).
The value of incorporating cognition into the study of judicial behavior is
underlined by the limited body of scholarship that has done so (Rowland &
Carp, 1996; D. Simon, 1998, 2004; Guthrie, Rachlinski, & Wistrich, 2001;
Braman, 2004; Braman & Nelson, 2007) and by several chapters in this book.
One important issue, especially relevant to this essays concern with motiva-
tion, is the issue of cognitive capacity. There is reason to question the implicit
assumption that judges can easily identify the course of action that will best
advance their goals.
The pure attitudinal model probably makes the most minimal cognitive
demands on judges, but even the task of applying policy preferences to the
alternatives in a case can be difficult. In the spatial metaphor that adherents to
the attitudinal model have always used, judges must locate both their pre-
ferences and the alternatives in a case along an ideological dimension. The
other models, in their simple and complex forms, require more from judges.
This is especially true of strategic models in which judges consider the
prospective responses of other policy makers to their courts decisions.
Psychologists and behavioral economists have amply demonstrated peoples
cognitive limitations (Kahneman & Tversky, 2000; see Mitchell, 2002). Judges
certainly are not immune to those limitations (Guthrie, Rachlinski, &
Wistrich, 2001).
Students of decision making in some other arenas have grappled with the
impact of human cognitive limitations. Behavioral economists, for instance,
have raised fundamental questions about the cognitive assumptions that
underlie orthodox models of economic behavior and probed the ways that
economic actors might depart from the predictions of the orthodox models as
a result (Thaler, 1991). Similar inquiries into the cognitive element in judicial
behavior are both necessary and potentially quite fruitful.
Second, in their treatment of judges motivations, the dominant models

of judicial behavior focus on proximate goalsgenerally the goal of making
good legal policy (a term that I use to encompass law, policy, or a combina-
tion of the two). The exponents of these models say little about the basic
motives that underlie proximate goals. In this respect, the scholarship on
judging is in sharp contrast with theory and empirical research in social
psychology.
Even though it narrows the scope of inquiry, a lack of attention to basic
motives is not inherently problematical. However, the judicial goal orienta-
tions posited by the dominant models can be contested on motivational
grounds. For one thing, these orientations are strikingly narrow. In these
models, among all the considerations that might influence judges choices,
only their interest in the substance of legal policy actually exerts much
influence.4 But without an inquiry into the motives that underlie judges
goals, it is difficult to establish why a wide range of other goals that judges can
be expected to hold should play no meaningful role in judicial behavior.
Further, the motivational basis for a strong judicial interest in making
good law and policy is not obvious. This is especially true in light of the
economic orientation that underlies strategic models of judicial choice (see
Schauer, 2000, pp. 620621). Legal and policy goals do not serve judges self-
interest in any conventional sense, so why are these goals so important to
judges? A satisfactory answer to that question is needed to provide a firmer
theoretical foundation for the dominant models of judicial behavior.
Third, social psychologists treat motivation as two-dimensional: it ener-
gizes behavior and sends the organism in a particular direction (Pittman,
1998, p. 549; see Chen & Chaiken, 1999, p. 76). Arguably, the two dimensions
are so different that different terms should be applied to them. But both
capture important elements of human behavior and sources of variation in
behavior.
The dominant models of judicial behavior focus on the directional
dimension, leaving the energizing dimension aside. More precisely, they
assume that judges will exert the maximum effort to advance their goals
to get decisions right, however they define right. That assumption is
especially clear in strategic models, in which judges engage in very time-
consuming labor to calculate their optimal strategic choices. Even if the
maximum effort were sufficient to overcome the cognitive limitations that
beset judges and other people, expending that effort would exact enormous
costs from judges.
The assumption of maximum effort is consistent with the premise that
judges care only about achieving good legal policy, but it is highly unrealistic.
It is true that judges have strong incentives to give time and effort to the task
of decision making, much stronger incentives than those of participants in
psychological experiments. Even so, judges have other things to do with their
time, and most (though not all) share the general human preference to
expend less effort rather than more.
Indeed, the anecdotal evidence that some judges make less than a full
commitment to their judicial work, a common subject of newspaper stories, is
compelling. A few legal scholars, including some who take an economic
perspective on the courts, posit that judges have a preference for leisure
(e.g., Posner, 1995, pp. 123126; Bainbridge & Gulati, 2002). These scholars
surely are right. In thinking about judges choices, we need to recognize their
interest in limiting the time and labor they devote to their jobs.
Finally, scholarship on judicial behavior generally treats that behavior as
homogeneous in important respects. This scholarship does allow for situa-
tional variation in the determinants of judges choices, primarily across courts
whose institutional characteristics vary in important respects. Scholars fre-
quently note another source of situational variation, the relative salience of
cases, and they have provided persuasive evidence of its impact (e.g.,
Maltzman, Spriggs, & Wahlbeck, 2000; Bartels, 2005, 2006; Unah &
Hancock, 2006; and McAtee & McGuire, 2007). But salience has not been
directly incorporated into the models that dominate the field, models that
typically treat every case as equal.
More striking is the implicit but deeply rooted assumption that, at least
within a particular court level, all judges act in the same ways. In attitudinal and
strategic models, every Supreme Court justice acts solely on personal policy
preferences. Justices also act on their preferences in uniform ways, responding
to cases in terms of the same ideological dimensions and (in strategic models)
adopting the same strategies. In these models the justices differ in their behavior
because their preferences differ, but in other respects they are alike.
There are good theoretical and empirical reasons to question the
emphasis on homogeneity in the dominant models. First of all, the potential
impact of situational variation on judging is greater than these models allow.
Though students of judicial behavior are sensitive to the effects of courts
institutional characteristics, those effects are likely more fundamental than
the scholarship recognizes. This is especially true of the differences between
trial and appellate courts (Rowland & Carp, 1996, chs. 67). And for judges
on a particular court, case salience is only one of the situational variables that
could affect decision making. In particular, judges might well act on different
bases in carrying out different taskscase selection and opinion writing, for
example.
Interpersonal differences are a more complicated matter. Research in
social psychology makes it clear that the situations in which people find
themselves play a powerful role in structuring their behavior. One effect is
to reduce interpersonal differences in the behavior of people who share the
same situations: inherent characteristics of individuals are less important
than most people think (Ross & Nisbett, 1991). Still, there remains consider-
able room for differences in behavior within a particular setting. The role of
appellate judge and the structure in which appellate judges do their work
constrain variation across individuals, but they hardly eliminate the potential
for variation.
Indeed, the evidence of differences among judges on the same court is

enormous in volume. Even a moderately attentive observer of the Supreme
Court can easily identify variation in the style of the justices opinions and of
their participation in oral argument. Journalistic and scholarly accounts of
the Court point to differences in the ways that the justices approach the tasks
of individual and collective decision making. To take one example, biogra-
phies of Harry Blackmun and Sandra Day OConnor point to significant
differences between the two long-time colleagues (Greenhouse, 2005;
Biskupic, 2005). Such differences are an essential part of the behavior that
scholars seek to understand.
This discussion of the homogeneity assumption underlines a theme that
runs through my consideration of the dominant models of judicial behavior:
our understanding of that behavior would benefit if we address complexities
that these models do not yet incorporate. At a different level, my prescription
is for greater attention to the insights that can be derived from psychological
theory. Psychological perspectives have played a limited role in research on
judging, and within political science that role has declined somewhat over
time.5 One effect of that trend has been to reduce attention to complexity.
Of course, there is a tradeoff involved in confronting complexities in
human behavior. While economic perspectives are applied most explicitly in
strategic models of judicial behavior, the other major models implicitly accept
the emphasis in economic theory on the use of simplifying assumptions to
make analysis more manageable (see Segal & Spaeth, 2002, pp. 4446). The
inherent value of such simplification is obvious. And in practice, simplifica-
tion makes empirical research more manageable and (not incidentally) more
publishable.
Even so, there are clear benefits to be gained from considering the
implications of human complexity. Even if scholars retain their commitment
to the current simplified models, recognition of the divergence between those
models and reality helps in understanding and interpreting empirical find-
ings. Further, at least some of the complexities left aside by the dominant
models can be built into those models and into empirical research. Thus the
complexities of human behavior that are identified by psychological theory
can and should be incorporated into the study of judging.
The Psychology of Judicial Motivation

Because the scholarship on judicial behavior has focused so much on judges
goals, motivation is a good place to start in thinking about how psychological
theory can inform our understanding of judges choices. In this section I
consider issues relating to each of the assumptions that I have identified as
problematical. These issues overlap with each other, and in the final discus-
sion (on motivation and cognition) I discuss one theoretical approach that
addresses multiple issues.
Goals and Motives
The assumption that judges seek primarily (or solely) to make good policy,
widely accepted in political science, has primarily an inductive basis.
Unsystematic observation of judges choices and systematic analysis of
those choices suggest to scholars that judges are acting on their policy
preferences. To take one example, the evidence of an ideological dimension
in the votes of Supreme Court justices on case outcomes lends itself to the
conclusion that the justices choices are strongly policy-oriented.
The assumption that the goal of making good law is important to judges,
accepted by most legal scholars and many political scientists, is more deeply
rooted in theory. But the theory on which scholars rely is primarily norma-
tive: trying to interpret the law accurately and well is a role prescription for
judges. Here too, little is done to connect the goals that scholars posit and the
motivations that must underlie these goals.
Some political scientists have developed a rationale for judges concen-
tration on legal policy, one that sidesteps the question of motivation. This
rationale rests on the exclusion of other goals. Focusing on the Supreme
Court, these scholars argue that institutional characteristics of the Court
render other goals irrelevant (Epstein & Knight, 1998, pp. 3649; Segal &
Spaeth, 2002, pp. 9296). One key attribute is the Courts ability to select the
cases it will hear, so that it typically hears difficult cases in which decisions
on both sides can readily be justified in legal terms.6
This rationale, though quite reasonable as far as it goes, is not entirely
satisfying. Most important, the scholars who offer it consider only a limited
range of goals that might compete with good legal policy, almost exclusively
career considerations. They point out that career goals are irrelevant to at
least the great majority of Supreme Court justices in the current era, since
other positions are seldom attractive enough to entice justices away from
the Court. But they do not consider other quite plausible goals, including
those relating to life at work (such as the quality of the working environ-
ment) and to the justices standing with colleagues and people outside the
Court.
Even if all potentially competing goals could be dismissed as irrelevant, it
would be useful to determine what basic motives of judges are satisfied
through efforts to make good legal policy. Because other goals cannot easily
be dismissed, that inquiry is even more important.
Judicial interest in good legal policy can be given firmer theoretical roots
by subsuming it within general inventories of motivation. Of the various
inventories that psychologists have offered (e.g., Beck, 2000, ch. 12; Fiske,
2003; Reeve, 2005, ch. 5), David Winters (2002, 2003a) typology is especially
relevant because it was developed to analyze political leaders.7 In Winters
conception, leaders act on their needs for achievement, power, and affiliation.
The power motive encompasses prestige, and it may be appropriate to treat
the need for prestige as a separate category.
Judges interest in making good legal policy might be fueled by each of

these needs. To take an obvious example, the need for achievement could
motivate efforts to advance what a judge sees as good public policy. Similarly,
judges may gain prestige within the legal community by fostering the percep-
tion that they are devoted to accurate interpretation of the law. In combina-
tion, the needs that Winter posits provide a potentially strong motivational
basis for the pursuit of legal and policy goals.
This does not mean that it is appropriate to rule out all other goals. The
motives that support the goal of achieving good legal policy support other
goals as well. For instance, the need for affiliation could motivate judges to
take positions that are popular with salient audiences rather than positions
they hold themselves. Further, because of Winters purposes, his typology
does not directly encompass basic needs such as economic security, needs that
could shape the choices of public officials such as judges.
Another issue in the linkage between goals and motives relates to cogni-
tion as well. Pursuit of goals may operate with varying degrees of conscious
thought, and psychologists have emphasized the role of nonconscious
motives in shaping behavior (Bargh et al., 2001; Gollwitzer & Bargh, 2005;
Moors & De Houwer, 2006). Almost surely, the motivations that underlie
judicial decision making operate more consciously than those of people in
most other situations. But even among political elites, motives that shape
behavior are not always conscious (Winter, 2003b, p. 121).
One implication is that the goals that judges consciously try to advance
and those they actually pursue through their choices may differ considerably.
Some judges proclaim that they seek only to make good law, even though
their patterns of votes and opinions indicate that their policy preferences have
a powerful impact on their behavior. Such judges are not necessarily dissem-
bling, because policy considerations may operate at an unconscious level. And
in all likelihood, judges are often unaware of their own efforts to win favor
from salient audiences through their decisions. For some analytic purposes, it
is unnecessary to identify the degree of consciousness in judges pursuit of
their goals. But nonconscious motives should be taken into account in efforts
to understand the purposive element in judicial behavior.
The Energizing Dimension

I have suggested that a full understanding of judicial behavior requires that
the energizing dimension of motivation be considered alongside the direc-
tional dimension. If the assumption that judges always devote the maximum
effort to advancing their goals is flawed, that inaccuracy has implications for
judicial decision making.
The two dimensions of motivation are interrelated. If judges devoted all
their time and energy to decision making, they still could not give every case
the scrutiny needed to ensure that their choices were the best means to
advance their goals (see D. Simon, 1998, pp. 8283). But if judges devote
some time and energy to other pursuits, the gap between the ideal and actual
levels of scrutiny is wider.
One way that judges might deal with this gap is to rely on the efforts of
others. In particular, they can defer to colleagues and delegate to law clerks
and staff attorneys. The costs to their goals of yielding some control over their
choices can be minimized if they rely on people whose judgment they trust.
Thus policy-oriented judges may consult like-minded colleagues, and they
may choose clerks who seem to mirror their own views (Ward & Weiden,
2006, pp. 99107).
Judges may also rely on heuristics in making choices. One example
concerns the selection of cases to hear in courts that have discretionary
jurisdiction. Some of these courts, such as the California and U.S. Supreme
Courts, face thousands of petitions for hearings each year. It is likely that
members of these courts (and the law clerks who assist them in screening
petitions) base their choices of cases to hear on a limited number of case
characteristics. Indeed, that was the conclusion of one study based on inter-
views with Supreme Court justices and law clerks (Perry, 1991, ch. 5).
The same is true of decisions on the merits. Whatever they seek to
accomplish with their decisions, judges must simplify the process by which
they reach those decisions. For example, some scholars have posited that
Supreme Court justices choose their positions in cases with the goal of
avoiding congressional overrides of their decisions through new statutes
(e.g., Eskridge, 1991; Bergara, Richman, & Spiller, 2003). The models that
scholars have used to test this hypothesis incorporate complicated calcula-
tions by the justices about the ideological placement of potential decisions
and of subsets of Congress (Segal, 1997; Bergara, Richman, & Spiller, 2003;
Sala & Spriggs, 2004). But even these models leave aside relevant considera-
tions such as the positions and activities of interest groups. If some justices do
try to avoid overrides, undoubtedly they make their task less time-consuming
by identifying a few variables that seem especially helpful in predicting
congressional action and ignoring others. This example underlines the need
to take into account both unavoidable and voluntary limits on judges efforts
to achieve their goals.
Variation by Judge and Situation

Despite the assumption of homogeneity that pervades most scholarship on
judicial behavior, there is abundant evidence of differences in that behavior
across judges and situations. Motivation is one important source of differ-
ences. In motivational terms, judges could be expected to differ in multiple
ways.
First, judges may have different motivational profiles. Winters empirical
research, based on analysis of leaders verbal outputs, has found considerable
variation within groups such as U.S. presidents in the relative importance of
the motives he describes (e.g., Winter, 2002). Barber (1965) found wide
differences among legislators. It seems highly unlikely that judges are

uniquely homogeneous in their mixes of motives. Indeed, two studies offer
strong evidence of interpersonal differences in motivation among trial judges
(Caldeira, 1977; Sarat, 1977), and Aliottas (1988) application of Winters
scheme to Supreme Court justices produced evidence of variation in the
justices mixes of motivations.
Second, the same motives may translate into different goals for different
judges. Supreme Court justices could seek prestige by taking policy positions
that accord with the values of groups outside the Court that share those
positions. Alternatively, they could demonstrate their skills in the legal craft
and thus appeal to people in the legal community who appreciate those skills.
A third way in which judges differ is in the energizing dimension of
motivation. There is considerable evidence of variation in the effort that
public officials give to their work in bodies such as Congress (R. Hall, 1996)
and state legislatures (Barber, 1965). Similar variation certainly exists in the
judiciary. The anecdotal evidence that some judges make a limited commit-
ment to their jobs sits alongside evidence that other judges work very hard to
get their work done.
It is difficult to distinguish empirically among these three types of
differences. That difficulty is illustrated by comparison of two Supreme
Court justices. William O. Douglas and William Brennan were both strongly
committed to liberal positions on civil liberties issues, and they voted together
in a high proportion of decisions. But they were quite different in the extent of
their efforts to win majorities on the Court. Douglas generally found it
satisfying to take his own positions, regardless of the collective outcome in
the Court (J. Simon, 1980). Exaggerating for effect, a colleague reported that
Bill Douglas is positively embarrassed if anyone on the court agrees with
him (Time Magazine, 1975). In contrast, Brennan worked regularly (and, in
general, effectively) to put together coalitions in support of his positions
(Eisler, 1993; Clark, 1995).
What underlay the differences between the two justices? Perhaps Douglas
and Brennan simply gained a sense of achievement in quite different ways.
Alternatively, Brennans interactions with his colleagues to win majorities
may have reflected a strong interest in affiliation or power that Douglas
lacked. For that matter, the differences between the two justices may have
resulted from Douglass relatively limited commitment to his work on the
Court (B. Murphy, 2003).
The recruitment process for judges undoubtedly limits variation in their
motivational profiles, because it narrows the range of people who are selected
for judgeships. However, as the examples of Douglas and Brennan suggest,
even judges who reach the same court may differ considerably. And differ-
ences among courts in recruitment, such as the contrast between appoint-
ment and election, can foster differences in the mix of judges motivations.
Of course, judges differ on many dimensions, not just those that involve
motivation.8 But variation in motivation is especially relevant to research on
judicial behavior, because that research emphasizes motivation so heavily.

Models built on the assumption that all judges want the same things are
highly questionable. Political scientists and legal scholars need to follow the
lead of psychologists in confronting interpersonal variation.
Two characteristics of the situations in which judges make their choices
have already been considered. Institutional differences among courts, such as
those between trial and appellate courts, can create quite different contexts for
judges choices. And the salience of cases could have multiple effects, the most
obvious of which is on the energizing dimension of motivation. We would
expect judges to expend more effort on behalf of their goals when they care
more about the outcome.
One situational characteristic that social psychologists have emphasized
is accountability. Their research has established the impact on behavior of
several forms and dimensions of accountability (Lerner & Tetlock, 1999).
Differences among types of cases in judges perceptions of accountability
could affect the weight they give to different criteria for decisions (Bartels,
2006) and the effort they devote to decision making. Whether or not a court
produces a signed opinion, for instance, might affect the care with which a
case is decided. In this and other respects, salience and accountability might
have reinforcing or conflicting effects.
Motivation and Cognition

In discussing issues that concern judges motivations, I have touched on
issues of cognition as well. Those issues are unavoidableor, at least, they
should be unavoidable. Indeed, as a corrective to the implicitly motivation-
centered perspective that dominates the study of judging, it makes sense to
follow the common practice in psychology of putting motivation within a
cognitive framework.
Of the frameworks and theories that psychologists have developed for the
study of cognition, the concept of motivated reasoning has the most obvious
relevance to judicial behavior as conceived by most legal scholars and political
scientists. The accuracy and directional goals that Kunda (1990) and others
(e.g., Baumeister & Newman, 1994; Lodge & Taber, 2000; see Hsee, 1996)
describe map well onto legal and policy goals.
Braman (2004) has analyzed judicial decision making in terms of
motivated reasoning. She demonstrates the insights that can be gained from
this framework while noting some problems in its application to judicial
decision making (see also Braman & Nelson, 2007). The concept of motivated
reasoning provides a way to understand how judges policy preferences
could influence their choices in a less than fully conscious way. It also high-
lights variation among cases in the balance between legal and policy con-
siderations: the more that the law supports one side in a case, the more
difficult it is for judges to reason their way to decisions that favor the other
side.
Referring to motivated reasoning, Segal and Spaeth (2002, p. 433) argue

that the legal ambiguity present in all Supreme Court cases frees the justices to
pursue good policy without regard to the legal merits of cases. As Braman has
shown, this argument can be contested. In any event, appellate courts that
lack discretionary jurisdiction hear a mix of cases that vary in how easy they
are to decide on a legal basis. The motivated reasoning framework seems well
suited to an understanding of variation in the processes by which judges reach
decisions in those courts.
Thus, the motivated reasoning framework addresses one important issue
in decision making. Ideally, a cognitive framework would take into account
each set of issues that I have considered: the motivational bases for goals, the
two dimensions of motivation, and motivational variation across individuals
and situations. One good candidate is the heuristic-systematic model of
information processing that Shelly Chaiken and her collaborators have devel-
oped (Chaiken, Liberman, & Eagly, 1989; Chaiken, Giner-Sorolla, & Chen,
1996; Chen & Chaiken, 1999). The heuristic-systematic model is one of
several dual-process theories of judgment (Chaiken & Trope, 1999). In such
theories, individuals are motivated tacticians who select different cognitive
strategies under different circumstances (Operario & Fiske, 1999, p. 67). In
the heuristic-systematic model, as is typical of psychological models of cogni-
tion, the dependent variables are attitudes and beliefs rather than authorita-
tive decisions. Even more than the distinction between ordinary people and
political elites, this attribute calls for caution in applying the model to judicial
behavior. Yet the basic insights of the model are clearly relevant to an under-
standing of judges choices.
In the heuristic-systematic model, heuristic information processing is
relatively effortless . . . characterized by the application of simple decision
rules, while the systematic mode is more effortful and analytic (Chaiken,
Giner-Sorolla, & Chen, 1996, p. 553). The model incorporates three goals
for individuals who make choices. As in the motivated reasoning framework,
one goal is accuracy. The others are defense motivation, an orientation
toward reinforcing important self-related beliefs, and impression motiva-
tion, an orientation toward holding and expressing beliefs dictated by
the current interpersonal situation (Chaiken, Giner-Sorolla, & Chen, 1996,
p. 554).
Both the second and third motivations are intriguing. Defense motiva-
tion can be understood as a basis for acting on ones policy preferences, but
for students of judicial behavior it is an unconventional formulation of the
linkage between preferences and choice. However, this formulation offers a
useful way of thinking about the weight of policy considerations in judicial
decision making. For judges to depart from their preferred positions might
detract from their sense of themselves. From this perspective, the popularity
of concurring and dissenting opinions in the current era is understandable.
There may be high personal costs to signing on to an opinion that does not
reflect a judges personal beliefs, and judges may gain considerable
satisfaction from taking a position that perfectly reflects those beliefs. This
satisfaction helps in understanding Justice Antonin Scalias (1994, p. 42)
lyrical description of writing an opinion to express precisely his own
position as an unparalleled pleasure.
Among the dominant models of judicial behavior, only strategic models
incorporate the impression management motivation, and they do so only in a
limited way. Strategic judges seek to affect the impressions of other people
only for instrumental reasonsto win support for their positions from
colleagues, to avoid negative reactions to their courts decisions from other
policy makers and the public, to win reelection or reappointment so they can
continue to make legal policy. Yet judges are not social isolates, so their
interest in impression management cannot be instrumental alone.
Inevitably, they also seek the regard of other people for its own sake. Thus
self-presentation is at least as important to judges as it is to ordinary people
(Goffman, 1959; Schlenker & Pontari, 2000). And judges are especially well
situated to present themselves to relevant audiences.9
Impression motivation can help to explain the goals posited by the
standard models of judicial behavior, in that judges interest in the approval
of salient audiences might reinforce their interest in good law or good policy.
But social motives can lead people to make choices that diverge from both
their conceptions of accuracy and their own directional preferences.10 For this
reason, separating impression motivation from other goals expands inquiry
into the motives that drive judicial behavior.
In combination, defense motivation and impression motivation call
attention to the role of individual and social identities in judging. Judges
act in part to produce desired self-concepts and favorable images of them-
selves among audiences that are important to them. It is doubtful that judges
choices can be fully understood in terms of identity. But that concept
provides a useful counterpoint to the usual ways of conceptualizing judicial
behavior, which leave the human element out of judicial choice.
As a dual-process theory, the heuristic-systematic model takes the ener-
gizing dimension of motivation into account. Under the sufficiency prin-
ciple (Chaiken, Giner-Sorolla, & Chen, 1996, p. 554), whether people engage
in heuristic or systematic processing depends on what level of effort is
sufficient to serve their needs. But very busy public officials must define
sufficiency in relation to the time and effort that are available for decision
making. One result is that judges sometimes engage in heuristic reasoning
even though they would prefer the level of confidence in their choices that
systematic reasoning would provide.
Undoubtedly, the dichotomy between two modes of reasoning over-
simplifies variation in the effort that judges give to their decisions. Yet it is
interesting that courts sometimes divide cases into two categories. Many
intermediate courts explicitly differentiate between two classes of cases. A
central staff of law clerks selects what appear to be easy cases, in the sense that
one side clearly is right under the law, and makes tentative decisions in those
cases before forwarding them to panels of judges who typically accept the staff
recommendation (Chapper & Hanson, 1990, pp. 1522; Symposium, 2002).
In the U.S. and California Supreme Courts, court personnel assign petitions
for hearings to two categories, and petitions in the two categories receive quite
different levels of scrutiny from the justices (Caldeira & Wright, 1990;
California Supreme Court, n.d.).11
In those examples, judges give more attention to classes of cases that they
care more about, and the discussion of salience suggested the same result. But
the impact of salience might be mediated by another factor. In decisions on
the merits, as suggested by Lawrence Wrightsmans essay in this book, judges
may engage in less systematic evaluation of cases that they can readily locate in
ideological space, because ideology serves as the critical heuristic. In other
words, issues that have clear ideological referents for judges tend to be easier
than those that lack those referents (see Pollock, Lilie, & Vittes, 1993, p. 30).12
On average, such ideological cases are probably more salient to judges than
other cases.
The distinction between two modes of reasoning is one way that the
heuristic-systematic model incorporates variation in motivation. As sug-
gested earlier, the care with which judicial decisions are made can vary
among judges as well as cases. Similarly, the absolute and relative importance
of the three motives in the heuristic-systematic model surely differs among
individuals and across situations.
Implications for Issues in Judicial Behavior

Most students of judicial behavior agree on some basic assumptions about the
bases for judges choices, but within the framework of those assumptions they
disagree strongly about certain issues. I have discussed specific ways in which
elements of psychological theory implicate both the disagreements and the
broader agreement. In this section I pull together those discussions and take a
broader look at the implications of psychological theory.
The Primacy of Legal Policy
The most widely shared assumption is that judges as decision makers act
primarily or entirely on the goal of making good legal policy. That assump-
tion is nearly universal in the study of the Supreme Court. The dominant
models of judicial behavior incorporate that assumption but provide only
limited justification for it. Can a justification be developed in psychological
terms?
The dominant models do not connect the posited goals of judges to their
basic motivations, to show the reasons that judges might be drawn to make
good law or good policy as they define it. This limitation is relatively easy to
overcome, and some scholarsprimarily among those who work outside
those modelshave suggested a means to do so. In their account, judges act

on an intrinsic motivation, the satisfaction they gain from efforts to make
good law or good policy (Landes & Posner, 1975, p. 887; Posner, 1995, pp.
131, 133; Higgins & Rubin, 1980, p. 130; Klein, 2002, pp. 1112; Cross, 2003,
pp. 14731476). This account seems quite reasonable in light of the strength
of lawyers socialization in the task of making good law and the strength of the
policy preferences that are held by most politically active people (including
judges).
It also seems likely that this intrinsic motivation is reinforced by an
extrinsic motivation: judges gain popularity and respect from people who
are important to them by pursuing good law or good policy (see Baum, 2006).
Other judges and lawyers respond positively to judges who demonstrate skill
in the interpretation of the law and adherence to the task of interpreting the
law faithfully. Salient audiences that range from political groups to circles of
personal friends react to the content of judicial decisions as public policy. Like
the satisfaction gained from pursuing good legal policy, this extrinsic motiva-
tion fits into Winters typology of needs and the typology of motivations in
the heuristic-systematic model.
Taking the motivational bases for legal and policy goals into account can
help in understanding how those goals play out in practice. For example, the
ideological dimensions that exist in judicial votes might be explained in part
by judges ideological self-identifications and the importance to them of
audiences whose members define good policy on an ideological basis.
Judges whose positions are more difficult to classify ideologically may stand
at a moderate position on the ideological scale, but they may also identify
themselves in less ideological terms.
A second motivational limitation of the dominant models is more diffi-
cult to overcome. It is far from obvious that judges devote themselves to good
legal policy to the exclusion or near-exclusion of other goals. For one thing,
judges audiences may influence them to adopt positions that differ from both
their reading of the law and their policy preferences. The effect of concerns
about reelection on death penalty decisions in state supreme courts is an
especially clear example of that phenomenon (M. Hall, 1992, 1995). Judges
interest in the regard of other people can shape their choices in more subtle
ways as well.
Psychological theories of motivation treat effort as a dimension separate
from the content of individual motives, but leisure can be considered a goal as
well. As I have noted, the implicit assumption that judges devote the max-
imum possible effort to achieving good legal policy is highly questionable.
This reality moves judges toward heuristic rather than systematic processing
of relevant information. The result is to increase the distance between the
actions that would advance a judges interest in good law or good policy most
effectively and the actions that the judge actually takes. Further, judges may
take other actions that enhance their leisure at the expense of their interest in
legal policy. One example is the large decline since the mid-1980s in the
number of cases that the Supreme Court accepts for decisions on the merits
(OBrien, 2005).
Most courts have considerably less control over their workloads than the
Supreme Court, and the inherent limits on the use of systematic processing
impinge heavily on judges in these courts. In some courts, the press of cases is
so great that simply disposing of them becomes a major concernsometimes
the central concern. Put differently, judges most immediate goal may be to
reach any conclusion rather than a specific conclusion (Kruglanski &
Webster, 1996; Dhami, 2003). In intermediate appellate courts at least some
judges appear to move consciously between systematic and heuristic proces-
sing modes, giving careful attention to law and policy in a minority of cases
and more limited scrutiny to cases that they characterize as easy (Wold, 1978;
Linder, 1985, pp. 498499).
Thus the widely shared assumption that judges act only on the goal of
achieving good legal policy applies to some courts and cases better than
others. In no context does it fully fit the reality of judicial behavior, but in
somedecisions on the merits in the Supreme Court, for instancethe fit
may be good enough to make the assumption acceptable. In any context,
however, scholars need to take into account the motivational bases for judges
interest in the content of legal policy. By doing so, they can better identify the
ways that this interest affects judges choices and the conditions that affect the
linkage between the two.
Law Versus Policy
Among the scholars who think that judges devote themselves to making good
legal policy, one continuing matter of disagreement is whether judges define
that goal solely in terms of policy or whether good law is also important to
them. Students of judicial behavior who espouse policy-only models treat
their position as more realistic than its alternative. As noted earlier, this
judgment is empirically based for the most part. In part, however, it is
based on an unspoken assumption that judges have stronger motivations to
pursue good policy.
The validity of scholars conclusions from the empirical evidence can be
contested (Baum, 1997, ch. 3), but the assumption about judges motivations
is my concern here. There is no inherent reason that judges should elevate
policy over law. They can get satisfaction from advancing what they see as
desirable public policy, but they can also do so by interpreting the law
effectively. Similarly, they may be rewarded by their audiences in concrete
or symbolic ways by taking either path.13 In the terminology of the motivated
reasoning framework, directional goals do not necessarily take primacy over
accuracy goals.
Situational variation may be more important than any general rule. If the
relative weights of accuracy and directional goals depend on the ease of
identifying the more accurate result, then the extent of legal ambiguity in
cases is a critical variable. Because the average degree of ambiguity is greater in

cases before the Supreme Court than in those decided by the federal courts of
appeals, directional goals can be expected to have greater impact in the
Supreme Court. The salience of cases to judges affects the level of effort
they give to decision making and perhaps the mix of considerations that
shape their choices.
Variation among individual judges is less obvious but quite likely. Even if
judges have the same motivational profiles, they may pursue their needs in
different ways. Some judges might seek to gain prestige within the legal
profession by developing a reputation for faithful and effective interpretation
of the law. Others might pursue prestige within groups of people who share
certain policy preferences by taking positions that accord with the prevailing
views in those groups. Similarly, judges could gain a sense of personal
achievement through either route. There is good reason to be wary of some
judges claims that they are more faithful to the law than their colleagues, but
this does not mean that all judges balance law and policy in the same way.
Strategy
The growing popularity of strategic models of judicial behavior reflects a

widespread belief that policy-minded judges must be strategic. If a judge has a
policy position, why would the judge not do everything possible to bring
public policy closer to that position? But from a motivational perspective, the
reality is more complicated in several respects.
First, strategic judges gain nothing concrete for themselves by moving
public policy closer to their positions. Thus they differ from people in many
other situations, such as economic actors. Judges can gain a less concrete
benefit from strategic action, the satisfaction of making a difference. But
they might also get satisfaction simply by taking positions that fully accord
with their conceptions of good policy. Indeed, judges would seem likely to
satisfy their defense motivation more effectively by acting sincerely rather
than strategically. To depart from the positions they favor for strategic
reasons, when the ultimate outcome is uncertain, may be unsettling.
Further, judges who are concerned with the impressions they make on
relevant audiences may find it easier to make favorable impressions by casting
votes and writing opinions that those audiences favor than by taking strategic
positions that require justification. For judges, like legislators (Denzau, Riker,
& Shepsle, 1985; Wilkerson, 1990), concern with the opinions of their audi-
ences may work against strategic behavior.
Strategic behavior can provide symbolic benefits to judges through the
satisfaction gained by influencing legal policy. But there is a complication
here as well: even the optimal strategies are likely to have limited impact,
thereby reducing these symbolic benefits. In general, neither judicial collea-
gues nor people outside the court are easy to move. One result is that strategic
judges may suffer more defeats than victories. There is only so much that a
liberal justice on a conservative Supreme Court can do to win liberal majo-

rities. Judges can console themselves with the hope that short-term defeats
ultimately will be reversed, but the tendency to discount the future reduces
the value of this consolation (Loewenstein & Elster, 1992). Given this reality, a
judge in an unfavorable situation may prefer not to play a game in which
defeats are so common (Tushnet, 1992, pp. 21092110).
In contrast, if a judges goal is simply to take the position that best reflects
the judges preferences, success is far easier to achieve. For such judges, as
suggested earlier, being on the losing side does not detract from a sense of
accomplishment and may actually enhance it. As a result, the nonstrategic
judge is likely to get more reinforcement than does the strategic judge (see
Shah & Kruglanski, 2000, pp. 118123).
Further, optimal strategies are difficult to achieve. Psychologists and
behavioral economists have amply demonstrated that even people with
strong incentives to adopt optimal strategies make major, systematic errors
(Barberis & Thaler, 2003). A strategic judge faces some relatively easy tasks,
such as calculating what kinds of compromises will be necessary to win a
colleagues support for an opinion. But other tasks are considerably more
difficult. One example, discussed earlier, is determining whether a prospec-
tive decision would be overridden by Congress. Another is estimating the
long-term effects of espousing doctrinal positions that command only min-
ority support in a current case.
The most obvious implication of this difficulty is cognitive: strategic judges
frequently will make mistakes. But the difficulty of achieving good strategy
affects judges motivations as well. Sincere behavior may be attractive on other
grounds, and it has the additional advantage that it is less effortful. The uncer-
tainties and frustrations of strategic behavior further reduce its attractiveness.
Undoubtedly, some judges are more willing than others to deal with the
disadvantages of strategic behavior. Judges who have a strong need for power
may gain considerable satisfaction from a perception that they are influencing
collective outcomes in their own court and the broader course of public
policy. But even those judges may be selective in their efforts at strategy,
concentrating on strategies that are easiest to accomplish and that produce
results most quickly. This is one reason that strategy aimed at shaping
collective decisions within courts seems far more common than other forms
of judicial strategy.14 And judges who are willing to face more difficult
strategic tasks can be expected to adopt heuristics as a means to limit the
effort they put into strategic choices.
Conclusion
As the discussion of strategy illustrates, psychological theory provides new
perspectives on judges motivations and thus informs thinking and research
on issues of motivation. This chapters survey of relevant theory has been quite
incomplete, but it should make clear the benefits of considering judicial

motivation in psychological terms. By analyzing judges motives more broadly
and more deeply, scholars can gain a richer sense of the bases for judicial choice.
Two themes derived from psychological theory are especially important.
The first is the limits that exist on the efforts of decision makers to pursue the
goals that students of judicial behavior have posited. By choice and necessity,
judges stop short of the comprehensive analysis that is implicitly assumed by
the dominant models of judicial behavior. That reality provides a possible
basis for some widely held conceptions of judicial behavior, such as the
importance of ideological dimensions in judges choices. It questions other
conceptions, such as some elaborate formulations of judges strategic calcula-
tions. Closer examination of the effects of limited efforts in decision making
can provide insights on the processes of judicial choice.
The second theme is variation among cases and judges. Some scholars
have recognized that judges may decide different kinds of cases on different
bases, but this insight should be incorporated more directly into models of
judicial behavior. Differences among judges in motivational profiles and in
the links between their motives and choices have generally been left aside
altogether. Taking into account differences among individuals and among
situations will complicate the analysis of judicial behavior, but doing so will
also lead to better conceptions of that behavior.
Notes
I appreciate the comments and suggestions by Kathleen McGraw and Tom Nelson.
1. Research in psychology concentrates on ordinary people who are engaged in
ordinary behavior, and the insights of that research do not always apply well to
judges and judicial decision making. Scholarship on political psychology gives
primary attention to the mass public rather than government decision makers,
and the portion of this scholarship that deals with government decision makers
is concerned primarily with officials in the executive branch and with foreign
policy decisions (see Sears, Huddy, & Jevis, 2003).
2. The discussion that follows draws from ideas in Baum (2006, ch. 1).
3. This does not mean that there is always a single correct decision in a case,
because judges legitimately might adopt different rules of legal interpretation or
apply them differently.
4. As noted earlier, studies of state judges who lack life tenure are an exception
(Langer, 2002), though even for those judges continued tenure is typically treated
as a means to the end of making legal policy. Students of Congress tend to give re-
election greater weight as an end in itself (Fenno, 1973; Mayhew, 1974).
5. The use of psychological theory in research on judicial behavior is discussed in
Baum (1997, pp. 135141). In the decade since that time, psychologists, legal
scholars, and political scientists have made additional use of theories in psy-
chology to understand judicial behavior (e.g., [To distinguish from J.F. Simon,
1980) D. Simon, 1998; Gruenfeld & Preston, 2000; Guthrie, Rachlinski, &
Wistrich, 2001). Wrightsman (1999, 2006) has surveyed issues in Supreme

Court decision making from a psychological perspective, identifying a variety
of ways that psychological theory can inform our understanding of judicial
behavior. During the same period, however, economic theory has been given
considerably more use within law and political science.
6. These institutional characteristics are shared to varying degrees by other appel-
late courts. Because of that variation, some scholars have offered cautions about
applying the dominant models to appellate courts other than the Supreme
Court. This is especially true of state supreme courts, whose members depend
on voters or the other branches of government to maintain their positions in
the great majority of states (see Langer, 2002).
7. Another typology, developed specifically for analysis of politicians, is Payne and
Woshinskys (1972) categorization of motives for participation in politics.
Caldeira (1977) and Sarat (1977) each used this typology in studies of trial
judges, identifying their dominant motives and relating those motives to ele-
ments of the judges behavior.
8. On variation in cognition, specifically cognitive complexity, see Tetlock,
Bernzweig, & Gallant (1985) and Gruenfeld (1995).
9. I have developed this argument in Baum (2006). Concern with reputation as a
judicial motive is also discussed in Miceli and Cosgel (1994) and Schauer
(2000).
10. Judges directional goals in the motivated reasoning framework could result
from a variety of considerations. But the ways that Kunda and others describe
directional goals relate most clearly to policy preferences as their source.
11. Trial judges who must impose criminal sentences or reach final resolutions of civil
cases typically give much less attention to cases in which the two parties have
agreed on a proposed resolution than they do to cases in which no such proposal
exists. This dichotomy is understandable, perhaps inevitable, but it is also
consequential.
12. In turn, this aspect of judges perceptions of caseslike othersmight be
subject to framing effects (see Nelson, Oxley, & Clawson, 1997). Lawyers
routinely engage in efforts to frame cases in ways that they perceive as favorable
to their positions, but we know relatively little about the efficacy of those efforts.
13. Indeed, to the extent that judges personal and social identities are based on a
conception of judging as adherence to proper interpretation of the law, they
may have powerful reasons to emphasize legal considerations in their decision
making. To borrow terminology from Simonson and Staw (1992, p. 421), legal
audiences may create perceptions of accountability for process as distinct from
accountability for outcomes.
14. Another reason is that efforts to achieve consensus on appellate courts are
regarded as desirable because they contribute to clarity and coherence in the
law. Even judges whose only goal is to make good law as they see it would engage
in those efforts (Kornhauser & Sager, 1993).
2
Multiple Constraint Satisfaction in Judging
Jennifer K. Robbennolt, Robert J. MacCoun,

and John M. Darley
Under our law judges do in fact have considerable discretion in

certain of their decisions: making findings of fact, interpreting
language in the Constitution, statutes, and regulations; determining
whether officials or the executive branch have abused their dis-
cretion; and, fashioning remedies for violations of the law, including
fairly sweeping powers to grant injunctive relief. The larger reality,
however, is that judges exercise their powers subject to very signifi-
cant constraints.
Hon. Alex Kozinski, U.S. Court of Appeals for the Ninth Circuit
Judges make decisions for a living, and their decisions are unusually conse-
quential, with direct effects on immediate cases, and a ripple of less direct effects
on future cases. Trial court judges must variously act as finders of fact in bench
trials, jury trial supervisors, and overall case managers. Appellate court judges
may, for example, make decisions about the merits of particular cases, deter-
mine whether to join an opinion and whether to write separately (see e.g., Taha,
2004), or participate in decisions about whether to grant cert or en banc review
(George, 1999). Chief judges face an additional set of administrative responsi-
bilities, such as managing the docket and the budget, assigning opinion writing,
coordinating visiting judges and judges on senior status, hiring and firing staff,
and handling issues related to building maintenance and equipment (George &
Yoon, 2007). The Chief Justice of the U.S. Supreme Court has additional
responsibilities, such as appointing judges to the committees of the judicial
conference (Chutkow, 2007). Judges may even need to decide when it is the best
time to retire (see e.g., George & Yoon, 2007).
27
Each of the varied decisions that judges are called on to make inevitably
evokes a range of possible goals. Different models of judicial decision making
tend to highlight particular goals. For example, traditional legal theory posits
that in making decisions judges strive to reach the correct legal decision as
dictated by precedent. There are various legal realist and critical realist
alternatives to this baseline account; of particular relevance here are the
attitudinal, strategic, and managerial models. The attitudinal model focuses
on the ways in which judges make decisions that further their preferred policy
objectives (Segal & Spaeth, 1993, 2002). Strategic models incorporate con-
sideration of the ways in which judges seek to effectuate their goals in the long
term (Epstein & Knight, 1998; Maltzman, Spriggs, & Wahlbeck, 2000). The
managerial model emphasizes the increasing caseload pressures that judges at
all levels face (Resnik, 1982). Thus, as they make decisions, judges must
balance their desire to reach the right legal result, their preferences for
particular outcomes, their need to manage their workload, and many other
objectives.
To view these models as competitive accountsone more valid than the
othersis probably misguided. We argue that each model accurately cap-
tures some of what every judge does some of the time, and that no single
model is likely to describe any judge all of the time. A sophisticated under-
standing of judicial decision making should explicitly incorporate the notion
that judges simultaneously attempt to further numerous, disparate, and often
conflicting, objectives. In this chapter we attempt a preliminary account of a
more psychologically plausible account of judicial cognition and motivation.
Goals in Judicial Decision Making

Traditional legal theory posits that judges ought to attempt to reach a correct
legal decision through the proper application of legal rules and precedent (see
review in Cross, 1997). As judge Harry Edwards (1985) argues, it is the law
and not the personal politics of individual judgesthat controls judicial
decision-making. Consistent with this approach, there is evidence that the
law does influence judicial decision making (see, e.g., Klein, 2002; Richards &
Kritzer, 2002). However, substantial evidence demonstrates a range of addi-
tional objectivesbeyond a correct legal holdingthat judges seek to accom-
plish. For example, the attitudinal model of judicial decision making holds
that judges make decisions that will maximize their policy preferences, voting
in ways that are consistent with their political ideology (see Segal & Spaeth,
1993, 2002; see also George, 1998).
An influential approach to the attitudinal model of judging appears in the
methodologically ambitious work of Andrew Martin, Kevin Quinn, and their
colleagues (see, e.g., Martin & Quinn, 2002), who argue that Supreme Court
votes across 47 consecutive terms are well represented by a single ideological
dimension, and that at any given time, a justice can be located at a position
Multiple Constraint Satisfaction in Judging 29
an ideal pointon that dimension.1 This model, or any attitudinal model

that accounts for a good deal of the variance in judicial decisions, if correct,
would not necessarily invalidate our constraint satisfaction account, but it
would render our goal management principles superfluous, at least for jus-
tices of the Supreme Court. However, based on methodological considera-
tions, we think it is plausible that the Quinn-Martin analysis could overlook
other meaningful dimensions in justices votes.2
We hope future research will attempt to cross-validate the Martin and
Quinn analysis. For example, their unidimensional ideological scores could
be validated against content analyses of how justices and decisions have been
characterized in op-ed essays in U.S. newspapers,3 and against ideal-point
congressional datafor example, data on which representatives endorsed or
opposed which justices. Without further validation, it is unclear whether
Martin and Quinn are accurately characterizing the dimensional structure
of the attitudes of Supreme Court justices. Still, we recognize that a uni-
dimensional model might be a reasonable (and usefully simple) first
approximation.
Related models have incorporated elements of strategic behavior in
judicial decision making (Epstein & Knight, 1998; Maltzman, Spriggs, &
Wahlbeck, 2000). These strategic models propose that judges do not simply
vote in ways that are plainly consistent with their attitudes, but make deci-
sions that take into account the ways that the predicted actions of other
players (such as their colleagues or Congress) influence the feasibility of
attaining their desired ends. Thus, judges may agree to decisions that do
not completely effect their policy preferences to avoid results that depart
even further from their preferences or may draft opinions in ways that do
not perfectly represent their preferences in the instant case, but that will
garner the necessary votes. Such models have primarily examined judges
use of strategy in effectuating their ideological goals, but strategic behavior
could be employed in the service of other judicial objectives as well (Baum,
2006).
In contrast to an account of judicial decision making based on ideology,
an account of judges as case managers highlights the effects of caseload
pressures on judicial decision making (see Resnik, 1982). Managerial judges
are thought to be concerned with saving time, reducing delays, and improving
efficiency.4 Thus, in this view, a primary goal for judges is to move the docket.
For trial court judges, this goal may manifest itself in a desire to settle cases
(Resnik, 2002), more judicial involvement at earlier stages of the case
(Galanter, 2004), and decreased opinion writing (Taha, 2004). But a need
to move the docket may be experienced at all levels of the judiciary. For
example, judges may choose to dispose of cases on procedural grounds to
limit the need to decide cases on the merits or to avoid the need to decide
cases in areas where they have less expertise (see Macey, 1994; Resnik, 2002,
describing the profoundly challenging problems of rendering judgment).
Judges as case managers may strive to limit their workloads, minimize the
amount of time they spend on aspects of their job they find less interesting in
favor of decision tasks (or cases) they prefer, or attempt to achieve control (or
a sense of control) over the nature and pace of their work.
One side effect of a focus on efficiency, coupled with the ability that
courts now have to collect more data on court operations, is increasing
opportunity for judges to attend to their statisticsfor example, the
number of cases they terminate, the number of motions they have ruled on,
or the number of days they have spent in trial (Resnik, 1982; see also Darley,
2001). Thus, judges may have goals that relate to improving their perfor-
mance on these types of measures.
Intertwined with the goals that are most central to legal, attitudinal,
strategic, and managerial models of judicial decision making, judges may
also be influenced by a range of additional objectives. For example, judges
may seek to make decisions that will not be overturned by a higher court or on
en banc review; they may seek to maximize their opportunities to exercise
discretion; they may seek to cultivate their reputation with their peers or
another constituency (e.g., the bar, academics, Congress, the press, particular
interest groups, or the public), aspiring to be respected, influential, and
frequently cited; they may seek to be reelected, to be promoted to a higher
court, or to move to another position beyond the court; they may seek to
build collegial relationships with their colleagues on the bench; they may seek
to make decisions that are consistent with their self-identity; and they may
seek to achieve a measure of consistency with their own past decisions (see
generally Baum, 1997, 2006; Cohen, 1991; Posner, 1993; Wrightsman, 2006).
While many of these disparate goals may be entertained consciously, others
may be adopted or pursued without conscious awareness (see, e.g., Bargh &
Chartrand, 1999; Shah, 2005; see also Guthrie, Rachlinski, & Wistrich, 2007).
Furthermore, variations in the decision contextfor example, whether and
how the judge will be accountable for the decisioncan serve to make
particular goals temporarily operable or salient (see, e.g., Lerner & Tetlock,
1999).
While judges as a group may share this range of objectives to one degree
or another, judges sitting on different courts or across jurisdictions face
different sets of tasks and demands. Similarly, different decision tasks may
evoke different judicial goals. For example, trial and appellate court judges are
called on to make different kinds of decisions and face differing constraints on
their decision making. U.S. Supreme Court justices enjoy a greater degree of
control over their agenda and more discretion than do judges on other courts
and may seek to effect a somewhat different set of objectives. For trial court
judges, ruling on a pretrial motion may be subject to different constraints
than reaching a verdict in a bench trial. Judges who sit in jurisdictions in
which judges are elected may face different pressures than do judges who are
appointed. This divergence in decision tasks and in the range of demands
faced by judges sitting on different types of courts or across jurisdictions may
lead to different (though overlapping) sets of salient goals.
Trial Court Judges
Trial court judges wear many different hats, variously serving as finders of
fact, trial supervisors, and overall case managers. These judges decide some
cases on the merits, but they also manage the trial processruling on objec-
tions and motions and instructing juries about the law. Trial court judges may
hold Daubert hearings to determine the admissibility of scientific evidence,
make determinations about the appropriate amount of bail, and conduct
posttrial assessments of damage awards. In addition, trial court judges now
spend much of their time managing the pretrial and case settlement processes
as well as overseeing the implementation of remedies posttrial (Resnik, 1982).
As fact-finders, judges may struggle to simultaneously accomplish
myriad goalsmaking accurate factual determinations and reaching a verdict
consistent with the evidence (see Pennington & Hastie, 1993); accomplishing
optimal deterrence (see Becker, 1969; Cooter & Ulen, 2007); awarding appro-
priate compensation (Darley & Pittman, 2003;); accomplishing some mea-
sure of distributive justice (see Deutsch, 1975); punishing when appropriate
and to the extent that is fitting (see Darley et al., 2000); using the appropriate
rules to guide decision making (see Robbennolt, Darley, & MacCoun, 2003);
or expressing their values (see e.g., Sunstein, 1996; Robbennolt, Darley, &
MacCoun, 2003). Indeed, as with jurors, judges have been shown to have
difficulties with some of the decisions required by legal and economic models
of decision making. Specifically, judges have been shown to have trouble
ignoring inadmissible evidence (see e.g., Landsman & Rakos, 1994; Wistrich
et al., 2005) and evaluating scientific, expert, or statistical evidence (Gatowski
et al., 2001; Kovera & McAuliff, 2000; Redding & Repucci, 1999; Wells, 1992).
In addition, judges have been found to be susceptible to a variety of cognitive
heuristics such as anchoring, framing, hindsight bias, the representativeness
heuristic, and the egocentric bias (Guthrie et al., 2001; Rachlinski et al., 2006).
Trial court judges, however, spend much of their time engaged in tasks
other than presiding over trials. Indeed, judges are presiding over fewer and
fewer trials (Galanter, 2004). In their role as case managers, trial court judges
have different tasks and goals than they do in their role as fact-findersthey
must negotiate with parties pretrial to settle cases, plan litigation, and manage
discovery, and supervise the implementation of remedies posttrial (Resnik,
1982). In addition, trial court judges may be called on to manage complex
class-action or multidistrict litigation (see, e.g., Galanter, 2004; Walker &
Manahan, 2007).
As noted above, these case management responsibilities give rise to
incentives to get cases resolved and off the docket. Judges may even utilize
the symbolism of procedural justice to get cases settled. MacCoun (2005)
relates an anecdote in which a judge conducted a settlement conference in
which the attorneys negotiated a settlement in the clients absence. When the
plaintiffs attorney complained that his client might not accept the settlement
without getting her day in court, the judge put on his robe, called her into
an empty courtroom, and sat her on the witness chair. After she told her story,
she assented to the settlement.
These case management pressures provide an additional set of goals with
which trial court judges, in particular, must contend.
Appellate Judges
Appellate judges face an overlapping, but somewhat different set of decision
tasks than do trial court judges. Rather than acting as fact-finders, appellate
court judges are primarily engaged in the business of judicial review.
Accordingly, they face decisionssuch as whether to grant cert to hear a
case or whether to grant en banc reviewthat trial court judges do not.
Similarly, appellate court judges must make decisions about whether to ask
questions at oral argument and what the nature of those questions will be.
Appellate court judges engage in more opinion writing and, in addition to
determining how they will vote in a particular case, must decide whether to
join a particular opinion, to write separately, or to author a dissenting
opinion. Appellate judges at different levels may face differently structured
decision tasks. For example, justices of the Supreme Court exercise more
control over their agenda than do intermediate appellate judges.
Elected Judges
Judges who face reelection or some form of retention election face additional
pressures attendant to such elections. There is evidence that judicial decision
making is influenced by such political concerns. For example, there is evi-
dence that in years in which they are up for reelection, judges are more likely
to sentence criminal defendants to death (see Brace & Hall, 1997; Brooks &
Raphael, 2003; see generally Bright & Keenan, 1995) or to sentence more
harshly in general (Huber & Gordon, 2004). Other hot button issues such as
tort reform also play a role in the politics of judicial elections and have the
potential to influence judicial decision making (see generally Champagne,
2005; Ware, 1999).
Judges as Goal Managers

It is clear that judicial decision making implicates a wide variety of objectives.
Judges may be required to balance, for example, a desire to follow precedent
against preferred policy preferences, or to balance the effort needed to act
strategically against a desire to limit workload, among other goal conflicts.
Moreover, in attempting to balance these varied goals, judges have at their
disposal a range of decision-making options, or choices about how to proceed
(see, e.g., Molot, 1998, discussing the wide array of tactics available to
judges as they attempt to manage their dockets). Models of decision
making that portray judges as pursuing single objectives and that do not
account for these intricacies are likely to miss important facets of the process.
We therefore propose a model, the parallel constraint satisfaction model, that
is explicitly designed to incorporate multiple objectives.
In an attempt to encourage the development of models of legal decision
making that capture these types of complexities, we have argued that legal
decision making might profitably be conceived of as a process of parallel con-
straint satisfaction that can be represented using connectionist models
(Robbennolt, Darley, & MacCoun, 2003; also see Simon, 2004). Connectionist
models endeavor to provide a framework for thinking about and modeling
decision-making tasks that require the integration of a range of disparate, and
potentially inconsistent, information and objectives (see Read & Miller, 1998;
Read et al., 1997). Accordingly, such models are well suited to modeling decision-
making by judges who must balance numerous, potentially inconsistent, goals.
Constraint satisfaction networks are made up of a set of nodes or
elements connected by links in a neural-like network. Each element comprises
a concept, item of evidence, legal proposition, or goal, and the links or
connections between elements vary in strength and valence (indicating the
degree of coherence or incoherence between elements) (Read et al., 1997).
The links that connect the elements represent the constraints faced by deci-
sion makerselements that are consistent, or mutually compatible, are said
to be coherent and are connected with positively valenced links, while ele-
ments that are negatively related or that inhibit each other are said to be
incoherent and are negatively linked (Read et al., 1997; Thagard, 2000). Such
models are sophisticated enough to take account of differential initial prio-
rities among goalsthis would be done by initially linking favored goals to
elements set to higher levels of activation and less favored goals to elements
with lower levels of activation (see Thagard, 2000).
Under this framework, decisions are made by finding the action that best
balances the constraints among the decision elements (Read & Marcus-
Newhall, 1993; Thagard, 2000). In a parallel constraint satisfaction model,
this balance is struck through a process of iterative updating of the model:
In a parallel constraint satisfaction connectionist model, each element
is assigned an equal initial activation value (e.g., .01). The central
aspect of the model is that the activation level of each element in the
model is then updated simultaneously based on four factors: (1) the
number of other elements connected to it; (2) the level of activation of
those elements; (3) the strength of the links to these other elements;
and (4) the valence of those links. This updating process is iterated
with activation of elements spreading through the network based on
the configuration of links between the elements until the activation of
each element stabilizes. Once the network settles, each element is
accepted or rejected based on its final degree of activation
(Robbennolt, Darley, & MacCoun, 2003, pp. 11491150).
Thus, a constraint satisfaction model is a mechanism for simultaneously

accounting for all of the relevant constraints on the decision, including both
consciously articulated and more intuitive objectives. Such models provide a
useful framework for considering judicial decision making and are broad
enough to encompass the range of decisions that judges have to make,
including decisions in the role of fact-finder, legal decisions at trial or on
appeal, strategic decisions, and administrative decisions.
For example, parallel constraint satisfaction models have been used to
model trial level decision making (see, e.g., Byrne, 1995; Holyoak & Simon,
1999; Simon, 1998; Simon & Holyoak, 2002 Simon et al., 2004; Thagard,
1989). Indeed, it has been argued that [p]rocesses of maximizing explana-
tory coherence are particularly well-suited for accounting for . . . decision
making where the task is to evaluate the coherence of accounts presented by
the prosecution and the defense (Thagard & Kunda, 1998, p. 13). In parti-
cular, fact-finders must engage in parallel constraint satisfaction as they
attempt to integrate and account for the array of evidence presented at trial
in an attempt to achieve explanatory coherence.5 That is, fact-finders attempt
to come to an understanding of the facts that fits with the available informa-
tion [i.e., trial evidence] better than alternative interpretations (Thagard,
2000, p. 16). Moreover, the best interpretation is one that provides the most
coherent account of what we want to understand, considering both pieces of
information that fit with each other and pieces of information that do not fit
with each other (Thagard, 2000, p. 16).
Importantly for our purposes here, parallel constraint satisfaction
models can also be used to model the ways in which judges and other legal
decision makers make decisions to maximize satisfaction of their varying
goals, that is, to achieve what is referred to as deliberative coherence. In a
model of deliberative coherence, decision makers faced with multiple, poten-
tially inconsistent goals, seek a course of action that accomplishes the greatest
coherence among competing goals. The competing goals and the potential
avenues open to the decision maker are linked together in ways that signify
the degree to which they are compatible or incompatible (Thagard &
Millgram, 1995). Decision making, then, is:
inference to the best plan. When people make decisions, they do
not simply choose an action to perform, but rather adopt complex
plans on the basis of a holistic assessment of various competing
actions and goals. Choosing a plan is in part a matter of evaluating
goals as well as actions. Choice is made by arriving at a plan or
plans that involve actions and goals that are coherent with other
actions and goals to which one is committed. (Thagard &
Millgram, 1995, p. 440)
In essence, parallel constraint satisfaction models provide a way to account
for the complex interplay among actions and goals. Specifically, we
(Robbennolt, Darley, & MacCoun, 2003) have proposed a set of goal
management principles that can describe the interrelations among the dis-
parate goals and actions pursued by legal decision makers:
Principle of equifinality: some goals may be alternately satisfied
through any one of a number of actions (see Kruglanski et al.,
2002; Anderson & MacCoun, 1999);
Principle of best fit: a particular action may better fulfill some goals
than others;
Principle of multifinality: a particular action may sometimes
accomplish multiple goals simultaneously (see Kruglanski et al.,
2002);
Principle of goal incompatibility: some goals will inevitably conflict
and, thus, be difficult or impossible to satisfy concurrently.
Connectionist models of parallel constraint satisfaction accommodate
these principles in various ways:
a goal might be connected by positive links to more than one action
(equifinality) and each possible action may be connected by positive
links to more than one goal (multifinality). At the same time, the
links between a goal and several different actions may have different
weights (best fit) and some of the links between two goals or two
actions may be negatively valenced (incompatibility). The
connectionist network updates activation of the elements (goals and
actions) in parallel until the network stabilizes. In this case, the final
activation of the elements represents the decision makers chosen set
of selected actions or goal valuations. (Robbennolt, Darley, &
MacCoun, 2003, pp. 11541155)
Thus, examining deliberative coherence through parallel constraint satis-
faction provides a way to formally model how decision makers such as judges
mediate among the influence of multiple, salient, and often conflicting goals
and do so in a way that results in reasonable behavior that is sensitive both to
the desires of the individual and the opportunities and constraints of the
environment (Read et al., 1997, p. 47).6 Judges may, for example, be able to
reach a particular desired outcome through two different analytic approaches
(equifinality), but one approach may be contrary to precedent (incompat-
ibility). A trial court judge may have at her disposal a number of case manage-
ment approaches (equifinality): one may be the most effective at speeding the
docket (best fit) but at the expense of party satisfaction (incompatibility),
while another approach may simultaneously move the docket and achieve a
good substantive result (multifinality). An appellate judge may weigh a desire
to write a detailed dissenting opinion against strategic objectives or against an
overwhelming workload (incompatibility).
Consider the following extended example as one illustration of how
parallel constraint satisfaction can be used to model the complexity of judicial
decision making.7 Imagine a judge who is considering a motion to suppress
key evidence in a criminal trial that raises a somewhat novel, but not unpre-
cedented issue related to the exclusionary rule. The judge is faced with the
related tasks of determining the content of her ruling and deciding whether to
rule from the bench or to issue a written ruling. The judge, having reviewed
the evidence, is aware of its strong probative value and (again, having seen the
evidence) is of the opinion that the defendant quite likely committed the
crime with which he is charged and ought to be punished accordingly. Given
the nature of the other evidence in the case, the judge understands that it is
unlikely that the prosecution will be able to proceed without the evidence.
The judge is also aware of the public sentiment surrounding the case and
predicts a strong negative public reaction if the case were to be dismissed. At
the same time, the judge is strongly committed to the principles underlying
the exclusionary rule and believes, as an empirical matter, that following the
rule and excluding evidence in appropriate cases has had and continues to
have positive effects on police procedure. The judge is strongly committed to
following legal precedent. The relevant legal precedent in the judges jurisdic-
tion clearly requires that she find the evidence inadmissible. On the other
hand, the prosecutor is urging her to follow an exception that another
jurisdiction has recently carved out that would support the admission of
the evidence. While the judge believes that following the approach suggested
by this exception is analytically stronger than the approach currently followed
in her jurisdiction and may apply to the facts of this case, she does not think
that such an approach will be accepted in her jurisdiction and predicts that
any ruling admitting the evidence stands a relatively high chance of being
overturned on appeal. The judge enjoys writing opinions and would welcome
the challenge of crafting an elegant analysis of an important legal issue; she
views the appellate court as a more attractive audience for these scholarly
efforts than she does the general public. She feels some need to explain her
reasoning (particularly to the appellate court if she admits the evidence and to
the public if she chooses not to admit it). However, as a busy trial judge, she
does not have the luxury of spending a week or a month crafting a nuanced
scholarly exposition of the issue.
Figure 2.1 diagrams these goals and constraints in a connectionist parallel
constraint satisfaction network. Solid lines represent compatible links; dashed
lines represent incompatible relationships. In the model, ruling the evidence
inadmissible would be consistent with the judges understanding of the
relevant precedent and with the general purposes of the exclusionary rule,
would minimize the chance of a reversal, and would move the docket (as the
charges would most likely be dropped). Conversely, such a ruling would be
inconsistent with the judges own best legal analysis, would be inconsistent
with the judges view of just deserts by resulting in the release of a probably
guilty defendant, and would inflame public sentiment. While ruling the
evidence admissible would be consistent with the judges notions of analytical
rigor, would allow the prosecution of the defendant, and would comport with
public sentiment and, therefore, build public confidence in the judicial
Goal: Follow precedent
Goal: Not be overturned
Goal: Further purposes of exclusionary rule

Not Admissible
Admissible
Goal: Allow prosecution of guilty defendant
Goal: Gain public respect
Goal: Move the docket
Write Goal: Make strong analytical argument Rule from

Opinion Bench
Goal: Explain reasoning to valued audience
Goal: Spend time writing interesting opinion
Figure 2.1 Parallel constraint satisfaction model of judicial decision.
system, the judge believes that it is contrary to the relevant precedent, opening
the judge up to the possibility of reversal, and would be inconsistent with the
aims of the exclusionary rule. Drafting an opinion on the ruling would further
the judges goals in being analytically rigorous and in spending time engaged
in the intellectual enterprise of thorough legal analysis, and might catch the
attention of the judges judicial colleagues. However, the judge is simulta-
neously aware of the ever-present pressure to move the docket and spending
time drafting such an opinion will not further this goal.
The judges decisions, then, involve the elaborate interplay of these
myriad goals, and subjectively, such decisions can feel rather mysterious.8
One struggles and struggles with a decision, and then all of a sudden an
internal threshold is crossed and the judgment is made. Though much of the
deliberation is in the form of conscious internal dialogue, the cognitive
process by which the various constraints are reconciled is largely unconscious,
because serial consciousness cannot represent the kind of parallel processing
required to reconcile all the conflicting positive and negative activations
among elements. We experience the struggle, then we feel something settle,
and we then begin a secondary process of trying to rationalize in words what
we have decided.
Considering judges as decision makers who must reconcile numerous
objectives in carrying out a variety of different decision tasks provides an
avenue toward a more nuanced view of the cognitive complexity of judicial
decision making and may lead to increasingly sophisticated hypotheses about
judicial behavior. Identifying the distinctive constraints faced by judges with
regard to particular decision tasks and settings can give rise to testable
predictions involving those constraints. For example, concern about being
overturned will be salient in some contexts and for some decisions, but not
others, and differences in the constraints faced may lead to predictable
differences in decision making across such contexts (e.g., differences in the
citation of precedent or in the scope of the decision). Alternately, one might
predict that alternate goals will have more influence on decision making when
legal precedent is unclear (i.e., less constraining). Or consideration of work-
load constraints might lead one to predict more intuitive processing by judges
facing greater workload pressures and more deliberation by those who are
relatively unconstrained by such pressures (see Guthrie, Rachlinski, &
Wistrich, 2007a). By guiding the generation of such predictions, conceiving
of judges as decision makers who attempt to simultaneously satisfy myriad
goals by engaging in a process of parallel constraint satisfaction offers a model
for incorporating the range of considerations that influence judicial decision
making and for understanding the interplay among them.
Notes
The authors would like to thank Matt Taksin for his helpful research assistance and
Margareth Etienne and Kevin Quinn for their helpful comments on parts of this
chapter.
1. The notion of ideal points has its origins in so-called spatial models of voting in
the political science literature on legislatures (see Downs, 1957; Enelow & Hinch,
1990; McCarty, Poole, & Rosenthal, 2006.
2. Martin and Quinn (2002) attempt to identify the dimensionality of Supreme
Court votes (to affirm or reverse) using a database of 3,450 cases from the 29
justices sitting on the Court during the 1953 through 1999 terms. They develop
an innovative Bayesian Monte Carlo algorithm to infer ideal points for each
justice. The details of the algorithm and its derivation are beyond the scope of
this chapter, but our concern is less with the method than with the available
data. In psychometric practice, the usual rule of thumb for fitting a latent
dimensional model is that one needs a minimum of 5 to 10 respondents per
measured item. The Martin-Quinn analysis reverses this inequality; with
between 41 (in 2003) and 108 (in 1972) decisions per term and only nine
seated Justices per term, there are far more items (41 to 108) than respondents
(9 for any given item).
It is difficult to correctly characterize an underlying multidimensional struc-
ture when the data are sparse (Fabrigar et al., 1999). For example, IQ tests are
carefully and painstakingly constructed from a very large set of items selected to
be homogeneous (high g loading), with thousands of respondents. Even so,
with factor rotation algorithms, or with a confirmatory factor analysis using
structural equation models, one can usually fit a multidimensional model to
these seemingly unidimensional datasets, and many psychologists believe there
are sound theoretical reasons to do so (see e.g., Cattell, 1963; Sternberg, 1999).
3. A new effort seems to proceed in the reverse direction, using Martin-Quinn

scores to validate the ideologies of media outlets (see Ho & Quinn, 2007).
4. This concern is not unique to judges; citizens also seem to expect the courts to
trade off efficiency and procedural thoroughness. MacCoun and Tyler (1988)
found that citizens strongly preferred juries to judges (and 12-person unanimous
juries to smaller or nonunanimous juries) for homicide cases, but preferred
bench trials or small, majority-rule juries for shoplifting cases.
5. This view is consistent with the story model of juror cognition, in which jurors
choose a verdict by constructing a story consistent with the trial evidence and
matching that story to the available verdict choices (see Pennington & Hastie,
1986, 1988, 1992, 1993).
6. We have focused here on the decisions of individual judges. However, parallel
constraint satisfaction network models can also be used to model the decisions
made by groups, such as panels of judges. See, for example, Thagard, 2000
(describing a model of consensus decision making).
7. For other examples of constraint satisfaction models, see Byrne (1995); Thagard
and Millgram, (1995).
8. For the sake of simplicity, we have focused on the links between goals and
actions, and have not depicted relationships between and among goals.
However, a full implementation of the model would include links showing the
ways in which goals facilitate or compete with one another. For example,
achieving the goal of following precedent facilitates achieving the goal of not
being overturned; a link between these two goals would represent such a facil-
itative relationship. These extra links would help illustrate why a simple linear
regression or cognitive averaging model is unlikely to accurately represent the
relative impact of each goal.
3
Top-Down and Bottom-Up Models of Judicial
Reasoning
Brandon L. Bartels
As is apparent from the other readings in this volume, the punchline of 50

years of behavioral research on Supreme Court decision making is that policy
(or ideological) preferences1 have a major, if not dominant, impact on
justices choices (Segal & Spaeth, 2002). Moreover, most perspectives
assume that policy preferencesas well as other ingredients of decision
makingexhibit generally uniform effects across all situations in which
justices make decisions and between justices as well. While research focusing
on the potent impact of policy preferences has increased our knowledge of
judicial decision making, the work brings up nearly as many questions as it
does answers. In other words, there is still a great deal we do not know about
how judges make decisions.
In this essay, I offer a perspective on how social psychological insights on
the cognitive processes of judgment can help enrich our understanding of
judicial decision making. By highlighting a cognitive perspective of judicial
reasoning, studies can move beyond black box models of decision making
that ignore the crucial cognitive processes mediating the relationship between
the judgmental considerations and the choices judges ultimately make. Such a
focus can fill in the gaps regarding what we do not know about judicial
decision making, namely, when ideology and legal considerations will exhibit
greater or lesser effects on judges choices. After reviewing some theories on
cognitive processes of judgment and decision making, I posit a theoretical
framework of judging focusing on top-down versus bottom-up reasoning
processes. In providing one possible explanation for judges reasoning pro-
cesses, the theory suggests hypotheses specifying the conditions under which
41
law and ideology will exhibit greater or lesser impacts on judges choices. The
theoretical perspectiveand this essay in generalis primarily aimed toward
explaining decision making by justices on the U.S. Supreme Court. However,
the ideas have implications for judges at other levels of the judiciary.
Cognitive Processes of Decision Making
Motivational and Behavioral Heterogeneity
As Lawrence Baums essay in this volume (ch. 1) highlights, political scientists

of various theoretical persuasions have posited theoretical frameworks spe-
cifying justices as motivated primarily by policy goals, that is, as single-
minded seekers of legal policy (George & Epstein, 1992). Yet it is possible
that under certain conditions justices may be motivated by policy goals,
under other conditions they may be motivated by legal goals, and under a
third set of conditions they may be motivated concurrently by both goalsa
desire to make both good law and good policy (Baum, 1997; Hausegger &
Baum, 1999). A focus on judgmental reasoning processes makes one think
more intently about what might be called motivational heterogeneity, or
the idea that under certain conditions justices might be motivated by some-
thing other than policy, or ideological, goals (Baum, 1994, 1997, 2006).
Attitudinal and strategic perspectives of Supreme Court decision making
have neglected such nuanced, multiple-goal frameworks and the broader
notion of motivational heterogeneity. But it seems reasonable to think that
context, case type, issue area, or other factors may determine which goals are
operative in a given case for a given justice.
Social cognition theorists place an explicit focus on motivational hetero-
geneity (e.g., Fazio, 1986, 1990; Petty & Cacioppo, 1986; Fazio & Towles-
Schwen, 1999; Kunda, 1990; Fiske & Taylor, 1991). Many social cognition
theorists are first interested in what types of motivations may be at play. Then,
ones motivation determines the nature of the cognitive process that produces
a decision or judgment. In short, if we want more nuanced and realistic
explanations of justices behavior, we should think more broadly about
justices goal structures, and the conditions under which certain goals
might become operative.
Behavioral heterogeneity is linked to motivational heterogeneity in that the
types of motivations that are operative determine the nature of ones cognitive
processes that will dictate the decision process, which in turn establishes the
extent to which particular considerations will influence decisions. In the judi-
cial context, Pritchett (1969, p. 42) alludes to a need to confront this particular
type of heterogeneity, arguing that [a]ny accurate analysis of judicial behavior
must have as a major purpose a full clarification of the unique limiting
conditions under which judicial policy making proceeds.
Top-Down and Bottom-Up Models of Judicial Reasoning 43
In the following section, I demonstrate how insights from theories of

social cognition can illuminate the study of Supreme Court decision making.
In particular, these theories are capable of explicating a more nuanced
portrait of decision making addressing the following issues: (1) the multiple
motivations that might be at play in the judicial context; (2) thinking about
the relationship between policy preferences and behavior as a process of
judgment as opposed to a stimulus-response relationship with an unex-
plained black box; and (3) specifying the conditions under which policy
preferences or legal considerations influence behavior with greater or lesser
force.
Top-Down and Bottom-Up Models of Reasoning and Judgment

Here, I describe and discuss two models of reasoning and judgmenttop-
down and bottom-up processesthat are prominent in various social cogni-
tion perspectives (Chaiken, 1980; Fazio & Towles-Schwen, 1999; Fiske &
Taylor, 1991; Petty & Cacioppo, 1986). These models lay the theoretical
foundation for my application of social cognition to the judicial domain.
For both processes, I assume that reasoning is systematic as opposed to
heuristic (Chaiken, 1980; Petty & Cacioppo, 1986). Systematic processing
occurs when individuals engage in active and effortful processing of relevant
stimuli and information in a decision context. On the other hand, heuristic
processing is low-effort, passive processing, where individuals may skim over
important stimuli and information and rely on more peripheral decision
cues. In the context of judging, it is reasonable to assume that judges engage
in systematic processing of the facts, briefs, oral arguments, and so forth,
when making decisions (though see Guthrie et al., 2001, 2002). While both
top-down and bottom-up processes involve systematic processing, the key
difference between the two models relates to the extent to which ideological
predispositions will bias the entire reasoning process. The two processes can
be treated as a continuum of biased processing, where top-down processing
represents the most biased reasoning process, and bottom-up processing
represents the most unbiased process.2
In a top-down reasoning process, the generic predispositions, percep-
tions, or theories people bring to a judgment context dictate how they process
the new information in front of them. Top-down processing is biased proces-
sing, and can be thought of as deductiveit is theory driven. In other words,
the theories and predispositions people bring to a judgment context produce
a biasing influence on how they process the relevant facts and information.
These predispositions, then, dominate the final judgment by providing a lens
through which the facts and evidence are evaluated and assessed. I use the
term theory to mean a set of beliefs, based on a directional predisposition,
that becomes an individuals story of how the world works or ought to
work. It is separate from the facts and evidence at hand. An example of top-
down, or theory-driven, processing involves a police investigation of a
murder case. The police may develop a theory early on about who the
murderer is, based on their prior knowledge about the particular type of
murder. For instance, if a child is killed inside the home, parents are typically
suspects. The police may develop a theory about how and why the parents
killed their child. In a top-down reasoning process, this theory would dom-
inate the investigation process, leading to a biased search for the truth. The
police will zone in on the parents, give less weight to alternative evidence
suggesting a different suspect, and exclude other suspects. They view all
evidence through the lens of the parent-centered theory they develop.
In contrast to top-down processing, bottom-up processing involves objec-
tive scrutiny of the information, facts, or evidence at hand. It involves objec-
tively assessing the relevant information and making a judgment based solely on
the facts, as opposed to predispositional biases. Bottom-up processing is usually
referred to as inductiveit is data driven. The theories or predispositions
people bring to the judgment context do not dominate the decision process.
Returning to the police investigation example, police would engage in bottom-
up reasoning if they are able suppress the biasing influence of a particular theory
that may develop during a murder investigation. They engage in an objective
search for the truth, considering all the relevant facts and evidence.
Fazios MODE model (1990; Fazio & Towles-Schwen, 1999), which
explains the processes by which and extent to which attitudes guide behavior,
sheds further light on top-down versus bottom-up processing. MODE stands
for Motivation and Opportunity as DEterminants, and these determinants
regulate whether people will enter into one of two attitude-behavior pro-
cesses: a deliberative process or a spontaneous process. The deliberative
process is a data-driven, bottom-up process, in which an individual closely
and systematically scrutinizes information, or the data, that is, the the
attributes of the behavioral alternative (Fazio & Towles-Schwen, 1999, p. 99;
see also Ajzen & Fishbein, 1980). Thus, it is an objective form of processing,
where attitudes may play a role in guiding behavior, but their influence is
diminished in the presence of the other attribute-based considerations.
A spontaneous attitude-behavior process is a top-down, theory-driven
process, where an automatically activated attitude is triggered, which then
biases how the individual processes the data and the attributes of the alter-
natives. In short, the attitude triggered by the immediate appraisal of the
decision context biases how one processes and perceives subsequent informa-
tion in an automatic, unconscious fashion. This strong attitude functions like
a theory, discussed above. The stronger the attitude, the more likely that
attitude will dominate the decision process at the expense of objectively
assessing the facts and evidence at hand.
Importantly, a mixed, controlled process may also occur whereby people
can overcome the potential biasing influences of even a relatively accessible
attitude when they [are] properly motivated; they can objectively process the
attributes of the alternatives instead of readily accepting the interpretation
implied by their attitudes (Fazio & Towles-Schwen, 1999, p. 102). Petty and
Wegeners (1993; see also Wegener & Petty, 1995) flexible correction model
is akin to this mixed model. Upon entering a judgment context, an attitude
may be automatically activated, but the activation of knowledge regarding
the normative requirements induces an individual to define the event as one
in which he or she needs to control and monitor impulsive behavior carefully
(Fazio & Towles-Schwen, 1999, p. 103). This controlled process means that
people will recognize their biases and, if motivated, will correct for those
biases, inducing one to engage in more data-driven processing, which will
attenuate the impact of the automatically activated attitude (Fazio &
Towles-Schwen, 1999, p. 103).
Conditions Under Which People Engage in Top-Down or

Bottom-Up Reasoning
Social cognition theorists have specified the conditions under which we might
expect people to engage in top-down or bottom-up processing, and the issue of
motivational heterogeneity is directly implicated. That is, the motivations, or
goals, of the actor determine which type of processing the decision maker will
engage in. First, when a fear of invalidity motivation is operative, people will
tend to process information more objectively, in a bottom-up fashion, and rely
less on their predispositions (Fazio & Towles-Schwen, 1999; Fiske & Taylor,
1991). Related to this motivation, when people feel accountable for their
decisions, they are more likely to be objective, bottom-up processors (Lerner
& Tetlock, 1999; Fiske & Taylor, 1991). The possibility of having to justify ones
decision to another person or group leads to more careful scrutinizing of the
attributes and information specific to the context, and less of a reliance on the
potentially biasing predisposition one brings to the case.
The motivated reasoning perspective also highlights conditions under
which people will engage in different reasoning processes. As Kunda (1990,
p. 480) states, People rely on cognitive processes and representations to arrive
at their desired conclusions, but motivation plays a role in determining which
of these will be used on a given occasion. Motivation is defined as any wish,
desire, or preference that concerns the outcome of a given reasoning task
(Kunda, 1990, p. 480). The motivations one possesses entering the decision
context bias the reasoning process toward achieving the end state specified in
that motivation. Kunda discusses two motivations in particular: accuracy and
directional goals. The core theoretical contention is that goals affect reasoning
by influencing the choice of beliefs and strategies applied to a given problem
(Kunda, 1990, p. 481). Braman and Nelson (2007) use the motivated reasoning
framework to explain when biases will occur in legal decision making.
A motivation to be accurate leads to a reasoning process akin to a bottom-
up, data-driven process. Accuracy goals lead to an elimination or reduction of
cognitive biases (Kunda, 1990, p. 481); they reduce top-down, biased processing
and induce objective, data-driven processing. Accountability, self-presentation
considerations, and fear of invalidity induce accuracy goals, which then
lead to more careful, objective processing of the information, evidence, and

data. Directional goals lead people to construct a justification of their desired
conclusion that would persuade a dispassionate observer . . . . In other words,
they maintain an illusion of objectivity (Kunda, 1990, pp. 482483). Unlike a
bottom-up reasoning process, directional goals bias memory search and belief
formation processes. One important constraint on the biasing role of direc-
tional goals is the presence of strong arguments. In the persuasion context,
bottom-up processing involves yielding to strong and influential arguments,
even if they promote a counterattitudinal position (Petty & Cacioppo, 1986;
Kunda, 1990).
According to the MODE model, motivation and opportunity determine
the type of attitude-behavior process one enters into. Fazio and Towles-
Schwen (1999) specifically discuss fear of invalidity as an important motiva-
tion that induces a deliberative, bottom-up attitude-behavior process.
Opportunity relates to the availability of time and resources. The more time
and resources one has, the more likely one will engage in a deliberative
process. The key variable that moderates the attitude-behavior relationship
is attitude accessibility, which is the strength of the association in memory
between an attitude object and its summary evaluation (Fazio et al., 1982;
Fazio & Willams, 1986). Accessibility ranges from nonattitudes, where there is
absolutely no association between an object and a summary evaluation, to
complete accessibility, where attitudes are automatically activated when one
encounters the object. According to Fazio, attitudes will guide behavior to the
extent that they are accessible in memory.
Schuette and Fazio (1995) report compelling experimental findings that
make several connections to the judicial context. Their findings support a
mixed, controlled process (see also Petty & Wegener, 1993; Wegener & Petty,
1995). They manipulate attitude accessibility and motivation (i.e., fear of
invalidity) and examine how each factor influences how subjects assess the
quality of a death penalty study. As expected, the low motivation (i.e., no fear of
invalidity), high accessibility subjects were more likely to assess the studies in
accord with their attitudes, evidence of attitudinally biased, top-down proces-
sing. Increases in accessibility enhanced this biasing effect. However, increasing
fear of invalidity reduced this biasing effect, inducing more bottom-up rea-
soning, even for those with highly accessible attitudes. Importantly, the findings
suggest that individuals are capable of controlling the potentially biasing role of
attitudes and predispositions when they are properly motivated.
Processes of Supreme Court Decision Making

Adopting insights from the work discussed above, I posit a cognitive model of
judging specifying the reasoning processestop-down and bottom-up
processesby which Supreme Court justices make decisions. The model
posits conditions under which justices will engage in either type of process.
It also produces empirical implications suggesting when policy preferences

and legal considerations will have greater or lesser impacts on justices choices.
Attitudes, Policy Preferences, and Ideological Values

Before launching into the theoretical framework, I discuss a conceptual issue
regarding the concepts attitudes, policy preferences, values, and predispositions.
Thus far, I have used these terms somewhat interchangeably, which follows the
tradition in judicial behavior scholarship (e.g., Segal & Cover, 1989; Gibson,
1991; Segal & Spaeth, 2002). In general, all relate to a justices ideological
predispositions toward legal policy issues, and I will treat themparticularly
attitudes and policy preferencesas synonymous. Many scholars tend to use
these terms without providing explicit definitions. Referring to Eagly and
Chaikens (1993) definition of an attitude, I will define justices policy prefer-
ences/attitudes as evaluative tendenciesin terms of favoring or disfavoring
toward legal policy. Note that the attitude object is legal policy. Using Fazios
(1995) definition, we could refer to justices policy preferences/attitudes as
associations in memory between legal policy and evaluative orientations.
Judicial scholars tend to think of justices policy preferences as global as
opposed to issue-specific. That is, policy preferences are thought of as more
global views toward broad legal policy areas, like civil liberties and economics
(e.g., Segal & Spaeth, 2002; Segal & Cover, 1989). In this sense, ideological
values may come closer to how scholars conceive of policy preferences both
conceptually and operationally (see Segal & Cover, 1989). Values can be
thought of as attitudes toward relatively abstract goals (Eagly & Chaiken,
1993, p. 270). The difference between attitudes and values, then, is in the
specificity of the object toward which the evaluation is directed. Justices
global policy preferences resemble abstract formulations regarding their
ideological tendencies toward broader issue areas, like civil liberties. Those
with more liberal values on the civil liberties issue area favor the protection of
individual liberties and rights, while conservatives favor the governments
capacity to impose certain restrictions on those rights. For this essay, I will not
distinguish between values and attitudes in terms of the attitude objects
degree of specificity. Following in the tradition of judicial behavior scholar-
ship (e.g., Segal & Spaeth, 2002), I will refer to policy preferences in more
global termsas an ideological orientation toward a broad legal policy area.
Top-Down and Bottom-Up Processes of Judicial Decision Making

The top-down and bottom-up reasoning processes I have discussed above
distinguish between theory and data in a judgment process. The theory a
justice brings to a decision setting can be thought of as a set of beliefs on a
given legal issue, rooted in ones general ideological predispositions. These
predispositions, or policy preferences, have the potential to bias the reasoning
process. The data in a decision setting are the facts of the case, past
precedent(s) and legal doctrine, the arguments in the briefs, oral arguments,
and other legal considerations.
The top-down model is a theory-driven reasoning process whereby the
policy predispositions a justice brings to a decision context determine how
the justice will both process the information at hand and make the judgment.
In this process, predispositions, in the form of policy preferences, dictate how
the data will be processed. For a top-down processor, ones predisposition
provide a lens through which one views the data, therefore biasing the
reasoning process. Instead of letting the data guide the decision-maker, the
decision-maker finds the data that best supports his or her desired conclusion
(a la Segal & Spaeth, 2002). Akin to the MODE models spontaneous attitude-
behavior process, ideological predispositions condition the entire reasoning
process by determining how one appraises the cases, how one processes
relevant information, and ultimately, how one makes a decision.
The bottom-up model is a data-driven reasoning process whereby the
evidence, information, facts, and legal considerations objectively guide the
decision maker. Bottom-up processing is objective, unbiased processing of
the information and facts, untainted by the ideological predispositions one
may possess about the attributes in the decision context. Accuracy, fear of
invalidity, and accountability motivations drive one to engage in bottom-up
reasoning, to let the data determine how the decision is made, as opposed to
finding the evidence that best supports an ideological predisposition.
Therefore, given sufficient motivation, the impact of even a relatively acces-
sible attitude will be attenuated throughout the reasoning process, with the
justice instead focused on the facts and legal considerations.
Before moving on, a few caveats are in order. First, legal scholars and
political scientists recognize that facts, legal rules, and precedent are never
completely self-evident. The discovery of these factors can often involve sub-
jective choices based on differences of interpretation. Braman and Nelson
(2007), for example, report how the ascertainment of case similaritychoosing
which precedent most closely resembles the current casein legal reasoning
can be biased by policy preferences. Thus, rarely would we ever witness a judge
engaging in pure bottom-up reasoning. This leads directly to the second caveat.
I do not mean to depict judicial reasoning processes as either strictly top-down
or strictly bottom-up. Instead, I view these two processes as endpoints of a
reasoning continuum, with various hybrid processes falling in between. Social
cognition perspectives, and my own perspective, tend to focus on factors that
serve to reduce the amount of top-down processing or increase the amount of
bottom-up processing. These perspectives recognize that under certain condi-
tions, bias may not be completely eliminated, but instead reduced, with ones
predispositions and the data guiding the decision process. I recognize this
nuance and do not necessarily examine the conditions under which the effects
of policy preferences in the decision making process are completely eliminated,
which would suggest a purely bottom-up process, or are completely determi-
nant, which would suggest a purely top-down process.
Thus, while a pure bottom-up process would suggest that facts and legal
doctrine are essentially self-evident, recall that it represents an endpoint on a
continuum. Processes that move away from that endpoint can more realisti-
cally capture the subjectivity of interpretation that is inherent in legal rea-
soning. It is helpful, however, to depict and describe the full range of variation
in order to proffer a comprehensive explanation of judicial reasoning.
Sequence of Justices Decision Processes
To understand justices reasoning processes, it is instructive to describe first

the sequence of processes justices go through when they are confronted with a
case. The sequence, depicted in Figure 3.1, begins at the merits stage. Upon
Justice exposed to case
Policy preference is activated

to some degree
Determine where alternatives lie in

Familiarize self with facts of the case
policy space
Motivated to control inclination to allow

preferences to dominate?
Study the data: facts, precedents,

briefs, oral arguments
Collegial interaction
Final vote on the merits
Figure 3.1 Sequence of justices processes of judgment.

exposure to a case, a justices policy preference is activated to a certain extent.

What I refer to as degree of activation is akin to Fazios conception of
attitude accessibility. I do not necessarily assume that a justices policy pre-
ference is automatically activated upon confronting a case. In a death penalty
case, for instance, we might assume that a justices policy preferences toward
this issue are highly accessible, and therefore, automatically activated.
However, in a case that involves a new, emerging issue area on which the
Court has not frequently decided, we might expect less accessible policy
preferences. For these latter cases, a policy preference is not strongly activated
because the justice has not had much experience with the issue. Since judges
are legal experts, though, and have seen a wide variety of cases, they can
probably access a policy preference on just about every case that comes before
them But this does not mean that the degree of preference activation is
uniform across all cases.
After the policy preference is activated to a certain degree, a justice
becomes familiar with the facts of the case. This contributes to the determina-
tion of where the alternatives (i.e., potential outcomes) lie in an ideological
issue space. Note that this stage is at the heart of the attitudinal model (Segal
& Spaeth, 1993, p. 65), which states that justices decide disputes in light of
the facts of the case vis-a-vis [their] ideological attitudes and values. Modern
conceptions of the attitudinal model are akin to a proximity spatial model,
where a justice possesses an ideal point in an issue space, determines where
the two alternatives are in the issue space, and votes for the alternative closest
to his or her ideal point. My perspective thus far subscribes to the process by
which justices come to realize the location of their policy preference relative
to the alternatives; they attain this information by consuming the case facts,
relevant precedents, and the arguments made by the parties.
In a cognitive account, the next stage in justices decision processes is
crucial. What justices do at this stage depends on the extent to which they are
motivated to control an inclination to allow their personal policy preferences
to dominate and bias the remainder of their decision processes. This stage
may occur either consciously or unconsciously.3 Fear of invalidity, account-
ability, or accuracy goals help motivate justices to control these predisposi-
tional biases, whatever their strength. Note the similarities to the mixed,
controlled processes discussed earlier (Petty & Wegener, 1993; Schuette &
Fazio, 1995; Wegener & Petty, 1995). This is both an original and controver-
sial way of thinking about motivations in the judicial context. To clarify,
when I talk about motivations, I am referring to the goals that push a justice to
reason in one way or another. Recall that typical treatments of Supreme Court
decision making have posited a one-goal framework assuming at the outset
that justices are single-minded seekers of legal policy. Adopting Baums
(1994, 1997, 2006) multiple goals framework, a cognitive approach can more
easily incorporate the idea of motivational heterogeneity. The motivation a
justice possesses under a particular condition will determine the relative
impact of predispositional biases versus data. In Supreme Court decision
making, this has implications for explaining the relative impact of policy
preferences versus legal considerations.
Following this stage, justices must go through the legal reasoning process:
study the facts, read the parties briefs and amicus curiae briefs, examine past
precedents and legal doctrines, and engage in oral arguments. During this
data processing stage the justices sort through the information and attri-
butes associated with the case. The degree to which a justice is motivated to
control predispositional biases will determine how the justice processes the
data. It is at this stage that theory and data collidea justice begins reasoning
either in a more top-down, theory-driven process or a bottom-up, data-
driven process. In a top-down process, a justice assesses the data through
the biasing lens of his or her policy preferences. In a bottom-up process, a
justice suppresses this bias and assesses the data through a more objective
lens. In a hybrid process, a justice processes information via a mixed
processa weighted combination of top-down versus bottom-up processing,
where the weights are determined by the operative motivation(s).4
Ideal Types of Justices Decision Processes
To summarize, I posit that there are two key stages in the sequence of a
justices decision process where there is variation crucial to explaining the
relative influence of policy preferences and legal considerations on justices
decisions. First, upon exposure to a particular case, a preference will be
activated to a certain degree, such that less than complete preference acces-
sibility will set the stage for a process where the biasing impact of policy
preferences in the decision process will be attenuated and the impact of
objective considerations will be elevated. At the second stage in the process,
a justices motivation to control bias is central. As this motivation increases,
bottom-up processing is more likely to occur, and the impact of policy
preferences in the decision process will be attenuated. Below, I construct
ideal types of justices decision processes based on combinations of these
two key factors. These ideal types provide predictions about the type of
reasoning process under various conditions, and predictions about the effects
of policy preferences and legal considerations.
The four ideal types summarized in Table 3.1 consist of all combinations
of whether or not a justices preference accessibility is high or low and
whether or not a justice is motivated to control an inclination to act solely
on the basis of policy preferences. As ideal types, these four models focus on
high and low values of both accessibility and motivation to control bias in
order to explore the full theoretical spectrum of processes of behavior. I begin
by discussing the two most extreme ideal types. The first ideal type encom-
passes a situation where a justices policy preference is highly accessible, and
moreover, the justice is not motivated to control biasperhaps the justice
does not feel accountable to another entity, possesses no fear of invalidity, and
is driven primarily by ideological goals. This motivational type produces a
Table 3.1 Ideal Types of Justices Decision Processes
Predictions
Motivation to Control Preference-

Preference Predispositional Type of Behavior Impact of Legal
Accessibility Biases? Process Relationship Considerations
High No Top- Intensified Minimal

Down
Low Yes Bottom- Significantly Significantly
Up Attenuated Large
High Yes Mixed, Attenuated Moderate
Controlled
Low No Mixed Attenuated Moderate
strongly top-down reasoning process. It is akin to Fazios spontaneous atti-

tude-behavior process, where an attitude is highly accessible and hence
dominates the information processing stage and serves as the predominant
influence on reasoning and choice. This top-down ideal type of justices
judgment processes can be considered a cognitive analogue to the contem-
porary attitudinal model, where a justice possesses fixed preferences over
policy issues and is uninhibited by legal, political, and normative constraints,
leaving unbridled discretion to decide cases in an ideological fashion. The
predictions that flow from this ideal type, then, are: (1) the preference-
behavior relationship will be very potent, and (2) the impact of legal con-
siderations will be minimal.
The second ideal type produces a polar opposite reasoning process from
the first type. In this situation, a justices policy preference is not highly
accessible. Low accessibility means that a justices preference will not dom-
inate the decision process to the extent that it will in the spontaneous process.
Moreover, the justice is motivated to control an inclination to act in a biased
fashion. The situation in front of the justice induces a fear of invalidity,
accountability, or accuracy motivation, whereby the justice suppresses pre-
dispositional biases and instead, processes the attributes of the case, legal
doctrine, and other relevant information in an objective manner. This process
strongly resembles a bottom-up reasoning process. The following predictions
emerge from this ideal type: (1) the preference-behavior relationship will be
significantly attenuated, and (2) the impact of legal considerations will be
significantly large.
The next two ideal types represent reasoning processes somewhere in
between top-down and bottom-up processes. First is a situation where a
justices policy preference is highly accessible upon exposure to a case.
Thus, the justices reasoning process is capable of being biased and dominated
by the justices policy preference. However, the justice simultaneously pos-
sesses a motivation to control such bias. As mentioned above, the decision
context could induce a fear of invalidity, accountability, or accuracy motiva-

tion that competes with a justices highly accessible policy preference. This
motivation causes a justice to suppress the inclination to engage in a pure top-
down process, and instead to focus on the legal considerations and the
particular attributes of the case in an optimally objective manner. This
mixed, or controlled, process falls somewhere in between a top-down and
bottom-up reasoning process since it is possible for a top-down process to
take over, but a motivation to suppress this top-down inclination also exists,
which increases the prospects for bottom-up type behavior. In the psycholo-
gical literature, this process resembles both a controlled, mixed attitude-
behavior process (Fazio & Towles-Schwen, 1999) and a flexible correction
model (Petty & Wegener, 1993; Wegener & Petty, 1995). Recall that in both
models, an individual recognizes his or her bias, is motivated to control that
bias, and implements a mechanism that corrects for the bias. From this
mixed, controlled process of a justices reasoning process, a prediction
emerges that the relationship between policy preferences and behavior will
be attenuated. Also, the impact of legal considerations of various objective
criteria will be accentuated.
The final ideal type is the case where a justices policy preference is not
highly accessible, but he or she also possesses no motivation to control the
inclination for biased reasoning. This combination is probably the least likely
to occur in reality, given the odd combination of low accessibility and a
high likelihood for biased processing. Nevertheless, it is a mixed reasoning
process since the justices policy preference is something less than completely
accessible, which suggests that a pure top-down reasoning process will not
take hold. Thus, the justice is likely to focus more on the attributes of the case,
including the legal aspects, in the absence of a strong policy predisposition.
However, the justice is also not motivated to control an inclination to behave
in a top-down manner, even though the capacity to do so is suppressed due to
the low accessibility of preferences. Thus, the low accessibility pushes the
justice to be more bottom-up, but the lack of a motivation to control bias
pushes the justice to reason via a top-down process. The predictions flowing
from this ideal type are similar to the previously discussed type: the prefer-
ence-behavior relationship will be attenuated, and the impact of legal con-
siderations will be elevated.
Conclusion
In this essay, I have presented a broad cognitive perspective of judging. In this
section, I touch on some more tangible issues and obstacles, particularly with
respect to testing some of the empirical implications of the model. First, what
factors are associated with preference accessibility and motivation to control
predispositional biases? In other words, what situations or conditions would
induce the accessibility of policy preferences and what situations would
induce accountability, fear of invalidity, or accuracy goals? In other work

(Bartels, 2005, 2006), I suggest that certain case-level, situational factors are
capable of shaping the degree of preference accessibility (or preference
strength) and accountability. In particular, I posit that increases in issue
salience and issue familiarity and decreases in case complexity activate
strong policy preferences among the justices, leading to the possibility of
more top-down processing and a stronger preference-behavior relationship.
Pertaining to accountability, I posit that the interest group environment,
participation by the solicitor general, and whether the case involves a statu-
tory or constitutional question will trigger varying levels of accountability.
Higher levels of accountability among the justices lead to a greater possibility
of bottom-up processing and a weaker preference-behavior relationship.
Another factor associated with preference accessibility relates to so-called
freshmen effects (e.g., Hagle, 1993). Do new justices have less accessible
preferences than veteran justices? Another factor associated with account-
ability includes the ideological configuration of Congress and the president
(Bartels, 2006), which would relate to debates about whether and how the
separation-of-powers structure constrains the justices (Epstein & Knight,
1998; Segal, 1997; Segal & Spaeth, 2002; Bergara et al., 2003).
The second empirically oriented issue pertains to the issue of observa-
tional, or behavioral, equivalence. In particular, one might find support for
the empirical implications of a cognitive perspective of judging, but one could
argue that these empirical implications are also consistent with implications
emerging from a rational choice, attitudinal, or some type of legal perspective.
If observational equivalence is an issue, as it almost certainly would be in this
context, empirical evidence in favor of the cognitive models empirical impli-
cations would not necessarily indicate support for the cognitive model
itselfthat is, the processes underlying the modelbecause this evidence
might also support, for example, a rational choice approach. Since the pre-
dictions would not be unique to the cognitive model, it would be necessary to
test empirically the processes underlying the cognitive model via experi-
mental methods.
This last point transitions to the third and final empirical issue I
discusshow to test the cognitive processes of judging. If legal researchers
had their way, we would recruit judges as experimental subjects, design an
experiment that manipulates some factors (e.g., preference accessibility and
motivation to control bias), randomly assign the judges to experimental
conditions, and test for causal processes underlying judging. Guthrie et al.s
(2001, 2002, 2007a) innovative experiments on judges provide a valuable
template for conducting such work. My theoretical framework requires an
experimental approach that would allow researchers to get inside the heads
of judges to explain how they reason about cases. Recruiting currently serving
judges as experimental subjects for this type of study might be improbable,
given judges sensitivity to rendering judgments on hypothetical cases. This
presents an obstacle for testing the cognitive processes of judging, but the
obstacles are not insurmountable, as Guthrie et al show. Moreover, Braman

and Nelson (2007) used law students as experimental subjects to explore
biases in legal reasoning. To test a theoretical framework like the one I have
proposed, one possibility is to recruit retired judges as experimental subjects
and perform a survey experiment.5 Retired judges may be more likely to
respond to such a survey experiment with considerable candor, given they no
longer have an active stake in the judiciary. Obviously, the pool of retired
Supreme Court justices is extremely limited, but recruiting retired U.S.
Courts of Appeals or District Court judges is a possibility.
In conclusion, the cognitive perspective presented in this essay has the
potential for providing a more realistic, nuanced explanation of judging. By
focusing on the cognitive processes inside the judicial mind, the cognitive
approach has the ability to improve on existing models of judicial decision
making that treat judicial reasoning processes as a black box. While I have
aimed the essay toward justices of the U.S. Supreme Court, the theoretical
framework I have presented has implications for judging at other levels in the
judicial hierarchy.
Notes
I am grateful to Lawrence Baum, Eileen Braman, David Klein, and Howard Lavine
for extremely helpful feedback and suggestions on this essay.
1. I use the terms policy preferences, ideology, and attitudes toward legal
policy interchangeably.
2. Posner (1992) has distinguished between top-down and bottom-up legal rea-
soning processes. While there are some general similarities between his approach
and mine (e.g., emphasis on theory-driven reasoning), some key differences
exist. Namely, his conceptual framework is not psychologically oriented. My
focus is on the extent to which, and conditions under which, ideological predis-
positions bias legal reasoning processes.
3. The issue of whether this is a conscious or unconscious process would require a
more in-depth discussion, which is beyond this essays central scope of inquiry.
What is most important in this discussion is whether and to what degree justices
are motivated to control their biases. It is quite conceivable that such a process
could be either conscious or unconscious, a question I leave for future research.
4. I exclude from this discussion any collegial interaction that occurs in the opinion
drafting stage (e.g., Maltzman, Spriggs, & Wahlbeck, 2000). Incorporating small-
group dynamics (see Martineks essay in this volume) would offer a compelling
addition to the framework I have laid out.
5. I thank Pete Rowland for suggesting this idea in a conversation.
4
Persuasion in the Decision Making of U.S.
Supreme Court Justices
The purpose of this paper is to provide some thoughts about the role of
persuasion in judicial decision making. Herein, persuasion is treated broadly
as the process of influence, which covers not only overt attempts at influence
(by the media, by the advocates, and by other judges) but also the judges
responses to attempts at influence. Emphasis is on the question: What
influences how a judge forms an initial opinion about a case? The coverage
and examples refer to the United States Supreme Court, but the ideas are
applicable to other appellate panels and in some respects to trial judges.
Basic Assumptions
This paper assumes that justices are more susceptible to persuasion
depending on the nature of the case; some cases deal with matters to which
the justices have given much prior thought. Certain cases may reflect issues on
which they have developed strong views long in their past. Research on
persuasion certainly indicates that the nature of the material affects the will-
ingness to be responsive to an attempt to persuade. But individual differences
may exist in justices readiness to be persuaded on all issues; some may possess
personality characteristics that inhibit change. Thus, this paper reflects two
assumptions:
1. Persuasion operates differently depending on the type of case.
2. Persuasion operates differently depending on the particular justice.
57
The body of the paper elaborates on these assumptions and presents some
data and examples to evaluate the papers claims.
Testing the First Assumption: Ideological Versus

Nonideological Cases
Cases that are granted certiorari by the Supreme Court can be classified in
many different ways: by their topic, by whether they reflect ordinary litigation
or political litigation (Baum, 2007), by the presence or absence of the govern-
ment as a party, by whether they come from the liberal 9th Circuit Court or
the conservative 4th and 5th Circuits, and so on. This paper hypothesizes
that persuasion, and thus the decision-making process, operates differently
in ideological cases than in nonideological cases. By ideological cases,
I mean cases whose content is related to an attitude or value held by a justice.
Hot-button issues lead to ideological cases. In general, agreement exists on
just where positions on salient ideological issues can be placed along a liberal-
versus-conservative political dimension. For example, in criminal cases,
liberals are relatively sympathetic toward criminal defendants and their
procedural rights, while conservatives give more emphasis to the effectiveness
of the criminal justice system in fighting crime (Baum, 1990, Table 13,
p. 13). This paper employs Lawrence Baums distinctions between liberal and
conservative positions on judicial issues as a convenient summary.
Nonideological cases are those in which the central issues do not tap
into deeply felt values of the justices. In the studies to be described subse-
quently I asked two raters to independently classify cases as nonideological if
their content did not deal with the various topics that distinguished between
liberals and conservatives, using Baums detailed distinctions. Thus, non-
ideological cases are less easily classified by content, but disputes between
two states, patent disputes, and conflicts between two businesses usually are
nonideological cases. The case of Wachovia Bank v. Schmidt (2006) dealt
with the question of whether a bank with many branches was a person in
every state where it had a branch or only in its headquarters state. The case
was decided unanimously and did not seem to provoke any ideological
biases. Another nonideological example is the case of Kansas v. Colorado,
to which the Supreme Court granted cert in 2001. The Arkansas River begins
in the Rockies, flows across Kansas, part of Oklahoma, and across Arkansas
before emptying into the Mississippi. A dispute between Colorado and
Kansas over diversion of river water had seemingly been settled by the
passage of the Arkansas River Compact by Congress in 1949. But for a
number of years Kansas had complained that Colorado had violated the
compact. After the Court had so ruled in 1995, it remanded the case to a
Special Master. But Kansas claimed that Colorado continued to violate the
Persuasion in the Decision Making of U.S. Supreme Court Justices 59
rules. In 2001 the Court again sided with Kansas, including a judgment that
Colorado should pay interest on the damages accrued. While the question of
an interpretation of the Eleventh Amendment was part of the issue at hand,
the Supreme Court ruled unanimously in favor of Kansas in 2001, and the
basis appears to be simply what the law says. Thus the case appears to be free
of ideological triggers.
How does persuasion operate differently in ideological cases than in
nonideological cases? First, it is proposed that justices are aware of ideological
cases earlier than they are about nonideological cases. Well before their
conference to decide which cases to grant cert, even well before a petition is
submitted, the typical ideological case has generated publicity. A state passes a
law that critically restricts abortions, a universitys admissions program to
increase diversity is challenged in a lower court, the Bush administration
places prisoners of war at Guantanamo Bay and denies them rights to a trial
such decisions draw attention from the media, and the justices read the
newspapers and watch television news as many members of the citizenry
do. In contrast, for many nonideological cases, the first awareness may come
when the justice reads the recommendations from a law clerk who has
processed the case as part of the cert pool.
Second, at the point of granting cert, justices know more about the issues
in ideological cases than nonideological ones. Issues of search and seizure, for
example, are frequently before the Court; in the October 2003 Term the Court
dealt with ten cases involving claims of a violation of Fourth Amendment
rights. More often, nonideological cases deal with a relatively obscure federal
law, so that during the process leading up to the decision to grant cert, most
justices have to do more review. For example, a case during the October 2003
Term (BedRoc Limited v. United States) dealt with the terminology in the
Pittman Underground Water Act of 1919, a legislative decision probably not
on the forefront of each justices awareness prior to preparing for the cert
conference.
Justices Values and Ideological Cases
But the most important difference is what makes the case ideological
whether its issues generate a value-based predisposition. Many years ago
psychologist Fred Kerlinger (1967) proposed that in conceptualizing attitudes
and values, it was useful not to think of each as a bipolar continuum, but
rather to focus on what he called criterial referents. Certain objects, topics,
or issues may serve as anchors, or criterial referents, which define ones values.
For each justice, different issues may serve as criterial referents; for some,
abortion; for some, the death penalty; and for some, racial or gender dis-
crimination. These serve as triggers to at least a preliminary leaning in one
direction. Sometimes it is more than a leaning; it is an irrevocable response.
For example, toward the end of Justice Harry Blackmuns service on the
Court, his position on any case coming before the Court that dealt with the
death penalty was clear. In 1994, his dissent in Callins v. Collins states:
From this day forward, I no longer shall tinker with the machinery of
death. . . . It is virtually self evident to me now that no combination
of procedural rules or substantive regulations can ever save the death
penalty from its inherent constitutional deficiencies. . . . The
problem is that the inevitability of factual, legal, and moral error
gives us a system that we know must wrongly kill some defendants, a
system that fails to deliver the fair. . . and reliable sentences of death
required by the Constitution. (p. 1145)
More recently, Justice Antonin Scalia has made it clear that on certain
matters before the Court, he had made his mind up before the process of
considering the issues had been completed. On one occasionthe issue was
the constitutionality of the words under God included in the Pledge of
Allegiancehe recused himself because of public statements he had made.
But in another notable case, that of Hamdan v. Rumsfeld, he did not. This
case required the Court to decide whether an enemy combatant who was
detained at Guantanamo Bay was protected by the articles of the 1949 Geneva
Conventions as a prisoner of war. In a speech a few weeks before this cases
oral arguments, Scalia ridiculed the suggestion that detainees captured on
the battlefield should receive a trial in civil courts; that proposition, he said,
was a crazy idea. He interrupted a subsequent question by claiming: If he
was captured by my army on a battlefield, that is where he belongs. I have a
son on that battlefield and they were shooting at my son. And I am not about
to give this man who was captured in a war a full jury trial. I mean its crazy
(Isikoff, 2006, p. 6).
Justice Scalias comments drew strong criticism and calls for recusal from
several sources (Isikoff, 2006). But Justice Scalia was not deterred; he participated
actively at the oral arguments and voted in the case. In fairness to Justice Scalia, he
is certainly not the only justice and this was not the only case in which a justices
eventual vote was fixed in concrete before the oral arguments; the matter came to
light because of his provocative style and willingness to express his opinion in
public. For him, if not every justice, this was an ideological case, and his actions
support the argument that in such cases, opinions are formed early and not
susceptible to persuasion. Thus, even though the focus of this paper is on
persuasion, it questions how much opportunity exists to persuade justices on
certain cases; their minds may be made up very quickly.
Ideological Cases and Automatic Responses

Thus, for some justices on some cases, it is proposed that their reaction is
emotional and their response is instantaneous. (For a sharp contrast in very
well formed values between two justices, read Justice Scalias opinion for
the Court and Justice Brennans dissent in the case of Michael H. v. Gerald
D., 1989.) Social psychologists have concluded that the evaluations of

stimulievents, persons, issuesare often automatic; that is, they are so
primed by the stimulus that they are given without further processing
(Bargh & Ferguson, 2000; Duckworth, Bargh, Garcia, & Chaiken, 2002).
There is even some evidence that we may have two systems
for evaluating various aspects of the social world; one that operates in
an automatic manner and the other that operates in a systematic and
controlled manner. . . . In fact, studies conducted from the perspective
of social neuroscience indicate that these differences exist. Certain
parts of the brain, especially the amygdala, may be involved in
automatic evaluative reactions, simple good-bad judgments that
occur in a rapid and nonconscious manner. In contrast, portions of
the prefrontal cortex (especially the medial prefrontal cortex and
ventrolateral prefrontal cortex) may play a key role in more controlled
executive reactionsthe kinds about which we think carefully and
consciously. (Baron, Byrne, & Branscomb, 2006, p. 54)
Are some values so entrenched and so relevant to case decisions that they
produce automatic reactions that do not receive review? If Justice Scalia
(rather than the ubiquitous freshman psychology student) were a respondent
in the kind of experiment used by psychologists to demonstrate automatic
responses, how would he respond if abortion were flashed on the screen?
We cannot say, but the emphasis on automaticity of responses is provocative.
Attitudinal and Legal Models of Judicial Decision Making

The distinction between ideological and nonideological cases may help to
understand the long-standing conflict between two models of judicial deci-
sion making. Both the legal model and the attitudinal model have been
extensively considered and debated in the political-science literature. (Other
chapters in this book describe shifts in the attitudinal model toward a more
strategic orientation.) Indirect support for the attitudinal model has also been
demonstrated in several extensive analyses of the voting records of federal
district court and circuit court judges; for example, Rowland and Carp (1996)
showed how district judges appointed by Republican presidents decided cases
on civil liberties and civil rights, for example, in a different direction from
similar cases decided by judges appointed by Democrats. More recently,
Sunstein, Schkade, Ellman, and Sawicki (2006) presented data on how the
composition of judges on federal circuit-court panels (specifically the poli-
tical party of the president who appointed them) affected their votes on what
we would call ideological issues.
Each of these models (especially the attitudinal model) has its adherents
(see, for example, Segal & Spaeth, 1993, 2002), but the possibility that each
may apply in certain types of cases is less often emphasized. Another
possibility is that in a particular case, the attitudinal model may describe the
reaction of some justices, while others may struggle to maintain a basis in the
legal model.
The wealth of empirical data used by adherents of the attitudinal model
(consistency in votes, correlation of votes of individual justices with their
ideological pronouncements in speeches, books, or articles) to support their
claim may apply in ideological cases, but in those cases where no personal
values are salient, the justices are more likely to examine the statutes, the
precedents, and, if applicable, the Constitution in a dispassionate manner.
(See, for example, the recent article by Lindquist and Klein, 2006, that revives
support for the role played by legal reasoning in Supreme Court decisions.)
Predictions
In summary, it is proposed that, compared to their response to nonideolo-
gical cases, individual justices in ideological cases are less responsive to the
persuasion expressed in merit briefs and oral arguments. Specifically, it is
hypothesized that they:
1. Form initial opinions more quickly.
2. Are less likely to change their opinions.
3. Have formed opinions by the time of oral arguments, so that their
questioning reflects their already-formed opinions.
4. Are less responsive to outside influences.
5. Are more predictable with regard to their final votes.
Furthermore, this paper suggests that the degree to which a case is ideolo-
gical can be reflected in several qualities of the dispensation of the case.
Specifically, in ideological cases, compared to nonideological cases:
1. The final vote will less often be a unanimous vote, because the
current composition of the Court is divided ideologically.
2. It will take longer for the Court to announce the decision in the case,
because the minority will more likely write dissents and the opinion
of the Court will go through more drafts.
As part of an ongoing project (see Wrightsman, 2008), data have been
collected from several recent terms of the Supreme Court that sometimes
directly and sometimes indirectly respond to these hypotheses.
For four recent terms each case has been classified as ideological, border-
line, or nonideological. The following variables have been determined:
1. In the oral arguments, the length of time (measured in words) that
the petitioner spoke and the respondent spoke before they were
interrupted by a justice.
2. The number of questions asked each advocate by the justices during
the oral arguments.
3. Whether or not a representative of the Office of the Solicitor

General participated in the oral arguments, and if so, on which side.
4. The length of time (measured in days) from the oral argument to
the announcement of the decision.
5. The announced vote.
6. For the October 2002 Term, the predicted votes and case outcomes
as determined by the Supreme Court Forecasting Project at the
School of Law at Washington University in St. Louis. (Although this
project calculated statistical predictions of the outcomes in some
cases in the next term, the October 2003 Term, the data analysis was
discontinued in the middle of this second term.)
Empirical Tests of the Predictions

Data from the Supreme Court website were used to test the above predictions.
For example, it was hypothesized that in ideological cases, justices form
opinions quickly, they are less likely to change their opinions, and their
questioning during oral arguments reflects their already-formed opinions.
How can these hypotheses be tested empirically?
Over the last three years, two findings have been published that have
concluded that during oral arguments, justices as a group ask more questions
to the advocates representing the eventual losing side than they do to the
advocates whose side eventually wins. The first was a study done by Sarah
Shullman (2004), who observed ten oral arguments during the October 2002
Term. She reported, All nine justices seemed to ask fewer questions of the
party in whose favor they would ultimately decide (2004, p. 278). But there
are numerous methodological problems in her study. With only one observer,
there is no test of inter-rater reliability. The task of doing these ratings on the
spur of the moment seems very challenging; a typical 30-minute presentation
by one side generates anywhere from 30 to 100 questions/comments by the
justices. And when my students and I set forth to replicate and extend her
findings, we found it is not easy to say just what is a question. Justices
interrupt each other, their questions are sometimes answered before they
are finished, there are numerous brief comments (Okay. I see. Right.)
which may or may not be counted as questions. But despite these, Shullmans
study drew attention. Tony Mauro devoted an article in the American Lawyer
to it, and it certainly provoked me to determine if the effect was genuine and
widespread.
Just about the same time as the publication of Shullmans article, John
Roberts (before he was named Chief Justice) addressed the Supreme Court
Historical Society at its annual meeting. Although most of his talk dealt with
the reemergence of a Supreme Court bar, he did report on an analysis he had
conducted. He took the first and last cases in each of the two-week argument
sessions for the October 1980 Term and for the October 2003 Term, and
determined, for each case, the number of questions directed at the petitioner
and the number directed at the respondent. Then he examined which side
won. He reported: In the 28 cases I looked at, 14 from the 1980 Term and
14 from 2003, the most-questions-asked rule predicted the winneror,
more accurately, the loserin 24 of those 28 cases, an 86 percent prediction
rate (2005, p. 75). Then he drily added, The secret to successful advocacy is
simply to get the Court to ask your opponent more questions (2005, p. 75).
Like Shullman, Roberts did not define what a question was, and neither
study distinguished between ideological and nonideological cases. But the
methodology seemed to provide an entry into determining if any difference
related to persuadability existed. Does the more-questions-to-the loser rule
hold when all cases in a term are examined? Does it hold more strongly for
ideological cases?
Testing the Accuracy of the More Questions to the Loser Rule
In our research (Wrightsman, 2008), we quantified questions by referring to

the transcript of the case and counting any comment or question by a justice,
even if it was interrupted or fragmentary. The virtue of this operational
definition is that it provides reliability. We investigated other ways of counting
questions, such as combining interrupted statements or questions; such pro-
cedures lacked reliability; it is not as easy to say what is and what is not one
question as one might think. The analysis my students and I did revealed the
following: For the October 2001 Term, the eventual losing side was asked more
questions in 29 of 41 ideological cases, or 70.7%; this success rate contrasts
strongly with those in the borderline ideological cases6 of 14, or 42.8%and
the nonideological cases8 of 16, or 50%. For the October 2002 Term, the
same differences were observed; the eventual losing side was asked more
questions in 21 of the 33 ideological cases, or 63.6%; in 9 of 16 borderline
ideological cases, or 56.2%, and in only 11 of 23 nonideological cases, or 47.8%
(Overall, in each of these terms, while the eventual loser did get more questions,
the percentage of outcomes predicted accurately, 60% in each term, was much
lower than what was found in the previous studies.)
For the October 2005 Term, the results were consistent with the
October 2001 and 2002 terms, although the differences between types of
cases were smaller. More questions were asked of the losing side in 28 of
43 ideological cases (65.1%), in 9 of 14 borderline ideological cases (64.3%),
and 9 of 15 nonideological cases (60%). Overall, the rate was 46 out of 72
cases, or 64%.
Thus, the rule does seem to have some validity, although not as great a
predictability as the early studies promised. The fact that it holds more
strongly in ideological cases (albeit the differences are not huge) indicates
that to some extent, the nature of the case has led to an earlier formation of
opinions by justices.
Agreement with the Office of the Solicitor General
Another measure that supports the distinction between ideological and

non-ideological cases is the degree to which the eventual decision was
consistent with the position advocated by the Office of the Solicitor
General. As has been well documented (Baum, 1997; McGuire, 1996), the
side supported by the solicitor general, as advocate or amicus curiae, more
often wins. In those terms studied in this project, this was true: in the
October 2002 Term, the solicitor generals side won in 43 of 61 cases, or
65%, and in the October 2005 Term, 38 of 56 cases, or 68%. But in each of
these terms, in ideological cases the solicitor generals side won less often
than it did overallonly 60.7% in 2002 and 61% in 2005. In the October
2001 Term, the Courts decision was consistent with the solicitor generals
position in a whopping 89% of the ideological cases and 93% of the
nonideological cases. Again, the differences are small, and subject to several
interpretations, but one possible interpretation is that in ideological cases
the justices are less susceptible to persuasion from the Office of the Solicitor
General, despite the high regard with which these attorneys are held,
because their minds have been made up.
It was also predicted that differences in the dispensation of cases
would be based on the degree to which their content was ideological.
Two measures were employed to test these predictions: percentage of
cases in which the final decision was unanimous, and length of time to
reach a decision.
Do Ideological Cases More Often Lead to Nonunanimous Decisions

by the Court?
During the October 2001 Term, of 72 decisions, only 26, or 36%, were
unanimous. As predicted, unanimous decisions were reached less often in
ideological cases (32.5%) than in nonideological cases (43.7%). Borderline-
ideological cases produced just about the same level of unanimity (33.3%) as
ideological cases.
While the overall degree of unanimity increased slightly in the October
2002 Term, the pattern was consistent with the earlier term. Only 12 of 33
ideological cases resulted in unanimous decisions (36.4%), compared to 7
of 16 borderline-ideological cases (43.7%), and 13 of 23 nonideological
cases (56.5%). Overall, 32 of 72 decisions, or 44.4%, were unanimous. For
the October 2005 Term, the effect of the type of case is quite similar:
unanimity in 20 of 43 ideological cases (45.6%), in 8 of 14 borderline-
ideological cases (57.1%), and 10 of 15 nonideological cases, or 66.7%).
Overall, in 20052006, 38 of 72 decisions were unanimous, or 52.8%. (The
latter term, Chief Justice Roberts first, led to a higher rate of unanimous
decisions than in any recent term, perhaps reflecting his aspirations for
greater consensus in decisions.)
How Long Does the Court Take to Reach a Decision?
It was also predicted that the justices would take longer to announce decisions
in ideological cases. Decision latency was measured by the number of days
between the oral argument and the announcement of the decision. For the
October 2002 Term, the average latency for ideological cases was 8l days, for
borderline ideological 75 days, and for nonideological 67 days. For the
October 2005 Term, these average latencies were 84, 78, and 66 days. In
both terms, ideological cases took longer to reach final resolution. But in
the October 2001 Term, the justices took less time to decide ideological
cases82 days versus 92 days for nonideological cases.
Reanalyzing Data from the Supreme Court Forecasting Project

Finally, with regard to predictions generated from the initial hypothesis that
persuasion operates differently based on the nature of the case, a further
analysis was done of some data generated by researchers at the School of Law
at Washington University in St. Louis (Ruger, Kim, Martin, & Quinn, 2004).
The Supreme Court Forecasting Project has been an exceedingly useful
vehicle for anyone interested in Supreme Court decision making. For all
cases in the October 2002 Term, the staff developed a prediction of votes
and decisions based on a statistical formula that employed only six generally
straightforward variables. These were the following:
1. The circuit court of origin.
2. The issue area of the case (using the 15 topic areas developed by
Harold J. Spaeth in his U.S. Supreme Court Judicial Database).
3. The type of petitioner (e.g., the federal government, an employer, a
defendant, etc.)
4. The type of respondent.
5. The ideological direction (liberal or conservative) of the ruling by
the lower court.
6. Whether the petitioner argued that the law or practice was
unconstitutional (Ruger et al., 2004, p. 1163).
This information was fed into classification trees and generated predictions
for the votes of each justice and hence the outcome of the case. The classifica-
tion trees differed from justice to justice; a variable that was prominent in the
decision tree of one justice might be relatively unimportant or altogether
absent in another (Ruger et al., 2004, p. 1165). But also, the decision trees of
the different justices are not independent of one another; for example, a
branch point in Justice Thomass decision tree is based on Justice Scalias
anticipated vote (Ruger et al., 2004, Figure 9, p. 1198).
The staff also identified a pool of 83 experts, 71 law professors, and
12 appellate attorneys (including 38 former Supreme Court law clerks),
and, for each case, asked as many as three who had specialized knowledge
in the type of case to predict the votes and outcome. Experts were asked to
predict the outcomes of cases within only their areas of expertise. Like the
predictions from the statistical formula, all predictions by the experts were
made prior to oral arguments. Experts were provided a copy of the lower
court opinion and citations to the parties Supreme Court briefs, but they
were free to consider any sources of information they considered relevant.
Overall, the statistical model correctly predicted 75% of the decisions in
the October 2002 Term, while the experts, as a group, were correct only 59.1%
of the time (Ruger et al., 2004, Table 1, p. 1171). It was hypothesized that
decisions in the ideological cases would be more predictable than those in the
nonideological cases, and so a further analysis was done of data that were
available on the projects website (www.wusct.wustl.edu). Considering the
statistical model first, it was found that it was correct in 26 of 33 ideological
cases or 78.8%, correct in 11 of 15 borderline-ideological cases, or 73.3%, and
correct in 16 of 23 nonideological cases, or 69.6%. Thus the predicted
difference was obtained. For the experts, in ideological cases, 53 of 88 were
correct or 60.2%, while 18 of 35 were correct in borderline-ideological cases,
or 51.4%, and 32 of 55 were correct in nonideological cases, or 58.2%. For the
experts, the differences are not linear, as they are with the statistical model,
and the experts did not do appreciably better in the ideological cases than in
the nonideological ones.
Interim Summary
In summary, persuasion does appear to operate somewhat differently based
on whether the case is an ideological one or not. Data from several terms
indicate that:
1. In ideological cases, the final vote less often is unanimous.
2. In ideological cases, the time it takes to reach a decision is longer, at
least in the majority of terms.
3. In oral arguments in ideological cases, justices direct more questions
to the advocate or advocates who represent what later becomes the
losing side, implying that to some extent justices have already
formed an opinion in ideological cases before the oral argument.
4. In ideological cases, the decision of the Court is less often consistent
with the position of the Office of the Solicitor General, again
reflecting the power of the justices own ideologies in deciding such
cases.
Most of the empirical tests of the hypothesis produce small differences,
although the pattern is strongly in the expected direction. The smallness of the
differences is perhaps not surprising, given the rather broad means of distin-
guishing between ideological and nonideological cases. For example, abortion
cases were classified as ideological; in actuality, the topic of abortion triggers
an instantaneous, value-drive reaction in some justices more than it does in

others. More work is needed in specifying the interaction between type of case
and the individual justice.
To conclude that justices react in different ways to ideological and non-
ideological cases may, to many readers, seem to be less than a surprising
conclusion. But this variable may be useful in helping us understand the role
of persuasion in judicial decision making. A next step is to examine the content
(rather than the frequency) of justices questions during oral arguments, to
determine if evaluative comments are made more often in ideological cases. In
an analysis of 24 oral arguments during the October 2004 Term, Jacqueline
Austin and I recorded 109 instances of a justices question or comment that was
unsympathetic to the advocates position. In 87 of the 109, the justice later voted
against that side. Do these reflect values that play a role in ideological cases?
Testing the Second Assumption

The second assumption offered at the beginning of this paper proposed that
persuasion operates differently depending on the particular justice. Here an
individual-differences approach may be fruitful. What makes certain justices
more persuasive and what make certain justices more resistant to persuasion?
Justices Who Were Effective Persuaders

If we consider justices over the last 50 years, certain ones stand out for their
ability to persuade their colleagues. Earl Warren is recognized for taking a
conflict-riddled Court that was divided on Brown v. Board of Education and
persuading its holdouts so that the Court was able to announce a unanimous
decision on May 17, 1954. Chief Justice Warren was not a legal scholar, but
the other justices were influenced by his charisma and his political skills.
During that period, and even after Warren had left the bench, William
Brennan was very influential, even bringing conservatives to his side in
some cases, because of his genuine interest in people and his willingness
and ability to craft majority opinions that reflected the wishes of justices
who did not completely agree with everything that Brennan would have
wished to achieve. Consideration of the Court during those years leads to a
conclusion that sheer brilliance is not, in and of itself, enough to make justices
effective in their attempts to persuade their colleagues (Rosen, 2007). Felix
Frankfurter came to the Court with everything going for him: a professorship
in constitutional law at Harvard Law School, a number of articles and books
on the Supreme Court, and a network of friends and former students in high
places in the government. Yet Frankfurters attempts to ingratiate and manip-
ulate other justices were largely unsuccessful, and, in fact, some justices
(Douglas and Black and even, eventually, Whittaker) came to ignore and
even ridicule his efforts to persuade.
Justices as Recipients of Persuasion
Judges, as recipients of persuasion, can be studied through several approaches.

For example, an appellate court is like any other small problem-solving group
in that pressures toward uniformity exist and that some group participants
succumb to them. Solomon Asch (1955, 1956), a social psychologist, demon-
strated in what has become a classic study, that it is very hard for a sole
participant in a group project to maintain his or her response when all the
other participants differ in response, even if he or she is the only participant who
is correct. That such pressures to conformity often cause the outlier to succumb
has been demonstrated in everything from jury deliberations to decisions by the
advisers to the United States president. Justice OConnor is quoted in a recent
book as saying that justices would never change their vote simply to be a part of
the majority (Greenburg, 2007) but it does happen. Certainly on occasion
justices join the majority opinion even when they have reservations. A memo
from Chief Justice Burger to Justice Black in a 1971 case said: I do not really
agree but the case is narrow and unimportant except to one man . . . . I will join
you in spite of my reservations (Maltzman, Spriggs, & Wahlbeck, 2000, p. 22).
The justices even have a name for this reaction, as illustrated in a communica-
tion from Justice White to Justice Marshall: I was the other way, but
I acquiesce, i.e., a graveyard dissent (Maltzman et al., 2000, p. 7).
Responses to Pressures Toward Uniformity

The claim that even justices frequently respond to pressures to uniformity is
illustrated in an analysis of the data generated by the Supreme Court
Forecasting Project (Wrightsman, 2006, Chapter 10). Recall that this project
generated predictions of each justices vote in each case for the October 2002
Term. The project then took these anticipated votes to make predictions of the
outcomes. But the procedure did not have any means to add in what might be
called a conformity correction. Thus the statistical model predicted that in
this term there would be only 7 unanimous decisions out of 72 cases. In
actuality there were 31. The model predicted there would be 20 8-to-1 decisions,
but in actuality there were only 4 (Wrightsman, 2006). Clearly, when most
justices vote one way, pressures exist on the holdout justice to go along, and
often they do. An analysis by Granberg and Bartels (2005) of voting patterns in
the Court extending back to the 1950s found that 8-to-1 votes were the
least frequent type, accounting for only 10% of the decisions. In contrast,
over this period of 48 terms, unanimous votes occurred in 35% of the cases
and 5-to-4 votes in 21%. William O. Douglas, who served on the Court for 36
years, the longest of any justice, also has the record for the most sole dissents,
106, or about 3 a year. But William Brennan, on the Court for almost as long as
Justice Douglas, had only 11 in 34 years, reflecting his desire to be conciliatory.
And Chief Justice Burger, who did not like other justices to write dissents, or
even concurrences, was a sole dissenter only 4 times in his 17 years on the Court.
Such individual differences are also apparent among current justices; in

the last five terms, Justice Stevens has been a sole dissenter 10 times, Justice
Thomas 6 times, and Justice Scalia 4 times. In contrast, Justice Ginsburg never
was a sole dissenter in that time period, and neither Chief Justice Roberts nor
Justice Alito have been sole dissenters in their briefer times on the Court. (As
noted, Chief Justice Roberts, during this term, has been advocating for more
narrow, unanimous decisions.)
What contributes to a justices being a sole dissenter? Strongly held
opinions and a relative lack of concern about the opinions of others would
seem to be determinants. In his 30 years on the Court, Justice Stevens has
always gone his own way. Kenneth Starr has written that he has taken on the
role of a naysayer (2002, p. 43). Jan Crawford Greenburgs recent book calls
him an iconoclast and has this description of him:
Stevens was a maverick who didnt ascribe to a particular theory. He
was fiercely independent in his writings and actions. When the
justices donned their robes before taking the bench, Stevens was the
only one who refused assistance from the aides in the robing room.
He always insisted on putting on his own robe. He took his own path
in his opinions, too. (2007, p. 180)
Justice Thomas has, of course, been subjected to intense scrutiny from
the time of his nomination to the Court; within the last year two books have
been published that direct attention to his style of decision making. Supreme
Discomfort (2007) by Kevin Merida and Michael Fletcher provides an insight
into Justice Thomass background and possible reasons for his adamant
position on ideological issues. Jan Crawford Greenburgs (2007) book illus-
trates that even from his first months on the Court, Justice Thomas had an
independent streak. At his very first conference after the oral arguments
(Foucha v. Louisiana, 1992) he chose to dissent from the majority, and it
initially appeared that he would be the sole dissenter. (Several months later,
after he had circulated his written dissentwhich was sharply critical of
the majority opinion by Justice Whitethree other justices shifted from the
majority to his minority side.) In his second week of oral arguments, he again
chose to be a sole dissenter, in the Eighth Amendment case Hudson v.
McMillian (2002). His dissent drew wide condemnation in the media (the
New York Times editorialized that he was the youngest cruelest justice) but
again it drew the support of one other justice (Scalia).
Greenburgs view is the following:
Though quiet on the bench during public sessions, Thomas wasted
no time sharing his views in conference. Pundits and analysts would
disparage Thomas as Scalias intellectual understudy, but from the
beginning that portrayal was grossly inaccurate. . . . Thomas made
clear that he was willing to be the solo dissenter, sending other
justices the strong signal that he would not moderate his opinions for
the sake of comity. (Greenburg, 2007, p. 115)
According to Mark Tushnet, Justice Thomass strong will and uncom-

promising positions created problems when he was assigned majority opi-
nions by Chief Justice Rehnquist. Tushnet writes:
Thomas was rarely unsure about the positions he took, so strategy
never counseled in favor of giving him an important opinion. Indeed,
his very certainty sometimes recommended against doing so. Thomas
tended to write strong opinions, drawing sharp lines and rarely
acknowledging that different circumstances might produce different
outcomes. OConnor and Kennedy sometimes preferred a more
nuanced doctrine than Thomass opinions articulated. Also, Thomas
was more reluctant that other justices to accommodate such concerns.
As a result, Thomas lost a Court more often than other justices given
opinion assignments. That is, instead of gaining the five votes that
seemed to be available when the opinions were assigned, his opinions
might get only four votes. (Tushnet, 2005, p. 86)
Tushnet makes an interesting observation. It seems to predict that, at
least when Chief Justice Rehnquist was assigning opinions, Justice Thomas
was more likely to be assigned safe ones. An examination of the opinion
assignments for the last two terms when Rehnquist was Chief Justice (October
2003 and October 2004 terms) finds that Justice Thomas wrote 16 opinions
for the Court, and 9 of the 16, or 56%, were unanimous. Of the remaining 7, 3
were 7-to-2 or 6-to-2, 2 were 6-to-3, and 2 were 5-to-4. The percentage of 9-
to-0 decisions in cases in which Justice Thomas authored the Courts opinion
was much higher than the 35% to 40% unanimity in those terms.
Justice Thomass response to his critics is to say, I dont care what they
think. I am free to live up to my oath (Greenburg, 2007, p. 121). And it is
certainly true that his independence is manifested in a number of ways, some
related to his work on the Court and some not. His failure to participate in
oral arguments has been noted here and elsewhere. During a visit to the
University of Kansas Law School, he was interviewed by the local newspaper;
when asked about oral arguments, he told the reporter, I dont see the need
for all those questions. I think justices, 99 percent of the time, have their
minds made up when they go to the bench (Rombeck, 2002, p. 5B). Justice
Thomas also demonstrates his nonconformity in his off-the-Court prefer-
ences, includingin what seems to be a deliberate act of perversity in
Washington, D.C., where everyone lives and dies with the results of the
Redskins football fortuneshis identification as a Dallas Cowboys fan.
Individual Differences Variables and Personality Variables
Psychologists interested in personality and individual differences have gen-

erated concepts applicable to resistance to persuasion, going all the way back
to work on dogmatism in the 1950s. Tetlocks (1983a; Tetlock, Bernzweig, &
Gallant, 1985; Gruenfeld, 1995) work on integrative complexity or cognitive
complexity led him to analyze majority and dissenting opinions by the

Supreme Court, and such concepts could be applied to resistance to persua-
sion. As discussed in the Mitchell and Tetlock chapter in this book, Isaiah
Berlins prototypes of the hedgehog and the fox can be used to identify
formulaic approaches to decision making, including that by judges and
justices. These and other related traits, such as need for cognition, dogma-
tism, and the need for cognitive closure are worthy of the attention of scholars
studying persuasion in judging.
Conclusion
In the preface to The Puzzle of Judicial Behavior, Lawrence Baum expresses his
belief that we are a long way from achieving explanations of judicial behavior
that are fully satisfactory (1997, p. xi). That was more than10 years ago; this
chapter has presented a modest effort toward an explanation, by identifying
ways that persuasion operates differently on judges as they form their opi-
nions, depending on their biases and the nature of the case. Beyond this,
scholars are beginning the hard work of examining the files of several recent
justices, especially the detailed records of Justice Blackmun, to illustrate the
role of persuasion during the process of moving from the initial decision draft
to the final opinion of the Court. The recent article by Johnson, Spriggs, and
Wahlbeck (2007) is an excellent example. We can look forward to a greater
understanding of the role of persuasion at all stages of judicial decision
making.
5
Judges as Members of Small Groups
Wendy L. Martinek
Though most judges are trial court judges, and most judicial decisions are
made by trial court judges, appellate courts and the judges who populate
them have attracted a considerable (disproportionate?) share of scholarly
attention. This focus is perhaps lamentable but understandable given that
judicial policymaking falls more centrally in the province of appellate
courts rather than trial courts.1 In fact, some legal scholars have gone so
far as to suggest that policymaking is the primary function of appellate
courts (Landes & Posner, 1979).2 The importance of the decisions appellate
courts make is reflected in the fact that virtually all appellate courts use
groups of judges to render decisions. Sometimes this includes the full
complement of judges on a court (e.g., the U.S. Supreme Court) and
sometimes merely a subset of those judges (e.g., the South African
Appellate Division) but, with exceedingly rare exceptions, appellate court
decisions are group decisions. The logic underlying the use of groups of
judges at the appellate stage is straightforward: deliberation among a set of
judges is intended to enhance the likelihood of arriving at the correct
decision; that is, reduce the likelihood of erroneously reversing a correct
lower court decision or erroneously affirming an incorrect lower court
decision (Drahozal, 1998).
The fact that appellate courts are collegial (that is, multimember) courts
has not been lost on students of judicial behavior. In particular, those scholars
who approach judicial choice through the lens of strategic behavior explicitly
recognize that, to achieve their most preferred policy outcome, judges on
73
collegial courts must consider the likely actions of their colleagues on the
bench to determine their best course of action:
[J]ustices may be primarily seekers of legal policy, but they are not
unconstrained actors who make decisions based only on their own
ideological attitudes. Rather, justices are strategic actors who realize their
ability to achieve their goals depends on a consideration ofthe preferences
of other actors, the choices they expect others to make, and the
institutional context in which they act. (Epstein & Knight, 1998, p. 10)
Though Epstein and Knight were writing about members of the U.S. Supreme
Court bench, their description of strategic behavior by members of that court
is equally applicable to judges on other appellate courts, both domestic (e.g.,
Hettinger, Lindquist, & Martinek, 2006; Langer, 2002) and foreign (Helmke,
2005; Vanberg, 2005).
Such scholarship takes seriously the fact that judicial decisions on collegial
courts are the product of group choices and, in that regard, takes into account
the small group environment of collegial court decision making. The focus of
the majority of this scholarship is unduly narrow, however, in that it almost
always presumes a single goal (policy) and neglects to consider nonstrategic
aspects of appellate court decision making (see Baum, 2006, pp. 68). But the
small group context of appellate court decision making has meaningful con-
sequences beyond serving as a venue for strategic calculations. This is where
psychology, especially the insights of social and organizational psychology
scholarship, focused on the behavior of small groups, can be profitably
deployed to further our understanding of how judges on collegial courts
behave. This is by no means the first time such an approach has been suggested.
Schubert (1964), Murphy (1966), and Ulmer (1971), among others,3 made this
same argument quite some time ago. But since then, researchers approaching
collegial court decision making as a type of small group behavior have been few
and far between.4 This is an unfortunate state of affairs.
Taking a small group approach to the study of judicial decision making is
very much an interdisciplinary activity. Sociologists, organizational behavior
researchers, social psychologists, and anthropologists have all applied small
group theory to their work.5 Regardless of the disciplinary context, however, a
key preliminary issue is determining what constitutes a small group. Levine
and Moreland offer a useful definition of a small group: a group of individuals
who interact on a regular basis, have affective ties with one another, share a
common frame of reference, and are behaviorally interdependent (1994,
p. 306). Notwithstanding their enormous institutional variation, this defini-
tion certainly fits the situation of virtually all appellate courts.6
First, judges serving on appellate courts must interact on a regular basis to
dispose of their caseloads. For example, many appellate court judges, such as
the justices of the U.S. Supreme Court and some members of state courts of last
resort, come together in conference to make decisions about which cases to
accept for review (Langer, 2002; Perry, 1991). And all members of an appellate
Judges as Members of Small Groups 75
court or appellate court panel must come together for oral arguments (Cohen,
2002, pp. 133136; Johnson, 2004) when they are held.7 Further, a written
opinion that formally disposes of a case may be the product of an intensive and
iterative process among the judges (Maltzman, Spriggs, & Wahlbeck, 2000) or a
less interactive process but, nevertheless, does entail at least some level of
interaction if for no other reason than a majority of the judges on an appellate
panel must sign off on an opinion to make it a majority opinion.
Second, judges serving on appellate courts may squabble like children,
bond like family, or behave toward one another in a more detached, profes-
sional manner, but both anecdotal (Cooper, 1995; Hirsch, 1981; Schick, 1970)
and systematic (Cohen, 2002; Howard, 1981) evidence make clear that there
is an affective component to the interactions between and among judges
serving on appellate courts. The personal closeness between Chief Justice
Warren Burger and Justice Harry Blackmun, which devolved into a relation-
ship that could be called strained, at best, is one well-known example. Further,
Justices Brennan and Marshall were considered close colleagues, both on and
off the bench, while Justices Ginsburg and Scalia, though ideologically dis-
similar, are known to be personally quite friendly. Even in the absence of
personal affecteither positive or negativefor one another judges on
appellate courts share affective ties to the institutions on which they serve.
Third, colleagues on a given appellate bench possess a common frame of
reference provided by the institutional environment within which they
operate. That institutional environment includes rules regarding case selec-
tion, oral argument, opinion assignment, and the like. It also includes a
common body of law, which they are charged with interpreting and applying.
While nonjudges may recognize and acknowledge this common frame of
reference, they are viewing it as outsiders rather than partaking of it as judges
themselves do. In other words, colleagues on the bench function as a true
peer group, people who share the same position and work in the same
situation (Baum, 2006, p. 54).
And, fourth, appellate court judges are behaviorally interdependent by
definition. No single judge on an appellate court, not even the majority
opinion author, can individually determine the winner and loser in a given
case or dictate the content of the legal rule embedded in a particular written
opinion. Those are functions of the collective choices of the judges on that
appellate court. While it is true that some judges may be more influential than
others both as to the winners and losers in a case and as to the reasoning
subscribed to by an appellate court in determining those winners and losers,
no single judge can be determinative in the disposition of an appellate court
decision. In short, [a]ppellate court decisions are inherently collective pro-
ducts. The outcome for the litigants and the legal doctrine that a court
promulgates are determined by where a majority of judges stand (Baum,
2006, p. 51).
Though individual collegial courts may differ in their proximity to the
archetype of a small group as defined by social psychologists and
organizational behavioral specialists, as a class they fall quite comfortably

under the rubric of small groups. This makes small group theory very inviting
as a tool for the investigation of a plethora of appellate court decision-making
processes. Two aspects of appellate court decision-making are particularly
ripe for the application of theories emerging from small group research. First,
scholarship devoted to group cognition can aid students of judicial decision
making in their quest to understand when and how legal versus nonlegal
factors influence judicial choices. Second, the small group literature can shed
valuable light on how group rolesboth formally and informally defined
occupied by members of collegial courts matter for the decision making of
both those who occupy a given role and those with whom such individuals
make decisions.
Legal Versus Nonlegal Factors and Collegial Court Decision

Making
A persistent debate in the law and courts community is over the relative
influence of legal and nonlegal factors in the decision calculi of judges.
The view of judging as a mechanistic legal process in which judges simply
match cases with the relevant legal factors (e.g., the language of the relevant
statute, the principle of law articulated in the pertinent precedent) was
unsatisfying to the group of judges and legal scholars who were the progeni-
tors of what became known as legal realism (Duxbury, 1995, ch. 1). Though
hardly all of like mind in terms of the proximate cause of judicial decisions,8
the legal realists did all agree that a focus on the law on the books was
uninformativeand, in fact, misleadingwhen compared to a focus on law
in action (Duxbury, 1995, pp. 6768). The legal realists ultimately inspired
the attitudinal model of judicial decision making (Pritchett, 1948; Schubert,
1965), which holds that the Supreme Court decides disputes in light of the
facts of the case vis-a`-vis the ideological attitudes and values of the justices
(Segal & Spaeth, 2002, p. 86).
Notwithstanding the fact that it has proven to be useful for under-
standing judicial decision making in a variety of courts, the attitudinal
model has not been without its critics.9 Some of the sharpest criticisms of
the attitudinal model have focused on its perceived failure to fairly evaluate
the evidence with regard to the influence of factors other than the attitudes of
judges; in particular, the influence of the law. Segal and Spaeth (1996; Spaeth
& Segal, 1999), the standard bearers of the contemporary attitudinal model,
offered an empirical test of one aspect of the traditional legal model: the role
of precedent as a determinative influence on judicial vote choice. They found
little evidence that precedent was a meaningful constraint on judicial choice
but were taken to task by a range of scholars for defining the influence of law
in narrow, mechanistic terms.10 Some of these critics have argued that the
influence of precedent, and of law more broadly defined, should be seen in a
nuanced fashion as drawing the attention of judges to particular aspects of

subsequent cases that merit special attention in the decision process (Richards
& Kritzer, 2002; see, also, Kritzer & Richards, 2003, 2005) rather than man-
dating particular outcomes.
Though numerous scholars have extended our collective knowledge
about the role of law in judicial choice, virtually none have paid attention
to how the small group context of collegial court decision making might
matter for understanding the influence of legal factors in appellate adjudica-
tion. The fact that appellate court judges are rendering decisions collectively,
however, means that those charged with crafting the written opinion of the
court must convince their colleagues (or, at least, a majority of their collea-
gues) as to the correctness of the opinion they have crafted. The work of
social psychologists that considers the effect of group membership on atti-
tudes and behavior is particularly promising in terms of understanding how
opinion authors might go about doing this. Though membership in a group
as denoted by a set of objective criteria is neither necessary nor sufficient to
give rise to an internalized sense of group membership (Abrams et al., 1990),
the fact that appellate court judges are readily recognized by themselves and
others as belonging to an unambiguously defined group (a court) may
enhance the likelihood that they will incorporate their group membership
into their concept of themselves. In other words, the fact that judges see
themselves as members of a distinctive institution (a court) and that others
(e.g., litigants, attorneys, other judges) see them that way, too, suggests that
judges might be especially attentive to the norms and expectations that attach
to the members of their small group.
In this regard, there is perhaps no other norm that has a stronger prima
facie claim on judges than the norm that the decision making of judges should
be governed by a consideration of the relevant legal factors. This norm is woven
tightly into the fabric of legal education and the legal profession. Judges, then,
come to the bench already well inculcated with this norm. A judges self-
identification as a member of a court may lead to an enhanced reliance on
conventional legal factors in arriving at a case disposition; not because she sees it
as an instrumental way to marshal the support of her colleagues, as the strategic
theorist might have it (Epstein & Knight, 1998; Maltzman, Spriggs, & Wahlbeck,
2000), or as a cloak to mask the brazen influence of personal ideology, as the
attitudinalists might have it (Segal & Spaeth, 1993, 2002), but because she views
legal factors as those most relevant for her decision making precisely because of
her membership in the small group constituted by her court.
Theorizing about if and how the law matters in the decision making of
appellate court judges would substantially benefit from an even more direct
consideration of how the members of the small group constituted by a court
affect one another when it comes to reasoning to resolve a case. In particular,
a natural question is whether a judges reliance on legal factors is conditioned
by the extent to which his colleagues on the bench rely on such factors. Small
group researchers have argued that the social exchanges among group
members produce shared cognitive products, including memories, norms,

scripts, schemas, and interpretations of shared events and activities
(Gruenfeld & Hollingshead, 1993, p. 384). This suggests the possibility of
an iteratively reinforcing process in which reliance on legal factors becomes
even more ingrained as the correct approach to the disposition of cases as
appellate court judges continue to decide cases together.
This conjecture gives rise to two questions for which research on small
groups and small group decision processes can provide important insights.
First, what initial conditions are necessary for legal (as opposed to nonlegal)
factors to dominate the group decision-making process? Is it sufficient for
one judge to be self-conscious about locating persuasive legal arguments to
guide the resolution of a case to prompt all judges on the court to do so or is
there a necessary critical mass of such judges on a given court? Second, does
the stability of the membership of the appellate court matter for the extent to
which any single judge or group of judges can influence colleagues to pay
special attention to legal factors in arriving at a case outcome? This question
takes on special importance for courts with frequent membership changes
and those in which decision making is done via panels of judges rather than en
banc; for example, the U.S. Court of Appeals (Howard, 1981) and the
Supreme Court of Canada (Hausegger & Haynie, 2003).
Group Roles and Collegial Court Decision Making

Another means by which a small group perspective may inform the work of
students of judicial behavior is the attention it directs to the roles members of
small groups occupy. Though any given judge may be formally fungible with
his colleagues on a particular court, in reality there are often tangible differ-
ences among judges operating on a court, differences of which the judges are
themselves aware and that are rife with potential behavioral consequences.
Some of these differences come in terms of a formally defined role, such as the
Chief Justice of the United States, the Chief Judge of the Appellate Division of
the Supreme Court of South Africa, or the chief judge of the U.S. Court of
Appeals for the Ninth Circuit. Other differences arise from roles that judges
occupy on a temporary basis. For example, there is considerable evidence that
judges new to the bench experience acclimation effects (e.g., Hurwitz &
Stefko, 2004; Hettinger, Lindquist, & Martinek, 2006). But, assuming a long
enough tenure, every freshman judge will eventually lose his newcomer status.
Moreover, some judges are not regular members of an appellate court but
participate on a temporary basis, such as certificated judges in the New York
judicial system11 and district court judges serving by designation on the U.S.
Courts of Appeals.12 All of these differenceswhether informal or formal,
subjective or objectivehave potential import both for the behavior of
individual judges possessing a given characteristic as well as the behavior of
those with whom that individual is making decisions.
Perhaps the quintessential group role is that of group leader. Leadership

is a staple topic in the study of judicial behavior (e.g., Atkins & Zavoina, 1974;
Haynie, 1992), and there is a voluminous set of organizational psychology
research devoted to leadership in small groups that can be usefully brought to
bear in theorizing about leadership on appellate courts. The concept of
leadership has been profitably parsed into task and social leadership (Bales,
1950): The former seeks to complete the present task in the most effective
and efficient manner; the latter seeks to provide the friendly atmosphere that
eases cooperation (Murphy, 1966, p. 1567). A single individualsuch as the
Chief Justicemay or may not exercise both types of leadership. Danelski
found, for example, that Chief Justice William H. Taft provided social leader-
ship while his colleague and friend, Associate Justice Willis Van Devanter,
provided task leadership (1960, pp. 490491), and that Chief Justice Charles
E. Hughes exercised both (1960, p. 491).
Thinking about leadership on a court from the perspective of organiza-
tional behavior, one question is whether the formal role of chief justice or chief
judge is sufficient for the individual occupying that role to exercise leadership
of either variety. Some of the existing judicial behavior literature suggests that
formally occupying such a role can have some effect on group deliberations. For
example, Hettinger and her colleagues (2006, ch. 5) found that the presence of a
chief judge of a U.S. Court of Appeals on a decision-making panel made the
reversal of a lower court decision more likely. They speculated that this was a
function of a chief judges being more attentive to the need to reverse errant
lower courts due to their institutional responsibilities as the head of their
respective circuits. They also found that chief judges were less likely to express
disagreement by filing separate opinions (2006, ch. 4), which was attributed to a
desire to maintain collegiality among the judges on the circuit. Other literature,
however, suggests that the magnitude of the effect of possessing a formal
leadership role on a court pales in comparison to the influence attributable to
characteristics such as collegiality, intellect, and administrative competence
(Wrightsman, 1999, pp. 83103). This is exactly where the psychology of
small groups has the potential to help students of judicial behavior determine
how and under what conditions those occupying formal leadership positions
on a court can wield influence on that court.
Particularly promising in this regard is work by Ridgeway and her
colleagues (Ridgeway & Berger, 1986; Ridgeway, Johnson, & Diekema,
1994) devoted to understanding how the external status of group members
affects their perceived legitimacy and, hence, the propensity of colleagues to
comply with the authority of formal leaders. Those who occupy formal
leadership positions on an appellate court typically have few coercive
mechanisms at their disposal for inducing compliance with (or deference
to) their wishes. Neither the Chief Justice of the United States nor the chief
judges of the U.S. Courts of Appeals, for example, play a role in the selection
or removal of their colleagues.13 Nor do such formal leaders possess the ability
to dock the pay of those colleagues they see as recalcitrant. This is not to say
that formal leaders such as chief justices have no punitive measures they can
impose. The opinion assignment authority of the Chief Justice of the United
States is an example of a power that could be used punitively through the
withholding of desirable opinion assignments, though the discretion to use
opinion assignment in this fashion is not infinitely elastic given efficiency
concerns (Maltzman & Wahlbeck, 1996). Compliance with or deference to
the wishes of an appellate court leader is secured in large part voluntarily or
not at all. The question remains, however: when can a person occupying a
formally defined leadership position on an appellate court effectively influ-
ence his colleagues? Small group theory can provide useful guidance in
answering this question.
Likewise, small group theory can profitably inform our work regarding
the behavioral effects of other types of roles. For example, the evidence to date
suggests that judges who are new to their positions are likely to face an
acclimation or socialization process (Alpert, Atkins, & Ziller, 1979; Hurwitz
& Stefko, 2004; Wood et al., 1998). Louis defines organizational socialization
as the process by which an individual comes to appreciate the values,
abilities, expected behaviors, and social knowledge essential for assuming an
organizational role (1980, p. 229). In other words, organizational new-
comers must gather a great deal of information about an institution before
they can become fully fledged members of that organization. A natural source
for that information is the other members of the organization. Freshman
judges, in effect, may cue off the behavior of their more senior colleagues.
Small group theory can guide our thinking about how freshman judges
determine which colleagues to rely on as cues for their own behaviors. Is
ideological proximity key or is it, perhaps, respect that comes from demon-
strated expertise? Alternatively, selecting a cue giver may be a function of
demographic and experiential similarities or the recognition of the status
accorded to a particular member of a court by other members of that court.
Research devoted to understanding conformity and status attainment in
small groups holds promise vis-a`-vis its utility for understanding when and
how new members of collegial courts select cue givers from among their
colleagues, as well as how more senior group members behave toward new
colleagues. Specifically, expectation states theory suggests that group mem-
bers have expectations for themselves and for other members of the group
regarding their ability to contribute toward the completion of the groups
tasks (Berger et al., 1977; Berger, Wagner, & Zelditch, 1985). These expecta-
tions are based on both external status characteristics and specific informa-
tion regarding task competency. External status characteristics include gender
and race, demographic characteristics that shape beliefs about task compe-
tence. For example, women are generally disadvantaged compared to men in
terms of their perceived competence in economics and foreign affairs but
advantaged when it comes to their perceived competence in social welfare
issues (Sapiro, 1983). Further, African Americans are seen as less able to
engage in abstract thinking compared to whites (Plous & Williams, 1995).
Specific information regarding task competency may come from the

personal observation of group members but may also be a product of
reputational factors.14 Newcomers to an appellate court generally have had
no prior means for acquiring information about the competence of their
peers through direct observation. They can, however, readily observe demo-
graphic characteristics and are likely to know something about the reputa-
tions of their new colleagues. Small group research can help us understand the
relative contributions of these two types of information to the expectations
judges new to a court have about their colleagues skills and abilities, and in
the process, can help us understand which of those colleagues are likely to be
influential on newcomer judges because of their status as cue givers.
Expectation states theory can also shed light on how senior members of a
collegial court interact with newly appointed members. Not all freshman
judges are created equal. Some judges new to an appellate court come with
a wealth of prior experience on another appellate court. Others come with
little to no experience as a judgeappellate or otherwisebut have other
experience as prosecutors, public defenders, legislators, or other elected
officials. These experiences have the potential to shape how newcomers are
received by their senior colleagues because they contribute to the expectations
those senior colleagues have about the competence of their newly arrived
brethren. Ignoring these differences in past experience and the expectations
they generate about how the competence of new members of a collegial court
are perceived by senior colleagues on the bench is risky. For example, one
strand of the literature devoted to the study of freshman effects considers
whether chief justices and/or senior associate justices on the U.S. Supreme
Court treat newly appointed members of the Court differently when it comes
to opinion assignment (e.g., Slotnick, 1979; Bowen & Scheb, 1993; Brenner &
Hagle, 1996). The inconsistencies in the empirical findings to date may, in
fact, be due to ignoring a consideration of how preappointment experiences
matter in terms of the reputations newcomers bring to the bench. A small
group perspective, especially that embodied by expectation states theory, may
help reconcile those findings as well as, more generally, refocus attention on
how reputations of group members might matter for appellate court
adjudication.
Appellate Courts as Small Groups and the Quality

of Adjudication
The utility of approaching the study of appellate court decision making as a
variety of small group decision making is not limited to understanding the
role of legal factors in the judicial calculus or how the roles appellate court
members occupy influence their behavior and the behavior of their collea-
gues. More generally, thinking about appellate court adjudication from this
perspective can help students of the courts consider how the nature of the
small group constituted by a collegial court might matter for the quality of
adjudication. As noted at the beginning of this chapter, appellate courts consist
of more than one member on the presumption that groups of judges will be less
likely to err than single judges. In theory, each member of a collegial court will
engage in a deliberative process with his colleagues in which the decision-
making biases and other cognitive deficiencies of each judge will be compen-
sated for by his fellow judges on the bench. The end result will be adjudication
that, though a function of the decisions of individual judges on the bench, goes
beyond the mere sum of those individual decisions. Work by Gruenfeld
and Hollingshead suggests that a group cognition process such as this ideal
conception of adjudication is possible: [I]ndividuals in social interaction do
more than trade individually produced cognitions. They also engage in active
reconciliation and integration processes, leading to the emergence of unique,
collectively produced conceptualizationsincluding ideas, representations,
solutions, and argumentsthat no individual had to begin with (1993,
p. 385). As a consequence, the quality of a group decision has the potential to
exceed that of even the most skilled individual member of the group under
ceteris paribus conditions. Ceteris paribus conditions, however, are rare rather
than common.
There are a host of factors that can compromise the quality of adjudica-
tion, factors to which our attention is directed by the extant body of small
group research. One of these factors is undue deference on the part of some
members of a collegial court to other members. Such deference may result
from a status differential between and among judges. For example, the nature
of the judicial hierarchy implies that judges serving on appellate courts
occupy more prestigious positions than those serving on trial courts. This,
in turn, suggests that U.S. District Court judges serving temporarily by
designation on the U.S. Courts of Appeals or with circuit court judges on
three-judge district courts may defer to the circuit court judges with whom
they serve (Brudney & Ditslear, 2001; Walker, 1973). And, as previously
discussed, judges new to the appellate bench may similarly be deferential to
their more senior colleagues (Hettinger, Lindquist, & Martinek, 2003).
Certainly there are conditions under which such deference may be desirable,
as when neophyte judges pay heed to senior colleagues with more expertise.
Regardless, the roles group members occupy can result in less than the full-
throated deliberations among equals that the ideal of appellate adjudication
suggests.
Even assuming equal status among all members of an appellate court,
there are reasons to be concerned about the quality of the adjudication
process that derive from the fact that it is decision making by a small
group. Specifically, individuals participating in group decision making pro-
cesses are susceptible to conformity effects (Sunstein, 2003). Part of this may
be attributable to the fact that members of a group care about the evaluations
of their fellow group members and, all things being equal, prefer higher to
lower status within the group. And at least some research on organizational
behavior indicates that conformity initially contributes to the attainment of

status within a group (e.g., Hollander, 1960; Warhman & Pugh, 1972;
Ridgeway, 1978). In the specific context of collegial court decision making,
the evidence is that appellate judges are profoundly affected by those with
whom they render decisions (Sunstein, 2003, ch. 8; Sunstein et al., 2006). In
particular, they are subject to conformity effects, another deviation from the
ideal of appellate adjudication.
Whether the goal is to extend our understanding of the choices made by
appellate court judges or to identify threats to the ideal appellate court
decision process, small group theory offers rich possibilities. Appellate
courts are small groups. Hence, what we know about small group dynamics
from social psychology and organizational behavior can (and should) inform
research on appellate court decision making. This is not an argument that
small group theory entirely replace the attitudinal or strategic theories of
judicial choice. The attitudes of at least some judges certainly matter at least
some of the time; and at least some judges can and do engage in strategic
calculations in at least some circumstances. The evidence is too overwhelming
to assert the contrary without being, at best, naive or, at worst, disingenuous.
But paying attention to the small group context within which appellate court
adjudication takes place will enrich our understanding of judicial choice well
beyond the confines of the attitudinal and strategic paradigms by offering a
more realistic view of the psychology of judicial behavior.
Notes
Special thanks are due to Paul M. Collins Jr. and Raymond V. Carman Jr. for their
thoughts about the utility of small group theories for understanding judicial
behavior on collegial courts and Harold J. Spaeth for his insights on this and related
projects.
1. There is also a very practical reason for the focus on appellate courtsand
appellate courts of last resort, such as the United States Supreme Court, in
particularthat has to do with the comparative ease of collecting the requisite
data to answer questions about how and why judges make the decisions they
make for appellate courts. The very fact that there are so many more trial courts
makes it a daunting task to gather information on a representative set of courts,
judges, and decisions sufficient for the purposes of inference.
2. Of course, trial court judges arguably make policy in the pattern of decisions
they render (Rowland & Carp, 1996).
3. Other notable work drawing on the psychology of small groups includes work
by Snyder (1958) and a series of articles by Walker (1973, 1974, 1976).
4. There is a similar dearth of recent political science scholarship that takes a small
group approach outside the context of judicial behavior, despite Kirkpatricks
(1976) exhortation and some promising work from the late 1970s and early
1980s (e.g., Dorff & Steiner, 1981; Fiorina & Plott, 1978; Hinckley, 1979).
5. Hare, Borgatta, and Bales (1965) contains an excellent set of readings that nicely
illustrates the interdisciplinary history of small group theory.
6. This is consistent with Murphy: Collegial courts. . . use small groups in a face-
to-face relationship that interact under an obligation to solve a specific problem
or set of problems (1966, p. 1565).
7. Of course, not all appellate court cases are disposed of with the aid of oral
argument. On some appellate courts, in fact, a sizable proportion of the cases
are disposed of without oral argument (Cohen, 2002, pp. 6062).
8. One school of thought among the legal realists focused on the role of social
forces, while the other focused on the characteristics of individual judges.
9. The symposium appearing in the Spring 1994 issue of Law & Courts and the
symposium appearing in the Summer 2003 issue of that same newsletter offer a
representative sampling of the criticisms of the attitudinal model.
10. Issue 3 of volume 40 of the American Journal of Political Science (1996) was
specifically devoted to Segal and Spaeths empirical test of the influence of
precedent and critiques of that approach.
11. Certificated judges are retired judges who are certified to be both willing and
able to serve on a New York State court.
12. Designated district court judges serve for temporary periods of time on the
appeals court bench. Those coming from within the circuit are so designated at
the discretion of the chief judge of the circuit. Those coming from outside the
circuit are so designated only with the permission of the Chief Justice of the
United States. Similarly, circuit judges from one circuit may serve temporarily
in another circuit with the permission of the Chief Justice. Service by these
judges visiting the appellate bench from outside the circuit or from the
district court bench provides considerable service to the U.S. Courts of
Appeals (Cohen, 2002, pp. 194195).
13. There are notable exceptions, of course, such as the influence of Chief Justice
Warren Burger in President Richard Nixons selection of Burgers boyhood
friend Harry Blackmun for a spot on the Supreme Court.
14. The empirical evidence suggests that judicial reputations do have tangible
effects. For example, the reputations of United States Courts of Appeals
judges make a difference in the extent to which individual judges are influential
in the development of intra- and intercircuit law (Klein, 2002). Further,
Caminker (1994) makes the case that judges care about their reputations for
reasons of advancement.
6
The Supreme Court, Social Psychology,
and Group Formation
Neal Devins and Will Federspiel
The justices of the Supreme Court function not only as individuals, but as
members of groups. One group of which they are part is the Court itself, as
discussed in Wendy Martineks chapter in this volume (ch. 5). But they can also
come together to form important subgroups. In this chapter, we examine a
particular type of subgroup that we refer to as a majority coalitiona group of
ideologically simpatico justices who are able to issue unambiguous, far-reaching
decisions, as opposed to fact-specific decisions of limited consequence. We
employ social psychology literature to better understand when the Court will
and will not function as a cohesive coalition. In so doing, we also comment on
the models political scientists use to describe Supreme Court decision making.
Our principal claim is intuitively obvious but in tension with much of the
political science literature. Political science models of Supreme Court deci-
sion making typically focus on the legal and policy goals of individual
justicesso that the key question concerns the legal policy preferences of
the median justice and the ideological gap between the median justice and
other members of the Court. We think the political science models focus too
much on the individual and not enough on the group (including the reasons
why individuals do or do not join groups). Specifically, when there is an
ideologically simpatico majority coalition, intragroup dynamics play a pro-
minent role in determining the reach of Supreme Court decisions. More to
the point, the individual preferences of the median justice are less consequen-
tial on a cohesive Courtsince the median justice will (up to a point) give in
to intragroup pressures to uniformity. In contrast, the preferences of the
median justice play a more prominent role on an ideologically diverse
85
Court. At the same time, these preferences may not mirror the policy views of
the median justice. In refusing to join forces with an ideologically cohesive
coalition, the median justice is likely to place a high value on personal power
and reputation. In other words, median justices on ideologically diverse
Courts have comparatively weaker legal policy preferences and are willing
(up to a point) to sublimate those preferences in order to pursue other goals.
We begin with a brief tour of the chief political science models, high-
lighting the ways in which those models focus on individuated legal and policy
preferences. We then turn to social psychology to examine both the importance
of and obstacles to group formation. Finally, by comparing differences in
decision-making styles of the (largely simpatico) New Deal Court and the
(very diverse) Rehnquist Court, we illustrate how social psychology can con-
tribute to an understanding of Supreme Court decision making.
The Political Science Models

The dominant political science models posit that Supreme Court justices are
principally interested in pursuing favored policies. The attitudinal model
assumes that judges vote reflexively in each case; that is, they cast their votes
based solely on their individual reactions to the facts and legal issues presented,
rather than by considering, in addition, how judges or institutions are likely to
react to the decision (Merrill, 2003, p. 591; Segal & Spaeth, 2002). A second
model, the strategic model, posits that judges take the reaction of others into
account when advancing their policy preferences. A Supreme Court justice, for
example, might calibrate a decision in order to secure the votes of other
justicesso that the Court will embrace a decision that most closely matches
the justices preferred policy outcome (Epstein & Knight, 1998; Maltzman,
Spriggs, & Wahlbeck, 2000). Alternatively, a justice might take implementation
concerns into account and, with it, potential resistance from either elected
officials (Epstein & Knight, 1998; Segal, 1997) or the American people
(Mishler & Sheehan, 1996). In recent years, some political scientists have tweaked
the attitudinal and strategic models. Institutionalists shift their focus away from
the long-standing question of how institutions are affected by the personal
characteristics of judges and toward the question of how judges are affected by
the institutional characteristics within which they are embedded (Gillman,
1999, p. 66). In this way, judges act strategically to pursue both policy and legal
goals (federalism, separation of powers, adherence to precedent). At the same
time, institutionalists focus on an individual justices pursuit of legal policy goals.
The most influential models of judicial behavior share not only a basic
assumption but also a limitation, the lack of a persuasive theory of judges
motivations (Baum, 2006, p. 19; see also Baum in this volume, ch. 1).
Notwithstanding their differences, the attitudinal, strategic, and institutional
models all assume that justices are single-minded maximizers of legal and policy
preferences. Differences between the models turn on whether justices act
The Supreme Court, Social Psychology, and Group Formation 87
strategically and whether justices are pursuing legal or policy objectives. For this
very reason, the median justice plays a central role in all three models. All
models, for example, think that power resides at the medianso that the
most powerful justice is the Justice in the middle of a distribution of Justices,
such that (in an ideological distribution, for example) half the Justices are to the
right of (more conservative than) the median and half are to the left of (more
liberal than) the median (Martin et al., 2005, p. 1277). To pick a simple
example, if the Court is split 5 to 4, the median justice would be the weakest
member of the majority coalition. Under the attitudinal model, the median
justice would only sign an opinion she agreed with and, as such, the majority
might move closer to her position (so long as they too agreed with the final
opinion) or, alternatively, the median Justice might write a consequential
concurring opinion that would limit the reach of the majority or plurality
opinion. The strategic and institutional models likewise see the medians view
as controlling. Not only might the median write a consequential concurring
opinion, but other justices in the majorityfearing possible defectionmight
move their opinion closer to the medians preferred legal or policy position.
The power of the median justice is variable, and that variability will call
attention both to commonalities and differences between the political science
models and a model that makes use of social psychology. For the political
science models, medians are most powerful when there is substantial ideolo-
gical distance between the median and other members of the Courtso that
the median sits between one group of justices substantially to the right and
another group of justices substantially to the left (Epstein & Jacobi, 2008).
During the 2006 term, for example, Justice Anthony Kennedy was a super-
median; among other measures, he was a member of the winning coalition in
each case decided by a 5-to-4 vote. In sharp contrast, medians are least
powerful when their preferences overlap with the preferences of justices to
their right or left. This convergence of preferences, moreover, makes it more
likely that there will be an ideologically simpatico majority coalition of
justices. When this happens, the Court is likely to issue consequential opi-
nions, for a majority coalition sharing great unity of mind has the ability to
adopt whatever rule it would like (Staudt et al., 2008, p. 369).
We agree with these conclusions but nevertheless feel that the political
science models are incomplete because their policy-preference-driven focus is
too narrow and ignores basic psychological concepts. As we discuss below, the
power of the median is diminished on an ideologically simpatico Court
because the median justice is a member of a majority coalition and pressures
toward uniformity will diminish the preferences of any individual justice.
Correspondingly, although median justices are more likely to assume power
on an ideologically diverse Court, the unwillingness of a median justice to join
one or another group is not simply a matter of ideological or jurisprudential
divergence. Median justices do not join groups because they are less interested
in the pursuit of some ideological or legal vision and more interested in
competing values, most notably power and image.
We are not the first to observe that justices think about more than their
legal and/or policy preferences. Lawrence Baum, both in his 2006 study Judges
and their Audiences and in his chapter in this volume, criticizes the leading
political science models for failing to take into account the desires of judges to
win approval from audiences they care about. Noting that the Spock-like
judges of the dominant models have no interest in public approval as an end
in itself, Baum argues that political scientists need to take into account the
commonsense notion that judges, like other people, care a great deal about
what people think of them (Baum, 2006, p. 22). We agree and will discuss
how impression management figures into the willingness of a justice to be
part of a coalition of justices. Unlike Baum, however, the approach taken in
this chapter also applies social psychology to describe the interplay between
the justices themselves.
Social Psychology and Coalition Formation on the Supreme

Court
Before turning to what the psychological literature teaches us about group
formation, let us begin by clarifying our central concept. By a coalition, we do
not necessarily mean a set of justices who vote together all or nearly all the
time. Instead, a coalition of justices is a set of justices who coalesce around an
issue or a set of issues that are highly important or salient to the justices
involved, and who vote and act together in the relevant issue space. This
coalescing need not be a conscious decision made by the justices in the sense
that they consciously choose to form a coalition on a particular issue, but is
rather a recognition on the part of the justices involved of a shared set of goals
or opinions that are salient for each individual justice. Unlike in the dominant
political science models, coalitions of justices are not simply individuals who
share a similar legal and/or policy preference. Instead, social psychology
indicates that where a coalition forms, the very presence of such a subgroup
will have profound effects both on the action of other coalition members and
on the development of the opinions and reasoning of other coalition mem-
bers (Stangor, 2004, p. 3; Cartwright & Zander, 1968, pp. 321). This, of
course, is not to say that legal policy preferences are irrelevant to the forma-
tion of subgroups of justices. An individuals personal beliefs are key to
coalition formation. At the same time, membership in a coalition transcends
the individualized preferences of coalition members.
Importance of Group Formation

When a majority coalition forms, group dynamics play a crucial role in the
Courts decision making. This is because when people align themselves as part
of a group, powerful psychological pressures begin to bear on the members of
the group. The most important of these pressures is the pressure to
uniformity that occurs in groups. Pressures to uniformity in group decision

making have long been recognized as a hallmark of group behavior and they
present themselves in several contexts (Festinger et al., 1968). First, and most
intuitively, membership in a group creates pressure to go along with the
group in order to achieve the goals for which the group was formed initially.
The more important a goal is, the more powerful this pressure is (Cartwright
& Zander, 1968). The amount of pressure to conform to a groups decision
also increases when the members are more dependent on one another in
order to achieve their goals (Festinger, 1968).
There is also evidence that the opinions of group members become more
influential for other group members. Some studies indicate that the opinions of
group members actually converge once the group has made a decision. Even in
situations where consensus among the group is not required, the opinions of
group members are influential to other members as they form their opinions
(Tinsdale et al., 2000, p. 10). Interestingly, group dynamics may actually push
group members to take more extreme positions than they might otherwise be
inclined to take (Stangor, 2004, pp. 202203; Forsyth, 1999, p. 320).
Experiments examining this phenomenon may have special relevance to the
Court as they examined the decision making of people in a judicial setting.
Mock jury experiments indicate that where a group is predisposed to a parti-
cular outcome, discussion of the issues presented to the group has a tendency to
lead the group to adopt more extreme positions than the average group
member held prior to discussing the issues (Stangor, 2003, pp. 202203).
Social Judgment Theory posits that people generally are most persuaded
by positions that are slightly different from the positions they already hold,
but that they are not particularly persuaded by positions that are very
different (Tindale et al., 2000, pp. 910; Kerr & Tindale, 2004, p. 635).
Because members of a coalition on the Court will tend to hold similar, but
not identical, views on a given issue, the opinions of the other members of the
coalition will tend to be more influential to each other than any opinions of
noncoalition members. In other words, when justices associate as a coalition,
the median member of the Court (i.e., the most moderate member of the
group) will be most susceptible to being pulled in a more liberal or conserva-
tive direction by the other members of the coalition instead of by noncoali-
tion members.
Taken together, these psychological processes suggest that Court decision
making may be substantially affected when the justices coalesce in a majority
coalition. First, pressures to uniformity in the group indicate that members of
a group are more willing to sublimate personal preferences, as long as the
members remain committed to the core purposes of the group. Thus, where
there is a majority coalition of justices, the members will be likely to join an
opinion that may be more reflective of the coalitions preference, not neces-
sarily the justices individual preferences.
Further, where there is a majority coalition on the Court, the members
should tend to show greater deference to the opinion writer. The members
should also be more likely to go along with the coalition without issuing a
consequential concurrence. This should be especially true when the issue at
hand is particularly important or salient to the groups core beliefs. And, the
opinions issued by a majority coalition of justices will not reflect the prefer-
ences of the median justice on the Court. Instead, it will be the product of the
group dynamics of the majority and may be a more extreme position than
some members of the group would have preferred on their own.
Finally, depending on the cohesiveness of the coalition, there may be a
willingness among the members to vote together on other issues, provided the
votes on those issues are not central and opposed to a justices personal
beliefs. The more group members see the group as significant, important,
the more likely it is that the group will bond together and stick together
(Stangor, 2004, PIN/24). With respect to Supreme Court justices, norms of
independence (which typically cut against the formation of a cohesive
majority coalition) are likely to limit the willingness of group members to
form a group that cuts across all issues. It is far more likely that the group will
coalesce around a set of core issues and that the justices will act in a more
disparate way on issues that are not central to the groups identity. For
example, the New Deal Courtas we will soon discusswas formed
around the core issue of governmental power to regulate economic condi-
tions. Civil rights and liberties issues were not core to the formation of this
coalition and, not surprisingly, the majority coalition broke apart on civil
liberties questions.
Indeed, the bitterness that subsets of New Deal justices expressed about
each other in connection with civil rights and liberties issues backs up the
central point of this chapter: Although justices can come together to act as a
coalition on one set of core issues, at the same time, the social psychology
barriers that stand in the way of group formation also make it likely that these
justices will splinter on issues that are not central to the groups mission. We
turn to a discussion of those barriers now.
Barriers to Forming a Majority Coalition
From a group dynamics perspective, attraction to a group for a given

individual will depend on his assessment of the desirable and undesirable
consequences attendant upon membership in the group (Cartwright, 1968,
p. 95). For reasons we will now detail, social psychology identifies numerous
roadblocks that stand in the way of a majority coalition forming on the
Supreme Court. The most obvious roadblock (and the one hurdle that
political scientists and social psychologist agree on) is ideological diversity
among the justices. An individual will not act in ways that are inconsistent
with matters central to their cognitive network. In particular, group member-
ship is a basic part of individual self-conception; it is a key component of how
we perceive our place in the world, throughout our lives (Forsyth, 1999,
pp. 6680). Because of the fundamental importance of group identification
in our lives, individuals are only willing to associate themselves meaningfully

with groups that are in sync with their core beliefs (Arrow et al., 2000, pp. 7077;
Stangor, 2004, p. 25).
A second potential barrier to group formation is tied to an individual
justices motivations, specifically, the need for power (Baum, ch. 1 in this
volume). An individuals need to influence others and to control or shape the
world around them, the need for power, is a basic psychological need; but it is
valued differently by different individuals, and it manifests itself differently in
different situations. In some settings, membership in a group may actually
provide an outlet for people with high needs for power because groups
present opportunities for leadership (Forsyth, 1999, p. 92). Also, an ideolo-
gically simpatico coalition may join together in order to decisively advance
the individual preferences of coalition members. However, people with a high
need for power may find it best to refrain from joining a group and instead
play the role of power broker, or decider, between rival factions. And, of
course, for some people the need for power is simply not a sufficient entice-
ment to join any group.
Consider, for example, the so-called swing justices who cast the deciding
votes on controversial cases. Swing justices exercise power by writing
consequential concurring opinions that limit the reach of the majoritys
ruling or by insisting that their legal policy preferences are reflected in the
majority opinion. Like any justice, a swing justice will not cast votes at odds
with core beliefs. But a swing justice might have comparatively weak legal
policy preferences and a comparatively strong desire to exercise power. To
exercise power meaningfully, however, the Court must be ideologically
diverse (Epstein & Jacobi, 2008). An ideologically cohesive Court (with a
majority coalition of 5 or more justices) will not need the swing justices
vote to advance their legal policy preferences. In this situation, the swing
justice might seek to exercise power by joining that coalition in the hopes of
playing a leadership role in that group (assuming that the coalition is acting in
ways consistent with her core beliefs). Alternatively, the swing justice might
not want to join that coalitioneven if that will mean fewer opportunities to
exercise power. For example, the swing justice (or, for that matter, any
justice) might place a high value on external variablesmost notably, how
she is perceived by audiences that she cares about. These audiences might
include journalists, law professors, lawyers groups, other judges and justices,
political parties, interest groups, and even the public (Baum, 2006).
In paying attention to audiences, justices engage in impression man-
agement, that is, the process of controlling how one is perceived by other
people (Leary, 1996, p. 2). Like group dynamics generally, impression
management is a universal phenomenon. Everyone engages in some form of
impression management every day. It is an essential component of social
interaction (Leary, 1996, p. 3). Like an individuals desire to exercise power,
the amount of impression management engaged in by individuals varies
significantly with the situation and the individual. For Supreme Court
justices, there are countless opportunities to take into account their standing
with various audiencesoral argument, opinion writing, the giving of
speeches and interviews, attending social gatherings, so on and so forth. In
other words, judging on the Court is in many ways an exercise in self-
presentation, and the behavior of the justices is shaped in important ways
by the opinions of outside groups that the justices care about. More than that,
the very process by which we select justices tends to favor those with an
especially strong interest in the esteem of other people (Baum, 2006, p. 32).
Accepting a judgeship entails accepting relatively significant constraints on
personal activities and behaviors as well as a significant reduction in monetary
compensation. The inducement for accepting these losses is an increase in
prestige (and an increase in potential power). As a result, the types of people
who end up with judicial positions tend to be those who care a great deal
about the esteem of others.
Impression management figures prominently in the willingness of a
Supreme Court justice to join forces with others and forge a majority coali-
tion. To start, a justice will not join a coalition if that will harm her reputation
among groups that are important to her. Just as a justice will not join a group
that would require her to vote in ways not in sync with her personal beliefs, a
justice will not hurt her standing with groups she cares about. And while some
of these groups may have identifiable ideologies (Federalist Society, American
Constitution Society), externally focused justices are well aware that the norm
of judging in the United States is that the judge is a neutral, impartial arbiter
of disputes. For some (but not all) justices, this norm tends to act as a
disincentive to be part of a unified, ideologically identifiable subgroup of
justices, because people try to project images of themselves that are consis-
tent with the norms in a particular social setting and with the roles they
occupy (Leary, 1996, p. 67). In this way, justices have incentives to act like an
independent judge and not a member of an ideologically identifiable group.
As such, an externally focused judgeespecially as compared to public
officials whose status is tied to political battles that play out in public
viewhas little reason to curry favor with one or another ideologically
identifiable constituency. Justices with strong ideological precommitments,
however, will place a higher value on winning the esteem of some ideologi-
cally identifiable group. For these justices, approval by such groups may
matter more than engaging in self-presentation that is aimed at reinforcing
the norm of neutral, impartial arbiter.
Consider again our so-called swing justice. If all she cared about was
power, she would pay no mind to her reputation. Her decision to join one or
another side of a dispute would simply be an exercise in powerher efforts to
wield as much as influence as possible (either by filing a consequential
concurring opinion or by joining one or the other side of a dispute). In
particular, she would want to maintain her swing justice statusso that her
vote would be critical to the resolution of any dispute. Along these lines, she
would want to locate herself at the Courts median (and, to the extent
possible, distance herself from justices to her immediate right and left)
(Epstein & Jacobi, 2008, p. 7481). An externally focused swing justice,
instead, would focus on how others perceive her. Perhaps she would cultivate
a reputation of neutrality; perhaps she would want to be known as the critical
vote; perhaps she would want groups with disparate ideologies to view her
vote as gettable. Whatever her methodology or motivation, the externally
focused swing justice will place a high value on cultivating a positive image
with groups that do not demand ideological conformity.
Indeed, the desire to appear independent may prompt some justices to
engage in a type of behavior known as reactance. Reactance speaks to the
desire of individuals to resist challenges to their autonomy (Brehm & Brehm,
1981). In particular, when people feel their independence is threatened, they
will take steps to demonstrate that they are in control of their own behavior.
For example, the Supreme Courts 1992 reaffirmation of abortion rights in
Planned Parenthood v Casey may well be tied to the desires of Justices
OConnor, Kennedy, and Souter to demonstrate that they were not the
political lackeys of the presidents (Reagan and Bush I) who appointed them
to the Court. Proclaiming that the Courts legitimacy is tied to its ability to
withstand political attacks, these justices made clear that they would not
facilitate efforts by the Reagan and Bush administrations to push for the
overruling of Roe v. Wade. Taken together, these psychological concepts
illustrate some of the difficulties of forming a majority coalition on the
Court. A justice, of course, will not choose to join a coalition if doing so
means they have to cast a vote on a core issue that does not match her central
beliefs. In addition to legal and/or policy preferences, the desire for power,
impression management, and reactance may all contribute to a justices
refusal to join a coalition. In other words, even if a justices legal policy
preferences are largely in sync with an existing subgroup on the Court, a
justice might not join it. Put another way: Without strong ideological pre-
commitments to a particular group, Supreme Court justices are likely to value
power and image in ways that make them resistant to forging a majority
coalition.
On the other hand, justices with strong ideological precommitments may
be especially likely to join coalitions. Members of ideologically simpatico
coalitions will agree with each other on issues of high salience to coalition
members; consequently, they will more likely seek to assume power by
forcefully advancing a shared agenda. In other words, members of such a
group have less interest in exercising individualized power by casting the
decisive swing vote; for them, the pursuit of a shared agenda is the most
important manifestation of power. Likewise, justices with strong ideological
precommitments may be less interested in fostering the norm of an impartial,
independent jurist. Rather, when it comes to impression management, the
outside groups they care about are those who share their values and objec-
tives. Compare, for example, Justices Anthony Kennedy and Clarence
Thomas. Kennedyconsistent with swing justice behaviorplaces a high
value on the opinions of the news media and other elites; Thomas identifies
closely with ideologically conservative groups (Baum, 2006, pp. 132, 142
144). If there are 5 or more ideologically simpatico justices, a majority
coalition may form. The key variable, as noted above, is whether these justices
have sufficiently strong ideological precommitments to overcome the basic
obstacles to group formation. For example, in determining whether a justice
will join a group, it may require more than the justice agreeing with other
members on the preferred outcome and legal reasoning in any given issue
space. A justice not strongly precommitted to the groups agenda may place a
higher value on the exercise of individual power or cultivating a reputation
for judicial independence. Perhaps for this reason, Justice Anthony Kennedy
broke ranks with the Rehnquist Courts conservative bloc by switching his
initial conference votes in high visibility school prayer and abortion cases.
(Greenburg, 2007, pp. 145160).
The appointments-confirmation process also stands as a substantial
obstacle to the formation of an ideologically simpatico majority coalition,
especially with regard to controversial, highly salient issues. Because justices
have life tenure, it is very unlikely that appointments to the Court will be
clustered closely together. Such clustering of appointments facilitates group
formation (Arrow et al., 2000, p. 69). In the case of the Court, this is both
because people who join an existing organization tend to identify with others
who join at the same time and because such clustering means that the same
president and Senate will be making the appointments, increasing the like-
lihood of clustered appointees being relatively closely aligned ideologically.
For example, as we will discuss near the end of this chapter, President
Roosevelts clustering of Supreme Court appointments from 1938 to 1943
figured prominently in the New Deal Courts dramatic expansion of govern-
ment power over the economy. At the same time, this perfect storm of closely
clustered appointments and other factors that would help overcome the
barriers to group formation rarely occurs.
Applying the Psychological Perspective

Social psychology explains both the ramifications of group formation on the
Supreme Court and the innumerable roadblocks that typically stand in the
way of group formation. When there is no dominant majority coalition on
the Court, social psychology suggests that concerns of power and image
(including reactance) stand in the way of justices voting their true legal
and/or policy preferences. And when there is an ideologically cohesive
majority, social psychology suggests that intragroup dynamics will play an
important role in defining the Courts decision as well as the willingness of
justices to stick with the coalition on issues that are not core to the groups
identity. This section will provide a preliminary test of the social psychology
model. We will compare the willingness of the Rehnquist and New Deal
Courts both to overrule precedent and to issue consequential rule-like (as

opposed to minimalist fact-specific) decisions. For both Courts, we will focus
on two issue setscongressional power and individual rights.
Before turning to our discussion of these two Courts, two clarifying
comments: First, even though this paper highlights significant differences
between the social psychology and dominant political science models, these
models overlap in significant respects. Most important, just as political
science models talk about the pursuit of legal policy preferences, social
psychology likewise talks about the importance of personal beliefs to an
individuals willingness to join a group. For this very reason, it is often the
case that the social psychology model and the political science models will
both point to personal beliefs as a principal motivation for a justices deci-
sions. More to the point, the social psychology and political science models
both anticipate that the Court is more likely to generate consequential pre-
cedents when there is an ideologically simpatico coalition of five or more
justices. Likewise, when there is no such coalition, each of these models
recognizes that the median justices views are often controlling. At the same
time, social psychology provides a much more nuanced explanation for
Supreme Court decision making. That explanation has strong empirical
foundations and, as such, we think that political scientists must do more
than demonstrate the predictive powers of their models. They must also
explain why Supreme Court Justices do not function like other individuals
who operate in a group dynamic. Second, in discussing the Rehnquist and
New Deal Courts, our objective is quite limited. Specifically, we want to see if
these two Courts superficially track the social psychology model discussed in
the preceding section. A more detailed, empirical assessment still needs to be
doneand we hope to do that in another paper. For reasons we will now
discuss, Rehnquist and New Deal Court decision making seem to follow the
social psychology model discussed in this chapter.
The New Deal Court

The New Deal Court (19371949) was, in critical respects, two Courts. On
issues involving Congresss power to regulate the economy, an ideologically
simpatico majority coalition operated as a cohesive group. Those issues were
central to the groups identity. On individual rights issues, however, the Court
was anything but coherent. These issues, while of great national significance,
were not central to the groups identity.
To start, the New Deal Court was forged by President Franklin Delano
Roosevelt. Roosevelt used his appointments power to nominate eight justices
during a five-year period, 19381943. More than that, Roosevelt used his
appointments power to celebrate the New Deals embrace of big government,
especially the power of government to regulate the economy. Roosevelt felt
compelled to do so because the pre-1937 Supreme Court had taken the
country back to its horse and buggy days by overturning several New
Deal initiatives; indeed, Roosevelt promisedwhen introducing his ill-fated

Court-packing planto appoint justices who will not undertake to override
the judgment of Congress on legislative policy (quoted in Devins & Fisher,
2004, p. 61).
Roosevelt did just that; his appointees were committed New Dealers who,
from the moment they joined the Court, advanced an expansive view of the
federal governments power to regulate the national economy. From 1937 to
1944, the New Deal Court had created a new constitutional order, over-
ruling thirty casestwo thirds as many as had been overruled in the Courts
previous history (Leuchtenburg, 1995, pp. 208215). Over the course of its
twelve-year tenure (19371949), the Court throroughly repudiated the
entire doctrinal system of constitutional limitations of federal power over
the national economy (Ackerman, 1999, p. 47). It handed down 42 rulings
that overturned at least 59 of its prior decisions. The majority of these
decisions had broad supportonly five were decided by a 5-to-4 vote (as
compared to 10 unanimous overruling decisions).
Group dynamics, as well as the legal policy preferences of the justices,
likely figured into New Deal Court decision making. As discussed earlier,
justices who are part of an ideologically simpatico majority coalition seek
power by voting with the coalition. Likewise, rather than cultivate an image of
impartiality by refusing to join a coalition, justices who are part of a majority
coalition pay attention to audiences that agree with the core agenda of that
coalition. Perhaps most significant, justices on an ideologically simpatico
majority do not necessarily vote their personal preferencesinstead, they
allow the group dynamic to shape their final vote.
Consider, for example, the New Deal Courts 1942 decision in Wickard v.
Filburn. Wickard concerned the power of the secretary of agriculture, acting
pursuant to the Agriculture Adjustment Act, to extend a quota on wheat
production to a farmer who grew wheat for home consumption. In upholding
the secretarys power, the Court issued a sweeping opinionruling that
Congress may regulate economic conduct trivial by itself so long as the
aggregation of similar activity by other actors affects interstate commerce
(Wickard v. Filburn, 1942, pp. 127128). For our purposes, Wickard is
especially instructive because some justices on the Court put aside personal
misgivings about the decisions reach in order to forge a pathbreaking ruling
that reflected the core beliefs of the New Deal.
Before Wickard, the Court encouraged Congress to make findings that
commerce indeed was affected. In this way, the justices placed the ball in
Congresss court, for once Congress found facts, it would be very difficult for
the Court to meaningfully check Congress. Nonetheless, in the years pre-
ceding Wickard, Congress contributed to the Courts approval of New Deal
initiatives through its sustained and thoughtful showing that there was, in
fact, an integrated national economy (Frickey, 1996, pp. 711712). When
Congress enacted the Agriculture Adjustment Act, however, lawmakers made
no factual findings. For this very reason, Justice Robert Jackson, who had been
tasked to write the decision, initially drafted an opinion that would have
remanded the case so that a trial court could make additional factual findings
(Cushman, 2000, p. 1138). Jackson nevertheless backed away from his ori-
ginal opinion and wrote a decision that effectively granted Congress carte
blanche power to regulate anything arguably economic. In private correspon-
dence, Jackson signaled his discomfort with his handiwork. Recognizing that
we no longer have legal judgment upon economic effects which we can
oppose to the policy judgment made by the Congress in legislation,
Jackson observed: I really know of no place . . . where we can bound the
doctrine (quoted in Cushman, 2000, pp. 1143, 1145).
Wickard exemplifies what a coherent Court can do. Committed to a
shared agenda, group members can work together to advance an expansive
vision of the law. Wickard also stands in sharp contrast to New Deal Court
decisions on individual rights. Unlike economic issues (which were core to
the groups formation), civil and individual rights were irrelevant to the
formation of the New Deal Court. Roosevelt wanted justices who would
validate the regulatory state; he was not especially interested in constitutio-
nalizing civil liberties and civil rights. At the time of Court-packing, the
Courts docket had almost no cases implicating civil and individual rights.
But with the Courts approval of sweeping legislative power over economic
issues, the Court inevitably turned its attention to other matters. Reflecting
both changing social conditions and their personal interest in asserting
power, judges created for themselves a new role in the political system, one
that involved identifying those preferred freedoms or suspect classifica-
tions that might provide a basis for trumping the otherwise unrestrained
power of the modern legislature (Gillman, 1993, pp. 202203). Here, the
New Deal Justices dividedreflecting the fact that groups organize around
clusters of core issues, that justices will not vote against their legal policy
beliefs on issues of consequence, and that the norm of impartiality pushes
justices away from groups that do not share their core beliefs. In other words,
just as social psychology helps explain why the New Deal Court acted as a
coherent group on economic questions, social psychology is also useful in
understanding why the justices were unwilling to forge a majority coalition
on issues involving civil and individual rights.
The Rehnquist Court

The Rehnquist Court (19862005) likewise exemplifies the forces that push
against group formation on the Supreme Court. Throughout its history, the
Rehnquist Court was fractured on issues involving civil and individual rights.
But even its much ballyhooed efforts to reinvigorate federalism-based limits
on congressional power proved to be a bustprincipally because a majority
coalition was never able to coalesce around these issues. The inability of the
Rehnquist Court to fundamentally transform doctrine, as we will now
explain, is to be expected. Without five justices strongly committed to the
pursuit of some shared agenda, concerns of power, impression management,

and reactivity stand in the way of group formation.
On civil and individual rights issues, the Rehnquist Court seemed destined
to embrace Reagans vision of judicial conservatism. When running for pre-
sident in 1980 and 1984, Ronald Reagan both pledged to appoint judges who
share our commitment to judicial restraint and reached out to social conser-
vatives by condemning Supreme Court decisions on school prayer, busing, and
especially abortion (Devins & Fisher, 2004, quoting Republican party platform).
But two of Reagans four nominees, Sandra Day OConnor and Anthony
Kennedy, refused to embrace the social conservative agendaso much so that
Republican domination of the Court did not result in the overruling of a
single revolutionary Warren [or Burger] Court decision (Nagel, 2006).
On social issues, Justices Kennedy and OConnor were anything but
precommitted to the social conservative agenda. Reagan picked OConnor
to fulfill his pledge to nominate the first woman to the Supreme Court.
Accounts of his decision to nominate her make clear that ideology was not
central to Reagans decision (Toobin, 2007, pp. 1718). Kennedys selection is
even more telling. Reagan initially nominated Robert Bork for that seatbut
civil rights and abortion rights groups strenuously objected to that nomina-
tion and the Senate rejected Bork. Reagans second choice, Douglas Ginsburg,
withdrew from consideration after newspapers revealed embarrassing per-
sonal details. Kennedy was selected to stave off further embarrassments;
ideology entered the calculus but it was not figural in Kennedys nomination
(Greenburg, 2007, pp. 3565).
Kennedy and OConnor repudiated the social conservative agenda by,
among other things, voting to reaffirm earlier rulings on school prayer and
abortion rights. Reactance may well have been a contributing factor to these
decisions. OConnor and Kennedy also acted in ways that expanded their
personal power and fostered their reputation for judicial independence.
Kennedy, in particular, seemed concerned with his public persona. His
decisions to reaffirm Court rulings on school prayer and abortion rights
may not have reflected his true preferencesbut, instead, his desire to
exercise power in ways that would distance himself from the Reagan admin-
istrations social conservative agenda. According to one of his law clerks,
Kennedy would constantly refer to how its going to be perceived, how the
papers are going to do it, and how its going to look (Tushnet, 2005, p. 176,
quoting an anonymous Kennedy clerk). On the very day that the Court
reaffirmed Roe, Kennedy told a reporter, [s]ometimes you dont know if
youre Caesar about to cross the Rubicon or Captain Queeg cutting your own
tow line. (quoted in Greenburg, 2007, p. 159). Kennedy, moreover, seemed
determined to occupy the pivot on the Court. According to one account,
Kennedy sought to maneuver himself to the centerand even boasted of
employing this strategy (Lithwick, 2004, p. 25; Lazarus, 1998, p. 515).
Kennedys concerns for power are further revealed in a 2005 interview; he
spoke about Supreme Court justices shap[ing] the destiny of the country,
noting that in any given year, we make more important decisions than the
legislative branch does (quoted in Rosen, 2007b, p. 17).
For her part, Justice OConnor made extensive use of fact-specific con-
curring opinions to keep her options open in future cases and, more impor-
tantly, to tell litigants that the outcome of a case goes through herso
much so that litigants spoke about writing for an audience of one when
crafting Supreme Court briefs (Brust, 2005, p. 37; Estrich & Sullivan, 1989,
p. 119). As Justice Sandra Day OConnor votes, so goes the Court, wrote
one commentator, and it is undeniable that OConnor was aware of both the
power she wielded and her legacy as the Courts first women justice (Lazarus,
2000). OConnors flexible, context specific approach was most pro-
nounced in cases implicating civil and individual rights (Maveety, 1996,
p. 31). In a prominent voting rights case, OConnor filed a concurrence to a
decision she authored (Bush v. Vera, 1996, pp. 990995). When concurring to
a decision rejecting a constitutional right to physician assisted suicide,
OConnors reasoning fundamentally limited the majority opinionso
much so that Justice Stephen Breyer joined the concurrence except insofar
as it joins the majority(Washington v. Glucksberg, 1997, p. 789). Whatever
her motivations, OConnor did not want to be pinned down. She wanted to
make her mark through individuated fact-specific decisions of limited reach,
decisions that would make her the focal point of subsequent cases.
Without a solid coalition of five ideologically simpatico justices, the
Rehnquist Courts civil and individual rights legacy was inconsequential.
The Court did not make a single move that would radically change or
unsettle existing constitutional doctrine (Friedman, 2002, p. 146). The
Rehnquist Courts federalism revival, for the most part, tells a similar story.
Unlike civil and individual rights, the Rehnquist Court did pursue doctrinal
innovations on federalism (Merrill, 2003, p. 58486). More than that, com-
mentators initially labeled a group consisting of Justices OConnor, Kennedy,
Scalia, Thomas, and Chief Justice Rehnquist as the federalism five. But the
federalism revival, ultimately, was more bust than boom; the Court over-
turned only one significant precedent and, ultimately, backed away from its
campaign to limit congressional power under the Commerce Clause and
section 5 of the Fourteenth Amendment. In cases decided in 2003, 2004,
and 2005, four of the five so-called federalism five distinguished earlier
Rehnquist Court rulings in order to back up congressional power. The only
justice to consistently vote in favor of limits on Congress was Clarence
Thomas.
The failure of the federalism revival is tied to the simple fact that feder-
alism-qua-federalism was never a core issue to the so-called federalism five.
Presidents Reagan and Bush never used federalism as a measuring stick
when screening candidates; the Senate paid no mind to federalism during its
confirmation hearings. The focus, instead, was on first-order policy issues
race, privacy, religion. Unlike the New Deal era (where Court limits on
congressional power frustrated Roosevelts pursuit of a fundamental
restructuring of the regulatory state), elected officials neither pushed for nor
resisted Rehnquist Court efforts to place some federalism-based limits on
congressional power (Devins, 2004). Against this backdrop, it is not sur-
prising that a core group could not form around this low salience issue and,
in so doing, invalidate laws that they otherwise supported.
For our purposes, the Rehnquist Court highlights the various roadblocks
that stand in the way of group formation on the Supreme Court. Groups form
around core issues and, in part, that requires the appointment and confirma-
tion of justices who are precommitted to the pursuit of some agenda.
Otherwise, median or swing justices will resist banding together with
other justicesfor these swing justices are likely to place a high value on
power and/or their image. Indeed, the Reagan administrations embrace of
the social conservative agenda may well have boomeranged, in that, swing
justicesconsistent with reactancefelt that their independence was threa-
tened by the administrations assault on the Court.
Conclusion
Social psychology provides important insights into group formation on the
Supreme Court. In particular, unlike political science models, which emphasize
the pursuit of legal and policy preferences, social psychology highlights the
importance of group processes and how issues of power and reputation also
contribute to group formation on the Supreme Court. In so doing, social
psychology suggests that political scientists overemphasize the median justice
benchmark. When a majority coalition forms, intragroup dynamics define the
scope of the Courts ruling. Those dynamics reflect group preferences, not
the preferences of the median justice. And when there is no majority coalition,
the median justice may well be influenced by concerns of power and reputa-
tionconcerns that may lead the median justice to vote in ways that do not
necessarily reflect her true legal policy preferences. Through limited case studies
on the New Deal and Rehnquist Courts, there is reason to think that justices
like other humansoperate within the boundaries of group dynamics. That, of
course, is not to denigrate the profoundly important role of legal policy
preferences. Justices, according to the social psychology model, will never cast
votes that do not jibe with their core beliefs. At the same time, the dominant
political science models offer a too simplistic picture of Supreme Court deci-
sion making.
Note
Thanks to David Klein, Greg Mitchell, Lee Epstein, Larry Baum, and especially John
Nezlek.
Part II
JUDGING AS SPECIALIZED ACTIVITY
7
Is There a Psychology of Judging?
Frederick Schauer
In the United States, as in most countries, judges share three prominent

characteristics. First, and tautologically, they are judges. Second, and with
the exception of the lay magistrates who hear small cases in many states, they
are lawyers. And third, the opinions of some attorneys and litigants notwith-
standing, they are human beings. My goal in this paper is to examine in a
preliminary way the relative contributions of each of these three character-
istics in explaining judicial cognition and judicial behavior.
The potential value of such an inquiry lies in its contrast with the (small)
existing literature (e.g., Guthrie, Rachlinski, & Wistrich, 2001; Wistrich,
Guthrie, & Rachlinski, 2005) on the psychology of judging.1 That literature,
with few exceptions, aligns itself with the conclusion that it is the third and
not the first or second of the items on the above listthe judge as human
being and not the judge as judge or the judge as lawyerthat has the greatest
explanatory power in accounting for judicial behavior, and that holds out the
greatest promise for setting a research agenda for law and psychology and for
the psychology of judging (Spellman, 2007). More often implicitly than
explicitly, the existing research tends to support the view that a judges
attributes as a human being reveal more about the psychology of judging
than does anything a judge might have learned in law school, acquired in the
practice of law, or internalized by virtue of serving in the judicial role.
The conclusion that judges share (some) important decision-making
characteristics with their fellow human beings is occasionally supported by
empirical findings (Guthrie, Rachlinski, & Wistrich, 2001; Wistrich, Guthrie,
& Rachlinski, 2005). More often, however, this conclusion lurks in the
103
104 Judging as Specialized Activity
background as an undocumented and unargued premise of the research on

the psychological dimensions of judicial behavior. Researchers commonly
assume that what is known about human decision making and cognition will
apply to judges, and thus conclude that nonjudge experimental results can be
applied to explain and predict judicial behavior. One survey of (nonjudicial
and nonlegal) analogy research (Holyoak, 2005), for example, asserts that the
legal systems use of precedent is but a formalized application of the nonlegal
and nonjudicial human practice of analogizing, while another study (Simon,
Krawczyk, & Holyoak, 2004) describes two of the authors earlier study using
undergraduate research subjects (Holyoak & Simon, 1999) as being about
legal decision making. And in the law review literature, it is routine to take
the teachings of contemporary cognitive and social psychology as substan-
tially applicable to the decisions of lawyers and judges (Arlen, 1998; Hanson &
Yosifon, 2004; Schauer, 2006a, 2006b; Simon, 1998, 2002, 2004).
Although applying the research on lay decision makers to judges is thus
relatively common, research on real judges has to date been quite limited.2
And even when there has been serious research on the psychological dimen-
sions of actual judicial decision making, it is of less pervasive value than it
might be because the research has focused almost exclusively on the fact-
finding3 and verdict-rendering dimensions of the judicial role. Judges are
indeed often required to determine simply what happened, and then, in place
of a jury, are often in the position of delivering a final verdict. Thus, judges
must frequently decide which of multiple opposing factual accounts is most
likely true.4 And in engaging in such tasks, judges perform functions similar
to those performed by a jury.5 So insofar as people tended to believe that
judges would be superior to jurors as fact-finders or verdict-renderers, or to
believe that judges would be largely immune from the cognitive biases of mere
mortals, much of the current research on the psychology of judging has
usefully cast doubt on the view that judges by virtue of their intelligence or
legal training or judicial position could significantly outperform juries with
respect to the same fact-focused inquiries.
By concentrating so dominantly on the fact-finding and verdict-
rendering tasks that judges share with jurors, however, the existing research
tends to slight those aspects of judgingmost obviously selecting the relevant
law, interpreting the law, and sometimes making lawthat are more or less
the exclusive province of the judge. Because judges thus appear to take on
many tasks that jurors and everyday decision makers do not, and also because
judges likely possess some characteristics that experimental subjects do not,6
perhaps the conclusion (or the assumption) that judicial decision making is
substantially similar to the decision making of those who are not judges is
open to question. Just as it would be a mistake to conclude very much about
the mathematical reasoning of the Harvard mathematics faculty from studies
about how ordinary people make mathematical calculations at the super-
market or when balancing their checkbooks, so too might it be a mistake to
draw conclusions about how judges reasons with rules and precedents and
Is There a Psychology of Judging? 105
authorities from the way in which the man on the Clapham omnibus7 deals
with similar inputs into and constraints on his decision-making processes.
And if it is a mistake to move too quickly from what we know about how lay
people perform certain tasks to conclusions about how judges perform some
of those same tasks, it certainly would be a mistake to draw conclusions about
how judges perform a range of judge-specific tasks from what we have found
about how lay people perform quite different tasks.
Thus, one questiona question and not a conclusionis whether the
experience of studying to be a lawyer and then of practicing law causes
decision making in law, especially about legal (as opposed to factual) matters,
to diverge in deep and cognitively substantial ways from the decision making
of human beings who do not possess such training and experience.8 And a
further question is whether those who self-select to be judges, who are selected
as judges, and who have the experience of serving as judges make decisions
differently from nonjudge lawyers, thus causing further gaps between judicial
decision making and the decision making even of similarly trained and
experienced people holding different roles.9 Consequently, there are hypoth-
eses worthy of investigation about whether in law-focused decision making
there are divides between lawyers and people in general, between judges and
nonjudge lawyers, and consequently even larger divides between judges and
people in general.
Indeed it is likely that multiple phenomena are at work. Self-selection
into law, subsequent legal training, subsequent legal experience, self-selection
into judging, and then finally serving in the judicial role may all interact with
each other to produce considerable differences between how judges and lay
people reason and decide. To the extent that this is so, the interaction among
legal training, legal acculturation, legal experience, and the judicial role may
even generate process- and not just content-based differences between the
cognitive mechanisms of judges and those of nonjudge humanity. If so, there
may be differences, at least with respect to some highly important judicial
tasks, between how judges and lay people think and not merely differences in
what they think about.
The battery of possibilities offered in the previous paragraphs is no more
than an array of testable hypotheses. If even some of these hypotheses turn out
to be true, however, then there actually may be a genuine psychology of
judging. But if on the other hand these hypothesized differences between
judges and lay people turn out not to exist, and if instead the assumptions and
premises of judge as human being lying behind most of the existing research
are sound, then research into the psychology of judging will be an interesting
application of larger psychological issues, but will not in any fundamental way
constitute a discrete area of inquiry. If the most important or only determi-
nant of judicial decision-making characteristics is the fact that the judge is
human, after all, then a psychology of judging will be little different from a
psychology of dentistry or a psychology of plumbing. It would be interesting
and possibly even important to know what psychology could teach us about
how dentists and plumbers think, but the application of research findings
about human beings generally to the human beings who fill these socially vital
roles is a long way from saying that there is a psychology of dentistry or
plumbing. Perhaps the psychological dimensions of judging are different
from those of dentistry or plumbing, but we will not know that unless we
depart from the assumption that what we know about people is necessarily
applicable to judges. And because I suspect that there might be more to the
psychology of judging than there is to the psychology of dentistry or
plumbing,10 and because I suspect as well that there are reasons to believe
that legal and judicial attributes may cause judicial decision making to depart
in relevant ways from the decision making of lay people, my aim in this paper
is to examine in a preliminary and nonempirical wayhypothesis offering
but not hypothesis testingwhat a genuine psychology of judging might
look like, and why, most of the existing literature on the psychology of
judging notwithstanding, we ought to take this possibility seriously.11
The Promises and Premises of Legal Reasoning

Lord Coke wrote of the artificial reason of the law (Coke, 1628/1985; Fried,
1981) hundreds of years before even the advent of university-based formal
training in law. Now that such training is ubiquitous, Lord Cokes premise is
more important yet, because the view that there is a special reason of and for
law has become the guiding principle for the vast numbers of American law
schools and their equivalents12 in other countries. These schools purport to
teach their students the mysterious art of legal reasoning, and they hold out
the hope that at the end of law study a student will have learned how to think
like a lawyer (Schauer, 2003, 2004a, 2009).
The belief that thinking like a lawyer is fundamentally different from
simply thinking has declined a bit in the past several generations, but not
much. Law schools these days pay more attention than previously to philo-
sophy, literature, economics, and the empirical social sciences, among others,
but they have scarcely abandoned their commitment to there being such a
thing as legal reasoning, to legal reasoning being a somewhat autonomous
skill, and to the responsibility of law schools to inculcate this skill in those
who would be lawyers and judges. Moreover, law schools subscribe to the
view that legal reasoning is not easily picked up on ones own, and that formal
training and subsequent experience in thinking like a lawyer can and char-
acteristically do produce a genuinely transformed method of thinking,
reasoning, arguing, and decision making.
Although I will discuss presently my view of what legal reasoning just is,
I want to be careful not to overstate the claim about the alleged distinctiveness
of legal reasoning. Law schools and the legal culture do not typically maintain
that legal reasoning is totally or even almost totally unconnected with
ordinary reasoning,13 in the way that Estonian is unconnected with English,
for example, or that literary interpretation is unconnected with multivariate

calculus. Nor could they. Rather, the most plausible version of the claim to
distinctiveness in legal reasoning would be the comparative statistical claim
that some of the methods of reasoning that are located outside of legal
reasoningarguments from precedent,14 reasoning from rules, and reliance
on authority, most prominentlyare more highly concentrated in legal
argument and decision making than in ordinary reasoning and decision
making, the difference being sufficiently great as to support the conclusion
that legal reasoning is, in the aggregate, substantially unlike the kind of
reasoning that takes place in other decision making domains. So although
lawyers and judges necessarily employ nonlegal forms of reasoning as they
argue cases and make decisions, and although nonlawyers occasionally make
use of the characteristic modalities of legal argument, the concentration of
these modalities in legal argument is so great, the law schools view of the
world appears to maintain, as to justify the claim that something very
different is going on when lawyers and judges tackle a problem or face a
decision.
So what then is legal reasoning, or at least what is it alleged to be? What is
it to think like a lawyer, as opposed to just thinking? These are neither easy
nor uncontroversial questions, and so we find in the literature on legal
reasoning and argument the claims that legal thinking is about the ability to
seek and do justice in the individual case (Bartlett, 1990; Burton, 2005
Henderson, 1987; Minow & Spelman, 1990; Solum, 1988; Sunstein, 1996,
1999), or about the capacity for self-critically seeing and appreciating view-
points opposed to ones own (Sherry, 2007), or about a tendency toward
clarity and analytic precision (Sherry, 2007; Vandevelde, 1996), or about a
talent for understanding and dealing with facts (Bandstra, 2005), or about a
facility in argument and debate (Bandstra, 2005), or about the capability of
engaging in analogical reasoning (Brewer, 1996; Levi, 1948; Weinreb, 2005).15
Yet although there can be little doubt that all of these skills and many more are
necessary for successful lawyering (and judging), and equally little doubt that
good lawyers tend to have them in greater abundance than poor ones, these
are not skills that seem especially of greater importance for lawyers than they
are for police officers, physicians, social workers, politicians, and countless
others. Most of these skills, even including the skill of analogical reasoning,16
are domain-general reasoning abilities, and while lawyers may on average be
better at some of them than other people, it is probable that any differential
ability with respect to these and similar tasks is explained almost entirely by
the fact that lawyers are on average somewhat better educated, smarter, and
possibly even more motivated than the population at large.
But although many of the posited components of legal reasoning
are neither unique to nor even much concentrated in lawyers and legal
argument, there is one form of reasoningor one cluster of associated
forms of reasoningthat can plausibly be understood to set lawyers apart
from others, and it is one that can be described as second-order reasoning (Raz,
1979; Schauer, 1991, 2004b; Sunstein & Ullman-Margalit, 1999). When

engaged in ordinary (first-order) reasoning and decision making, people
tend, not surprisingly, to try to make the best decision for the problem or
task at hand. Their aim is typically to reach the right result for this casethe
present case. That this is so for ordinary people, however, is not to say that it is
so for lawyers and judges, for one of the things that law schools attempt to
teach their students is precisely to avoid thinking that the right result for this
present case is necessarily the right result all things considered. So consider,
for example, the typical allegedly Socratic17 dialogue that takes place between
student and teacher in the first year of law school. After eventually being
coaxed into reciting the facts of some reported case correctly and accurately,
the student is then asked what the correct result should be for the present case,
and she commonly responds by announcing what she believes to be the most
fair or just outcome as between the opposing positions of the particular
parties. At this point the student is asked to give the rule or principle that
would support this outcome, and here the characteristic pattern of Socratic
inquiry begins. By a series of patterned and well-planned hypotheticals, the
professor challenges the students initially proffered rule, with the aim of
demonstrating that the rule that would generate a just or fair or efficient
outcome in the present case would generate less just, less fair, or otherwise less
satisfactory results in other cases. And in taking the student through this series
of uncomfortable applications of the students initially chosen rule, the
professor attempts to get the students to understand that the best legal rule
may be one which produces an unjust result in the present case, but which will
produce better results in a larger number of cases, the result in the present case
notwithstanding.
This form of Socratic inquiry is not restricted to the law school class-
room, and it is noteworthy that it is the common form of judicial questioning
in appellate argument.18 Because appellate courts see themselves as setting
forth rules that will control other and future factual situations and as writing
opinions that will serve as precedent for future cases, appellate judges often
focus as much on the effect of this ruling on future cases as on reaching the
best result in the present case. As a consequence, appellate advocates often
find themselves asked how the rule or result they are advocating will play out
in various hypothetical cases. As in the law school classroom, these hypothe-
tical situations are offered against the background of the view that the right
result in the particular dispute before the court will only be the actual out-
come if it can be justified in a way that will not produce the wrong outcomes
in too many expected future cases (Golding, 1963; Greenawalt, 1978).19
In seeking to demonstrate to the hapless student or struggling advocate
how the best legal outcome may be something other than the best outcome in
the immediate case, the prototypical Socratic interrogation aligns itself with
an even more important dimension of legal reasoning and argument, the way
in which the backward-looking, constraining, and limiting dimensions of law
(Levi, 1948; MacCormick, 2005; Raz, 1979; Wasserstrom, 1961, pp. 2526)20
often mandate a result other than the one that is optimally fair or maximally
wise, all things considered, in the particular casea result that will some-
times be wrong for the particular dispute (Sherwin, 1999). It may seem
unfair on the balance of all reasons to deprive a person of property (United
States v. Locke, 1985) or a place on a ballot (Hunter v. Norman, described in
Schauer, 1988a) just because he has missed a statutory deadline for under-
standable, innocent, and ultimately inconsequential reasons, but the law
characteristically even if not universally enforces the literal meaning of
authoritative language even when such an action produces a bad outcome
in the particular case (Manning, 2003; Schauer, 1992). And it may seem
equally unfair to take the existence of clear precedent as commanding a
suboptimal result, especially from the vantage point of a decision maker
who thinks the precedent mistaken, but following even a precedent perceived
to be erroneous is what, under the traditional understanding, the law expects
its decision makers to do.21
The second-order reasoning I describe here is not about what is, but
instead about what to do. The law must frequently engage in factual inquiry to
determine who fired the gun, how much toxic waste was discharged into the
river, whether someone was in possession of inside information when they
purchased a quantity of securities, or whether the driver of some car had
consumed alcohol prior to being involved in an accident. But in the law such
factual determinations are typically precursors to a judgment about what the
law requires to be done on the basis of these facts; and what the law requires to
be done may be something other than that which a nonlegal decision maker
would decide, all things considered, should be done. So although the legal
system engages in factual inquiry, it is precisely in moving from factual
inquiry to action-producing consequence that legal reasoning potentially
differs in fundamental ways from the reasoning of other action-producing
agents. These other agents, we typically think, are focused on producing the
right decision for this decision making event, but if the characteristic deci-
sion-making modality of law is different from the decision-making modal-
ities of other domains, than legal reasoning and decision making may be
different as well. Legal reasoning, on this widespread account, is artificial
not only because of the way in which it is deliberately not focused on reaching
the best for just this case, but also because decisions having legal consequences
differ for just this reason from the practical reasoning in which nonlawyers
ordinarily engage.
The Realist Challenge

Although a widely believed traditional conception of legal reasoning is con-
sistent with the foregoing account,22 the descriptive accuracy of the tradi-
tional conception has hardly gone unchallenged. More particularly, an array
of perspectives collected under the heading of Legal Realism23 can be
understood as presenting not only a challenge to the traditional conception of

legal and judicial reasoning, but also, and more germane here, a challenge to
the view that judicial reasoning is substantially different from the reasoning of
ordinary people.24 For the Realists, judges were less different from people in
general, and from other public and private decision makers, than the tradi-
tional artificial reason view maintained. The Realists saw legal reason as
human reason, and thus traditional claims for the distinctiveness of legal
reasoning were to the Realists largely pretense. Because the Realists therefore
implicitly (and sometimes explicitly) subscribed to the view that judges were
best seen simply as human decision makers with few distinctive methods,
most of the existing research on the psychology of judging can be understood
as incorporating an unexpressedand typically unresearchedLegal Realist
outlook on what judges do and how they do it.
The connection between Legal Realism and the psychology of judging can
be traced to the earliest days of Realism. In Law and the Modern Mind (Frank,
1930), Jerome Frank, then in the aftermath of his own recent psychoanalysis,
claimed that it was impossible for judges to engage in the second-order
decision making then and now associated with the traditional account of
legal reasoning.25 Frank argued that judges, like other human beings, invari-
ably trained their attention on the facts and details of this particular case, and,
moreover, unavoidably strove to make the best decision for this case and this
dispute. Having done that, Frank insisted, the judges would then seek to find
conventional legal materialscases, statutes, regulations, constitutional pro-
visions, maxims, canons, and so forththat would provide ex post justifica-
tions or rationalizations for decisions that were causally uninfluenced by
those materials. For Frank and others (Hutcheson, 1929; Radin, 1925), the
key challenges to the traditional account lay first in the way in which they saw
judges as focused on reaching the best result for the particular case, and,
second, in the fact that the law was used by judges not to produce decisions,
but instead to justify ex post decisions made on decidedly nonlegal grounds.26
Not all of what ordinarily rides under the banner of Legal Realism fits this
mold. When Karl Llewellyn, for example, suggested that judges often made
decisions based on rules that diverged from the rules that one would find in
the law books (Llewellyn, 1930, 1960; Twining, 1973), he was denying neither
the possibility nor even the desirability of rule-based second-order decision
making. What he did deny, however, was the view that formal official written
law provided the source for the actual rules that judges and other decision
makers employed in making their decisions (Dagan, 2007). There were rules,
Llewellyn agreed, but those rules came not from the law books or the decided
cases, but instead from the judges own policy views and from the social and
professional culture within which the judge operated.
Even for Llewellyn at times, however, and for other Realists more perva-
sively, the challenge offered to the traditional model of legal reasoning was a
challenge to the possibility that judges could avoid what they saw as the
best result for the case at hand in the service of other and larger law-based,
rule-based, precedent-based, or process-based goals. In this regard, therefore,

the Realists saw judges first and foremost as human beings, and saw the
(natural) human desire to reach the best outcome for this case as the primary
determinant of judicial behavior and judicial decision making. What judges
learned in law school or in practicing law might make some difference at the
margins, and so too would what judges internalized in taking on the role and
duties of a judge, but these minor differences, many Realists insisted, were
overshadowed by the particularistic proclivities which judges shared with
their fellow human beings. And it is precisely in this respect that most of
the existing research on the psychology of judging, research that also sees the
pervasively human characteristics of judges as the primary determinant of
judicial behavior and judicial decision, is best understood as embodying a
Realist outlook on adjudication in particular and law more generally.
The Issue Joined

The contrast between Realist and traditional views of legal reasoningbetween
Frank and Coke, if you willis an important window through which to view
questions about the psychology of judging. And although I have stressed the
contrast between Realist and traditional views in terms of particularism and
generality, and in the distinction between first-order and second-order rea-
soning, my larger point hinges on neither of these distinctions. Rather, the
central claim is that the Realists, or at least many of them, were concerned with
challenging the larger view that there is something distinctive about legal
reasoning. The traditional view, captured well by Cokes appeal to artificiality,
is that lawyers and judges are engaged in demonstrably different cognitive
processes from other reasoners and decision makers. And the Realist view,
exemplified by Frank, is that the alleged cognitive differences between judges
(or lawyers) and the rest of humanity are exaggerated, with judges engaged in
forms of cognition not appreciably different from those of the human species in
general, a species of which judges are of course a part.
If the Realists were right to trivialize the differences between judges and
the rest of us, then the psychology of judging is, simply, psychology. What we
can learn about the psychology of human cognition, human reasoning,
human perception, and human decision making will accordingly serve us
well, with few modifications, in describing, predicting, and understanding the
psychology of judging. But if the Realists are wrong, then what the existing
research tells us about how ordinary people use analogy will tell us little about
how judges decide according to precedent, what the existing research tells us
about how people make decisions will tell us little about how judges make
decisions according to rules, and what the existing research tells us about the
inputs into human decision making will tell us little about how judges make
decisions by following the dictates of even those authoritative sources with
which they disagree.
At the very least, the contrast between the traditional and Realist
accounts of legal reasoning suggests that a research agenda could be aimed
at answering the precise question of the extent to which, if at all, judges
actually do engage in second-order reasoning and actually do refrain from
reaching what they believe to be the correct outcome in this case because of
the perceived (or actual) constraints of precedent, rule, or authoritative
source. Moreover, such a research agenda need not be limited to examining
the hypothesis that judges engage in second-order reasoning. It could also
usefully test the hypothesis that judicial second-order reasoning is substan-
tially different from, more frequent than, or more effective than the second-
order reasoning of nonjudge decision makers, even assuming that nonjudge
decision makers engage in second-order reasoning at all.
Even the foregoing sketch of a research program is far too crude. In
addition to attempting to control for intelligence, education, motivation, and
other attributes that judges likely possess to a greater degree than the popula-
tion at large, such a program would attempt to disaggregate the components
of second-order legal reasoning in order to determine whether there was a gap
between judges and lay decision makers for each of those components.
When judges are expected to make a decision consistent with a previous
decision with which they disagreethe central case of decision according to
precedentwill they follow precedent and reach what they think is the wrong
result more often than ordinary people assigned the same task? If judges are
told that the only sources on which they may rely are part of an artificially
constricted array of sources that in this instance might support an erroneous
result (Schauer, 2004c), will they limit their attention to this suboptimizing
and error-producing (from their lights) array to a greater extent than the
nonjudges? If commanded to follow a bad rule or a good rule which in this
instance produces an unfortunate outcome, will judges more than others
simply follow the rule and swallow the unfortunate outcome? And if
instructed to refrain from doing the right thing because doing the right
thing is in someone elses jurisdiction or is someone elses responsibility,27
will judges more than others remain passive in the face of an opportunity to
do the right thing, or will they, like most others, treat jurisdictional and
similar limitations as inconsequential?28
Even if such research were to indicate that judges really were different
from lay people for some or all of these tasks, additional research would still
be necessary to determine whether it was simply legal training that produced
the difference, or whether it was something about the role of judge as judge.
We can imagine a research design that might, for example, assign similar tasks
to judges, practicing lawyers, and law students, in order to determine whether
an identified difference between judges and lay people was explained by some
difference between judges and lawyers, or was explained instead by a differ-
ence between judges and lawyers and (advanced) law students, on the one
hand, and those without legal training, on the other. And even more fine
grained research could attempt to locate differences even among classes of
judges, as for example in the differences between elected and appointed

judges, or between trial and appellate judges, or among judges with different
varieties of prejudge backgrounds.
Other hypotheses and research possibilities abound, but like the ones just
sketched they would identify a task other than fact-finding or verdict-ren-
dering, and then seek to determine whether for this task judges were genu-
inely different, either in outcome or in method, from some relevant nonjudge
class. An affirmative answer to this question would not, of course, exclude the
likelihood of relevant similarities existing alongside the genuine differences.
And that is why the existing research showing that judges are susceptible to
many well-discussed cognitive failings and biasesanchoring and avail-
ability, for exampleis highly important. Even though important, however,
this research is incomplete. Even if judges when acting as finders of fact or
when reaching verdicts are prone to all or most of these familiar reasoning
failures,29 the question remains entirely open whether there are also areas in
which judges think quite differently, even supposing that with respect to those
areas judges would be similarly afflicted with the same or analogous cognitive
deficiencies. The existing research tells us little about whether there are such
areas of differential thinking, and, if so, what they look like, but until we can
answer this question we cannot know whether the conclusions of Legal
Realism are correct, and whether the hidden Legal Realist premises of the
existing psychological research on judging are sound. Much of the existing
research on the psychology of judging takes the Realist view of judging as
axiomatic, but that conclusion is hardly inevitable and hardly based on
systematic research directed precisely at that question.
The Question of Expertise

Although a substantial psychological literature explores the nature of exper-
tise and the differences between expert reasoning and that of novices,30
surprisingly little of that literature is especially relevant to the question
whether, if at all, judges reason differently from ordinary folk. And that is
because the psychological literature on expertise tends to be focused on the
question of comparative expertise within a single area of knowledge rather
than on the hypothesized cognitive differences across different areas of
knowledge. The existing research examines what experts at x do that novices
at x do not, but almost none of it looks at whether people who know how to x,
whether experts or novices or somewhere in between, tackle problems and
make decisions differently from people who do not know how to x at all, or
from people who know only how to y, again regardless of whether they are
expert or not.
Yet although little research addresses this question of cross-domain
expertisewhat we might call specialization rather than expertisethis ques-
tion is central to examining the psychology of judging. We could determine
what expert fact-finders do that novice fact-finders do not, just as we could ask
what expert judges do that novice judges do not. But if we are interested in
testing the hypothesis that there is a psychology of judging that differs from the
psychology of decision making simpliciter, then we want to see whether there
are some tasks that judges perform differently just because they are judges. We
know from the existing research on judging that in some tasksmany aspects
of fact-finding, principallyjudges do differ less from nonjudges and non-
lawyers than the conventional wisdom has appeared to suppose. But to take
these findings, important as they are, as answering the central question about
the psychology of judging is like imagining that because auto mechanics
approach the finances of their own small business in the same way that
psychiatrists approach the finances of their own small businesses that auto
mechanics are importantly similar to psychiatrists. That the two are similar
with respect to accounting says nothing about whether they are similar with
respect to fixing cars and fixing heads, and similarly the discovery that judges
and jurors (or people generally) are similar with respect to fact-finding skirts
the question whether there is something else that judges do for which their
training and expertise might actually produce important differences.
If I am right that an important component of judging is something other
than fact-findingarguably true for trial judges and self-evident for appellate
judgesthen we can understand the importance of focusing on law-finding,
law-applying, law-interpreting, and, yes, law-making, for these are a large part
of the judicial task. But when judges perform these tasks, do they perform
them in the same way that those without legal or judicial training or experi-
ence would approach them, which is what many of the Realists argued, or do
they employ a different skill set, to use an infelicitous and fashionable but not
inappropriate term from contemporary management-speak? When it comes
to tasks other than fact-finding, do judges think like human beings, or like
lawyers, or like judges? Addressing this question should be one of the central
items on a research agenda for the psychology of judging, but it is, surpris-
ingly, an item that up to now has been almost completely absent.
Conclusion
Jerome Frank understood the traditional claim about legal reasoning, but he
argued that judges were psychologically unable to do what the traditional
theory demanded.31 Frank is treated nowadays as a bit idiosyncratic, but the
psychological lens through which he viewed judging points to the importance
of distinguishing three questions about second-order reasoning. The first, the
answer to which is embodied in Franks own point of view, is whether people
are naturally particularistic. When engaged in decision-making tasks, do
people just because of the makeup of the human mind think in terms of
this task, thus being psychologically averse to making the wrong decision on
this occasion in the service of larger or more distant goals?
In insisting on a deep human proclivity toward the particular, Frank may

well have been wrong. After all, delayed gratification is hardly beyond the
capacities of most people, and there is more than a remote possibility that
Franks speculationsand they were hardly more than thatabout the raw
material of human psychology were guided less by psychological fact than by
Franks own normative views about what lawyers and judges ought to be
doing.
Even assuming that Frank was right about what humans start with, he
may nevertheless have been unduly pessimistic about the possibility that these
antecedent particularistic instincts could be changed. So even if humans are
temperamentally, physiologically, or genetically averse to second-order rea-
soning, there is little reason to believe that this aversion is so hard-wired as to
be incapable of change. Perhaps one form of education, including one form of
moral education, is aimed, at least in part, at fostering various forms of
second-order reasoning, and to the extent that such education is at times
successful Frank and his compatriots may have given up too quickly on the
possibility that anyoneand not just judgescan both grasp and perform
the basic skills of reasoning from rules, making decisions constrained by
precedent,32 and taking the commands of an authority as providing reasons
for action. To the extent that humans in general can be taught to engage in
such reasoning, then there would be reason to believe that lawyers and judges
could be trained to do the same thing, to do it more often, and to do it better.
Implicit in the traditional picture of the artificial reason of the law, therefore,
is a story about the possibility that this artificial reason can be inculcated in
and internalized by even those humans for which it would initially seem
artificial.
Because Frank believed that human particularism was so hard-wired as to
be unchangeable, however, he was never forced to reach the normative
question. If we assume that the basic tools of second-order reasoning are
learnable, we then face the question whether it would be good for lawyers and
judges to learn them. Weber sneered at his (erroneous, as a matter of Islamic
law) image of the qadi, making the best decision all things considered for
each case. But as some voices in contemporary philosophy (Dancy, 1993)
feminist theory (Bartlett, 1990), and legal theory (Sunstein, 1999) have
insisted, making decisions for the particular casedeciding things one case
at a timeis supported by influential arguments, and has much to recom-
mend it, even for those of us (Schauer, 2003) who in the final analysis see
fewer virtues in particularism than others. But the point here is not to
rehearse these familiar debates. It is instead to emphasize that an inquiry
into the possibility of judicial second-order reasoningan inquiry into a
central but understudied dimension of the psychology of judgingis impor-
tant not only as description and explanation of how judges behave and decide,
but also as the precursor to a normative inquiry into how judges should
behave and decide. Such inquiries dominate legal scholarship, often to
unfortunate exclusion of almost everything else, but they are hardly without
import. But before we can intelligently decide what judges should do, we need
to see both what they are doing and what they can do. This inquiry can be
usefully informed by serious empirical inquiry into the psychology of judging,
but little progress will be made even on this dimension until the research
agenda begins to take seriously the possibility that there might actually be a
psychology of judging, a possibility that is surprisingly absent from almost all
of the existing literature.
Notes
An earlier version of this chapter was presented at the Workshop on the Psychology
of Judging, University of Virginia, Charlottesville, Virginia, March 2931, 2007.
I am grateful for the comments of the participants on that occasion, for thoughtful
and challenging written comments from Barbara A. Spellman and Dan Simon, and
for research support from the Harvard Law School and the Joan Shorenstein Center
on the Press, Politics and Public Policy, Harvard University.
1. For a useful analysis of the research that is genuinely focused on judges and
judging, see (Robinson & Spellman, 2005).
2. There are numerous logistical and ethical impediments to research using real
judges as experimental subjects, and thus the conclusion that nonjudge and
nonlawyer research subjects are representative of judges is facilitated by the ease
of locating the former and the obstacles to doing serious experimental work on
the latter.
3. Fact-finding is not to be confused with fact-locating or fact-discovering. Fact-
finding is the legal term of art for determining what actually happened based
solely on the evidence presented in court by the parties.
4. Indeed, the preoccupation with the jury in much of the psychological research is
itself curious in light of the fact that the institution of the jury does not much
exist in civil law countries, is not used outside of criminal cases (with the
occasional exception of libel trials) in any common law country other than
the United States, and is a rapidly declining institution even for criminal cases
in the United States and elsewhere (Guthrie, Rachlinski, & Wistrich, 2001;
Schauer, 2006b).
5. On the implications for social science research of the distinction between the
tasks typically performed by trial judges and those typically performed by
appellate judges, see (Rowland & Carp, 1996).
6. The statement in the text is not intended to express even the slightest sympathy
with the hoary but misguided cavil that experiments on university under-
graduates are of limited value in learning about the behavior of people who
are not undergraduates. In the absence of identifiable and germane differences
between undergraduates and people in general, there is no good reason to
doubt the generalizability of findings about undergraduates to conclusions
about people as a whole. But when we are drawing conclusions about the
decision-making characteristics of individuals who are in theory specially
trained to make decisions of a certain kindas are judgesit is far more
appropriate to question whether research using people without that special
training can tell us much about the way in which people with special training
make the very decisions for which they are supposedly specially trained and for
which they are specially selected by virtue of possessing the requisite training
and skills.
7. The man on the Clapham omnibus being the quaint British equivalent of the
American reasonable man (Hall v. Brooklands Racing Club).
8. Those who become lawyers may self-select, or may be selected, on the basis of
their possession of attributes that are relatively rare in the general population
but that are not only germane to success as a lawyer, but are also germane to
success as a judge even among the class of lawyers.
9. Or it may be that the causal mechanism operates in a different direction, with
lawyers being selected for the judiciary, or self-selecting into the judiciary,
because they possess skills or proclivities to forms of reasoning and decision
making that are different from those of the mine-run of practicing lawyers.
10. This says nothing about the respective abilities or intelligence of judges, den-
tists, and plumbers. One need not be smarter (or dumber) than the average of
humanity in order to engage in a cognitively specialized task.
11. It is worth emphasizing that nothing I say here denies that judges share some or
perhaps even many decision making psychological characteristics with ordinary
people (Simon, 1998, 2002, 2004), and that many of those shared characteristics
are useful in understanding what judges do. My concern is that by focusing so
heavily on the shared characteristics, researchers have slighted the nonshared
characteristics to the detriment of a fuller understanding of what judges do.
Tiger Woods and I both play golf, and I am reasonably sure that Woods and
I share some number of decision making pathologies on the golf course, such as
exaggerating the probability of making (for our skill level) low probability
shots, or taking the most recent (and thus most cognitively available) shots as
more representative of the array of outcomes on shots of that variety than they
in fact are. But to focus only on these shared characteristics and to ignore the
numerous ways in which Woods and I differ as golfers, mentally as well as
physically, is to ignore something seemingly quite important. Without investi-
gating the ways in which judges might be able to do things that lay people
cannot, we run the risk of overgeneralizing from the ways in which judges
assuredly have decision-making characteristics they share with lay people.
12. Outside of North America, the study of law takes place largely at the under-
graduate level, although additional and postgraduate law study is common.
A potentially valuable research project, although not my focus here, would be to
examine whether studying law from the age of eighteen or nineteen, and in
place of some other undergraduate specialization, produces a significant dif-
ference in reasoning and decision making from that which exists in those who
do not commence the serious study of law until at least the age of twenty-two,
and who already have as undergraduates studied another field.
13. And if they do, they shouldnt.
14. To forestall a potential objection, I signal here (and address at somewhat greater
length below, and see also Schauer, 2007) that I do not take the use of analogy
and the constraints of precedent as being especially similar. Lawyers use analogy
frequently, but so do other professionals and most lay people. Feeling obligated
to follow a previous decision that one believes to be erroneous, however, is
arguably far less prevalent outside of law than in, and may thus comprise part of
the array of reasoning and decision-making modalities that collectively can be

thought of as legal reasoning.
15. For skeptical views about the alleged distinctiveness of analogical reasoning, see
Alexander (1996), Posner (2006), and Schauer (2008).
16. On the use of analogy in numerous occupations and endeavors, see Blanchette
& Dunbar (2001a, 2001b).
17. There is scant connection between the question-centered methods of teaching
employed by Socrates in the Platonic dialogues and the type of questioning that
has traditionally taken place in the law school classroom. Even apart from the
enormous advantage that Plato had over the rest of us in being able to write the
answers as well as the questions, Socrates goal was to extract from his inter-
locutors some latent but nonspecialized insight, rather than to inculcate in
them a specialized skill that they hitherto did not possess. Now it may be that
the ability to engage in just this kind of second-order reasoning is latent in
everyone, but if it is sufficiently latent that it takes law professors and three years
of law school to extract it for most people, then there is no difference of
consequence between an inculcation and an extraction model of legal educa-
tion, for in either case that student develops the ability actually to do something
she could not do before.
18. A common mistake is to assume that legal argument is about persuasion, and
that in seeking to persuade lawyers act similarly to politicians, editorialists,
teachers, clergymen, and countless other persuaders. But legal argument under
the traditional account is persuasion of a special kind precisely because it is
parasitic on how the judge will make her decision. So if the traditional account
of legal reasoning is soundand it may not bethe lawyer is not attempting to
persuade the judge that such-and-such is a good outcome simpliciter, but is
instead trying to convince the judge that some good outcome is not precluded
by contrary precedents, or that precedent commands a result without regard to
the precedent-independent desirability of that result.
19. On the use of hypothetical cases to ensure principled decision making in just
this way, see (Christie, 1969).
20. The looking backward aspect of law makes judges think at least as much
about conformity to an announced principle as about the right and justice and
social utility of the case they are about to decide (Ulman, 1933).
21. [I]n most matters it is more important that the applicable rule of law be settled
than that it be settled right, (Burnet v. Coronado Oil & Gas. Co., Brandeis,
dissenting; see also Alexander, 1989; Alexander & Sherwin, 2001; Schauer,
1987).
22. For earlier works articulating the traditional account even more boldly (and,
perhaps, crudely) than the various sources cited in the previous section, see
Black (1912), Cross & Harris (1990), and Wambaugh (1894).
23. Or, sometimes, American Legal Realism, not only in acknowledgment of its
provenance, but also to distinguish it from the largely unrelated Scandinavian
Realism of, for example Hagerstrom (1953), Olivecrona (1971), and Ross
(1958).
24. On Legal Realism generally, see Kalman (1986), Leiter (2004), and Rumble
(1968). There are competing conceptions of what ideas lay at the core of
Realism, but I disclaim the role of arbiter among the multiple views of what
Legal Realism was really all about. So although various Realists and their
fellow travelers advanced a cluster of different claims all in the name of Legal
Realism, it is uncontroversial that the particular part of Realism I stress in the
text is at least among the positions advocated by some of the more prominent
Realists.
25. It is fashionable these days to marginalize Franks contribution to Realism
because of his belief that the judges personal idiosyncrasies were a significant
determinant of judicial outcomes (Dagan, 2007; Leiter, 1997). But although
Franks (and Hutchesons) views about the source of the judges decision may be
unrepresentative of so-called mainstream Realism, he remains a seminal figure
for the view that the judges determination of the right outcome precedes the
judges consultation of formal legal doctrinal sources, and also for the view that
judicial decision is substantially particularistic, both being central tenets of the
broad Realist program.
26. For a modern and sophisticated version of this form of Realism, see Kennedy
(1986).
27. See Blanchflower v. Blanchflower (2003), concluding that same-sex adultery
ought to constitute grounds for at-fault divorce, but that such a change was
for the legislature and not the courts.
28. Consider the question of federalism. Although lawyers and judges spend much
time wrestling with the respective jurisdictional competences of the federal
government and the states, there is little indication that either the public or
the political world that caters to that public takes the principles of federalism as
constituting an independent second-order constraint on either the states or the
federal governments adopting what the public believes to be a desirable first-
order policy.
29. I bracket here the important debates about whether patterns of reasoning falling
short of optimal or perfect rationality are better understood as desirable
adaptive strategies (Gigerenzer, 2000; Gigerenzer & Selten, 2001) or instead as
potentially correctable errors whose correction would, in general, be desirable
(Kahneman & Tversky, 1981, 1984; Kahneman, Slovic, & Tversky, 1982).
30. For an introduction to the psychological literature, see Ericsson et al. (2006)
and Chi, Farr, & Glaser (1998).
31. It is interesting that when Frank became a federal judge, a role he occupied from
1941 until his death in 1957, he wrote opinions that hardly differed in style from
those of other judges. Much better examples of Realist judging can be found in
the opinions of Justice William O. Douglas.
32. This is a good place to point out the important difference, off-hand remarks in
the psychology analogy literature notwithstanding, between analogical rea-
soning and the legal constraint of precedent. When people, including lawyers
(Levi, 1948; Weinreb, 2005), seek to persuade others, or seek guidance in
making a decision, they often rely on analogies. They think it is good to take
some action now because it is similar to some action in the past that has worked
out successfully, or they think it wise to avoid some decision because the
circumstances resemble circumstances of the past. But in such cases the deci-
sion-maker is using the analogy (Spellman & Holyoak, 1996) to help reach the
right decision now. The analogy is a tool, and in theory a friend. The constraint
of precedent in law, however, which is not coextensive with lawyers use of
analogical reasoning, is more foe than friend. Having concluded that the right
thing to do now is f, the lawyer or judge will sometimes find that f is precluded
by some previous decision, often a decision that the present judge thinks
mistaken. But insofar as the constraint of precedent actually constrains
(which it likely does outside of hard Supreme Court cases far more than it
does in the Supreme Court (Segal & Spaeth, 1996, 2002), then the judge is not
looking for the analogy that helps, but instead seeking, often unsuccessfully, to
avoid the analogy that hurts. Whether such constraint by precedent is desirable
is itself debatable, as is the empirical question about its frequency in legal
decision making. But the importance of these questions should not lead us to
think that being bound by a similar but erroneous decision from the past is
similar to choosing to be guided or persuaded by analogous circumstances from
an earlier time (Schauer, 2008).
8
Features of Judicial Reasoning
Emily Sherwin
Adjudication of legal disputes depends inevitably on human decision making.

It is not surprising, therefore, that legal scholars sometimes turn to psychology
for help in understanding law. Importation of psychological insights to law
leads naturally to the question whether certain aspects of human psychology are
special to, or especially prominent in, legal decision making. From a psycho-
logical point of view, does legal decision making, and particularly judicial
decision making, differ from decision making in nonlegal contexts?
Larry Alexander and I (2008) have argued at length that the methods of
reasoning judges use to decide cases are no different from the methods of
reasoning used by nonjudicial decision makers. Nevertheless, certain aspects
of the psychology of decision making have special importance in law because
of the role they play in the process of adjudication. In this sense, there is a
psychology of judging, although there is no reason why it should not be
informed by psychological research performed in other settings. My conclu-
sions on this point are consistent with the conclusions Frederick Schauer
reaches in his contribution to this book (ch. 7).
In this brief essay, I examine the various and sometimes conflicting psy-
chological tasks our legal system assigns to judges. I begin with first-order
judicial decision making, meaning reasoning about how particular disputes
should be resolved in the absence of authoritative rules. I then turn to what
Schauer usefully describes as second-order judicial decision making, meaning
decisions judges must make about the bearing of authoritative rules on their own
process of decision. From a psychological point of view, the most interesting
feature of second-order judicial decision making is that law often requires judges
121
to suppress reason in order to give authoritative effect to rules. How judges can
achieve the mental state necessary for this purposea mental state consisting of
abstinence from reasoningis one of the great psychological mysteries of law.
What I say in this essay is nothing new. In particular, the analysis of rule-
based decision making I provide is much indebted to Schauers excellent work
on the subject. Nor do I provide independent psychological insights: I write as
a legal theorist without expertise in the field of cognitive science.
First-Order Reasoning About Legal Disputes

When judges decide cases, their immediate task is to decide what outcome is
best in a particular, concrete dispute. For the moment, I shall set aside the
problem of legal rules and assume that the dispute in question is a case of
first impression, not governed by existing law. In the absence of governing
law, identifying the best outcome of a dispute involves empirical observation,
induction, and moral reasoning.
Empirical and inductive reasoning generate the factual premises for
adjudication.1 Research in the field of cognitive science indicates that these
forms of reasoning are subject to various types of bias: human reasoners rely
on mental shortcuts that normally are useful but can distort their judgment
about facts, and especially about probabilities (Tversky & Kahneman, 1982;
Plous, 1993; Gilovich, Griffin, & Kahneman, 2002; Symposium, 2005). For
example, they respond to irrelevant cues when making quantitative estimates,
they overlook the effects of hindsight (Rachlinski, 1998; Guthrie, Rachlinski,
& Wistrich, 2001), and they assume that vivid or emotion-provoking facts are
more typical than they are of the statistical classes to which they belong. Work
by Chris Guthrie, Jeffrey Rachlinski, and Andrew Wistrich (2001) confirms
that judges are susceptible to these biases, although experienced judges may
be somewhat better at controlling their effects than the population at large.
Given a set of facts (and assuming still that no law applies), judges must
engage in moral reasoning to determine how the state should to respond to the
parties dispute. Philosophers commonly equate moral reasoning with reasoning
to reflective equilibrium.2 To achieve reflective equilibrium, the reasoner con-
structs a tentative moral principle applicable to the problem at hand, then tests the
principle against moral intuitions about particular instances within the principles
scope and against background theories about the world at large. The reasoner then
adjusts the principle, the reasoners moral judgments, or both, until conflicts are
satisfactorily resolved. The principle that emerges in equilibrium provides the
answer to the problem case. As a method of justifying action or decisions,
reasoning to wide reflective equilibrium is open to some devastating logical
criticisms (Haslett, 1987), yet it may be the only method practically available.
Moral reasoning by the method of reflective equilibrium raises inter-
esting questions about the nature of moral intuition (see Guthrie, Rachlinski,
& Wistrich, 2007a). The process of constructing moral principles may also be
Features of Judicial Reasoning 123
affected by cognitive biases, particularly those that make salient examples

appear more representative than they are. Biases come into play because
moral principles necessarily are generalizations; meanwhile, the reasoners
objective is to resolve a particular problem. As a result, the specific facts that
generated the reasoners problem (in the case of a judge, the facts of a legal
dispute) may loom large in the reasoners mind and obscure other conse-
quences of the moral principle the reasoner hopes to test.
Neither moral reasoning nor empirical and inductive reasoning, how-
ever, is peculiar to, or specially salient in, law. They operate in adjudication
just as they operate in any decision-making context. The important psycho-
logical questions for law are about the extent to which judges can resist or
counteract the biases that affect ordinary reasoners.
Second-Order Reasoning About Legal Disputes

In his essay in this volume, Schauer points out that judges make not only first-
order judgments about the outcomes of disputes but also second-order
judgments about the rules that govern their first-order decisions. Second-
order reasoning of this kind is not limited to judges; moral decision making of
any kind involves the application of general principles. Yet the nature of legal
decisions as both sources of authority and products of authority makes
second-order reasoning about rules a particularly important feature of law.
Judges engage in two kinds of second-order reasoning, which involve quite
different mental tasks. First, judges must consider what authoritative rules their
decisions will generate. In doing so, they must consider the probable future
consequences of different decisional rules in cases not now before them.
Second, and perhaps more interestingly from a psychological point of view,
judges must submit to the authority of preexisting legal rules. At least according
to traditional conceptions of adjudication, judicial decisions are constrained
both by legislative rules and by rules announced by past courts. This means that
if a rule applies, the judge must decide as the rule requires even if, by the judges
own assessment of relevant reasons for decision, the result the rule requires is
wrong. Thus, to give authoritative effect to legal rules, judges must suppress
either their normal reasoning processes or substantive conclusions they have
already reached by means of normal reasoning. The mental feat of suspending
reason in order to follow authority is not unique to law, but, at least according
to one understanding of law and legal decision making, it is vital to an effective
system of law. In this sense, it presents a psychological problem of special
concern to the legal domain.
Rule Making
One prominent institutional circumstance of judicial decision making is that
judges make rules of law as they adjudicate disputes. Judges may decide cases
fairly narrowly, but in a legal system such as ours, in which judges commonly
explain their decisions in opinions, some degree of generalization is unavoid-
able and generality leads to rules.3 Moreover, in a system in which the out-
comes of adjudication are accessible to the public, judicial decisions are
studied by an audience of potential actors, who naturally generalize from
past decisions to probable legal treatment of their own activities.
One might imagine a system in which judges say nothing about their
decisions and the conclusions they reach are understood to have no prece-
dential effect. Only legislative rules would have implications for future deci-
sions. In our own legal system, however, this is not the case. The public
expects a fairly high degree of consistency in judicial decision making over
time and judges respond to this expectation by recognizing, to some extent,
the authority of prior decisions. Moreover, the publics expectation of con-
sistency has social value, because it enables the members of society to coordi-
nate their actions and to treat controversies as settled (Postema, 1982,
pp. 172186; Raz, 1986, pp. 4950; Hurd, 1999, pp. 214221).
Thus, implicitly or explicitly, judges announce rules of decision as they
decide cases. It follows that the complete set of reasons for or against a
particular decision includes the future effects of the decisional rules on
which it rests. To reach a fully reasoned decision in any case, the judge
must generalize from the specific problem at hand to the class of problems
governed by the operative decisional rule.
Research on cognitive bias, as described above, is pertinent to this aspect of
second-order reasoning. Assessing the future effects of a decisional rule typi-
cally requires a calculation of statistical probability over a range of possible
cases, some of which may be remote from the case before the court. At the same
time, the judge must attend to the immediate task of achieving a fair outcome
for the parties to the current dispute. These simultaneous demands on the
judges attention implicate at least two of the biases documented in cognitive
studies: the availability bias and the affect bias, which lead reasoners to err in
assessing probabilities when their attention is focused on salient or emotionally
charged facts (Devins & Meese, 2005; Rachlinski, 2006; Schauer, 2006a). As a
consequence of these biases, the facts of current cases are likely to appear more
representative than they are of the class of cases covered by a proposed decision
rule, causing judges to miscalculate the overall effects of rules.4
Rule Following
Following rules is more complicated analytically, and perhaps psychologi-
cally as well. Suppose a judge is called on to decide a case that falls
indisputably within the terms of a rule announced by a prior court. One
possible approach for the judge is what Schauer (1991) has termed rule-
sensitive particularism. Simple particularism is the process of deciding
what outcome is best in a particular case, all things considered. The reasoning
involved is ordinary reasoning (empirical observation, induction, and moral
reasoning), applied to the problem at hand. When a rule applies, simple

particularism is inadequate because it fails to consider what secondary conse-
quences might result from following (or ignoring) the rule. Rule-sensitive
particularism describes a process of decision making that takes account not
only of the good or bad results the rule will bring about if applied to the case at
hand, but also of the benefits that follow from compliance with an established
rule (Schauer, 1991, pp. 94100). A rule-sensitive particularist seeks the out-
come that is best (or least bad), given the existence of a rule.
Much has been written about the benefits of rules (e.g., Raz, 1986;
Schauer, 1991; Alexander & Sherwin, 2001). Rules settle controversy
(Eisenberg, 1988, pp. 47; Raz, 1994, pp. 187192). They support coordina-
tion among actors whose reasons for action depend on the actions of others: if
most actors follow the rules, each can predict what others are likely to do.
Rules designed by experts also can reduce error by individual actors, and
procedural rules can allocate decisional authority or rule-making authority to
experts. Each of these potential benefits, however, depends on regular com-
pliance with the terms of the rules. A decision to disregard a rule in a
particular case may have negative effects on the future effectiveness of the
rule and, consequently, the social benefits that follow from an effective rule.
The process of reasoning entailed by rule-sensitive particularism is
similar to the process of reasoning required to assess the future consequences
of a proposed rule, although it adds some complications. A rule-sensitive
particularist judge deciding a rule-governed dispute must first determine
what outcome is best between parties. If the outcome the judge selects is
contrary to the rule, the judge must make several further calculations.
Specifically, the judge must determine how likely it is that his own judgment
is wrong and what effect a decision to depart from the rule will have on other
judges or private actors (whether or not the judge is wrong).
These are not easy calculations. At a minimum, judges acting as rule-
sensitive particularists must determine the likelihood that others will observe
their own defections; the extent to which others who observe their defections
will discount the benefits of the rule; and the likelihood that others who
observe their defections, discount the benefits of the rule, and consequently
decide to defect themselves, will err in their own calculations of the relative
costs and benefits of following the rule. Yet, despite the difficulty of the
calculation, the reasoning involved is ordinary inductive reasoning, subject
to the same psychological impediments that affect all inductive reasoning.
Specifically, facts that are in the forefront of the reasoners mind, and facts
that evoke emotional responses in the reasoner, are likely to appear more
representative than they are and thus to distort the reasoners assessment of
probabilities. As I have noted, this is a particularly serious problem for judges
because judges take up the question of whether to follow a rule with the facts
of particular cases outstanding in their minds.
Rule-sensitive particularism is a rational approach to rule-governed
decision making. From the point of view of a legal rulemaker, however, it is
not the ideal judicial response to rules. This may seem odd: rules are blunt,
and as a consequence are likely to require the wrong result in some of the cases
they cover. In a world of perfect reasoners, therefore, it might be best if rule-
followers were rule-sensitive particularists who sometimes decided not to
follow applicable rules. Judges, however, are not perfect reasoners: they lack
the perfect information required for perfect reasoning5 and, as noted, they are
subject to cognitive biases that may cause them to systematically undervalue
the long-term costs of deviating from rules. Consequently, rule-sensitive
particularists will sometimes deviate from rules when they should not. In a
community of rule-sensitive particularists, each judge will anticipate that
some other judges will miscalculate in this way, and accordingly will discount
the benefits of the rule. Each such discount, by each rule-sensitive particu-
larist judge, means a corresponding decrease in the judges estimate of the
harm that will result from a decision to defect from the rule. Ultimately, the
rule has no value at all, and rule-sensitive particularism collapses into simple
particularism.6
Rule sensitive particularism may be more meaningful when some judges
consistently follow rules. Yet it is still less than ideal from the rule makers
point of view. Imagine that within a given legal system, some judges are rule-
sensitive particularists and others are rule followers. In this situation, a
decision to deviate from a rule has adverse consequences that the rule-
sensitive particularist must take into account. Specifically, other judges
who observe the deviation will know that not all judges are consistent
rule-followers. As a result, rule-followers may change their attitude, and
rule-sensitive particularists may discount the benefits of the rule. These
possibilities give the decision maker a reason to be cautious about deviating
from the rule. Nevertheless, although rule-sensitive particularist judges now
have some reason to follow the rule when it produces an outcome they
believe to be wrong, they may still err in calculating the balance of harms,
and, because of cognitive bias, their errors will tend systematically to favor
deviation from the rule.
Because rule-sensitive particularism is unstable in this way, a rational
legal rule maker would prefer that judges follow rules according to their
terms. A well-designed rule, consistently applied, will prevent errors of judg-
ment more often than it generates errors of overinclusion.7 From the point of
view of rule makers, therefore, a rule applied without reflection is preferable
to case-by-case judicial reasoning, even if judges are sensitive to the value of
rules. At the same time, when a judge believes that, in a particular case, the
harm to substantive values from following the rule is greater than the harm to
rule-based values from disregarding the rule, the only rational course is to
disregard the rule. The result is a logically unbridgeable gap between the way
in which rational rule makers would want judges to approach rules and the
approach that is rational for judges.8
This leads to the second possible judicial response to rules, which is to
follow them without further reflection about the justification, or lack of
justification, for the particular results they require. Legal rules, on this
approach, operate as the exclusive reasons for judicial decision in cases to
which they apply (Raz, 1979, 1986).9 If the terms of the rule call for a certain
result, that result is correct.
In ordinary life, rules appear to function in this way, at least at times. For
example, we may set schedules for ourselves to fend off procrastination,
although the schedules will never control our actions if we pause to think
about whether the time designated for each action is in fact the best time to act
(Fumerton, 1990, pp. 178188). Somehow, we manage not to think, and thus
to follow the schedule. Judicial rule-following, in the manner necessary to
avoid the pitfalls of rule-sensitive particularism, is similar in nature. Judges
may in fact follow rules, but they can do so only by abstaining from thought
about what is best, all things considered.
Thus, rule-following in the true sense presents a psychological enigma: in
what circumstances, and by what processes, do human decision makers
suspend their powers of reason and instead defer to authority?10 This is not
a phenomenon unique to adjudication: it occurs in daily life when people
follow nonlegal rules, and it occurs whenever nonjudges follow legal rules.
Nor is it something judges are especially likely to be good at: there is no reason
to think that selection (or self-selection) as a judge corresponds to a heigh-
tened capacity to suspend reason. Yet, if in fact there are psychological
mechanisms that permit blind rule-following, they have special social sig-
nificance in the context of law because they enable judges to give serious effect
to legal rules.
A further point is that when judges follow rules without reflection, they
must suspend reason in a special way. Judges typically do not decide legal
questions in the abstract; they resolve live disputes and enter judgments that
penalize or impose liability on the losing parties. This means that when a rule
calls for a result that the judge deems to be wrong in the case before the court,
the judge not only must decide the case against his best judgment, but also
must impose a penalty on a party he believes has acted correctly and does not
deserve to suffer harm (Hurd, 1992, 1999, pp. 253293; Alexander & Sherwin,
2001, pp. 7886). Even when the judge believes the result of the rule is correct,
enforcing the rule may sometimes entail penalizing a party who did not act
culpably, but simply misjudged. Thus, following rules according to their
terms requires judges to suppress both reason and moral censorship of their
own actions toward others. This aspect of adjudication adds to the signifi-
cance of the normative questions about rule following. We have reason to
prefer that judges enforce legal rules as written, but we also have reason to
question the morality of full enforcement of rules.
Assuming that blind rule-following is at least sometimes desirable from a
social point of view, a further psychological question is whether the capacity
to suppress reason can be taught or cultivated. Research indicates that to
some extent, reasoners can learn to curb some forms of cognitive bias. The
studies conducted by Guthrie, Rachlinski, and Wistrich (2001) confirm that
this holds true for judges: certain common types of cognitive bias are less
pronounced in experienced judges addressing legal problems. It seems unli-
kely, however, that absence of reason is a mental state that reasoners could
develop deliberately in themselves. By definition, abstaining from reasoning is
an unthinking, if not unconscious, process. It is possible that reasoners could
become accustomed by force of habit to reaching decisions without reflection,
but it is difficult to see how a propensity to decide without reflection could be
instilled in a reasoner with the reasoners conscious assent.
Rule Making and Rule Following
I have mentioned two second order judicial tasks that play an important
part in the psychology of judging. First, judges must formulate rules of
decision and calculate the effects of those rules will have when applied to
future cases. In making this calculation, judges are subject to cognitive biases
that can lead them to undervalue statistical probabilities. Second, judges are
expected to treat established legal rules as authoritative. To give authoritative
effect to rules, judges must decide cases mechanically, without reflecting on
the relationship between the outcomes the rules require and the values the
rules are supposed to promote.
The states of mind necessary to perform these two judicial tasks effec-
tively are quite different. Designing sound rules is a highly deliberative
process involving empirical observation, induction, and moral reasoning.
Judges engaged in rule making must, among other things, remain alert to
background probabilities and guard against the biases that might prevent
them from accurately assessing the future effects of potential rules. Thus, for
the purpose of rule making, the more judges know about their own psycho-
logical proclivities the better they will do. When judges are called on to follow
rules, the opposite is true. To follow rules consistently, they must suppress
their own judgment about the outcomes of the cases before them. The more
judges reflect on the process of decision, the more difficult this will be.
The tasks of rule making and rule following are distinct. Judicial rule
making, and the empirical and inductive reasoning it entails, come into play
when no rule applies. Rule following, of course, is reserved for cases governed
by rules. Accordingly, there is ordinarily no need to perform both these
functions at once. Yet, judges must frequently shift from one mode of
decision making to the other from case to case, or even in the course of a
single litigation. This will not be easy: rationality and self-awareness, which
conduce to good rule making, are naturally at odds with a habit of unre-
flective obedience that will enable judges to follow rules.
A further complication is that if a legal system hopes both to maintain
effective rules and to command public respect, it must provide some relief
from rules. Rules may be misconceived, and even well-conceived rules may
become obsolete as conditions change. Legislatures can intervene to repeal
unsound rules, but for various reasons they may fail to respond. When this
occurs, the only avenue of relief is judges own power to overrule judicial
rules.
In our own legal system, the power of judges to overrule precedent rules
established in prior judicial decisions is widely recognized. Ideally, however,
judges should overrule rules only when the rules are unjustified as rules.11
Rules are justified as rules if their benefits (settlement, coordination, and
preemption of faulty judgment) outweigh their costs (errors of overinclu-
sion) over the range of cases to which they apply. When a rule passes this test,
it is best, from a societal perspective, that judges leave the rule in place and
enforce it according to its terms, even when they believe that a particular
outcome of the rule is a mistake. When a rule does not pass the test of net
benefit over the full range of its applications, it should be overruled.
The difficulty is that when rules are justified, judges must combine the
two types of mental tasks I have described to solve a single legal problem. To
determine whether a rule is justified as a rule, the judge must calculate the
future consequences of the rule. This requires both careful reasoning and
attention to cognitive biases that may make a current bad outcome appear
more representative than it is (Schauer, 2006a, 906912). Then, if the balance
of error favors the rule, the same judge must cease reasoning and follow the
rule, whether or not the judge believes the outcome is correct. This is a
difficult psychological feat.
Conclusion
The authoritative nature of law generates some psychological puzzles that,
although not limited to law, take on special social importance in the context
of adjudication. One set of questions relates to the rule-making role of judges:
what sorts of errors are judges prone to make in designing or evaluating rules
of common law and what mechanisms, if any, can a legal system use to control
their errors? A second set of questions relates to judicial compliance with
rules: do judges follow legal rules against their own best judgment, and if so,
how do they disengage from the process of deliberating about the relationship
between purposes and outcomes of rules? The mystery deepens when the
functions of rule making and rule following are closely juxtaposed.
Of course, theoretical analysis of the type presented here can at best
provide a map of the problem. As Schauer observes, the challenge is to find
some means of empirical access to legal reasoning. Judicial opinions provide
an immense source of information, but as Legal Realists are happy to point
out, they are written after their authors reach decisions. Opinions also may be
products of negotiation that do not reflect any single individuals process of
decision. Experiments involving judges may be more promising, but they
require heroic efforts and even then it may be difficult to recreate the moral
pressures of adjudication in an experimental setting.
Notes
Thanks to Jeffrey J. Rachlinski and Frederick Schauer for helpful comments.
1. For present purposes, I am adopting a broad definition of empirical reasoning
as the process of reaching conclusions about the world through observation or
experiment. A cautionary point is that empirical observations may incorporate
elements of induction, and both empirical and inductive judgments may be
intertwined with normative judgment. There is no perfect line of demarcation
between judgments of fact and judgments of law or between empirical and
moral reasoning.
2. For explanations of the method of reasoning to wide reflective equilibrium, see
Rawls (1971, pp. 1421, 4353, 578582) and Daniels (1979).
3. For discussion of the generality of rules, see Schauer (1991, pp. 1734).
4. Studies focusing specifically on judges, while they do not address the particular
problem of rule making, confirm that judges are not immune to these biases.
See Guthrie, Rachlinski, and Wistrich (2001, pp. 807811) .
5. Part of the problem is lack of coordination: even a very wise judge cannot know
with certainty what other, less wise, judges will decide and how many errors
they will make.
6. For extended analysis of this problem, see Alexander and Sherwin (2001,
pp. 6168).
7. Of course, rule makers can also make mistakes, and judicial rule makers are
susceptible to mistakes for the reasons mentioned above. My analysis, however,
assumes sound rules. Overruling unsound rules is a separate, although not
unrelated, question.
8. Larry Alexander and I have written in considerable detail about this gap,
concluding that it cannot be reasoned away. See Alexander and Sherwin
(2001, pp. 5695). For similar analyses, see Alexander (1991); Schauer (1991,
pp. 128134).
9. For an analytical (as opposed to psychological) argument against the view that
rules have an exclusionary effect, see Hurd (1999, pp. 6294).
10. Stanley Milgrams (1974) famous experiments demonstrate that people cer-
tainly are capable of obeying authority, either contrary to reason or without
engaging in reason. Interestingly, there are indications that Milgrams subjects
did not follow authority thoughtlessly, but rather struggled with the problem
(Milgram, 1974, pp. 4143).
11. For extended discussion of this point, see Alexander & Sherwin (2008,
pp. 6163, 115).
9
In Praise of Pedantic Eclecticism: Pitfalls and
Opportunities in the Psychology of Judging
Dan Simon
Though not always acknowledged, the subject matter of psychology

namely, human cognition and behaviorhas long played a role in important
jurisprudential debates. How do judges make decisions (e.g., Cardozo, 1921;
Holmes, 1881; Posner, 2008)? Are judicial decisions determined by the law or
are they driven by judges predispositions (e.g., Edwards, 1998; Frank, 1930)?
Are assertions of judicial constraint genuine (e.g., Altman, 1990; Kennedy,
1986)? Are judges better fact finders than jurors (e.g., Kalven & Zeisel, 1966)?
How do judges weigh the numerous incommensurable and conflicting con-
siderations involved in their decisions (e.g., Cardozo, 1921; Llewellyn, 1960)?
Yet, as a discipline, psychology is rarely recognized in the debate. Though
hardly a nascent field (see, Frank, 1930; Schroeder, 1918), the psychology of
judging remains an underdeveloped body of research.
The reluctance to apply psychological research to the domain of judicial
decision making can be understood to be based on at least two concerns.1 First,
unlike other disciplines that have successfully influenced legal discoursemost
notably, philosophy and economicspsychological insights tend not to be
deduced from overarching concepts or axiomatic characterizations of human
behavior. The multidetermined nature of human behavior defies parsimony
and makes experimental psychology a rather messy field. Psychological findings
are made piecemeal, one finding at a time. Properly presented, psychological
claims ought to be narrowed to certain experimental circumstances, and
qualified by boundary conditions and counter influences.
A second, and more familiar, concern stems from the fact that psycho-
logical findings are generated mostly in the highly controlled environment of
131
the laboratory, which are starkly different from worldly human practices in real
life. The concern pertains to the external validity of experimental findings, that
is, their generalizability to settings outside the laboratory. Psychologists, who
are ever so sensitive to situational effects on human behavior (e.g., Lewin, 1935;
Ross & Nisbett, 1991), are the first to acknowledge that results obtained in any
given study could have come out quite differently under different variations of
their experimental design. It is not hard to see why critics question the relevance
of findings obtained by testing a relatively small group of lay subjects (say, sixty
psychology undergraduates at a midwestern university), performing hypothe-
tical tasks, with limited knowledge, and under specific instructions.
Psychological studies have been criticized for the nonrepresentativeness
of the subject samples, the artificiality of the experimental setting, the dis-
connectedness from institutional contexts, the glossing over of individual
differences, the lack of appropriate incentives, the inconsequentiality of the
tasks, and more (e.g., Konecni & Ebbesen, 1986; Mitchell, 2002; Sears, 1986;
Yuille & Cutshall, 1986). These reservations warrant a healthy dose of skepti-
cism toward even cautious applications of experimental findings to real world
situations (see Mitchell, 2003). This concern seems doubly warranted when
applying the findings to specialized domains such as judicial decision making.
Concerns over the external validity of psychological research appear to
have animated Fred Schauers discerning and provocative essay in this
volume (ch. 7). The core of his essay is a call for a unique and genuine
psychology of judging. The proposal starts with an appealing proposition
that a psychology of judging ought to be what it claims to bean examina-
tion of judging, as performed by judges. Explanations that rely on cognitive
processes that are performed also by nonjudges and in nonjudicial domains
might make for an interesting application of psychological research, but they
do not constitute a discrete area of inquiry. Merely applying basic findings
to the work of dentists and plumbers is a long way from saying that there is a
psychology of dentistry or plumbing.
Schauer proposes to distinguish decision making by judges from decision
making by other people, even lawyers, for the same reason that one ought not
to equate the mathematical reasoning of a Harvard mathematics professor
with that of ordinary folks balancing their checkbooks. He contends also that
the field ought to concentrate on decision making in domains that are unique
to judging, such as finding, interpreting, and making law. Even if auto
mechanics and psychiatrists approach the finances of their small business in
similar fashions, it does not follow that they do the same when they fix cars or
provide psychiatric care. The underlying intuition is that judicial training,
acculturation, experience, and role may lead to deep cognitive effects that
generate process- and not just content-based differences between the cogni-
tive mechanisms of judges and those of nonjudge humanity.
Schauers proposal is constructive in that it actually suggests an experi-
mental project to test his proposition. The central hypothesis focuses on the
role of second-order reasoning in judicial decision making. Second-order
In Praise of Pedantic Eclecticism 133
reasoning stands for higher order decision rules that are supposed to trump
reasons that would otherwise provide a sufficient basis for a decision. Second-
order reasoning is deemed central to the judicial function in that it forces
judges to abide by a hierarchy of reasons, and specifically, to yield to higher
order considerations even when they feel that doing so leads to suboptimal or
unwise outcomes for the case at hand. Schauer intuits that judges engage in
second-order reasoning differently, more effectively, and with greater fidelity
than nonjudges. Specifically, the studies would test whether judges are more
inclined to follow a precedent or rule with which they disagree, and to limit
their attention to a narrow range of permissible sources even when other
sources lead to what they consider to be superior results. The studies are
designed to compare judicial decisions with decisions made by lawyers and
law students, and to compare decisions made by different classes of judges,
such as elected and appointed judges, trial and appellate judges, and judges
with different personal backgrounds. Undoubtedly cognizant of the com-
plexity of the issues involved, Schauer emphasizes the tentative and explora-
tory nature of his proposal, characterizing it as a set of hypotheses, an
invitation to consider a possible line of research.2
Schauers proposal also has a distinctive critical component. He addresses
extant research that tests real judges in a variety of experimental tasks that have
previously been tested with lay subjects. These studies tend to demonstrate that
judges are indeed prone to most of the same biases and errors as people in
general (Guthrie, Rachlinski, & Wistrich, 2001; Wistrich, Guthrie, &
Rachlinski, 2005). Schauer considers this research to be of marginal interest
to his project because it focuses on the fact-finding and verdict-rendering
dimensions of the judicial role, which are performed also by lay jurors.
Because the studied tasks are not performed exclusively by judges, this research
fails to meet the proposed standard of uniqueness. Schauer is far more critical
of the literature that applies basic psychology to judging.3 This research fails on
both dimensions of the proposed uniqueness in that it uses ordinary people as
subjects in the performance nonjudicial decisions. In other words, the external
validity gap is seen to render this work invalid. Schauer briskly dismisses the
application of this work for being axiomatic, unargued, and unresearched.4
Schauers chapter provides a fortuitous opportunity to closely examine
the concerns over external validity that hamper application of psychological
research to legal theory. This chapter will focus on both the prescriptive and
critical approaches of his approach. First, it suggests that the design of
psychological experiments can be far more complex and subtle an endeavor
than it appears. Along these lines, it should be appreciated that the intuitively
appealing proposal to run experiments with judges is burdened by methodo-
logical pitfalls. External validity is just one piece of a larger set of validity
issues, which tend to be intricately intertwined. Attempting to fix one aspect
can be overwhelmed by greater compromises elsewhere, thus resulting in
a net loss in validity. To prevent this, one needs to adopt a pedantic approach
to the design of the study.
At the same time, the concern with external validity should not be
exaggerated. The discrepancy between the experimental environment and
real world settings does not automatically bar all applications of findings
from the former to the latter. It does, however, require cautious work and
oftentimes also more data. When experimental findings meet the heightened
demands of external validity, they can be applied safely to real world domains,
including specialized ones. This possibility opens up the field of judging to a
wide range of methodological approaches and thus offers the benefit of
insights originating from eclectic perspectives. To demonstrate the possibi-
lities in this regard, the discussion will center on the application of a particular
body of researchcoherence based reasoningto judging.
In Praise of Pedantry
Before expending the limited experimental resources on the proposed studies
(there is no abundance of judge-subjects), one ought to ensure that the
studies are capable of providing the insights they aim to discern. In this
vein, a brief methodological detour would be helpful.5 It is imperative to
acknowledge that external validity does not exhaust the validity challenges
facing experimental research. The foremost criterion of the validity of any
psychological experiment is that it be internally valid, which stands for the
degree to which the experimental treatment explains the observed results.
Studies are said to be internally valid when they demonstrate that the variables
that were set up or manipulated by the experimenter (independent vari-
ables) were indeed the cause of variation in the focal point of the hypothesis
(the dependent variables). Internal validity rests heavily on the researchers
ability to design the study so that it provides maximal control over the
experimental environment. Control is necessary for the restraining of unin-
tended factors that might affect the dependent variable.6 There seems no
reason to believe that Schauers proposal would be lacking in internal validity.
A finding of differences in decisions made by judges and nonjudges in a well-
controlled environment could be fairly understood to be related to differences
in the manner in which the respective groups make decisions.
Detecting differences, however, can be a far cry from understanding their
underlying mechanisms at work. To bridge this explanatory gap, studies need
to be shown to have construct validity, which stands for the degree to which
one can correctly identify and explain the operative psychological constructs.
This important and somewhat neglected facet of validity stands for the degree
to which studies accurately operationalize their theoretical constructs, which
is essential for the correct explanation of the relationship between the experi-
mental treatment and the observed phenomena. Absent this validity, one
cannot draw reliable inferences from the observed results.7
It should be acknowledged that the various forms of validity are often in
tension with one another. Notably, the high levels of experimental control
that are essential for ensuring internal and construct validity cut against the
generalizability of the findings. Likewise, tests that are designed to have a large
degree of external validity, such as archival research and field studies, typically
lack important aspects of control, most notably, random assignment of
subjects to the various conditions. Herein lies the tension in Schauers
proposal. Recall that the proposed research seeks to discern differences in
the underlying cognitive processes between judges and nonjudges. This
aspiration can be problematic, especially since such deep constructs are
most difficult to study. Even if the proposed studies yielded the expected
resultsnamely, that judges decisions were better aligned with second-order
principles than decisions by nonjudgesthe underlying operative
mechanism would remain unknown. A finding that judges show greater
deference to a precedent might or might not be indicative of superior
second-order reasoning. It could also be due to the fact that lay people lack
a sense of the judicial conventions and practices involved in assessing case
similarity, distinguishing precedents, and more. By the same token, a finding
that judges display a heightened respect for jurisdictional limitations need not
stand for the proposition that judges engage in different cognitive processes.
It could be readily interpreted as standing for the proposition that lay people
lack familiarity with the constitutional principles that underlie the rules of
federalism,8 lack the nuanced knowledge of how to weight them against
competing considerations, and the like.9 Lay people are bound to be unfami-
liar with the professions conceptions of the hierarchy of reasons, the reputa-
tional damage of being overturned, and the personal commitment to the
judicial role.10
One plausible alternative explanation for hypothesized findings of judi-
cial superiority is that judges are experts at judging, whereas lay people are
novices. Psychological research on expertise shows that experts perform
differentlyin certain ways, betterthan novices (Chi, 2006; Ericsson &
Ward, 2007). The possibility that judges have a better grasp of the conven-
tions and practices of the judicial role, and are better in weighting and trading
off the competing rules is consistent with the advantages of expertise.11
Moreover, the determinants of judicial uniqueness noted by Schauer
namely, selection, training, and experienceclosely resemble the factors
that have been found to develop expertise (Feltovich, Prietula, & Ericsson,
2006).12 While Schauer explicitly rejects the explanation based on expertise,13
the proposed studies do not seem to provide a way to discriminate between
the two explanations.14
Underlying the proposed hypothesis is the belief that judges are consid-
erably superior to lay people in following second-order rules. Schauer is
skeptical of lay peoples ability to abide by second-order rules, and goes so
far as to suggest that they might not know how to engage in this form of
reasoning at all.15 The explanation offered for lay peoples low capabilities is
that they have little experience making decisions of this kind. Yet, there is
reason to doubt whether second-order reasoning is rare or undeveloped in
nonjudicial settings. People engage in some form of second-order reasoning

every time they walk by a store window displaying a coveted item that exceeds
their budget. Dentists oblige their patients requests even when they might
condone a different course of treatment. Plumbers heed the building code
and the architects plans even when a different solution seems to them to be
more effective. Second-order reasoning plays a central role in other nonjudi-
cial walks of like.16
While it is clear that judges do indeed engage in a considerable amount of
second-order reasoning, there is reason to suspect that their fidelity is less
than perfect. A substantial and growing body of quantitative analyses of
judicial behavior shows that judicial decisions are systematically biased by
judges personal attitudes (Klein, 2002; Segal & Spaeth, 1993, 2002), the
composition of panels (Cross & Tiller, 1998), personal prestige (Klein &
Morrisroe, 1999), and more.17 These observations suggest either that judges
do not fully adhere to the second-order reasons, or that they interpret and apply
those reasons in a biased manner that effectively undermines the rules norma-
tive import.18 It should be noted that one experiment that tested judicial
adherence to second-order rules revealed a rather lackluster level of compli-
ance. In this study, the judges decisions tended to be influenced by information
which they themselves ruled to be inadmissible (Wistrich et al., 2005).19
Moreover, it is worth noting that judicial experience might actually make
judges feel less bound by second-order rules than one might otherwise
believe. To a large extent, judging requires deciding not only which rules
and precedents ought to be followed, but also which ones ought not to be
followed. In most complex cases, judges are faced with multitudes of powerful
reasons (Schauer, 1988b), some of which are likely to pose a conflict between
two or more second-order rules. When such a conflict occurs, the judicial
dilemma requires a determination as to which second-order principle ought
to be followed and which one ought to be rejected. In such situations, the
fidelity to second-order rules necessarily entails a rejection of (other) second-
order rules, as indicated in Karl Llewellyns conception of dueling canons.20
It appears, then, that finding acceptable ways to ignore, dismiss, or interpret
away second-order rules is yet another facet of judicial expertise.
In Praise of Eclecticism
This brings us back to the ubiquitous concerns over the external validity
of basic psychological research. With so many degrees of freedom separating
the laboratory environment from real world contexts of human action, one
might wonder how experimental research can ever be deemed to bear any
practical relevance to real life. Yet, an array of experimental findings are
notably present outside the confines of the laboratory. As it turns out, the
gathering of intelligence by the CIA in preparation for the Iraq War21 bears
an eerie similarity to various forms of biased reasoning generated in the
laboratory (e.g., Frey, 1986; Klayman, 1995; Kunda, 1990; Nickerson, 1998).
The behavior of nations and ethnic groups entangled in real conflicts corre-
sponds closely to the behavior of arbitrarily formed groups in the laboratory
(e.g., Brewer, 1979). Marketers and political consultants routinely exploit
human judgment processes gleaned from the laboratory (e.g., Rozin &
Royzman, 2001; Nisbett & Wilson, 1977). Prejudicial behavior by employers
(Bertrand & Mullainathan, 2004) mirrors behavior observed in experimental
settings (e.g., Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997).
Likewise, jury decisions to send convicted inmates to their death (Eisenberg,
Garvey, & Wells, 2001) appear to be influenced by the same attitudes that affect
mock jurors decisions in hypothetical cases (Thompson, Cowan, Ellsworth, &
Harrington, 1984). A meta-analysis of experimentation is social psychology has
shown a rather strong correspondence between findings obtained in the
laboratory and in the field.22
Thus, it seems that applying psychological research to capture real world
phenomena is a complicated feat, which can be neither accomplished nor
controverted offhandedly. As discussed below, subject to a careful and meth-
odical examination, psychological research can be safely applied to some real
life situations, but not to others. Applications to the stylized domain of
judicial decision making require a heightened level of scrutiny.
For the research to be deemed useful outside the confines of the con-
trolled laboratory setting, it must first be shown that the observed phenom-
enon is not an artifact of the specific experiment. One important way to allay
this concern is by replicating the finding under similar and different experi-
mental settings. Validity increases when the same finding is observed using
different populations of subjects, stimulus materials, instructions, and tasks.
It increases also if replications are conducted in different laboratories.
Validity is further increased by the robustness of the finding, that is, its
recurrence under various manipulations, across wide ranges of values, and
in the presence of counterforces.23
Still, external validity does not guarantee that the findings apply equally
to every domain of human behavior. To be deemed applicable to a particular
real world practice, the finding must not be trumped, weakened, or distorted
by particular features of the domain, as these were not present in the experi-
mental setting and their potential influence on the finding is unknown to the
experimenter. This last link in the applicability chain has been coined con-
textual attentiveness (Arlen & Talley 2008), which stands for the degree to
which the experimental findings map onto the rich context of the real world.
Contextual attentiveness can be deemed an additional layer of external
validity. A threshold criterion for establishing contextual attentiveness is the
facial similarity between the behavior captured by the laboratory finding and
the behavior observed in real life. It is enhanced by the extent to which the
psychological theory that underlies the phenomenon is deemed germane to
the domain, particularly in the absence of competing theories. Strong support
can be derived from corroborative evidence derived from sources other than
the experiment itself, such as when archival or field data reveal behaviors that
are consistent with the experimentally observed phenomenon. Finally, one
might also look, with caution, to self-reports by people working in the
domain, particularly to those who are considered to be introspective.
It is important to note that there is no established gold standard for
determining external validity. None of the abovementioned features can carry
the day by itself, nor can any body of research be expected to fit them all. The
guiding principle is convergent validity: the more of the noted features that
converge toward validity, the more reliable the conclusion.
With these methodological guidelines in mind, we can return to assess
Schauers objection to the application of basic psychological findings to
judging. It is beyond the scope of this chapter to examine the merits of this
charge with respect to the various bodies of work Schauer mentions (in which
he candidly includes his own previous work, Schauer, 2006a, 2006b; see also
Arlen, 1998; Hanson & Yosifon, 2004). It is, however, feasible to assess the
objection as it pertains to the applicability of one of the bodies of work,
coherence based reasoning.24 To do so, it would be helpful first to review this
line of research.
Coherence based reasoning seeks to explain an enigmatic, yet prevalent
mode of reasoning in judicial opinions. Even the casual reader of judicial
opinions is likely familiar with the experience of being strongly persuaded by
an opinion, with all of its components converging to provide overwhelming
support for the outcome. The facts of the case, authoritative texts, governing
precedents, legal principles, public policies, as well as sheer logic and common
sense, all come together in a coherent whole to make for the inevitable and
undeniably correct result. The sense of correctness is bolstered by the dearth
or absence of arguments to the contrary. By the culmination of the opinion,
one might wonder how the decision could be considered to have been any-
thing but obvious in the first place. This sense of obviousness, however,
quickly dissolves upon turning to the opinion of the dissenting judges.
Dissenting opinions too tend to be strongly coherent and persuasive in
defending their conclusion, which is invariably antithetical to the majoritys
conclusion. Thus, while the opinions are exceedingly coherent internally, they
are radically inconsistent with a slew of seemingly plausible arguments con-
tained in the opposing opinion. The divergence between opposing judges
views of a case can lead them to stake remarkable positions. For example,
when interpreting statutes, it is not unusual for judges to deny outright that
the there is any ambiguity in the statutory text, notwithstanding the fact that
similarly positioned judges read the text to mean the very opposite.25 This
mode of reasoning is apparent in almost every appellate case.
As a matter of legal policy, this phenomenon has mixed effects that
exceed the scope of this chapter.26 As a theoretical matter, it offers an
opportunity to peer into the judicial decision-making process and better
understand the extent to which it is constrained by the law, as judges routinely
claim it to be. If indeed the coalitions of reasons mustered by judges
accurately represent the state of the law, one ought to infer that judicial
decision making is indeed tightly constrained by the law, and more impor-
tantly, that the law is determinative of single right answers. If, however, we
find an alternative explanation for this mode of reasoning, one might call into
question the professed constraint and, by implication, the binding nature of
the legal materials.
There are good reasons to suspect that the legal materials are considerably
less determinant than they are portrayed in judicial opinions. For one, most
cases decided by appellate courts are truly complicated and difficult, as they
contain sound arguments supporting each side of the issue (Schauer, 1988b).
The suspicion intensifies once opinions are dissected and stripped down to
their constitutive arguments. Relatively short U.S. Supreme Court opinions
typically contain a handful of core issues, each of which is supported by an
array of arguments, which can easily total fifty arguments or more.27
A notable feature of the opinions is that virtually every one of the dozens of
arguments supports the corresponding conclusion and contradicts the oppo-
site one.28 This perfect alignment of reasons is plainly implausible. Assuming
that the soundness of the opposing arguments are roughly similar, the
mathematical probability that each of the fifty or more arguments line up
perfectly is astronomically minute. This observation suggests that the overall
conclusion of the case plays a role in determining which arguments are
endorsed and which are rejected. While judicial decisions are most likely
affected by their underlying reasons, there appears also to be an effect in the
opposite direction, by which decisions affect the reasons that are claimed to
support them. This calls into question the avowed unidirectional relationship
between reasons and conclusions, namely that the former should affect the
latter, but not the other way round. Judicial reasoning, it would seem,
operates bidirectionally, from reasons to decisions, and back in reverse.29
This feature of judicial reasoning cannot be explained by the conven-
tional theories of decision making (e.g., von Neumann & Morgenstern, 1944;
Edwards & Newman, 1982), which seem ill suited to handle complex
decisions of the kind that judges facewhere the variables are numerous,
conflicting, ambiguous, and incommensurable. The phenomenon could,
however, be consistent with a body of psychological research that shows
that certain cognitive tasks are driven by coherence-maximizing processes.30
This line of research follows the tradition of cognitive consistency theories
notably balance theory (Abelson & Rosenberg, 1958; Heider, 1946, 1958) and
cognitive dissonance theory (Festinger, 1957)which are based, in turn, on
Gestalt psychology (Wertheimer, 1923/1938). Cognitive consistency theories
were animated by principles of structural dynamics, which posit that relevant
cognitive processes are determined holistically, rather than elementally. The
holistic structural properties are deemed to be dynamic, so that interrelated-
ness of the elements generates forces that determine the configuration of the
structure. Some things go together, that is, they are related by cohesive
forces, whereas other things tend to disperse. These forces determine the
stability of the structure and drive it toward a state of equilibrium, or Gestalt.

Perhaps most importantly, the dynamic forces that occur at the structural
level involve changes, or reconstructions of the cognitive elements
(Rosenberg and Abelson, 1960), that is, by distorting the state of affairs
(Asch, 1940, p. 454).
A series of experiments conducted by Keith Holyoak, Stephen Read, and
myself was designed to explore the possibility that a theory of coherence-
maximization would apply also to the domain of decision making, in parti-
cular to complex decisions like the ones judges make. To understand the
concept of coherence based reasoning, it would be helpful to briefly describe
the common design of the studies. In the first study, participants were first
presented with a pretest that contained a number of apparently unrelated
vignettes, that were followed by a statement or two that could be inferred
from them. Participants were asked to rate their agreement with the total of
twelve such inferences. Some vignettes involved factual judgments, and some
involved more abstract issues such as analogies and issues of social policy. In a
separate phase of the experiment, participants were asked to play the role of a
young judge, assigned to decide a civil case in which Quest, a software
company, filed a libel lawsuit against one of its shareholders, Jack Smith.
The company alleged that Smith posted a libelous statement about the
company that caused it to go bankrupt. The case revolved around six core
points of dispute, with each party making an argument on each of the six
issues. The key feature in the design was that the case was constructed from all
of the vignettes that were used in the first phase of the experiment, and the
litigants arguments were virtually identical to the inferences that followed
those vignettes. Participants were asked to render a verdict and to report their
confidence in the verdict. They were also asked to rate their agreement with
the twelve arguments made by the parties.
The central finding in these studies is derived from the comparison
between the ratings on the virtually identical items obtained at the two
different phases of the study. Consistent with the prediction from coherence
based reasoning, participants were found to have made confident decisions
despite the complexity and difficulty of the case. In comparison to the
moderate and noisy ratings given in the first phase of the study, at the point
of decision, the ratings manifested polarized states of coherence: participants
who voted for Smith reported strong agreement with the arguments that
supported his case and disagreement with the arguments that supported
Quests position, while opposite ratings were reported by participants who
decided for Quest (Holyoak & Simon, 1999).
The findings from this and other studies support the conclusion that
complex decisions are driven by coherence-maximizing processes, by which
peoples perceptions of the tasks shift during the decision-making process
from an initial state of conflict to a final state of steadfast coherence. By
the culmination of the process, the arguments involved in the task have
shifted and spread apart into two or more coherent subsets, one providing
overwhelming support for the emerging decision, and the other providing
depressed support for the rejected decision choice. This process is understood
to be adaptive in that it enables people to make decisions and conduct their
affairs even in the face of stifling complexity. It follows, then, that the state of
coherence is not a property of the arguments themselves, but rather an
artificial cognitive state imposed by the decision-making process. In other
words, attaining a state of coherence entails a certain distortion of the factors
involved in the decision.
Though inspired by judicial decision making, these experimental find-
ings were borne by a basic-psychological research program, and thus cannot
automatically be said to pertain to the domain of judging. To support the
claim of applicability, the research must be shown to withstand the test of
external validity.31 Coherence effects have been replicated repeatedly, without
fail, in a wide range of studies (Simon, Krawczyk, & Holyoak, 2004; Simon,
Krawczyk, Bleicher, & Holyoak, 2008; Simon, Pham, Le, & Holyoak, 2001;
Simon, Snow, & Read, 2004. For reviews, see Simon, 2002, 2004). The studies
have been replicated also by other researchers in the United States, Canada,
and Europe. One study replicated the findings using the same materials as
used in the original research (Glockner, 2007), while others tested a variety of
different tasks including financial auditing decisions (Lundberg, 2004, 2007;
Phillips, 2002), judgment and decision making (Glockner, Betsch, &
Schindler, under review), legal-economic behavior (Landeo, under review),
and evidence evaluation (Lundberg, 2004). Across the various studies, the
subjects have role-played young judges, jurors, arbitrators, auditors, and job
applicants, while other studies involved no role-playing at all. In all, the
studies have been tested with some 3,000 subjects, including undergraduate
students, graduate business students, a general sample of Internet users, and
experienced professional auditors.32 The studies have tested a wide range of
reasoning tasks, including high level inferences, analogies, rule application,
policy decisions, factual judgments, social judgments, probability assess-
ments, and personal preferences. The robustness of coherence effects is
further bolstered by the fact that they have been manipulated in numerous
ways, always yielding the hypothesized results.33
Recall that to apply a body of research to the real world, it must be shown
also that the research is contextually attentive to the particular domain.
Support for the applicability of coherence based reasoning to judging is
derived from the close resemblance between the coherence that is present in
judicial opinions and the coherence that is found in the laboratory: complex
and taxing decision tasks are resolved successfully, resulting in lopsided and
coherent sets of arguments, accompanied by high levels of confidence.34 The
theory underlying the laboratory resultsnamely, that the cognitive system
imposes coherence to facilitate choiceis consistent with the judicial func-
tion of producing compelling decisions even for the most close and contested
of cases. It is noteworthy that key components of coherence based reasoning
appear in the theorizing of the some of the notable commentators on
the judicial practice, including Holmes (1881, 1897), Cardozo (1921), and
Llewellyn (1960).35 Furthermore, the effect of coherence based reasoning on
appellate judging has been supported by a study that examined data from
actual court decisions (Beebe, 2006).
There are three additional ways in which the research on coherence based
reasoning maps onto judicial decision making. First, the research shows that
coherence-maximizing processing operates mostly at a low level of awareness.
People do not consciously manipulate their mental representation of the task.
Their views of the task shift toward coherent states spontaneously and
imperceptibly.36 This lack of awareness offers a plausible retort to the
Realist charge that judges consciously misrepresent the law by stacking their
opinions with whichever arguments support their preferred choices. The
research indicates rather that the excessive coherence is a natural by-product
of the decision-making process.37
Another overlap with judging emanates from a recent study that shows
that coherence can dissipate soon after the decision has been completed
(Simon et al., 2008). Coherence seems to be an ad hoc construct that appears
around the time of making the choice but does not linger on to limit
the decision maker in future cases where the constellation of variables will
not necessarily align in the same manner. Thus, while coherence tends to be
very strong within each decision, it can be rather weak across cases. This
finding suggests that people are capable of applying a particular rule or
principle in one case, and not following it in the next. This observation is
consistent with the view that judges apply differenteven opposingrules,
policies, and interpretive principles from case to case (Llewellyn, 1950, 1960;
cf. Schauer, 2007).
The research also seems to provide insight into the enduring question of
freedom and constraint in judicial decision making, that is, why critics view
the law as replete with indeterminacy and room for judicial discretion, while
judges persistently describe the legal materials as constraining. Coherence
based reasoning indicates that even though the legal materials in hard legal
cases are not constraining, they can be experienced as such once the judges
cognitive system has imposed coherence on them and shifted the vying
conclusions apart. Thus, the judicial characterization of constraint is best
understood not as a reflection of the legal materials, but as an artifact of the
cognitive process that people employ in the making of complex decisions,
judicial and otherwise.
In conclusion, Schauer is of course correct in objecting to facile applica-
tions of basic research to judicial decision making. Yet, there seems good
reason to conclude that coherence based reasoning meets the requisite stan-
dards of applicability. Whether one is persuaded by it or not, this application
ought not to be regarded as unargued, axiomatic, or unresearched. Moreover,
given the difficulties involved in understanding the judicial process, one
ought not rush to discard a body of valid and pertinent basic-psychological
research. Rather, what is needed is an eclectic stance, based on any valid and
informative research. An illustrative example can be borrowed from the

application of psychology to medicine, specifically, the growing field of
medical decision making. This field employs a variety of methodological
approaches, including testing of physicians (e.g., Leblanc, Brooks, &
Norman, 2002; Wallsten, 1981), applying basic-psychological findings
(Croskerry, 2002; Graber, Franklin, & Gordon, 2005; Pines, 2005;
Redelmeier, 2005), and more. The reliance on generic psychological research
in this highly specialized field manifests a belief that any type of valid
psychological research ought to be brought to bear to better understand
how medical decisions are made and how they can be made better
(Groopman, 2007). This prioritization of the usefulness of the research over
its uniqueness could likewise benefit the study of judicial decision making.
Conclusion
The psychology of judging is poised to benefit much from a call for more
experimentation, especially when it comes from a scholar of Schauers stature.
Schauers essay provides a good opportunity to force people working in the
field to think hard about the strengths and limitations of their methodological
choices. Hopefully, it will also attract others to engage these issues and
contribute to the development of the field.
Still, the specifics of Schauers critiques and the proposed research are not
free of objections. The attempt to increase the external validity of the experi-
mentation does not come without costs. The insistence on uniqueness ends
up compromising the proposed studies construct validity, and thus muddies
up the conclusions that could be drawn from them. Methodological tradeoffs
of this kind hound experimental psychologists on a regular basis.38 It is
possible also that like many other important aspects of human behavior,
the hypotheses posed by Schauer simply do not lend themselves to experi-
mental testing. The insistence on uniqueness might also undermine the
contributions from nonunique research, especially basic psychology.
Instead of uniqueness, the field has most to gain from a pedantic attention
to the experimental design coupled with open-mindedness to the range of
useful methodologies.
The answer to the question posed in the title of Schauers essay ought to
be: yes, there is a psychology of judging. Admittedly, it is underdeveloped. To
foster its growth, the field should be guided by the understanding that the
practice of adjudication comprises a wide and diverse range of decision-
making processes. In some facets of their work, judges exert judicial expertise,
in others they behave just like ordinary people, and it is possible that in some
facets they engage in processes that are unique to them. Researchers should
opportunistically employ whichever methods are best suited for the subject of
inquiry. Looking forward, the field stands to be enriched by carefully vali-
dated findings from all strands of psychology: basic psychology, applied
psychology, the psychology of expertise, and perhaps also by a unique psy-

chology of judging.
Notes
1. In this chapter, the term psychological research refers to research based on
experimental studies.
2. Note that the proposal bridges two ingredients of external validity, namely, the
population of participants and the type of decision tested. It does not bridge
other aspects, such as realism, consequentialism, and the like.
3. Basic psychology focuses on phenomena that are relatively generalizable across
people, situations, tasks, and contexts. Basic psychologists research phenomena
like memory, reasoning, and persuasion. Applied psychology research focuses on
testing psychological phenomena as they are performed in particular contexts.
For example, applied psychology tests memory performance in the context of
witness testimony, reasoning in police investigations, and persuasion in poli-
tical campaigning.
4. Specifically, the essay states that the premise underlying this literature is an
undocumented and unargued premise that lurks in the background (p. 2). It
is described as something that researchers merely assume (p. 2); an unex-
pressed and typically unresearched outlook (p. 14); and a viewpoint that is
taken as axiomatic but hardly based on systematic research directed precisely
at that question (p. 20).
5. For useful discussions on methodological aspects of experimental psychology,
see Aronson, Wilson, and Brewer (1998) and Mitchell and Jolley (2007).
6. In the following examples, imagine a jury simulation that is intended to test the
effect of gruesome photographs from the crime scene on verdicts in the murder
trial. The hypothesis in this example is that the exposure of the fact finders to
the photographs will result in an arousal of negative emotion, which will
increase their tendency to convict.
A typical violation of internal validity is the failure to control for alternative
potential causal effects, also known as confounds. In this example, assume a
comparison between one group that receives a case that contains gory photo-
graphs and another group that receives a different case that does not contain
photographs. Given the discrepant stimuli, different rates of conviction (if
obtained) could not be explained as driven necessarily by the exposure to the
gory photographs. They could readily be explained by the fact that the two
groups received different cases. The evidence in the former case might have
been more incriminating.
7. In the abovementioned example of the study testing the effect of gruesome
photographs on jury verdicts, observing the heightened conviction rates in the
presence of gruesome evidence does not in itself provide a satisfactory under-
standing of the effect. While it is possible that the effect was driven by the
arousal of emotion (the hypothesized cause), it is also possible that it was driven
by the fact that the photographs contained incriminating information that tied
the defendant to the crime.
Construct validity is needed, first, to overcome the fact that human cognition
is notoriously multidetermined. Thus, to reliably interpret experimental
findings, studies need to be able to isolate the hypothesized mechanism and rule
out alternative explanations. Studies designed to have this capability are said to
have discriminant validity. Second, the experimental design needs also to over-
come the fact that psychological phenomena are generally not directly obser-
vable. To attain construct validity, the design needs to be able to identify the
psychological mechanisms that drive the observations. Studies designed to be
capable of identifying the correct construct are said to have content validity.
Content validity is enhanced also by the extent to which the construct fits into a
broader underlying theory. Theories are particularly useful when they are
capable of explaining other related constructs. The content validity in the
study testing the effect of gruesome photographs could be enhanced by showing
that the finding can be explained by a theory that explains other effects of
emotional arousal, such as in judgments of tort liability (Lerner, Goldberg, &
Tetlock, 1998) and aggressive behavior (Bushman, 1995).
8. Schauer actually discusses the potential problem with lay understandings of the
concepts of federalism and jurisdictions (fn. 29), but does not treat it as a
potential methodological problem with the study.
9. For an insightful demonstration of evaluating and weighting judicial goals see
Robbennolt, MacCoun, and Darley (this volume, ch. 2).
10. A tempting solution to these difficulties would be to provide lay subjects with
special training about these matters in preparation for the experiment. To the
extent that the training would be feasible and effective, it could jeopardize the
studys internal validity. For example, lay subjects might interpret the instruc-
tion as a cue for a desired decision.
11. It must be noted, however, that expertise is also characterized by subpar
functioning. Experts tend to display overconfidence, fail to notice details, and
are less agile in adapting to change. Expertise is also no guarantee against the
effects of bias (Chi, 2006; Koehler, Brenner, & Griffin, 2002). It is also important
to note that expertise is typically narrow in scope. While experts perform
differently on one type of task, the do not necessarily perform differently on
adjacent tasks, even within the domain of their expertise.
12. The expertise explanation would also explain the examples Schauer uses to
illustrate the uniqueness of judging: the difference in the mathematical skills of
a Harvard professor of mathematics and lay people, and differences between a
psychiatrist and a mechanic when it comes to providing psychiatric care or to
fixing cars.
13. See section 4, The Question of Expertise.
14. The proposed comparison of judges to nonjudges parallels what researchers in
the field of expertise call the relative line of inquiry. A different type of research
takes an objective approach, namely, focusing on how well the experts perfor-
mance stacks up against predetermined measures of excellence (Chi, 2006). By
adopting the former approach, the proposed studies will, at best, indicate that
judges are somewhat better at some aspects of judicial decisions than lay people.
But that finding falls short of answering what is arguably the crucial questions:
whether judges are good enough relative to some objective expectation, and
whether they fulfill their constitutional role satisfactorily.
15. Referring to second order reasoning, Schauer characterizes judges as people
who know how to x and contrasts them with lay people who do not know how
to x at all (p. 20).
16. For example, human resource personnel are habitually confronted with
considerations such as employment laws, company policies, maintaining con-
sistency with prior cases, and setting an example for future ones. It should be
noted that the article by Sunstein and Ullman-Margalit (1999) cited by Schauer
pertains to second-order decisions made in nonjudicial contexts.
17. This body of research is a good example of the productive use of cross-
disciplinary research. While the constructs underlying these findings are psy-
chological, much of this research has been performed by political scientists, in
nonexperimental settings.
18. A recent experiment conducted with law students demonstrated how second-
order rules are distorted by the ideology of the participants. See Furgeson, J. R.,
Babcock,. L., and Shane, P. M. 2008a), Behind the mask of method: Political
orientation and constitutional interpretive preferences. Law & Human Behavior.
19. True, the tasks involved in these studies were more akin to fact-finding and
rendering of verdicts, but there is no obvious reason to believe that the
performance would be better in appellate-like decision making.
20. Llewellyn, K. N. (1950), Remarks on the theory of appellate decision and the
rules or canons about how statutes are to be construed. Vanderbilt Law Review,
3, 395.
21. See Senate Report 108-301. Report of the Select Committee on Intelligence on
the U.S. Intelligence Communitys Prewar Intelligence Assessments on Iraq.
July 7, 2004 (http://intelligence.senate.gov/pub108thcongress.html).
22. The correlate coefficient of the findings was found to be about 0.73. See
Anderson, A. A., Lindsay, J. L., and Bushman, B. J. (1999), Research in the
psychological laboratory: Truth or triviality? Current Directions in Psychological
Science, 8, 39. The similarity between laboratory findings and field findings
does not ensure that the findings are applicable to real world applications, but it
does allay some of the concerns about the artificiality of the laboratory setting.
Another aspect of validity is ecological validity, which captures the similarity
between the experimental setting and the real life domain. It is undeniable that
the bulk of basic psychological research has little ecological validity with respect
to the practice of judging, or to any other real world practice for that matter.
Still, external validity ought not be confused with ecological validity. High
ecological validity naturally increases external validity, but the latter is not
dependent on the former.
23. To follow the abovementioned example, the external validity of the study of the
effect of gruesome photographs will be increased by showing that the same effect
is observed in other experimental variations, such as using different photographs,
media, populations, factual patterns, judicial instructions, and the like.
24. Schauer criticizes the application of coherence based reasoning to judging
(Simon, 1998, 2002, 2004), and also singles out some of the underlying
empirical work, namely Holyoak and Simon (1999) and Simon, Krawczyk,
and Holyoak (2004).
25. See Simon (1998), p. 71.
26. While it can be said to increase the acceptability of the opinions, this mode of
argument can be deemed detrimental to adjudication in that it obscures the
complexity of the issues involved, blunts the thoroughness of judicial analysis,
and unduly devoids the validity of the losing side. For a discussion, see Simon
(1998), pp. 129134.
27. There is of course more than one way to break down a case and to enumerate its
components. Alternative ways will always be possible, though the differences
should not bear on the substantive conclusions of the analysis.
28. This form of inquiry was first demonstrated in the analysis of Ratzlaf v. United
States, 510 U.S. 135 (1994), a relatively mundane case chosen almost at random.
See Simon (1998, pp. 6172, and 73102 passim). For another example, ana-
lyzing Rogers v. Tennessee, 532 U.S. 451 (2001), see Simon (2002). The Bush v.
Gore case provides a more familiar manifestation of the phenomenon (draft on
file with author).
29. One explanation for the excessive coherence of judicial reasoning is that the
strong alignment of arguments is a product of padding of opinions (Posner,
1995a). The concept of padding implies that not all reasons play the same role
in the decision, as some are deemed to actually exert power on the decision
maker, whereas others merely serve as ex post justifications. This seems true,
though it is doubtful that padding could fully account for the observed
coherence (see Simon, 1998, pp. 3536). To illustrate, even if all but the
handful of core issues served as mere justifications, one would still need to
explain how the handful of core arguments lined up to cohere with the
decision. For example, in the Ratzlaf case there were six core issues.
Assuming that they were about equally plausible, the probability of all six
lining up to support the respective conclusion is one in sixty-four. A precise
assessment of likelihoods would depend on the degree to which the issues and
arguments are independent of one another.
30. The processes include vision (McClelland & Rumelhart, 1981), social reasoning
(Read & Miller, 1998; Read, Vanman, & Miller, 1997), analogical reasoning
(Holyoak & Thagard, 1989; Spellman & Holyoak, 1992), relational inference
(Hummel & Holyoak, 2003), text comprehension (Kintsch, 1988), and more.
These strands of research are based on a connectionist architecture, and are
resolved via Parallel Constraint Satisfaction Mechanisms (Holyoak & Thagard,
1989; Read, Vanman, & Miller, 1997).
31. The studies internal validity and construct validity are beyond the scope of this
chapter. Concerns over internal validity can be put to rest in light of their
salience in the extensive peer review that the studies have undergone.
32. The professional auditors were tested performing an actuarial task (Lundberg,
2007).
33. Manipulations can be found in Holyoak and Simon (1999), studies 1, 3;
Simon et al. (2001), studies 1, 2, 3; Simon et al. (2004a), study 1; Simon et al.
(2004b,) studies 1, 2, 3, 4.
34. See Simon (1998), pp. 61102.
35. See Simon (1998), pp. 102121.
36. Holyoak and Simon (1999), study 2. This lack of awareness further enables the
endurance of this type of reasoning in the judicial practice (Wilson & Brekke,
1994).
37. See Simon (1998), pp. 134137.
38. A key to the success of research projects is the critical examination that takes
place in lab meetings at the early stages of experimental design. Of the many
seemingly good ideas proposed at these meetings, only a small fraction make it
to the phase of experimentation, not to mention onto the pages of scientific
journals.
10
Judges, Expertise, and Analogy
Barbara A. Spellman
One appellate case, three courtsand seven disparate opinions. Clearly,

different judges reach different decisions based on the same facts and same
legal doctrine. Why? Political scientists have shown that one can anticipate
how a judge will decide a case more often than chance, or a reading of the
facts, might allow. Using various predictorsparty affiliation, party of
appointment, the judges own decisions on earlier similar casesregression
analyses can demonstrate that judges are behaving in a manner consistent
with their explicit prior beliefs (e.g., Segal & Spaeth, 1993, 2002). The simplest
explanation for such behavior is that judges first decide what they want the
outcome of the case to be, then go back to find the precedents that justify their
opinions.1 The more complicated claim that I want to make is this: people
(and judges) may choose relevant analogies (or precedents) as better or worse,
applicable or inapplicable, not because of any particular desired outcome but
rather because of their own preexisting knowledge. The influence of such
knowledge on the decision process may be entirely unconscious; therefore,
judges may, in fact, be following the idealized decision-making process to the
letter, and be unmotivated toward finding a particular result, yet may usually
still reach the predicted result.
To understand this argument, I first present an overview of the analogical
reasoning research done by cognitive psychologists. Next I address the ques-
tion of whether judges are experts at analogical reasoning. If they are experts,
then the large body of empirical analogy research conducted with nonexpert
subjects might not be relevant to judges analogical reasoning. However, I
conclude that although judges might be expert at many things, analogical
149
reasoning is not likely to be one of them. Accordingly, I turn to other

researchincluding some from analogy and some from other areas relevant
to analogy like similarity and categorizationto show how nonattitudinal
and nonteleological factors (especially preexisting knowledge) can affect
analogy use. Finally, I link these arguments back to the initial question:
whether judges, or anyone, can be making predictable decisions while still
following an idealized analogical reasoning process.
Overview of Analogical Reasoning

Analogical reasoning is a core component of intelligence. Most intelligence
tests, not to mention general standardized tests like the GREs and LSATs,
incorporate some kind of analogical reasoning tasks. Performance on analo-
gical reasoning tasks correlates very highly with performance on almost all
other components of IQ tests (Salthouse, 2005; Snow, Kyllonen, & Marshalek,
1984).
On those tests, analogical reasoning often consists of verbal four-term
problems (like lawyer:client :: doctor:?) and geometric versions of such
problems. Although those simple tasks have been studied in the laboratory,
research using more complicated real world analogy materials has proven
more informative.
Analogical reasoning involves taking a situation that is well understood
(the source) and using it to help explicate a situation that is less well
understood (the target). Two important distinctions are (1) between the
processes of retrieval and mapping, and (2) between the attributes and
relations within analogs.
Steps in Using Analogies
Analogical reasoning typically involves several steps including retrieval and

mapping. To illustrate: Suppose you are a lawyer and a potential client comes
to you wanting to know whether she has a good negligence claim against a
cruise line. She had been asleep in her locked cabin when someone reached
through a window and stole her handbag including $500 in cash.
To figure out whether she has a good claim, you first need to retrieve
find potentially analogous source cases in memory (or by doing some legal
research). First, you recall a case in which a businessman was asleep on a train
berth in an open sleeping car and had his expensive cell phone stolen from the
pocket of the coat he was using as a blanket. Second, you recall a case in which
a man in a resort hotel had his wallet stolen from his room while he slept.
Third, you recall a case in which a woman on a cruise ship was hit by another
woman on the ship who used her handbag as a weapon.
The next step is to create a mappingfind a set of appropriate corre-
spondences between elements of the source and target. You might think of
Judges, Expertise, and Analogy 151
your client as the business traveler, the cruise ship as the train, and handbag as
the cell phone. Alternatively, you might think of your client as the man on
vacation, the ship as the hotel, and the handbag as the wallet.
If you think your case is most like that of businessman on the train (who
lost), you will expect the same outcome as in that case; if you think it is most
like that of the vacationing man in the hotel (who won), you will expect that
result. But despite the fact that the third case involves a woman, a cruise ship,
and a handbagjust like your ownit probably doesnt seem very similar to
your case because the objects that are the same dont stand in the same
relations to each otherand relations are the key to analogy.
Similarity in Using Analogies

The second important distinction is between attributes and relations within
the analogs (Gentner, 1983; Holyoak & Thagard, 1989). Attributes are objects
or qualities of objects (or events or people); they are one-place predicates like: is-
a-planet or is-valuable. Relations are two (or more) place predicates. Relations
may link objects, for example: is-bigger-than, revolves-around, owns, has-a-
duty-of-care-toward. But relations may also link other relations, for example,
the relation cause is important in linking propositions like: (1) The sun is
bigger than the earth and (causes) (2) the earth revolves around the sun. Or:
(1) An innkeeper has a duty of care toward those who rent rooms (plus some
other stuff) and (causes) (2) the innkeeper is liable for the theft.
Note that attributes are often referred to as surface or superficial
features because they are usually physically visible or explicitly described,
whereas relations must often be inferred.2
The difference between attribute and relational features is illustrated
wonderfully in an experiment in which subjects were asked to find similarities
between pairs of pictures (Markman & Gentner, 1993). In one pair, the top
picture showed a tow truck towing a car to the left along a road; the bottom
picture showed a (very similar looking) car pulling a motorboat to the right
along a road. Subjects were asked which object from the top picture
matched the car from the bottom picture. There are two obvious answers.
Subjects who had to answer the question quickly were more like to match the
car on the bottom to the car on the top because those two objects were very
similar in looks; that is an attribute match. Subjects who had more time were
more likely to match the car on the bottom to the tow truck on the top
because those two objects filled the same role (i.e., of pulling something else
that could not move over the road on its own); that is a relational match.
Attribute similarities between the source and the target play a greater role
in the retrieval of analogs, whereas relational similarities play a greater role in
mapping (Gentner, 1993; Holyoak & Koh, 1987). When people are asked to
judge the goodness or similarity of analogies, the depth and structure of
the relational similarities matters much more than the attribute similarities
(Gentner & Kurtz, 2006; Gentner, Ratterman, & Forbus, 1993).
Note that the difference between attribute and relational similarity is key
in the use of analogical reasoning in the law. A useful precedent is not usually
one in which the parties themselves (or the property in question) are similar
but rather one in which similar (legal) relations hold between the relevant
parties or property.
So, in the earlier example of the theft on the ship, the relevant similarities
are not whether the victim was a woman or man, whether it was a business or
pleasure trip, or what was stolen;3 rather what matters is the relation between
the victim and the owner of the ship, train, or hotel. And although most
people believe that a ship is more similar to a train than to a hotel, the court
in Adams v. New Jersey Steamboat Company (1896) ruled that for liability
purposes a ships cabin is more like a hotel room, where there is an expecta-
tion of privacy and protection, than like an open berth in a sleeping car, where
there is not.
Developing expertise in law is (at least in part) learning to ignore
irrelevant attribute similarities, learning what counts as a relational similarity,
and understanding which relational similarities are likely to matter in a given
case.
Analogy and Expertise

What would it mean to say that judges are experts at analogical reasoning?
Because analogical reasoning is a core component of IQ, and because judges
are likely to be a more intelligent group than a random collection of folks,
judges are more likely to be better than average at analogical reasoning. But
are they experts? And why is it important?
What Is an Expert?
Due to study, training, and practiceoften in addition to talent and

motivationexperts are better than nonexperts in some domain of perfor-
mance. Expert chess and golf and bridge players routinely beat nonexperts;
expert surgeons perform difficult surgeries more successfully than nonex-
perts; expert violinists create truer sounds and make fewer mistakes than
nonexperts.
One clear characteristic of expertise is that it is quite limited in domain.
Expert golf players are not experts at all sports or even all sports requiring a
swing or a ball. Expert neurosurgeons are not expert cardiac surgeons nor are
expert violinists expert cellists. Much expertise involves being good at a
procedure that involves a very specific content.
Within their domain of expertise, experts tend to be faster and more
accurate than novices, tend to have superior short-term and long-term
memories for information, see deeper relations in the structure of informa-
tion, use less cognitive effort, and have more accurate monitoring skills
(Chi, 2006; Glaser & Chi, 1998). These qualities are thought to reflect not just
that experts have more knowledge but also that experts have a qualitatively
different representation and organization of knowledge (Ericsson, 2006).
A potentially important characteristic of experts for the law is that, in a
variety of domains, being an expert means seeing past attribute similarities
to underlying relational similarities. For example, when shown index cards
depicting different physics problems and asked to sort them into piles, novice
physics students sort them based on the type of objects involved in the
problems (e.g., pulleys, inclined planes) whereas expert physicists sort them
based on the underlying principles involved (e.g., conservation of
momentum) (Chi, Feltovich, & Glaser, 1981). When using analogies, experts
are more able to retrieve previous analogs based on structural similarity and
avoid interference by those exhibiting only surface similarity (Novick, 1988).4
Thus, someone who is legally trained is less likely to be distracted by the hit-
by-a-handbag-on-a-cruise-ship case than someone who is not.
Another general characteristic of expertise is that it only develops given
specific conditions. One is that the person spends a lot of time at itbut, of
course, time itself is not sufficient. We all know people who have played golf
or bridge for years, and who seem to enjoy it, but who arent any better than
they were years ago. More important than just the amount of time is how that
time is spent: expertise develops out of a process called deliberate practice
which involves thousands of hours of specific types of practice and training.
Deliberate practice requires focused programmatic study with appropriate
feedback about performance. It includes identifying errors and working on
procedures to eliminate them (Ericsson, 2006; Horn & Masunaga, 2006).
Why Is Expertise Important

In his essay in this volume (ch. 7), Schauer poses the question: Is there a
psychology of judging? His answer is essentially no. He states that so far all
the experimental evidence shows that judges demonstrate the same cognitive
failings as mere mortals. But he correctly points out that psychologists have
not (often) studied what judges do when they are doing what judges often do.
He argues that it is therein their domain of expertisewhere a psychology
of judging would exist. Of course that is true of other expertsspecial
performance is only found within the domain of expertise; thus it is ironic
that Schauer perfunctorily dismisses most research on expertise as being
irrelevant to his conjecture.
In the first generation of theories of expertise (Holyoak, 1991), psy-
chologists believed that experts in any particular cognitive skill should be
experts at general reasoning and, therefore, should be experts at other rea-
soning tasks. But the next wave of research consistently found that experts,
despite specialized skills and virtuoso performances, made the same errors as
nonexperts on all sorts of general reasoning tasks; their expertise was limited
to their domain of detailed knowledge. However, no one found that judges
made those particular errorsbecause hardly anyone studied judges.5 And

now they have (e.g., Guthrie et al., 2001; Wistrich et al., 2005).
Some people seemed surprised that judges show the standard errors and
biases on the standard cannon of reasoning tasks; however, for those of us
who believed the findings that doctors and Indian chiefs were not special with
regard to all kinds of reasoning tasks, and were only special with regard to
tasks related to their expertise, the fact that lawyers and judges are not special
with regard to those tasks is not the least bit surprising (or interesting).
However, as Schauer argues, if we can define what judges expertise is, we
can (1) begin a smarter inquiry into the psychology of judging and (2) argue
that within their limited domain(s) of expertise, judges should show the kinds
of enhanced performance of experts and should not fall prey to the errors that
nonexperts would generate. Clearly, one candidate for judges expertise is
analogical reasoning.
Are Judges Experts at Analogical Reasoning?

Elsewhere others and I have argued that judges are not experts in several tasks
that might be viewed as components of judging. For example, it could be
argued that judges are neither expert fact-finders (Robinson & Spellman,
2005) nor expert at appropriately weighting evidence (Spellman, 2007). One
reason for the theorized lack of expertise is that although (some) judges may
often do those tasks, they are not trained to do them with extensive super-
vision and feedback.
In contrast, law school does train students (whether explicitly or impli-
citly) to reason analogically. In law school we had the pleasure of years of
reading cases, abstracting rules and similarities, drawing analogies to other
cases or hypotheticals, and being given corrective feedback about our ana-
lyses.6 And, in fact, law schools often (explicitly or implicitly) use techniques
in teaching that have been shown to improve analogical reasoning in the
laboratory. Thus, it seems as though the conditions for developing expertise
at analogical reasoning might be met.
Law School Techniques and the Possibilities of Improving

Analogical Reasoning
Although the psychology literature is fairly glum about peoples ability to take
what they have learned in one domain and use analogy to transfer that
knowledge to another domain (see Barnett & Ceci, 2002, for a review),
there are, in fact, ways to improve peoples performance on analogical
reasoning tasks. In the laboratory, bad performance on analogical reasoning
tasks is often to due a failure to retrievesubjects trying to solve a problem do
not find relevant analogs in memory; however, once they are told that a prior
experience might be relevant, they are good at mapping.
In a basic laboratory procedure for studying analogical reasoning, subjects

first learn about a way of solving a hypothetical problem. For example, a
general wants to attack a well-guarded fortress but if he sends his entire
army down one road, land mines will explode and significantly deplete his
forces. A better plan is to send smaller groups of soldiers down different
roads toward the fortress simultaneously. This divide-spread-and-converge
approach is called the convergence solution (e.g., Gick & Holyoak, 1980,
1983; Holyoak & Koh, 1987).
Time passes and subjects are later asked to solve an analogous problem,
typically Dunckers Radiation Problem (Duncker, 1945). A man has an inop-
erable tumor in his stomach. A type of radiation can destroy the tumor but if it
is used at sufficiently high strength to destroy the tumor it will destroy the
healthy tissue that it passes through and the man will die. What can be done?
There is a convergence solution to the radiation problem: use many less-
powerful rays simultaneously from many different angles. About 10% of
subjects will come up with that solutionand that is the same percentage
whether or not they have previous read the fortress problem. Why doesnt
having seen the obvious analogy help?
The main obstacle to using the earlier solution is that people do not think
of itthat is, they fail to retrieve it from memory. However, if people are
reminded of itfor example, if the experimenter tells them to think back to
something they did earlier in the studythen most will think of the source
analog, easily draw the mapping, and solve the radiation problem. Without
explicit reminding (or expertise), however, people are only likely to think of
superficially similar source analogs (Holyoak & Koh, 1987).
Laboratory studies that demonstrate ways to increase or improve the use
of analogical reasoning therefore tend to address the accessibility of source
analogs. But, of course, the source analog is a given; how can it be made more
accessible? Although the analog is fixed, peoples mental representations of
the analog can differ. For example, rather than storing the fortress problem in
memory as a story about a general attacking a fortress, people could store it
more abstractly, as a story about how a too-strong force can be split up into
smaller forces and then converge to accomplish a goal. Later, when a new
target situation comes along that has a similar abstract relational structure,
the abstract version of the source analog is more likely to be retrieved because
the two analogs seem more potentially related (e.g., if they are both about
converging forces rather than one being about an army and the other about
radiation).
Note that in some ways this characterizes the task of the law student,
lawyer, or judge searching for relevant precedentwhen facing a new fact
pattern to find good analogous cases in memory. However, the legal task is
also different. For one thing, in the legal arena people know that they should be
trying to find an analogous case. For another, except for law students on an
exam, the search for such cases is not just a search through memory but rather a
search through a database or through cases presented as relevant in legal briefs.
Below I describe three techniques that were designed to improve remind-

ing in the laboratory but are similar to techniques used in law school pedagogy.
They are also what is needed for developing expert knowledge within a content
areamaking important similarities more obviousin particular, making
relational features as obviousand as superficialto experts as attribute
features are to novices.
Creating More Abstract Source Representations Indirectly

by Comparing Analogs
One way to improve analogical reasoning in the laboratory is to have subjects

compare and abstract from multiple analogs. For example, subjects who read
both the fortress story and a story about firefighters (who used many small
hoses simultaneously from different directions) before trying to solve the
radiation problem were more than twice as likely to come up with the
convergence solution as subjects who had read only one source analog. In
addition, if subjects are asked to explicitly compare the two analogs to each
other, those who abstract the convergence solution from the comparison are
more likely to use it later than subjects who do not have a good representation
of the relational similarities between the stories (Catrambone & Holyoak,
1989; Gick & Holyoak, 1983). This compare-and-abstract technique has been
shown to benefit business school students in negotiation classes who, like law
students, participate in case-based learning (Loewenstein, Thompson, &
Gentner, 1999). In law school, hypotheticals may provide the grist for com-
paring analogs.
Training People to Abstract Principles from Single Analogs

Another way to improve analogical reasoning is to train people to represent
single source analogs at an abstract level. Mandler and Orlich (1993) had
subjects read the fortress story and then describe the story at one of three
different levels of abstraction: summarizing the story details; summarizing
the main points by stating the generals goal, dilemma, and solution; or
abstracting a general solution principle. When subjects later tried to solve
the radiation problem, those who had produced a good abstract solution were
much more likely to use the convergence solution and solve the radiation
problem than subjects who had not. (Note, however, that very few subjects
actually succeeded in creating a good abstract summary, suggesting that such
a representation is hard to create without either training or practice.) In law
school, students learning to extract abstract rules from single cases.
Teaching the Names of Relations
A third way to improve analogical reasoning is to use consistent relational

labels when people learn the analogs. Although the laboratory data with
adults is sparse, the idea is consistent with various kinds of developmental and
anecdotal evidence. People tend to use the same labels for objects (e.g., car,
tow truck) but different labels for verbs and relations (e.g., pulls, tows,
drags)making it easier to use objects/attributes in retrieval and also
making relations more difficult to learn (Gentner & Kurtz, 2007; Gentner &
Loewenstein, 2002). Relational categories seemed to be learned by progres-
sive alignmentby first comparing examples that are similar, then com-
paring more distant ones (Gentner & Kurtz, 2007).
Law students learn the names of many legal relations: contracts, torts,
negligence, standing, jurisdictionall are about the relations between parties
and/or actions that create legal rights or obligations. Certainly, there are some
legal categories that are attribute-based: there are laws that apply only to
people over 18 years old and there are laws that apply only to ships. However,
much of law school is about learning, by contrasting many examples, the
requirements and limits of legal relations.
But Does a Law Students Analogical Reasoning Actually Improve?
In effect, all of the above techniquescomparing multiple analogs, abstract-

ing from single analogs, learning the names of legal relationsare techniques
used in law school to teach the content of the law. Psychologists, however, do
not have any measures that demonstrate that law school improves general
analogical reasoning. In a study of the effects of graduate training on rea-
soning, law students, medical students, and graduate students in psychology
and chemistry took tests involving statistical, methodological, conditional,
and verbal reasoning during the first and third years in the programs
(Lehman, Lempert, & Nisbett, 1988). The verbal reasoning test included
verbal analogical reasoning (as on the GRE or LSAT). The first-year law
students had higher initial verbal reasoning scores than the other groups
suggesting self- (or law school) selection. However, after three years of
schooling, the law students improved only about 5 percent on average
(a statistically nonsignificant difference) in verbal reasoning; all of the other
groups average scores improved more.
Granted, these data showing no improvement in analogical reasoning are
not the bestamong other flaws they only include law students (at the
University of Michigan) after three years of training rather than experienced
judges and, of course, the verbal reasoning tasks are not the same as the type of
full-blown analogical reasoning done when reasoning about cases. However,
these data are consistent with a wide variety of other data showing limitations
on both the transfer of training and the generalization of expertise.
Expertise and the Process/Content Interaction
The best way to think of what judges may have developed is that it depends on
both process and content: it is using analogy in a domain in which they have
specialized knowledgeknowledge that enables them to quickly understand

which features of a case are the relevant ones for analogical mapping. Thus,
within the legal context (or, more likely, within a subset of that context),
judges are experts at using analogy; however, when reasoning outside their
knowledge base, although they may be more fond of using analogy than most
people (because of practice or precocity), they will not be any better than
equally intelligent and informed others.
To return to the cruise ship example, probably no one who was legally
trained would think that the sex of the victim, the nature of the trip, or the
particular items that were stolen would matter in that case; even if those
features bring to mind similar cases, those that do not have an underlying
structural similarity (e.g., the women being hit by the handbag on the
cruise ship) would be easily rejected as irrelevant. And those who are
legally trained should be less flustered by the surface similarity that boats
and trains move whereas hotels do not. Rather, those who know that the
law protects those who are justified in expecting privacy and security
whether passengers or hotel guestswould be more likely to recall,
recognize, and use the analogy between the obligation of a ship to a
passenger in a private cabin and the obligation of a hotel to a guest in a
private room.
Nonattitudinal (Unintentional) Factors Affecting Analogy

Selection and Use
To the extent that judges are not any different at analogical reasoning
from nonjudges, their use of analogical reasoning should be affected by
similar extra-logical influences. Below I describe several factors that influ-
ence the selection and interpretation of analogies that are not driven by a
conscious motivation to find one analog more relevant or persuasive than
others. I focus here on knowledge and representation; that is, how what
someone knows, and the relations they consider between the things they
know, affects analogy use. However, there are other cognitive factors that
might also work to make the choice of precedent seem teleologically
motivated.7
Knowledge and Unconscious Reminding

People may show unconscious influences of irrelevant parts of the source
analog on selecting and using analogies. (Of course, that assumes that
people are familiar with the source analog.) For example, when subjects
read about a hypothetical political crisis and were asked whether they
thought the United States should intervene, those who read a version with
several superficial reminders of World War II (e.g., a briefing in Winston
Churchill Hall) made more interventionist recommendations than those

who read a version with superficial reminders of Vietnam (e.g., a briefing
in Dean Rusk Hall). Thus, the superficial features (unconsciously) affected
the retrieval of similar analogs and those (unconsciously) influenced the
interpretation of the target analog. Interestingly, however, when the subjects
were later asked to explicitly rate how similar the hypothetical crisis was to
both World War II and to Vietnam, the differences in superficial details had
no effect (Gilovich, 1981).
Knowledge and Interests

A persons knowledge or interests can unconsciously influence which of
several equally good analogical mappings will be chosen. For example, subjects
read a science fiction story describing two different planets (Spellman &
Holyoak, 1996). Planet 1 had three countries: Afflu was economically strong
and gave economic aid to Barebrute; Barebrute was economically weak but
militarily strong and gave military aid to Compak; Compak was militarily
weak. Planet 2 had four countries: Grainwell was economically strong and
gave economic aid to the economically weak Hungerall; Millpower was
militarily strong and gave military aid to the militarily weak Mightless.
Subjects first made military or economic recommendations for each
country. Then they matched the countries of Planet 2 to the countries of
Planet 1. Which country was like Afflu? Easy, Grainwell. Like Compak? Also
easy, Mightless. But which was like Barebrute? There are reasons to pick both
Hungerall and Millpower.
Subjects mapping choices depended on the recommendations they
made. Subjects who made military recommendations saw Barebrute as
more like the militarily strong Millpower; in contrast, subjects who made
economic recommendations saw Barebrute as more like the economically
weak Hungerall.
Thus, knowledge or interests may drive mappings within an ambiguous
analogy.
Planet 1:
Afflu ! aids ! Barebrute ! aids ! Compak
(economically (economically weak; (militarily
strong) militarily strong) weak)
Planet 2:
Grainwell ! aids ! Hungerall Millpower ! aids ! Mightless
(economically (economically (militarily (militarily
strong) weak) strong) weak)
Figure 10.1 Representation of materials from Spellman and Holyoak, 1996.

Incorporating Structure in Levels of Abstraction
Which (of many potential) analogs one regards as best may depend on how
much of the abstract analogical structure one incorporates (Hofstadter &
Mitchell, 1994). Consider the following question:
If the string of letters abc is changed to abd, how would you change kji in
the same way?
This question is, in effect, a four-term analogy question (like lawyer:
client :: doctor:patient): find the relations in the first pair of letter strings, map
abc to kji, and then apply the relations to create the fourth term.
The following two solutions are based on superficial features and are very
literaland people find them very unsatisfying: kji could be turned into abd
or into kjd. The first ignores all internal properties of both abc and kji (and
ignores how kji might be similar to abc); the rule is: turn any letter string into
abd. The second also ignores all internal properties of kji but does consider the
relation between abc and abd; the rule is: change the rightmost letter to d.
A less literal solution is to turn kji into kjj. That solution means con-
sidering the relation between abc and abd and also understanding that c is part
of an alphabetical sequence and that d is one step lower in the sequence; the
rule is: change the rightmost letter one step lower.
People generally prefer the two analogical solutions that incorporate the
internal relational structure of the letter strings. Generating the answer kjh
means seeing that abc is a downward sequence from which the rightmost letter
is moved one more step down; however, kji is an upward sequence, therefore,
the rule is: move rightmost letter one step more in the direction of the sequence.
Generating the answer lji also means seeing that abc is a downward sequence
and kji is an upward one. Then abd means changing the lowest (rightmost)
letter down one; lji means changing the lowest (leftmost) letter down one.
One of the fascinating things about the Hofstadter letter-string analogies
is how many different ones can be sensibly generated from such seemingly
simple stimuli. Legal cases are similarly complicated in that they may contain
relations that are or are not incorporated into the litigants arguments and
that analogies may be drawn at very literal or abstract levels. For example, the
same case may be viewed as being about the right to engage in homosexual
sodomy or the right to be let alone (Bowers v. Hardwick, 1986).
Coherence
The selection of a particular analog, or relevant precedent, might also (unin-
tentionally) emerge out of a general pressure for cognitive coherencethat is,
the tendency for people to be consistent in their reasoning. As a consequence,
peoples views of the applicability of a source analog change in line with other
changes in their opinions (for overviews of this research see Simon, 2004;
Simon & Holyoak, 2002). In the basic study (Holyoak & Simon, 1999),
subjects (undergraduates) read a semifictional legal case. The plaintiff, a
software company named Quest, sued Jack Smith, an investor in the com-
pany, for libel. Smith had posted a negative message about the company on an
electronic bulletin board directed at investors, and soon after the stocks price
dropped drastically and the company went bankrupt.
Before reading the case, subjects were asked whether they thought mes-
sages posted on electronic bulletin boards should be treated like items pub-
lished in newspapers or like messages sent over a telephone network. Later, as
part of the case, subjects learned that defamatory messages published in a
newspaper could give rise to a cause of action for libel whereas those trans-
mitted by telephone could not.
Subjects were about equally divided in verdicts. But whereas before
reading the case, subjects found the newspaper and telephone analogies
equally compelling, after rendering their verdicts, they widely diverged.
Those who found for Quest believed that the newspaper analogy was much
better than the telephone analogy; the opposite was true for those who found
for Smith. Thus, belief in the quality of an analogy shifted coherently along
with other beliefs that led them to their decision.
Legal Knowledge
An experiment comparing law students to undergraduates demonstrates how
legal knowledge can affect analogical reasoning (Braman & Nelson, 2007, Exp.
2). Subjects (96 undergraduates and 77 law students) read an article summar-
izing the target case facts (but not the result) of Wazereud-Din v. Goodwill
Homes and Mission, Inc. (1999) in which the plaintiff, an Islamic man, was
denied admission to a Christian-administered drug treatment program. They
also read about a potentially relevant previously decided case. The precedent
case varied in its similarity to the target case and involved one of three
different plaintiffs: Islamic man, gay man, or black man; one of three different
types of defendant: religious treatment program, community service organi-
zation, or insurance company; and one of two holdings: discrimination or
not. Subjects rated how similar the precedent was to the target case. They had
also previously been asked whether they agreed that faith-based organizations
should have the right to exclude people who did not share their beliefs from
receiving their services.
Several findings demonstrate the effects of legal knowledge: (1) overall,
undergraduates rated the precedents as more similar to the target case than
did the law students, and (2) undergraduates treated precedents involving
Islamic and gay plaintiffs similarly whereas law students tended to treat
precedents involving Islamic and black plaintiffs similarly. These results
suggest that the law students were using their prior knowledgeof the
difference between strict scrutiny and rational basis testsin their
judgments.8
But Braman and Nelson (2007) want to go further than merely stating
that knowledge matters to analogy use; they argue that there was (sometimes)
evidence that subjects prior beliefs mattered: in general, those who opposed
exclusion based on religion were more likely to rate the target case as similar
to the precedent when the holding was that there was discrimination, but
those who favored exclusion were more likely to rate the target case as similar
to the precedent when the holding was that there was no discrimination. This
pattern was stronger and more consistent for the law students.
Therefore, Braman and Nelson (2007) also conclude in favor of moti-
vated reasoningthat peoples policy preferences (e.g., regarding exclusion)
influence legal decision making (p. 954). However, the causal conclusion
that preferences influence reasoning, and, especially, that it might have influ-
enced reasoning in a consciously motivated way, is not justified. Subjects are
not randomly assigned to favor or oppose exclusionthus, subjects who start
out holding different beliefs may differ from each other in other systematic
ways (creating a so-called third variable problem). For example, subjects
who are more tolerant of different groups and less likely to see differences
between them might both (1) be more likely to oppose exclusion and (2)
more likely to see similarities between the cases. Other types of preexisting
knowledgenot necessarily directly related to the casecould have similar
effects.
Note that the authors hasten to add that nothing we have found suggests
a conscious effort to twist the law to serve ones preferences (Braman &
Nelson, 2007, p. 954). It is easy to see how these results could emerge
simply from different knowledge and from the (unconscious) pressure for
coherence in reasoning.
Judges and Analogy
What can we conclude? Judges have had lots of practice using analogy; yet,
they might not actually be experts because just as there is no real generalized
expertise in problem solving it is not clear that there can be a generalized
expertise in analogy use. More important, however, judges (like laypeople)
know that when using analogies it is important to look for relational simila-
rities andbecause of their specialized training in legal contentthey know
which relational similarities matter within their domains of expertise.
Many of the limitations on using analogies described above have to do
with finding or retrieving the proper analogs to use. Judges dont have to try
to retrieve from memorythey have briefs and law clerks to find the relevant
sources. Yet, as the WWII/Vietnam study shows, unconscious remindings of
known analogs that are not present can affect judgments even though, when
made explicit, the analogs are not viewed as any better or worse than other
ones. In addition to this automatic retrieval of analogies, judges knowledge
and interests may influence how they mentally represent and use different
analogs. When judges know more about some issues than others, or, in the
past, have drawn analogies to one kind of outcome, they might be more likely
to unintentionally find in a direction consistent with past judgmentsin part
because of what they see as more (or less) similar, in part because of the level
of abstraction (i.e., how deep the relations) they use, and in part because of an
effort to maintain coherence in their beliefs.
Thus, although judges might decide consistently with predictions, it is
possible that they do so not for any of the intentional (and sometimes
seemingly nefarious) reasons suggested by legal realism. Regression ana-
lyses can reveal that it happens but understanding how analogical reasoning
works, and how judges might use it, is necessary for understanding why it
might happen.
Notes
1. This position is the most extreme version of the legal realist view. A more
nuanced view is that they are sensitive to both attitudinal and jurisprudential
concerns (Lindquist & Klein, 2006).
2. The terms surface and superficial are often mixed both within and across
articles.
3. Note that although irrelevant in this case, such factors could be relevant in other
types of cases.
4. When acquiring analogical skills, children parallel this development (Ratterman
& Gentner, 1998).
5. But see Lawrence (1988) for an early study of how Australian magistrates impose
sentences; and, Dhami (2003), with a larger data set showing that British judges
bail decisions can be well described by a simple heuristic model.
6. In a sense, when lawyers write briefs, and when judges read and rule on them,
they are engaged in a similar activity. Of course, the kind of feedback that
lawyers and, especially, judges get is much more erratic and sporadic than that of
the law student.
7. Other potential factors include context effects (see Hunter, 2001); whether
people set out to look for similarities or differences between cases (see Gentner
& Markman, 1994; and Medin, Goldstone, & Gentner, 1990); and beliefs about
the causal structure of relevant legal categories (see Kim & Ahn, 2002).
8. Under current constitutional doctrine, actions that treat people differently based
on race and religion merit strict scrutiny, whereas those based on sexual orienta-
tion do not.
11
Thresholds for Action in Judicial Decisions
Len Dalgleish, James Shanteau, and April Park
Many decisions that people are called on to make can be thought of as involving
thresholds for action. Should one buy a new computer, attempt a left turn into
oncoming traffic, excuse a late submission from a student? In each case, we can
understand the decision maker to be answering two questions: (1) How strong
are the arguments in favor of taking this action? (2) How strong must the
arguments be in order for me to take the action?
Decision makers in court cases, whether judges or jurors, are commonly
required to make this kind of decision. Take a hypothetical, but fairly typical,
drug trafficking case in a U.S. court, where the police uncovered one piece of
evidence by stopping the defendant on the street for questioning, subjecting
him to pat-down search, and finding a bulky packet of marijuana in his coat
pocket. In one of several motions to the judge, the defendants lawyer argues
that the marijuana should not be allowed into the trial as evidence because the
search was unconstitutional. The judge must decide, among other things,
whether the police reasonably suspected that the defendant was, at the time of
the search, armed and presently dangerous.1 Her decision will depend on
both her assessment of the evidence available to the police at the time and her
understanding of reasonable suspicion. At the end of the trial, the jury (or
judge, if a jury trial has been waived) must decide whether the prosecution
proved the defendants guilt beyond a reasonable doubt. Again, the decision
will require judgments about both the evidence itself and the threshold the
evidence needs to cross.
These are just two examples of a great many. Appellate judges may have
to decide whether a decision made by a trial judge constitutes an abuse of
165
discretion or, in another case, whether a trial judges mistake is harmless.

Judges at either level may have to decide whether an administrative agency
has taken an arbitrary and capricious action or whether a state law is
narrowly tailored to serve a compelling governmental interest. This list
could be extended for pages.
The aim of this chapter is to set out a framework for analyzing decisions
to take action in a judicial context. We begin by outlining a general model,
continue with a description of several studies of mock-juror decision making,
and conclude with implications for studying judges.
Model for Assessing Situations and Deciding to Take Action

The key to the model is a distinction between (1) the forming of an assess-
ment given the evidence available to the decision maker and (2) the decision
to take action or not, given that assessment. Consider a motion hearing, in
which a judge hears the arguments, reads the submissions, and makes a
decision. It is assumed that all the information presented to or available to
the judge is combined to form an opinion or an assessment about the strength
of evidence favoring approval along a continuum from low to high. An
assessment is formed through a psychological process of integrating the
information.
This assessment is then turned into a decision about action to deny the
motion or not by a decision rule. It is assumed that on this continuum
the judge has a personal threshold for action and this is in place before the
judge forms an assessment of the amount of evidence. The threshold is an
internal cutoff level for taking action. The decision rule is this: If the
assessment is above the threshold the judge approves the motion. This
model formalizes the question: Is the evidence sufficient to grant the
motion? The threshold is the minimal amount of evidence a judge
requires to grant the motion. Although it may feel like the assessment and
the decision occur at the same time and cant be separated, it is useful to
distinguish them since different factors influence the assessment and the
decision (Dalgleish, 2003).
The left hand side of the model in Figure 11.1 shows that the factors
affecting the formation of the assessment of strength of evidence are from
information in the current situation presented to the judge. The right hand
side of the model shows that factors from the experiences and history of the
judge influence the placement of the threshold. These factors are not from the
current case, and the threshold is in place before the current situation is
assessed. For example, a judge may have a reputation for being tough in that
he requires a lot of evidence before granting a motion. Another may have the
reputation for being easy to convince.2 Of course, a judges knowledge and
experience influence the assessment. For example, their experience may have
Thresholds for Action in Judicial Decisions 167
HIGH
Factors
Factors Influencing
Influencing Assessment Threshold
Assessment.
Information from
Information from Experiences and
Current situation History of
being Judged Yes Decision Maker
(The Past)
Threshold
No
LOW
Assessment Dimension
e.g. Strength of Evidence
Decision Rule:
If the Assessment of the current situation is above the threshold, decide Yes.
If the Assessment of the current situation is below the threshold, decide No.
Figure 11.1 A general model for assessment and decision making (GADM).
taught them to give more weight to certain types of evidence than others. The
case factors and to some extent their weight are reasonably easy to articulate.
Cooksey (1996) reviews how to elicit and analyze the factors influencing the
assessment. This chapter focuses on how to make explicit the factors influen-
cing threshold placement. Since they are not case or situation factors, they are
less easy to articulate.
An implication of the model is that it points to two different explanations
for why different people when presented with the same evidence come to
different decisions. Consider two judges presented with the same information
in a motion hearing, where one chooses to grant the motion and the other
denies it. According to the model, the difference in outcomes might come
from differences in the assessment of the strength of evidence, but it may also
arise from divergent thresholds for action, or from both causes combined.
This chapter focuses on the implications of variation across people in their
personal thresholds for action.
The left hand side of the model derives from Signal Detection Theory
(SDT) (McNicol, 2004), which has had wide influence beyond its initial
development for the interpretation of sensory processes in the 1950s. In the
example above, the judges task is to discriminate between two states: the
motion should be granted and the motion should be denied. For a juror in a
Table 11.1 The Four Categories of Response in the Context of Juror Decision
Making in Criminal Trials
Should have convicted Should have acquitted

because defendant is because defendant is
truly guilty truly innocent
You decide to convict Correct Conviction Wrongful

(Hit) Conviction (False
Alarm)
You decide to acquit Wrongful Acquittal Correct Acquittal
(Miss) (Correct No)
criminal case, the task is to discriminate between two states: the defendant is
guilty and the defendant is innocent. The signal the juror has to detect is
the guilt of the defendant. This decision is made under uncertainty.
Hammond (1996) argued that in all decision making there is irreducible
uncertainty. This means that no matter what decision is made, deny or
grant, convict or acquit, there is always the possibility of an error. The
fourfold table makes explicit that there are two types of error. Table 11.1
illustrates this for the jurors task. The two types of error are Wrongful
Acquittals, or misses, where the juror has missed the signal, and Wrongful
Convictions, or false alarms. For the judges motion hearing example, the
two types of error are wrongful denial of the motion and wrongful
granting.
There has been much interest in the relationship between the two types
of errors. Volokh (1997) demonstrated that people have expressed a great
many opinions about what the ratio of Wrongful Acquittals to Wrongful
Convictions is or should be. One of those opinions is by the English jurist
William Blackstone, who said, Better that ten guilty persons escape than that
one innocent suffer (cited in Volokh, 1997). The ratio 10:1 has become
known as the Blackstone Ratio (BR).
There has been considerable debate on the meaning of the numbers in
the BR (Dekay, 1996). One view is that the numbers in the BR indicate
frequencies in a population of cases. Assuming that jurors assess the like-
lihood of guilt as a probability, the numbers in the BR can be used to
establish a beyond reasonable doubt probability against which they com-
pare their assessed likelihood of guilt. Another view of the BR is that it
captures the worth, utility, or value of one persons being wrongly con-
victed. Signal Detection Theory (SDT) combines these two views. In SDT,
the decision maker is assumed to maximize the percentage of correct
decisions and to maximize the utility of the decision. Using Bayes theorem
and Subjective Expected Utility theory, Coombs, Dawes, and Tversky
(1970) give the formula for threshold placement in terms of the prevalence
of signals and the utilities for the four possible outcomes in Table 11.1.
Dekay (1996), in the context of juror decision making, points out that the
merit of SDT is that it separates two aspects of decision performance. These
are the ability of the person to discriminate between the two states in Table
11.1 and the persons willingness to take action, their threshold placement.
It can be equally helpful in analyzing judges decisions on which side to err
in considering a motion.
Studying Thresholds and Decisions to Take Action

In this section, we demonstrate studies designed to tap the factors influen-
cing the threshold placement. Although the studies reported below use
undergraduate students in simulated jury decision making, they could be
adapted to use with judges making decisions such as motion decisions or
any other decision able to be modeled by the GADM approach. The studies
indicate that thresholds and the factors influencing them can be measured.
This is important both theoretically and practically. The GADM model
provides a theoretical account of disagreement between decision makers
and provides a methodology for making the locus of the disagreement
explicit. Practically, procedures can be developed to reduce the amount of
disagreement and to make explicit the importance of values in the setting
of thresholds for action.
Identifying Thresholds
The first study to be discussed was designed to test whether giving different
definitions of the standard of proof shifted the threshold for beyond reason-
able doubt, but we ignore that element of the study here.3 Participants were
given general information about a particular murder trial from South
Australia, Pfennig v. The Queen (1995). This case was chosen because it was
largely based on circumstantial evidence and was appealed through to the
High Court of Australia. Table 11.2 gives a brief summary of the case.
Table 11.2 General Description of the Pfennig Case
Briefly, Pfennig v The Queen involved Peter Pfennig being charged with the murder of a
10-year-old boy named Michael Black. At trial, the Crown prosecutor argued that Pfennig had
abducted Michael for sexual purposes and then killed him. Michaels body was never found.
The Crown presented the court with much circumstantial evidence, including evidence of
Pfennig talking to Michael at the last place that Michael was seen alive. The Crown also
provided evidence that Pfennig had pleaded guilty to abducting and raping a 13-year-old boy
almost one year after Michael disappeared. In that case Pfennig had placed his victims
belongings at the top of a cliff to encourage speculation that the victim had accidentally fallen.
Similarly, Michaels belongings were found arranged neatly beside a river so as to suggest an
accidental drowning.
Our approach to measuring thresholds was to present participants with

100 cases based on the Pfennig case. The hypothetical cases showed different
combinations of seven different types of evidence that varied in amount. That
is, each case was one way in which the evidence could have been. The types of
evidence in the case were elicited from a group of lawyers who reviewed the
trial and appeal transcripts. Figure 11.2 shows an example case. The cases were
constructed so that there were two types of case that had overlapping dis-
tributions of strength of evidence. One set of 50 cases had an average of a high
amount of evidence, 60 out of 100, and the other set had a low average
amount of evidence, 40 out of 100.
Those with the high average were arbitrarily designated as guilty, the
signal, and those in the other set as innocent. This is accord with the
assumption that if the legal process works at all, truly guilty defendants
will, on average, appear more guilty than truly innocent defendants (DeKay,
1996, p. 101). This procedure is analogous to studies on auditory signal
EVIDENCE
Strength of Evidence
low high
0 20 40 60 80 100
Cuel: degree of match
between the description of 1
the man and Pfennig
Cue 2: degree of contact

with Michael that Pfennig 2
admitted to his daughter
Cue 3: Byrons degree of

certainty that Michael did 3
NOT ride across the
bridge to Thiele Reserve
Cue 4: Mrs Goulds
degree of certainty that 4
the commotion was caused
by Pfennigs van
Cue 5: Mr Jones belief 5
that he saw Pfennigs van
at Thiele Reserve on 18
January
Cue 6: the chance that 6
Michael did NOT go
swimming
Cue 7: the success of 137

7
prior searches for bodies
in the river
Figure 11.2 Example case vignette used in Study 1.

detection where the experimenter defines which stimuli are the signals and
which are no-signals (Ryder, Pike, & Dalgleish, 1974). The important thing
about this procedure is that it enables measurement of the threshold.
Three groups of 10 students from the University of Queensland partici-
pated as part of their course requirements. All participants were registered on
the electoral roll as required by s.4 of the Jury Act 1995 (Queensland).
Each participant received information on the general case, definitions of
the evidence cues, and practice cases. Each also received one of three different
definitions of reasonable doubt. For each of the 100 hypothetical cases they
made a judgment of the overall strength of evidence on a nine-point rating
scale and made a decision whether to convict or acquit.
Table 11.3 shows a cross tabulation of the decisions made by two partici-
pants, chosen because they were given the same definition of reasonable doubt
and their evaluations of the strength of evidence in different cases were quite
similar. Despite these similarities, Juror C made 79 convict decisions whereas
Juror A made 25 convict decisions. Their rates of Wrongful Convictions and
Wrongful Acquittals differ considerably. Juror C is a Convictor because for a
given case they need little evidence before they convict. On the other hand Juror
A is an Acquitter and for a case needs much more evidence before they
convict. Calculations from a Signal Detection Theory (SDT) analysis, Stainslaw
and Todorov (1999), indicate that the index of threshold placement, ln(b), is
1.5 for Juror C and 1.2 for Juror A. Note that ln(b) is the natural logarithm of
the threshold (or response criterion, to use SDT terminology). This index is
useful because it is symmetrical around zero and is independent of the accuracy
of the decision maker. It is negative when the participant has a decision
tendency toward convicting. The differences in the decisions of these two
Table 11.3 Results for Two Participants
Juror C Juror A
True State True State
Guilty Innocent Guilty Innocent
Actual Decision: Hit False Hit False

YES Convict Alarm Alarm
(WC) (WC)
31 3
48 62% 22 6%
96% 44%
Actual Decision: Miss Correct Miss Correct
NO Acquit (WA) No (WA) No
2 19 28 47
4% 38% 56% 94%
participants are due to threshold differences, that is, differences in their

threshold for beyond reasonable doubt.
Although these differences are particularly striking, we found consider-
able variation in willingness to convict among the participants. This is con-
sistent with early experiments by Simon and Mahan (1971) showing a large
variance in the way jurors interpret the phrase beyond a reasonable doubt.
Judges could participate in a similar study. The instruction condition
could be dropped and a relevant case to the sample of judges used and
developed into a set of hypothetical cases. Of interest is the degree of varia-
bility in threshold across judges. An advantage of this methodology is that
sufficient data is collected from each judge to enable the regularities in their
decisions to be estimated (Cooksey, 1996). From the analysis of each judges
data their threshold and its standard error is estimated for each judge. Thus,
a direct comparison of threshold placement between any two judges is
available.
Exploring Differences in Thresholds

Consider Jurors A and C making a decision about the same case. Assume that
they have worked hard on agreeing on the evidence in the case and they both
assess the strength of evidence as middling on the overall strength of
evidence scale, JA = JB. Figure 11.3 illustrates how differences in threshold,
TA and TB, explain why Juror C decides to convict and Juror A decides to
acquit.
SDT offers suggestions for accessing the locus of this variability. In SDT
the threshold is determined by the base-rate or prevalence of the signal and
the utilities for the outcomes in the fourfold table. In judicial decision making
the true prevalence is not known. For example, in decisions about motions
the proportion of motions that should be denied is not known and cannot
High
Threshold JA
Acquit
JA JC
Assessment. Assessment.
Convict
Threshold JC
Low
Assessed level of Strength of
Evidence for Guilt
Figure 11.3 Conflict in decision making when there is agreement on the
evidence.
be known. For the criminal trial, the proportion of truly guilty defendants is
also not known. However, decision makers have beliefs about these propor-
tions based on their experiences and history either directly or vicariously
gained. That is, they have in mind a perceived prevalence that is their prior
beliefs about the need to take action. This influence on the threshold is in
place before the decision maker has any information about the current case or
situation. Let p(Na) be the proportion of cases for which the decision maker
believes they need to take action for a particular type of decision. If a decision
maker has a high p(Na) then given the knowledge that they are about to see a
case, they need less strength of evidence to decide yes than someone with a
low p(Na).
The second theoretical influence on the threshold is a function of the
utilities the decision maker has for each outcome in the fourfold table, Table
11.1. These utilities are subjective in that each decision maker has their own
set of utilities and these derive from their experiences and history either
directly or vicariously gained. If a decision maker wishes to avoid misses
(Wrongful Acquittals), then they need less strength of evidence to decide yes
than someone who wants to avoid False Alarms (Wrongful Convictions). In
SDT both these influences are combined mathematically and captured in an
index of threshold placement, ln(b).
In the next three studies to be briefly reported we attempt to measure the
relative utilities and beliefs about the need to take action directly by eliciting
peoples prior beliefs about the prevalence of guilt. One view of the Blackstone
Ratio is that it reflects the value people put on a Wrongful Acquittal com-
pared to a Wrongful Conviction, that is, 10:1. We wanted to find out how
people replied when we asked them for their own version of the BR directly
and to look for variation in the elicited BR.
Thirty-six undergraduate Psychology students at Kansas State University
participated. They were randomly assigned to three groups that received
different labels for the standard of proof. Each participant was given a careful
instruction individually. Participants were asked to imagine themselves on a
jury, given a case to read, shown in Table 11.4, and asked to make a decision
about whether to convict or acquit.
Table 11.4 Case Used in Study 2
A 32-year-old, Paul, was shot to death in his car outside Brooklyn, New York, at 11:30 pm on
November 15, 2005.
The witness saw a man wearing black clothing who walked up to Paul and shot him several
times as he was entering his car.
Paul was a restaurant manager. According to colleagues, he was friendly and successful.
However, he had financial problems with his ex-wifes boyfriend, John.
Police accused John of the murder, and a shoeprint matching one of Johns shoes was found
near the crime scene.
However, John said he was in bed with Pauls ex-wife at the time, and she confirms the fact.
Think of 1000
defendants on trial
for murder
Of these, how
Of these, how many
many are
are convicted?
acquitted?
Of these, how
Of these, how Of these, how Of these, how
many are
many are Guilty? many are Innocent? many are Guilty?
Innocent?
Figure 11.4 Diagram to elicit prior beliefs about guilt.
They were then asked to think about murder cases in general and asked
this question to elicit their Blackstone Ratio, HOW MANY truly guilty
defendants charged with murder do you believe should be acquitted to
avoid ONE truly innocent defendant being convicted? To elicit their prior
beliefs about guilt, they were asked to think about 1000 defendants before the
court charged with murder. They filled out the boxes in Figure 11.4 by first
thinking how many of the 1000 were convicted or acquitted. For those that
were convicted they were asked to state how many did they think were truly
guilty and how many were truly innocent. They did the same for those they
thought were acquitted.
Results and Discussion

To calculate the participants prior beliefs about guilt, the numbers in the first
and third boxes from the left in Figure 11.4 were added and divided by 1000.
There was large variability in the prior beliefs. The mean was .62 with a
standard deviation of .19. The range was from .23 to .96. The interpretation
is that when thinking generally about murder cases participants on average
believed that 620 out of 1000 would be guilty. More worrying is the range
with values of .80 or above for 7 participants and .40 or below for 5 partici-
pants. That is, without any evidence some participants had a strong tendency
to convict whereas other had a strong tendency to acquit. The implication
from the GADM model is that a sample of these participants serving on a jury
would show decisional conflict as illustrated in Figure 11.3.
Table 11.5 presents the frequencies for the elicited Blackstone Ratio. The
findings are quite variable with only 3 participants having a Blackstone ratio
of 10 to 1 or greater. What is interesting is the fact that participants wanted
no guilty defendants to be acquitted to avoid one wrongful conviction and
that just over half the participants gave this as their answer. That is, most
Table 11.5 The Number of Wrongful Acquittals to Avoid One Wrongful

Conviction
Elicited Blackstone Ratios
0:1 1:1 2:1 3:1 4:1 10:1 100:1
Frequency 19 9 2 2 1 2 1
Percentage 53 25 5.6 5.6 2.8 5.6 2.8
participants wanted no truly guilty people to go free. It may also indicate

that they do not understand the trade-off between Wrongful Acquittal and
Wrongful Conviction. This may be because our participants probably had no
experience with judicial processes. It would be very interesting to replicate
this study with both judges and lawyers since they have directly experienced
the judicial process. Of interest would be the variation in elicited prior beliefs
about the prevalence of guilt within lawyers and judges as well as their elicited
BR. We would expect that their elicited BR would be closer to spirit of the BR
and have a high proportion of ratios greater than or larger than 10:1.
One difficulty with the design of this study is that participants did read a
case and this may have contaminated the eliciting of prior beliefs about the
prevalence of guilt and the Blackstone Ratio. In a third study, we elicited prior
beliefs about the prevalence of guilt and the Blackstone ratio from partici-
pants without presenting any case material (or vignette). SDT suggests that
the value of the consequences of Wrongful Acquittal compared to those for
Wrongful Conviction influences the threshold for beyond reasonable doubt.
That is, we wanted to know how potential jurors responded to the explicit
tradeoff between Wrongful Acquittal and Wrongful Conviction and answer
the question of how they value the two types of error.
Thirty-four students from an intermediate-level forensic psychology
course in 2006 at Kansas State University participated. We elicited prior
beliefs using the method from Study 2, Figure 11.4. The fourfold table for
jury decision making was presented and explained in the context of irredu-
cible uncertainty (Hammond, 1996). They were asked to write down on a
sheet of paper in two columns the consequences of a Wrongful Conviction
(defendant was convicted but was really innocent) and a Wrongful Acquittal
(defendant was acquitted but was really guilty). They were asked the question:
Given that you cant avoid the possibility of an error, which error do you
want to avoid the most? This taps the relative utility of the two types of error.
We then elicited their Blackstone Ratio as in Study 2.
Prior beliefs of guilt were distributed very similarly to those in Study 2,
with a mean prior belief of .69, standard deviation of .15 and a range from .34
to .94. Seventy-nine percent (27/34) wanted to avoid Wrongful Convictions;
but this means that 21% wanted to avoid Wrongful Acquittals, contrary to
Blackstone. The results from the elicited Blackstone Ratio are similar to those

Conviction from Study 3
0:1 1:1 27:1 10:1 10:1
Frequency 10 4 10 4 4
Percentage 31 12.5 31 12.5 12.5
from Study 2 with a sizable percentage not wanting to let any guilty people to
go free, see Table 11.6. As in the other studies, there is wide variability.
The key message from a Blackstone Ratio of 10:1 is that the interests of
the defendant are paramount. However, Volokh (1997, p. 211) cites the story
of a Chinese law professor listening to a British lawyer state it is better that 99
guilty men go free than that one innocent man be executed. The Chinese
professor asked Better for whom? This may help with an explanation for
the number of participants who wanted no guilty men to go free and the
participants who wanted to avoid a Wrongful Acquittal. It may be that some
participants have the defendants interests as paramount but others may focus
on the interests of society. Study 4 explores this explanation.
Consider the situation of a judge deciding to grant or deny a motion. For
a particular type of motion, perhaps a frequently encountered type of motion,
they will have beliefs about the prevalence with which such a type of motion
should be granted. They will also have beliefs similar to those expressed in the
BR. For example, how many motions should be wrongfully granted to avoid
one wrongfully denied motion? Further, apart from the two adversaries
before the judge, there are other stakeholders in the decision to grant or
deny a motion, for example, the judge, society, and the law.
There are many stakeholders in the criminal court, for example, the
defendant, the lawyers, the judge, the jury members, the victims and their
families, and society, to name some of them. The decisions made in court will
have different consequences for different stakeholders. We think that when
participants focus on the defendant as stakeholder, more will want to avoid
Wrongful Convictions and that when focusing on society as stakeholder
more will want to avoid Wrongful Acquittals. This could explain some of
the variability in our studies reported so far. Further, using the idea that
people with a conservative approach to justice would focus on the society as
the main stakeholders and that those with a liberal approach would focus on
the defendant as the main stakeholder, (Tetlock & Mitchell, 1993), we suggest
that those who want to avoid Wrongful Convictions would identify them-
selves as more liberal.
Thirty-four students from an intermediate-level forensic psychology
course in 2007 at Kansas State University participated. Prior beliefs were
elicited as in Studies 2 and 3. The fourfold table was explained as in Study 3,
and participants were asked to think about the defendant as stakeholder and
write out the consequences for the two types of error. They were then asked to
think about society as the stakeholder and wrote out the consequences for the
two types of error, Wrongful Acquittal and Wrongful Conviction. Following
this, we asked them which error they wanted to avoid the most with the
defendant as stakeholder and then with society as stakeholder. They were also
asked which stakeholder they thought was more important. We asked them to
place their political views on a scale from 1 to 100 labeled liberal at the low
end and conservative at the high end of the scale. We then elicited their
Blackstone ratio as in Studies 2 and 3.
The results for prior beliefs for guilt were very similar to those from
Studies 2 and 3. The mean was .623 with a standard deviation of .18 and a
range from .13 to .91. This range does not seem to be a function of being an
undergraduate student because it matched the range of prior beliefs elicited
from seminar attendees at seminars given by Dalgleish for faculty and grad-
uate students at some universities in the United States and the UK.
Table 11.7 presents the frequency with which participants wanted to
avoid the two types of error by stakeholder. When considering the defendant,
76% wanted to avoid a Wrongful Conviction, whereas 26% wanted this when
considering society. These two proportions are different using a McNemars
test, w(1) = 73.5, p < 0.001. Thus, the consideration of different stakeholders
matters. The most important stakeholder was society for 21/34 (64%) of parti-
cipants. The majority, 18/34 (53%), wanted to avoid Wrongful Convictions
when considering the defendant and avoid Wrongful Acquittals when consid-
ering society.
What is the relationship between self-reported political views and the
value placed on the consequences for different stakeholders? There were some
weak relationships but all were in the expected directions. When thinking of
the defendant as stakeholder those that wanted to avoid WCs had a more
liberal political view (N = 8, M = 52.2) than those who wanted to avoid WAs
Table 11.7 Frequency of Desire to Avoid the Two Types of Error Depending
on Stakeholder
Society as Stakeholder
Wrongful Wrongful
Convictions Acquittals Total
Defendant Wrongful 8 18 26
as stakeholder Convictions
76
Wrongful 1 7 8
Acquittals
Total 9 25 34
26
(N = 26, M = 62.3), t(30) = 1.42, one-tail p = .08. This test has low statistical
power because the sample sizes are small and the measure of political views
was a simple one-item scale. It would be interesting to see if this effect
is stronger if better measures of political views were used, such as those
developed by Evans, Heath, and Lalljee (1996) and Mehrabian (1996).
Those participants who said the defendant was more important had more
liberal views, (M = 50.2 versus M = 55.2, t(32) = 1.07, one-tail p = .15) but,
while in the expected direction, this difference was not statistically significant.
The pattern for Blackstones Ratio is similar to those for Studies 2 and 3.
However, it is useful to tabulate the elicited BRs by which stakeholder was
considered more important. Table 11.8 does this. There were 5 participants
who stated their BR as not wanting any guilty defendants being free. We
believe this category is separate from BRs that state that the participant wants
at least one guilty defendant to go free. The participants with a BR of 0:1 may
not have understood the question eliciting the BR or they have a particular
interpretation of WA or WC or they really wanted no guilty defendants free.
We calculated the mean political view for these 5 participants (M = 72.3,
SD = 21.8) and tested it against the mean for those 25 who said they wanted
at least one guilty defendant free, (M= 48.24, SD = 23.26). This difference is
statistically significant, t(29) = 2.36, p = .025) showing that those who had a
0:1 BR rated themselves as more politically conservative. Combined over our
2nd, 3rd and 4th studies, 35 out of the 100 participants had an elicited BR of
0:1. The implication is that these people would be more conservative in their
political views. Another implication is that in future research, the BR should
be elicited using a different question that allows participants to state the ratio
of WC to WA. For some of the participants who stated a 0:1 BR, their BR
elicited using a different question may have been 1:10 or some ratio opposite
to the direction of the ratio stated by Blackstone.
Excluding the 0:1 group, there were no differences in political view for
those participants whose BR were <10:1 and for those with BRs 10:1.
However, for those 10 participants with BRs 10:1 seven of them (70%)
stated that the defendant was more important while for those that favored

Conviction from Study 4 by the Most Important Stakeholder
Most important stakeholder 0:1 1:1 27:1 10:1 >10:1
Defendant 2 1 2 5 2
Society 3 2 9 4 1
Equal 1 0 0 0 0
Total (N = 32) 6 3 11 9 3
Percentage 18.8 9.4 34.4 28.1 9.4
society 5/16 (31.3%) had BRs 10:1. These proportions were close to being
significantly different, w2(1) = 3.72, p = .054.
So when participants consider either the defendant or society, this influ-
ences their threshold since the relative costs of WC and WA change. However,
which stakeholder is more important is not related to prior beliefs but is
related to their Blackstone Ratio. From these results we could infer that the BR
is more about values of consequences than about frequency (prior beliefs)
(cf., Dekay, 1996).
From the GADM model there is the prediction that if two members of a
jury have different thresholds, for example, one considers the defendant as a
more important stakeholder and the other favors society, and they assess the
strength of evidence as being the same in the case, then they will disagree
about the verdict. Going back to the case evidence will not resolve this
disagreement. There is a further implication of the GADM model from
these findings. Consider a person from the group who wanted to avoid WC
when considering the defendant and wanted to avoid WA when considering
the society as stakeholder (the majority of participants). Let us assume one of
them has a prior belief of guilt of 0.50 (neutral) and let them consider a case
that does not have clear-cut evidence of guilt. Figure 11.5 illustrates this
example with two possible thresholds for this juror. When considering the
defendant the threshold is Tdef and the assessment of the evidence is not
sufficient to convict. When considering the society the threshold is Tsociety
and the evidence is sufficient to convict. This juror would be in a state of
internal decisional conflict and would vacillate while trying to reach a
verdict. This may be seen as indecision but searching the evidence is not
going to resolve this internal conflict. The person needs to go with the
stakeholder they evaluate as the most important and use the threshold for
that stakeholder.
What are the implications of these findings for judges ruling on motions
or other judicial decisions? The link between judicial decisions and political
orientation, given the tentative evidence from Study 4, seems to lie not with
High
Threshold TDef
Acquit
J
Assessment.
Convict
Threshold TSociety
Low
Assessed level of Strength of
Evidence for Guilt
Figure 11.5 Internal decisional conflict within a juror.
the assessment of the evidence but with the relative importance of different
stakeholders and the relative values of Wrongful Convictions and Wrongful
Acquittals to the judge. The relative influence of the assessment of the
evidence and prior beliefs about prevalence, the values of the two types of
errors and the importance of the different stakeholders could be tested using a
vignette study similar to our Study 1 together with the elicitation of the other
factors developed for Studies 2, 3 and 4. More robust measures of political
orientation would need to be used.
Conclusion
The data from the four studies indicates a tremendous amount of variation in
(1) measured thresholds, and (2) measures of the influences on threshold.
These were the prior beliefs about guilt and the error they wanted to avoid the
most. Further there was considerable variation in the elicited Blackstone
Ratios. Some of this variation was related to which stakeholder they thought
most important or were considering and their self-reported political views.
From the GADM model this variation in threshold for guilt, implies that,
within jury trials, there will be considerable differences in verdicts reached
that are not due to differences in the interpretation and assessment of the
evidence.
While the data presented in this chapter has been with undergraduate
students in simulated jury decision making, the aim has been to show that
the methods could be generalized to judicial decision making. Whether the
results of our studies generalize to other judicial decision making tasks is an
empirical question. Throughout this chapter we have drawn parallels between
the judicial task of deciding to deny or grant a motion and the jury task of
deciding whether to convict or acquit a defendant. Further, we have suggested
that our convict/acquit decision tasks could be used with judges and lawyers.
The aim of these proposed studies would be to investigate the amount and
variability in their thresholds for beyond reasonable doubt and in the theore-
tical influences on thresholds among judges and among lawyers and between
judges and lawyers.
Similarly, the aim of the proposed studies on the judicial decision to deny
or grant motions would be to investigate and document variation in the
thresholds for action. If, for example, both trial judges and appellate judges
did the same tasks, we could compare thresholds between pairs of judges.
What are the implications of this? Let us consider a trial judge who has a
tendency to deny motions of a particular type. Let us assume that some of the
decisions are contentious enough to provide grounds for appeal. Now con-
sider an appellate judge who reads the original submissions and decides
against the original decision. Within the GADM model approach to the
assessment of the original submissions and the decision to take action, the
trial judge and the appellate judge may differ in their assessment or in their
thresholds or both. It is important to isolate whether such differences in
decisions are due to differences in the interpretation of the law or to differ-
ences in the values attached to the consequence of the decisions or to the
judges prior beliefs about the need to take action. In addition to the studies
reported in this chapter, there seems anecdotal evidence, for example on
websites such as Decision of the Day4 that such differences of opinion
exist. Investigating whether there is variation in the influences on the
threshold is important because making such influences explicit to people
enables discussion about them and provides ways of devising training to
reduce the effects of decisional conflict. One of our studies found links
between the relative importance of different stakeholders and political orien-
tation and if such effects were found among judges and lawyers, it may inform
and extend debate on the role of political orientation and judicial decision
making. The GADM model and its associated methodologies provide a way to
empirically study such issues.
Notes
Preparation of this chapter was supported by funds provided by the Scottish
Funding Council to HealthQWest; A research consortium for the West of Scotland.
1. U.S. v. Terry, 392 U.S. 1 (1968).
2. Similarly, in applying the GADM model to verdicts, it is assumed that an
assessment of the strength of evidence is made and a decision to convict is
made if the assessment exceeds their threshold for action. This threshold
depends on the level of proof needed to convict; for example, beyond reason-
able doubt would be higher than the preponderance of the evidence.
3. Anna Rickard (1998) completed the study as an honors project at the University
of Queensland.
4. http://www.enotes.com/blogs/decision-blog/
12
Every Jury Trial Is a Bench Trial: Judicial
Engineering of Jury Disputes
C. K. Rowland, Tina Traficanti, and Erin Vernon
For decades virtually every scholarly work on civil trial courts and trial judges
began by lamenting the unfortunate tendency of students of judicial behavior
to concentrate almost exclusively on appellate courts generally and on the
U.S. Supreme Court in particular. Fortunately, although this imbalance is still
present to a degree, a nascent body of research has ameliorated the disparity
and enhanced greatly our understanding of civil trial courts, trial judges, and
the psychology of trial judging over the last decade.
This growth in attention to trial courts has, of course, included tradi-
tional legal analysis and over 25 years of sophisticated, multidisciplinary
attention to all levels of criminal courts (Jacob, 1983; Gibson, 1980); how-
ever, studies of civil trial judges behavior have, generally speaking, either
pursued a political track, which studied trial judges as political actors by
reference to appellate judicial behavior (Baum, 2007), or a juror track
marked by explicit or implicit comparisons with juror behavior
(Robbennolt, 2005).
Political studies, especially judicial comparisons have, not surprisingly,
been primarily (but not exclusively) conducted by political scientists, and
have borrowed heavily from the study of appellate judicial behavior and the
influence of extralegal factors on that behavior (Rowland & Carp, 1996;
Buchman, 2007). Relying almost exclusively on federal trial judges published
opinions (but see Rowland & Carp, chap. 4 re unpublished opinions), judicial
scholars have found that for politically salient issues, such as abortion-related
disputes, federal trial judges published decisions can be predicted to an
impressive degree by extralegal factors, such as the appointing president
183
(Carp & Rowland, 1983; Buchman, 2007). Moreover, these appointment

effects tend to be most pronounced for ideologically relevant categories of
cases that were important to the appointing president and served as litmus
test appointment criteria (Carp & Rowland, 1983). Although still subject to
appropriate criticisms regarding theories and methods (see below), the con-
sistency of these findings is impressive, suggesting a value-based exercise of
judicial discretion that may reflect intentional or unintentional bias in trial
judging when the dispute involves ideological issues important to the judge
and the appointing president.
Juror comparisons, most of which have been conducted by attorneys,
law-school professors, and psychologists working in the law and psychology
realm, have relied on a much wider variety of methodologies, ranging from
archival analysis to surveys, to experimental designs to compare explicitly or
implicitly the decisions of jurors and judges faced with similar decision
stimuli and criteria (Blanck, 1991; Eisenberg et al., 2005; Robbennolt, 2005;
Kalven & Zeisel, 1966). In the civil arena, this approach is exemplified by
comparisons of judges and jurors willingness to award punitive damages
(Eisenberg et al., 2002), ignore potentially biasing information (Landsman
& Rakos, 1994), and the difficulty of ignoring inadmissible evidence
(Wistrich et al., 2005). Although the results of these comparisons differ in
their details, the key, consistent finding across multiple issues is that judicial
decisions and juror decisions made under similar constraints are remark-
ably similar. Particularly interesting in this context is research that
implicitly compares jurors and judges by focusing on judges ability to
understand complex evidence and/or apply legal standards that rely on
scientific or statistical concepts. These studies are remarkably consistent in
their finding that when faced with science-based standards (Gatowski et al.,
2001), statistical evidence (Guthrie, et al., 2001), and other risk-assessment
tasks, judges are susceptible to many of the same limits on objectivity and
accuracyfor example, hindsight bias and framing effectsthat plague
jurors and all human decision makers faced with difficult judgments
(Dhami, 2005).
Call for Synthesis and a Refocused Research Program

It is interesting and instructive to compare the political and juror tracks
summarized above. Most striking is an obvious anomalythe trial judges
studied by political scientists turn out to resemble closely appellate judges and
other political actors, but the trial judges studied by comparison with jurors
turn out to behave much like jurors. This apparent anomaly is in large part
the product of different research targets, methods, and theoretical perspec-
tives (Baum, 2006; Braman, 2006b). It serves, however, as a reminder that the
parallel research paths summarized above have left a large lacuna in our
Every Jury Trial Is a Bench Trial 185
understanding of trial judges and trial-judge behavior, in part because one

approach rarely acknowledges the contributions of the other (for a recent
exception, see Buchman, 2007).
The political approach has done a good job of identifying variation
among trial judges (at least federal trial judges) in natural settings and the
extralegal correlates of that variation, but this approach has largely limited
itself to atypical, politically relevant published decisions and has paid precious
little attention to myriad unpublished decisions made by thousands of trial
judges in thousands of trial courtsstate and federalevery day (see Rowland
& Carp, 1996, for an exception). Equally limiting has been the tendency of
political studies to simply ignore or assume away the psychology under-
lying the pattern of extralegal influences identified by their research. Juror
comparisons, while they have done a much better job of addressing a wide
variety of decisions that define everyday trial judging and are much more
sophisticated regarding psychological theories of human judgment, have
paid precious little attention to variation among judges in these decisions,
especially to the legal and extralegal correlates of these decisions in natural
settings.
The anomalous state of current research suggests that the time has come
for synthesisa research program that will, like the juror path, focus on
decisions that define what most trial judges do most of the time and, like the
judicial path, examine variation and the reasons for variation in these
decisions in natural settings.
Trial Judges and Jury Trials

We believe that the most fertile ground for this synthesis is a research
program aimed at understanding the behavior of trial judges in the context
of jury disputesthat is, civil disputes in which one or both litigants exercise
their constitutional right to a jury trial. In the remainder of this essay, we
justify our call for a research program focused on jury disputes in natural
settings and describe our expectations regarding what we will learn about
judging and trial-judge behavior from the research.
Justification for Refocus on Jury Disputes
As suggested by the title of this essay, every civil trial is in important ways a
bench trial, with evidence and, to some degree, outcomes engineered by the
trial judge. Yet, with notable exceptions discussed below, most recent research has
devoted little attention, especially in natural settings, to judicial participation
in the resolution of civil disputes in which litigants demand a jury trial.
Indeed, unlike students of the criminal justice process, who have for years
focused on judges as participants in a courtroom workgroup whose members
share an interest in efficient resolution of criminal disputes (Jacob, 1983),

most students of civil trial judges and civil trial judging have explicitly or
implicitly drawn a bright line between bench trials and jury disputes,
regardless of whether the dispute is ultimately resolved by a jury or resolved
prior to a jury trial.
We believe that the current vacuum is ironic and unfortunate for several
reasons. First, in thousands of disputes in thousands of trial courts every day,
what the constitutional right to a jury trial actually guarantees citizens is the
right to a trial in which a jury decides a dispute largely defined and engineered
by the exercise of judicial discretion. These discretionary exercises of judicial
judgment answer key questionsfor example, what evidence the jury will
hear and who among their fellow citizens will constitute their juryand
almost always favor one side or the other and greatly influence the likely
resolution of the dispute. Because they favor one side or the other, judges
jury-dispute rulings define the likelihood of trial outcomes and are instru-
mental in determining whether a case will settle and, if so, which party gains
the most favorable terms of settlement. Thus, the architecture of the right to a
jury is in the constitution, but the engineering is left largely to the discretion
of trial judges.
Second, resolving disputes in the context of an anticipated or actual
jury trial is what most trial judges in state and federal district courts do
much of the time, whether that resolution is via dismissal, settlement, or a
jury verdict. This means that in the aggregate these decisions have a
potential social importance that parallels their importance to individual
litigants. Moreover, in the context of modern mass torts, where one trial
judge may manage thousands of similar cases, judicial decisions in a
crucial bellwether case engineer the legal and evidentiary context for
thousands of potential jury trials. From a research perspective, the ubi-
quity of judicial engineering focuses inquiry in an unprecedented variety of
legal and extralegal contexts, all of which serve as potentially important
institutional constraints on behavior. Exemplar contextual variation includes
differences in standards for cause strikes or admissibility of expert testimony,
differences in how judges are selected, and differences in the extent to which
the jurisdiction has been a target of tort reform and countertort reform
campaigns.
Third, the study of judicial behavior in the context of a jury trial is
important because most engineering judgments require judges to exercise a
great deal of discretion, which maximizes the potential influence of sub-
jective factors on those judgments and creates a unique opportunity to
examine the interplay of legal and extralegal influences on trial-judge
behavior. Moreover, the potential for extralegal factors to influence the
exercise of engineering discretion is exacerbated by the fact that many
engineering decisionsfor example, rulings on motions for cause
strikesare reviewable only if there is evidence of judicial abuse of
discretion.
The scope of this discretion and the plethora of opportunities to

exercise discretion in the course of a jury dispute make judicial engineering
of jury disputes an ideal avenue for more insight into trial judges behavior.
Key individual variation includes, for example, whether the judges pre-
bench practice was primarily on behalf of civil plaintiffs, defendants, or
neither, and the extent to which the judges selection (whether via appoint-
ment or election) was supported by the plaintiffs bar or by corporate
defendants, or neither. (It is no exaggeration to say that, given the highly
politicized nature of judicial selection in the federal and in many state
systems, many federal and state trial judges would be struck for cause and
most others would be subjected to a peremptory strike by one side or the other
in many jury disputes.)
Finally, the evidence that trial judges are susceptible to biased assimila-
tion (Lord et al., 1979, p. 119), hindsight bias (Guthrie, Rachlinski, &
Wistrich, 2001), anchoring (Chapman & Bornstein, 1996) and other cogni-
tive illusions (Robbennolt, 2005), suggests strongly that when faced with
complex decisions and ambiguous evidence, trial judges will make decisions
(whether consciously or not) that reflect the biases and motivations asso-
ciated with their prebench experiences. A focus on trial judges behavior in
thousands of natural settings will not only let us test empirically the extent to
which findings associated with atypical published decisions are replicated for
more typical exercises of discretion but will also help us learn more about the
utility of social science for explaining the exercise of everyday judicial
discretion.
Specific Ways in Which Every Civil Jury Trial Is a Bench Trial
Experienced litigators accept as a truism that the exercise of judicial discretion

can essentially obviate the distinction between jury trials and bench trials at
every stage of the jury dispute, ranging from pretrial gate keeping, to rulings
during trial, to posttrial modification of a jury verdict. However, it may be
appropriate to specify for a social science audience several specific typical
engineering decisions that createindeed, requirethe exercise of essentially
unbridled judicial discretion. Jury disputes follow a predictable sequence
requiring numerous judicial rulings on key issues ranging from jurisdiction
and evidence to jury selection and posttrial modification of jury decisions. As
reflected in the examples below, most of the opportunities for judicial engi-
neering are issues hotly contested by the parties, and rulings tend to favor one
side or the other. Key opportunities for judicial engineering and the exercise
of judicial discretion include, but are not limited to, the following:
Pretrial
1. Jurisdiction. The trial judges initial decisions regarding jurisdiction issues

can exert a tremendous influence on likely outcomes, especially when the
choice to be exercised is between a federal and a state jurisdiction. Federal and

state rules tend to be very different regarding key elements of a jury trialfor
example, the conduct of voir direand federal jurisdictions tend to be larger
and include more diversity of attitudes and demographics among jurors than
do many state jurisdictions.
2. Discovery. The trial judge in a jury dispute can begin to shape the scope of
the trial and the evidence jurors will hear from the outset of the discovery
process by ruling, for example, on the number of depositions to be allowed;
time limits for depositions and for the completion of discovery overall;
whether to require further answers to written discovery or deposition
responses; and whether access to medical records of the claimant is limited
in any way.
3. Trial Date. Something seemingly as innocuous as the setting of the trial date
can also benefit one side over the other, given, for example, issues with
witnesses and party availability or the greater need of one party for more
extensive discovery.
4. Pretrial Publicity and Change of Venue. Although disputes over pretrial
publicity and motions for a change of venue are commonly associated with
criminal disputes, these issues can be important in civil disputes as well. For
example, local publicity in Houston regarding the Enron scandal generally
and the treatment by Enron of its employees in particular raised the same
publicity and venue issues in the civil claims brought by former Enron
employeesmost of whom had lost their life savingsas were raised by the
defendants in the Enron criminal case.
5. Motions to Dismiss or for Summary Judgment. Perhaps the most obvious
judicial intervention, at first blush, is the courts rulings on dispositive
motions, such as Motions to Dismiss or Summary Judgment Motions. An
extremely important component of this decision set is the decision to
include or not include claims for punitive damages. Because the presence
of punitive damages increases dramatically the defendants potential expo-
sure, this decision literally defines the stakes of the case. Moreover, it
usually includes significant judicial discretion based on ambiguous stan-
dards, such as malice or willfulness and a projection of likely impact of
evidence on a jury. Dismissal of some causes of action or partial grants of
summary judgment also narrow the scope of the dispute in ways that
typically favor one side or the other.
6. Motions in Limine. Motions in limine serve, by their nature, to narrow the
scope of evidence presented at trial and may exclude key testimony or
arguments, thus limiting what the jurors will hear and how they assess
other evidence and witnesses. Many of these decisions are made on the eve
(or morning) of trial and have a tremendous impact on the context and likely
outcome of the trial itself. Not surprisingly, many cases settle shortly after
these rulings are made.
7. Motions regarding admissibility of expert witnesses and testimony. There are a
multitude of preliminary decisions, leading up to the courts ultimate ruling, that
can determine how the court will handle its gate-keeping function as it relates to
expert testimony, all of which may dramatically alter the course of the trial. The
judge has within his or her power to decide whether to hold a hearing at all,
whether the hearing is evidentiary in nature and should include witness testimony,
and which factors to employ in determining reliability of testimony. Because the
motions are usually filed by defendants seeking to exclude plaintiffs expert
testimony, the rulings often define at an early stage which party has the higher
evidentiary ground. Given the influence of these admissibility rulings for both
parties, it is no accident that they are usually hotly contested, with extensive
briefings and adversarial argumentation.
During Trial
1. Jury Selection. There can be little dispute that the demographics, back-
ground, education, experience and attitudinal bent of each individual juror is
a critical component of how a jury dispute will eventually be resolved. The
judge exercises great power and influence over the jury selection process, the
ultimate makeup of the jury, and therefore, the verdict they will likely render
by deciding: whether or not to strike jurors for cause; the number of per-
emptory challenges allowed to each side; the reasons accepted or rejected for
hardship excuses; whether multiple plaintiffs or defendants are treated singly
or as a group for peremptory challenges; how many strikes each side gets
when the dispute involves multiple plaintiffs and/or defendants; whether a
supplemental juror questionnaire is allowed, and, if so, what questions and
what response options are permitted; whether attorneys can conduct their
own voir dire or not; how much discretion is given to attorneys in conducting
voir dire regarding scope of questions and latitude to introduce case themes
and facts. These jury-selection judgments, virtually untouchable at the appel-
late level, define the right to a jury trial in practice perhaps more than any
other exercise of judicial discretion.
2. Evidentiary Rulings. Evidentiary rulings do not stop once the trial begins.
Interpretation of the rules of evidence and the amount of leeway given on issues
like hearsay, laying proper evidentiary foundations, relevance, prejudicial
effect, and to what extent, if at all, an expert witness can testify beyond the
material contained in the four-corners of his/her report are judicial opportu-
nities to impinge on the province of the jurors to decide a dispute, by limiting
or expanding what it is those jurors will hear as evidence in the case.
3. Rulings on the Use of Demonstratives and Graphics. Another arena for
judicial engineering is in the area of visual evidence: whether graphics or
demonstratives, including highly sophisticated animations, proffered by either

side, depict accurately what it is they are purporting to demonstrate; when the
attorney is permitted to introduce the demonstratives (opening statement?);
limitations on which witnesses can use them to illustrate their testimony, and so
forth, can color jurors views of key evidence in the case. These rulings are
especially important when, as in the case of many torts, one or both sides submit
animations that essentially reify disputed eventsfor example, automobile
accidentsin ways that significantly benefit one side or the other.
4. Rulings on Courtroom Etiquette. The form and content of objections,
whether speaking objections are permitted, where attorneys must stand in the
courtroom, whether they must use a podium to address the jury, and to what
extent they can invade the personal space of the witness on the stand would,
by all appearances, be neutral, and yet such procedural formalities may very
well play to the strengths or weaknesses of the individual attorneys presenting
the case for each side.
5. Cues from the Bench. There are myriad ways in which the trial judge,
either consciously or unconsciously, can reveal his or her views or opinions of
either sides witnesses, evidence, and arguments during the course of the trial.
The manner of ruling on objections with each sides attorneystone of voice,
facial expressions, sarcasm, and whether that manner is visited equally on
each litigants counsel; whether and to what extent the judge admonishes
participants in the trial in front of the jurywitnesses and attorneys alike;
whether and to what extent the judge visibly reacts to evidence presented
verbally or through facial expressions and/or other physical cues (body
language) have the ability to influence juror reaction to those same witnesses,
evidence or arguments. Moreover, most of the nonverbal cues are not
recorded in the transcript and are, therefore, immune from appellate review
(see Blanck et al., 1985).
6. Jury Instructions and Verdict Form. The language and content, emphasis,
tone, and inflection of the trial judge when reading the jury instruction, the
discretionary instructions that are utilized, and the Verdict Form itself
general versus special questions, the order of the questionsare all ruled on
by the trial court and are critical to shaping the way in which jurors deliberate
about the case and the verdict that they reach. At the risk of stating the
obvious, even subtle differences in the questions the jury is asked to answer
can go a long way toward determining the probable answer, again in ways that
favor one side or the other.
7. Juror Deliberations. Once the jurors begin to deliberate, the trial court
judges influence does not end. The judge controls how long jurors deliberate,
the schedule on which they are required to be in court, and how to respond to
juror questions. Perhaps most importantly, the judge determines whether and
when an Allen Charge, or other less incendiary alternatives suggested by the
American Bar Association, is warranted and can alter the language of the
charge in subtle but important ways that determine the likelihood of a verdict
or a mistrial.
Posttrial
Through posttrial motionsJNOV (Judgment Notwithstanding the
Verdict), Motions for a New Trial, and Motions to Modify Damagesthe
trial court can nullify the jurys conclusion or alter it substantiallyfor
example, by significantly reducing damages. At the extreme, such decisions
have the potential to nullify not only the jury decision, but also to nullify the
litigants right to a jury trial.
It is obvious from this partial listing of discretion points that most of the
decisions made in this context offer the trial judge a great deal of discretion to
engineer process, content, and outcome probabilities with limited appellate
guidance or threat of appellate reversal, a condition that maximizes the
potential for variation in judges behavior and the extralegal correlates of
this behavior and demonstrates the importance of research that helps us
understand judicial behavior in this context. We next propose a theoretical
framework that we believe will most effectively guide inquiry into judicial
engineering of jury trials.
Theoretical Framework
In this section we outline a proposed theoretical framework for a research
program that focuses on the behavior of trial judges in jury disputes and the
extralegal influences on that behavior. At first blush, one would expect our
proposal to focus research on judicial behavior in natural settings to rely on
theories of judicial behavior associated with political inquiry into extralegal
influences on federal judges in natural settings. Unfortunately, with a few
notable exceptions (Baum, 2006a; Rowland & Carp, 1996; Braman, 2006b),
most of these studies have been either atheoretical or have relied rather
uncritically on adaptation of preference-based theories and rational-choice
models of strategic decision making (Epstein & Knight, 2000; Baum, 2006)
developed in the context of appellate judicial behavior (Carp & Rowland,
1983; Buchman, 2007). Although political scholars have engaged in a healthy
debate regarding the extent to which judges voting their preferences may
be bounded by institutional constraints (Baum, 2006; Epstein & Knight,
1998), their common assumption is that, to the extent possible under institu-
tional constraints, judges are motivated by their policy preferences and their
votes are strategic reflections of these preferences. Indeed, two of the leading
scholars in the field have speculated that when judges explain their decisions
by reference to legal reasoning, they may be posturing for the benefit of

external audiences (Epstein & Knight, 1998).
It is, of course, one thing to say that judicial policy makers on the
Supreme Court, who are expected to make policy for lower courts, are
motivated by their policy preferences. It is quite another thing to assume
that a state trial judges jury-dispute rulings, many of which are reached in a
context that affords the judge a great deal of discretion, are nothing more than
the judges expression of his/her preferred outcomes and personal biases. To
do so in the context of fact-finding is quite literally to accuse a judge who has
taken an oath against bias of violating his/her oath of office (Braman, 2006a;
Rowland & Carp, 1996). Therefore, setting aside debate regarding the utility
of preference-based models for the study of appellate courts and appellate
judicial votes, we believe that attempts to extend this paradigm to the study
of extralegal influences on trial judges decisions in jury disputes are funda-
mentally flawed and correctable only by abandoning the paradigm in favor of
social psychological approaches that take seriously and respect the common
insistence by trial judges that they are not biased in favor of their personal
preferences (Edwards, 1998).
It is therefore incumbent on us to offer an alternative that takes seriously
the judicial oath and the underlying psychology of human judgments but also
accommodates the empirically based expectation that the exercise of judicial
discretion in jury disputes will vary significantly among judges and that this
variation will be correlated significantly with extralegal factors, many of
which will be unacknowledged by the judge or the record.
Motivated Reasoning and Judicial Engineering
Rowland and Carp (1996) took a first step in this direction a decade ago. We
started with the premise that any useful theory of trial judging had to
reconcile the anomaly outlined abovethat is, how trial judges decisions,
at least in their published opinions, could be predicted with remarkable
accuracy by extralegal, ideology-driven variables when these same judges,
including judges who acknowledged and were sensitive to their own policy
predilections, insisted that their decisions were based on unbiased legal
reasoning subject to the constraints imposed by the law and the facts before
them (Baum, 1997). To accommodate this apparent anomaly, we turned to
social psychology and adapted concepts from social cognition and social
judgment theory to create an eclectic, rather primitive, general model of
trial judging based on the assumption that trial judges, like jurors, are
cognitive misers who, when faced with ambiguous decision criteria that
exceed their cognitive capacity to resolve episodically, must process the
evidence via existing schemata and rely on judgment heuristicsfor example,
the availability heuristic. Motives, according to this formulation, were essen-
tially irrelevant in the context of judgment tasks whose complexity and
ambiguity forced trial judges, like all human judges, to rely on cognitive
shortcuts to make sense of otherwise unmanageable decision cues. The

apparent anomaly, therefore, was theorized to be the product of what
Robbennolt (2005) has appropriately labeled cognitive illusionsthat is,
unrecognized psychological biases such as anchoring or hindsight bias
(Guthrie et al., 2001).
More recently, Eileen Braman (2006a; Braman & Nelson, 2007) has
improved significantly on our initial effort by adapting a more refined
social psychological model of motivated reasoning to judicial settings.
Bramans work does not examine trial judging per se, but her adaptation
of motivated reasoning; unlike the assumed preference-based motives in
previous political studies, recognizes implicitly the complexity of the inter-
face between motives and the kinds of engineering judgments required of
trial judges in jury disputes. Moreover, a basic extension of motivated
reasoning to judicial engineering of jury disputes is straightforward. As
articulated succinctly by Kunda (1990) and Lodge and Taber, 2000) theories
of Motivated Reasoning assume that human decision makers faced with
complex, ambiguous judgment criteria are motivated by two sets of goals,
both of which are potentially present when judges exercise judgment in the
context of a jury dispute. On the one hand, human judges are motivated by
accuracy goalsto arrive at an accurate conclusion, whatever it may be
(Kunda, 1990). This goal set is, of course, consistent with the fact that trial
judges share legal training, which includes training in legal reasoning and
socialization to legal norms. Thus, when they participate in jury disputes,
trial judges should be motivated in part by the shared legal constraints and
norms that define their profession and are expressed in their oath of office
that is, they are committed to unbiased decisions that accurately reflect the
evidence and law on which those decisions are based. A large body of
research has demonstrated empirically the importance of accuracy goals in
a variety of settings. Moreover, several characteristics of accuracy goals are
particularly important for application of motivated reasoning models to jury
disputes. For example:
Accuracy goals tend to be most prominent in situations where

subjects judgments are made in public and/or subjects are expected
to justify their judgments to others (Kunda, 1990). No matter how
much discretion a trial judge has, his/her most important jury-
dispute decisions are made and/or announced publicly and recorded
in the court record; moreover, regardless of the standard for review,
they must be justified to the litigants and their attorneys, and an
appellate court must be convinced that the judge did not abuse his/
her discretion.
Decision makers whose judgments are public are less susceptible
to biases and cognitive illusions, such as priming (Tetlock,
1985); however, there is no indication that accuracy goals
attenuate the effects of other, more fundamental illusions, such
as hindsight bias or inappropriate reliance on the availability

heuristic when decision criteria are complex and/or ambiguous
(Kunda, 1990; Lodge & Taber, 2000). Thus, we have every reason
to expect that trial judges will be motivated by accuracy goals, but
that, no matter how committed they are to accuracy goals, trial
judges in jury disputes will at the same time be susceptible to many
of the same cognitive illusions identified in extant research.
Moreover, these illusions are likely to be most apparent in the
resolution of complex issues and the evaluation of ambiguous
evidence.
The influence of cognitive illusion is accentuated when subjects are
required to make judgments under time pressures. As discussed
more fully below, trial judges are always under formal or informal
pressure to move their docket and dispose of disputes rather than
clogging the system and creating backlogs. Thus, we would expect
biases to potentially compromise accuracy goals and influence
judgment more in courts with greater time pressures than in courts
with less pressure.
Research associated with accuracy goals is certainly consistent with a view

of trial judges as motivated to judge the relevant evidence and law as accu-
rately as possible but limited in their ability to do so by circumstance and their
propensity to cognitive illusion. However, models of Motivated Reasoning
also assume that human decision makers are motivated by what Kunda calls
directional goalsthat is, the motive to arrive at a particular, directional
conclusion (p. 480). Kunda points out that such goals are seldom uncon-
strained; rather, directionally motivated judges attempt to be rational and to
construct a justification that would persuade a dispassionate observer.
(p. 483) They maintain an illusion of [their own] objectivity (Pyszczynski &
Greenberg, 1987) by searching their memory for beliefs and rules that could
support their desired conclusion. The illusion is maintained because subjects
engaged in motivated reasoning, do not realize that the process is biased by
their own goals, that they are accessing only a subset of their relevant knowl-
edge, that they would probably access different beliefs and rules in the
presence of different directional goals, and that they might even be capable
of justifying opposite conclusions on different occasions (p. 483).
Importantly for the research program proposed here, research has shown
that directional goals may bias some of the judgments trial judges are called
on to make in jury disputes. For example, directional motivation affects the
evaluation of scientific evidence because, in Kundas words,
people motivated to disbelieve evidence are less likely to believe it, and
there is some evidence that this outcome is mediated by differential proces-
sing of the information (Ginossar & Trope, 1987). This prediction is, of
course, consistent with Buchmans (2007) finding that appointees of
Republican presidents are more likely to rule in favor of defense motions to
strike plaintiffs scientific evidence and witnesses than are the appointees of
Democratic presidents.
In sum, models of motivated reasoning predict that when human judges,
with or without robes, are faced with difficult decisions and discretionary
decision criteria, the interface between accuracy and directional goals, and
cognitive illusion is complex and situational. Moreover, these goals and the
interface between them should be especially salient where, as in the case of
jury disputes, well-defined alternative accuracy and directional outcomes are
often argued in briefs or motions and verbally in open court, as in the court
should do (directional) because (accuracy). However, to maximize the
utility of these models for trial judging in jury disputes, we propose an
extension of motivated reasoning in trial judges to include a third potential
set of goals and an elevated attention to the audiences associated with trial
judges and trial judging.
Administrative Goals
As with many political actors, all trial judges are subjected to important
time constraints on their discretion as gatekeepers and judicial engineers,
and on their pursuit of accuracy goals. All are expected by their appel-
late and administrative principals to manage their docket, and to
achieve a case disposition rate that does not overburden their peers or
deny a speedy trial to litigants. Indeed, since many states have adopted
unified systems of judicial administration, federal and state judges case
disposition records are published and evaluated annually (Administrative
Office, annual). The most important and obvious incentive created by
administrative goals is the incentive to settle cases. Nothing wreaks more
havoc with a trial judges case backlog than a long jury trial, or series of
jury trials, each of which is recorded at the end of the trial as one case
resolution. Nothing alleviates this problem more than a settlement or
dismissal, both of which are typically also recorded as one case disposi-
tion. Therefore, if models of Motivated Reasoning are adapted to studies
of participation of trial judges in jury trials, we believe they should direct
specific attention to the relative influence of administrative as well as
accuracy and directional goals. Having said this, we recognize that admin-
istrative goals could in theory be accommodated by a judges directional
goal setthat is, rapid case disposition as desired direction. However, this
is qualitatively different from motivation to arrive at a particular sub-
stantive conclusion. Moreover, as with plea bargaining in the criminal
context, these administrative goals are, of course, shared by judges of
diverse policy predispositions and may dominate other directional goals
and accuracy goals, especially when judges are faced with external case-
disposition pressures. These administrative goals may conflict with sub-
stantive directional goals because the desire to move his/her docket may
conflict with a judges desired substantive outcome, and with accuracy
goals because the time necessary to maximize accuracy may cause

unwanted delays. Therefore, for purposes of the research program pro-
posed here, we believe that the assumption of administrative goals should
be given independent status and introduced into the mix with directional
and accuracy goals.
Audience
In our adaptation of motives to jury disputes, we expect that the trial judge
involved in a jury dispute will have multiple motives, and that the absolute
and relative salience of the three motive sets will vary significantly among
jurisdictions, among judges within jurisdictions, and among dispute categories.
Moreover, we believe that the relative saliency of each motive will be influenced
by the relative importance of multiple audiences to the judge. Larry Baum
has introduced this important concept of audiences to the general study of
judicial behavior (Baum, 1997, 2006).
Borrowing heavily from psychological models of self-presentation, Baum
assumes, axiomatically:
1. People want to be liked and respected by others who are
important to them;
2. The desire to be liked and respected affects peoples behavior;
3. In these respects, robes notwithstanding, judges are people.
For Baum, potentially important judicial audiences are numerous, ran-
ging from appellate justices, to the general public, to members of the
executive and legislative branches, to social and professional groups, to
the media. Thus, some audiences, which we will label task audiences,
may be important for professional validation, while others, which we will
call expressive audiences, may be important for elevating the individuals
self-esteem. For our purposes, audiences may be divided analytically into
accuracy, administrative, and directional audiences. However, the
distinction among these audience types is frequently blurred, and all three
types of audiences may be important influences on judicial engineering of
jury disputes.
Baums adaptation of a multiple audience perspective to judicial set-
tings is particularly helpful for inquiry into the behavior of trial judges in jury
disputes. Because judicial decisions are a special case of expressive choice
their decisions link judges to particular values and to others who share those
values. . . . By taking positions, judges align themselves implicitly with groups
whose positions are consistent with theirs (2006, p. 47). Moreover, trial
judges at jury trials arguably have more potential audiences than do judges
higher in the judicial hierarchy. Second, the importance of audience and
audience approval may be more important for trial judges than for their
appellate brethren. For example, the importance of audience is most apparent
in jurisdictions where judges are elected, where the electorate and sources of
campaign support are crucial audiences. Indeed, in the course of a jury trial,
representatives from multiple audiences, including sources of electoral
success or failure, are present in the courtroom and are eyewitnesses to
many of the crucial public decisions that make a jury trial a bench trial.
Finally, the context of jury disputes also enhances the potential importance
of audience relative to appellate settings. Whether at pretrial hearings or
during trial, the judge must make judgments and justify those judgments
face-to-face with audiencesranging from law-school classmates, to impor-
tant donors, to officers in the state bar associationswhose success or failure
may depend on those judgments.
In our adaptation of audience to judicial engineering of jury disputes,
we expect that the trial judge involved in a jury dispute will have multiple
audiences; that the quantity and qualitative importance of these audiences
will vary significantly among jurisdictions, among judges within jurisdic-
tions, and across disputes. We expect that this variation in audience will
help explain variation in jury-dispute judgments, including the reasons
given for these judgments, and we expect that variation in audiences will
interact with variation among accuracy, directional, and administrative
goals in ways that help us understand variation in judicial behavior in jury
disputes.
Methodological Opportunities
Although the juror track has produced more methodological sophistication
and variety than has the judicial/political approach to the study of judicial
behavior, both tracks have relied almost exclusively on quantitative methods
of analyzing their data, whatever the datas source. Indeed, this emphasis on
quantification is an important part of what distinguishes social science
scholarship in this area from traditional legal analysis. In this section we
propose an extension of quantitative methods to the study of judicial engi-
neering in jury disputes; however, we also outline some important opportu-
nities created for a sophisticated application of underutilized qualitative
methods.
Qualitative Methods
Applications of quantitative methods in current scholarship have established at
high levels of confidence that, for disputes that are politically salient and offer
judges sufficient discretion, trial judges can vary significantly in their responses
to similar disputes and that this variation correlates significantly with key
value-based background variation, such as the judge and his/her appointing
presidents support for, or opposition to, a womans right to choose. The most
recent efforts in this area have included logistic regression models, which have
helped us understand the relative influence of value-based variables and other
variables, including institutional variables (Buchman, 2007). Scholars working
in the law and psychology tradition have appropriately quantified findings
ranging from descriptive analysis of survey responses (Gatowski et al., 2001) to
aggregative comparison of judge-juror decision proclivities (Hersch & Viscusi,
2004), to multivariate analysis of data from experimental designs (Robbennolt,
2005). Each of these methodologies, including quantification of results, has
obvious and potentially productive application to studies of judicial engi-
neering of jury disputes. For example, one particularly intriguing possibility
is an experimental design reversing aspects of judge-jury comparisonsfor
example, how do retired judges rule on Daubert motions when viewing alter-
native argumentation or offered alternative information regarding the identi-
ties of the parties. However, we believe that the research program proposed
here also creates some important opportunities to benefit from underutilized
qualitative approaches.
A research program that examines judicial behavior in the context of jury
disputes, especially if framed in ways that respect judges own descriptions of
their judgments and the reasoning underlying those judgments, creates a rich
set of unmined opportunities for harvesting insights available via qualitative
analysis. Three exemplar avenues of opportunity are:
Content Analysis. Many jury disputes produce a voluminous,

publicly available written record of the dispute generally and of the
opportunities for judicial engineering in particular. Depending on
whether the dispute is resolved at a jury trial or how ripe the dispute
was if resolved short of a jury verdict, this written record will include
adversarial briefs interpreting facts and law in support of their
position on key motions, trial transcripts of oral argument and
verbal judicial response to key motions, and in some cases, the
judges written justification for his/her ruling on key motions.
Analytical techniques available would, of course, vary from
sophisticated, computerized content analysis to more traditional
examination of legal and evidentiary content. However, each of
these sources and techniques offers important opportunities to
analyze qualitative variation, or the lack thereof, in aggregate
patterns of behavior, and to identify specific examples of the
presence or absence of cognitive illusion in judicial reasoning.
Interviews. It is remarkable that with so much recent attention to
trial judging, so little effort appears to have been made to interview
judges and the parties to jury disputes engineered by trial judges.
Traditional wisdom from the political perspective is that judges
would be so unable or unwilling to acknowledge extralegal
influences or even the possibility of cognitive illusion that such
interviews would be unproductive and probably misleading.
However, this perspective derives in large part from the implicit

assumption that judges are strategic actors, consciously driven by
value-based preferences. If, as in our proposed theoretical
framework, this assumption is dropped, a number of interview
opportunities are apparent, ranging from identification of the
judges key audiences to questions asking the judge to
recreate his/her rulings on key motions, perhaps by reference to
these audiences. Moreover, addition of interview techniques to
the methodological quiver also suggests the importance of
interviewing not only judges but also the parties to a dispute and
the authors of key briefs and arguments. These paired interviews
would create a unique opportunity to, for example, reinforce the
content analysis of motions and opinions by reference to each
partys discussion of the evidence and law utilized in written
reasons for the judges ruling.
Case Studies. Case studies have long been a mainstay of more
traditional approaches to the study of courts and judging. As with
most research in the judicial area, these studies have focused
primarily on seminal Supreme Court casescauses, consequences,
and reasoning. Our proposed research program offers a unique
opportunity to extend the case-studywhether the case is an
individual judge, a single dispute, or combinations thereof
approach to judicial engineering of jury disputes. Specifically, it will
be interesting to examine the consistency, or lack thereof, in judicial
engineering for a single case. Do the judges rulings consistently
favor one party over the other? Does the judge consistently invoke
the same or similar decision criteria, or is he/she less predictable?
Are the judges rulings consistent with accuracy, expressive or
administrative goals? The case study methodologies will be
particularly productive for mass tortsfor example, asbestos,
Vioxx, breast implantsin the context of Multi District Litigation
(MDL) assignments in which a trial judge who administers the MDL
may be called on to make important rulings that define the
admissibility of evidence for literally thousands of cases and,
perhaps more important, the likelihood that these cases will settle
rather than go to trial and place thousands of burdens on case
docket(s).
Exemplar Research Targets

Having called for a new research program, including new theoretical and
methodological approaches, to study trial judges behavior in the context of a
jury dispute, we conclude this essay by suggesting two avenues of judicial
engineering that we feel offer particularly fertile ground for examination of

judicial behavior in the context of jury disputes: (1) judicial intervention in
the jury selection arena, and (2) judicial decisions regarding the admissibility
of expert testimony.
The exercise of jury-selection discretion is particularly appealing
because it is a call to examine judgments that, while extremely important
in defining the right to a jury trial, have been almost entirely ignored to date.
Given the influence of jury composition on trial outcomes, this dearth of
attention is puzzling and represents a major void in our understanding of
key judicial behaviors and the implications of those behaviors for civil
litigants right to a jury trial. This puzzlement is buttressed by the fact that a
judges jury-selection decisions are virtually bullet proof, rarely the subject of
appellate review unless the trial judge has violated the plain language of
eligibility statutes or has demonstrably abused his or her exercise of discre-
tion. Fortunately, jury-selection decisions are, for the most part, recorded in
daily transcripts by court reporters and become part of the record for every
jury trial presided over by a given judge; thus, the study of these engineering
decisions can benefit from a rich combination of the methodological
approach discussed above.
Having said that jury-selection decisions are appealing because they are
virtually unexamined, the study of decisions regarding admissibility of
expert testimony from the perspective proposed here is appealing precisely
because it is the natural extension of productive recent efforts on both
research fronts to understand trial judges judgments regarding the admis-
sibility of scientific expert testimony (Groscup, 2004; Buchman, 2007).
Buchmans work offers a particularly inviting foundation because he has
looked at federal district judges published Daubert opinions and concluded
that, at least for his sample of published decisions, judges ideology influ-
enced their decisions to grant or deny motions to admit or exclude expert
testimony under the Daubert standards, and his work suggests implicitly
numerous opportunities for expansion via the research program proposed
here. Most obvious would be expanded attention to unpublished decisions
generally and to state decisions regarding admissibility of scientific testi-
mony in particular. This expansion could include reexamination of these
decisions from the theoretical and methodological perspectives defining the
research program proposed herefor example, examination of alternative
motives and audiences, including donor groups in states that elect their
judges, analyzing the briefs and arguments, and possibly interviewing a
subset of judges and litigants. Less obvious is the possibility of synthesizing
Buchmans political-science findings and Gatowskis findings (Gatowski et
al., 2001) regarding the failure of most judges surveyed to understand most
of the Daubert decision criteria outlined in the Supreme Courts Daubert
decision. Moreover, the opportunities and incentives to expand this avenue
of inquiry have been expanded by recent decisions that extend judicial
gatekeeping and, therefore, judicial engineering to nonscientific expert

testimony (Groscup, 2004).
Jury-selection and admissibility decisions notwithstanding, virtually
every example of judicial engineering outlined in this proposal can be
subjected to analysis that in the aggregate will help us understand not only
how trial judges judge when they engineer the key elements of a jury trial, but
also help us understand what the right to a jury trial means in a system where,
constitutional architecture notwithstanding, judicial engineering defines that
right in practice.
13
Searching for Constraint in Legal
Decision Making
Eileen Braman
This essay is about constraint in legal decision making: how we should con-
ceptualize it, how we should study it, and why psychological theory and
methods provide such a promising avenue for doing so. I treat constraint as a
democratic good, necessary to justify the distributional decisions of unelected
judges and as an empirical question: Does meaningful constraint exist? If so,
where are we likely to find it? What are the potential sources of constraint in
legal decision making? Asking these questions should lead to what I hope will be
the next generation of empirical research on legal reasoning involving a wealth
of theoretically based questions of interest to scholars in numerous disciplines.
Constraint itself is defined as something that limits the freedom to act
spontaneously; or some physical, moral or other force that compels some-
body to do something or limits their freedom of action. When legal types talk
about constraint they are usually referring to professional norms and/or
obligations that require judicial actors to use legal presumptions, rules and
authority in reaching decisions.1 Traditional legal approaches hold that the
reasoned use of these tools will often compel particular outcomes.
The argument for requiring that decision makers use and cite accepted
sources of authority is at least twofold. First, where judges use their training to
apply the logic set forth in sources of legal reasoning, they are encouraged to
think about specific disputes in light of generally applicable rules and larger
societal values that legislative actors and other judges have deemed applicable
in similar circumstances. In this way referencing legal considerations helps
imbue judicial decisions with continuity and predictability. Second, judges
utilizing tools of legal reasoning are prevented from calling on their own
203
values, beliefs, and preferences in making decisions between adverse parties;

the fact that judges are engaged in the seemingly objective task of legal analysis
helps legitimate the distributive choices they make in our constitutional
system.
The conception of law as a meaningful constraint on decision makers has
taken a serious blow in the last century with the rise of legal realism in legal
academia and studies of judicial behavior in the social sciences. The over-
whelming weight of empirical evidence demonstrates that judges vote dis-
proportionally for outcomes that are consistent with their political policy
preferences. This casts significant doubt on the constraining force of legal
authority jurists cite as determinative in their decision making.
These findings are extremely important from a normative perspective
because attempts to justify judicial authority without reference to meaningful
constraint are fundamentally flawed. As such, scholars in all disciplines, but
particularly political science, should be deeply concerned with identifying
constraint in legal decision making. If we come up empty in this search, or
even empty in some types of cases we, as a discipline, will need to do some
very serious rethinking about why judges are among the political officials that
get to say who gets what, where, and how.
How to Look for Constraint: Refining Our Empirical Approach

The search for constraint has been elusive in empirical studies. I think this is
in part because scholars have inadequately conceptualized constraint in
studies of judicial behavior. Empirical investigators commonly look for
evidence of constraint in the wrong place (Segal & Spaeth, 1993, 2002) and
they often use exceedingly hard or unrealistic tests of the concept (Brenner &
Spaeth, 1995; Segal & Spaeth, 1996; Spaeth & Segal, 1999).2
Large-sample studies that correlate case votes with broad indicia of
judicial preferences (like ideology, political party, or the party of a judges
appointing president) tend to treat law in the abstract as a constraining
force. Critics of these studies argue they fail to take account of individual case
facts and circumstances. Behavioral scholars counter that these differences
wash out in the aggregate. Still, where results demonstrate the influence of
attitudinal factors in legal decision making, we are left with debates about
whether findings reveal the influence of attitudes and/or ideology in the raw
or ideologically influenced legal considerations (Gillman, 2001).3 Thus, for
all our efforts we still know very little about how decisional norms and
institutional context operate to constrain (or fail to constrain) judges that
are the subject of our investigations.
Meaningful legal constraint may or may not exist; this is an empirical
question that has yet to be adequately tested. Here I argue that constraint is
not undiscoverableand legal models of decision making are not unfalsifiable
(cf., Segal & Spaeth, 1993, pp. 3334). What we need are fair, theoretically
Searching for Constraint in Legal Decision Making 205
based, tests of constraint.4 Specifically, empirical scholars need to be more

explicit about the competing goals and influences that act on judges in the
process of decision making. Moreover, we need to consider how we investi-
gate constraint; this means adopting (or at least being open to) appropriate
methods and thinking more carefully about the decisions and legal decision
makers we analyze in our investigations.
In formulating hypotheses researchers should ask the following questions:
What are the sources of constraint in legal decision making? How do those
sources operate to influence the choices of decision makers? How would a
constrained decision look different from one where decision makers are acting
in accordance with their personal preferences? What specific decision contexts
allow us to observe behavior where the law and preferences are in conflict? Under
what conditions should we expect decision makers to make choices consistent
with the law? Under what conditions might preferences play more of a role?
This is where theory and findings from cognitive and social psychology
have such great potential to inform our research. Here I identify various
normative and contextual sources of constraint and discuss what psycholo-
gical theory suggests about how we should approach studying these aspects of
legal decision making. Knowledge of research on psychological topics like
persuasion, group decision making, accountability, analogical reasoning, and
the role of motives in the evaluation of evidence can significantly aid in our
theory building about where judges are relatively free to make decisions and
where they are constrained by norms of legal reasoning and/or the institu-
tional context in which they act. Rather than perpetuate the all-or-nothing
debate between legal and nonlegal determinants of behavior, we should not
only consider, but allow for, the interaction of legal and nonlegal influences in
our hypothesis testing.5
The concept of motivated reasoning is particularly useful for thinking
about the interaction between legal and nonlegal factors in decision making
(Kunda, 1990; MacCoun, 1998). Arguably, the theory has the potential to
resolve the disconnect between empirical evidence of ideological influence in
judges outcome choices and their own subjective accounts of legal determi-
nants of their own decisional behavior (Braman, 2006a; Braman & Nelson,
2007). As several other essays in this volume focus on judicial goals and
motivations, I will not say much about motivated reasoning except that this
line of literature can be particularly useful to empirical scholars in specifying our
assumptions, operationalizations, and hypotheses. We should be thinking more
about what decision makers are trying to achieve and how particular decision
rules and institutional constraints facilitate and frustrate these decisional aspira-
tions. Using motivated reasoning as an empirical framework can help us to be
more explicit about all of this in theory building and hypothesis testing.
Although I am confident creative minds can think of ways to do this with
regression analyses that judicial scholars have grown so comfortable with over
the years, we should not be afraid of embracing methods cognitive psycholo-
gists have used to understand decision making processes. Indeed there already
are studies employing content analyses (Tetlock, Bernzweig, & Gallant, 1985;
Gruenfeld, 1995) and experiments (Guthrie, Rachlinski, & Wistrich, 2001;
Braman, 2006a; Furgeson, Babcock, & Shane, 2008b; Braman & Nelson,
2007) to understand aspects of legal decision-making behavior.
Also, we may have to move away from analyzing final case votes toward
discrete choices judges make in the process of legal reasoning in order to get a
realistic idea of how and where constraint operates in their decision making.
For instance, judges make many component decisions in the context of
deciding a case: what issue is determinative? Does a particular statute
govern? Is a cited precedent applicable or can it be distinguished from the
current case? Understanding how the law shapes these discrete choices seems
an important aspect of investigating how decision makers are or are not
constrained by controlling authority.
Where to Look for Constraint: Potential Sources in Legal

Decision Making
In thinking about the sources of constraint in legal decision making, two
broad categories come to mind: (1) constraint imposed by legal authority and
application of appropriate rules of decision making, and (2) constraint that
results from the institutional context in which judges make decisions. I will
refer to the first type of constraint as normative constraint and the second
type as contextual constraint. In describing each potential source of nor-
mative and contextual constraint I point to literature in cognitive and social
psychology that should help our theorizing about how these constraints work
(or fail to work) to limit individual discretion.
The list I propose is not exhaustive. Moreover, not all the sources of
constraint I mention have easily identifiable correspondents in the psycholo-
gical literature. It is meant as a starting point to help us think about where
psychological theory can inform our research. In the spirit of hypothesis
building, I suggest ways researchers may attempt to test the constraining force
of norms and context when looking at these aspects of legal decision making.
Normative Sources of Constraint
Issues Related to Argumentation

In thinking about the normative sources of constraint we should start with
what stylized norms of decision making require of judicial actors choosing
between litigants in an adversarial context. One of the most basic assumptions
in our legal system is that justiciable disputes meet criteria that make them
appropriate for judicial resolution: there must be a case or controversy
involving a plaintiff who has suffered an injury in fact; moreover, the
conflict must be current and redressible by means of a judicial pronounce-

ment. Where cases meet these criteria, judges are supposed to follow certain
rules in the context of decision making.
We assume that as long as a dispute meets criteria demonstrating the
necessary averseness, litigants will have the incentives to fight hard and bring
the best legal arguments for their position to the attention of the court. As
such, there is a norm that limits judges to the issues and arguments raised by
litigants. In legal jargon the rule is that judges are not supposed to raise issues
or arguments sua sponte. The rule is important to maintain fairness in
decision making so litigants are not surprised by having cases decided on
the basis of facts and/or arguments they have not briefed or considered in the
context of their dispute. In this respect, judges are constrained by relevant
case facts and arguments raised by the parties to litigation.
There is a lively debate in political science about the extent to which
judges at the highest level of our judicial system comply with this norm, based
on comparative content analyses of case briefs and opinions (McGuire &
Palmer, 1995, 1996; Epstein, Segal, & Johnson, 1996). This is one area where
research on normative constraint that has already caught the attention of
empirically minded scholars may be aided by research methods and ideas
from psychology. For instance one question that has not been adequately
addressed in this literature is what motive(s) judges might serve by raising
additional issues and arguments. One could image a range of potential
motives ranging from the altruistic (e.g., to help litigants with inadequate
representation) to the strategic (e.g., to aid a decision maker in reaching the
ideological outcome the judge prefers).6
To get at this question, one might be able to create an experiment with
different legal briefs related to an identical factual scenario that vary in terms
of their comprehensiveness. If the norm is violated by decision makers there
may be systematic differences in the pattern of results that help reveal their
reasons for doing so. For instance, one could ask: are new issues or arguments
more likely to be raised in conditions where legal argument is sparse (sug-
gesting altruistic motives)? Or are they more likely to be raised by decision
makers with certain ideological views or by decision makers who decide the
case in a particular direction (suggesting more strategic motives)? Identifying
why decision makers may violate this particular constraining norm may help
us to sharpen our statistical analyses by identifying the class of cases where
such behavior is most likely to occur.
Another issue related to argumentation has been raised by Stone Sweet
(2002), who argues that we should reconceptualize our thinking about con-
straint to encompass the range of plausible arguments adversaries could raise
in litigation rather than the more narrow choice sets created by selected
arguments. On this view, argumentative choices themselves represent an
important part of the constraining influence of law. Framing literature from
psychology and political psychology (Nelson & Kinder, 1996; Nelson, Oxley,
& Clawson, 1997; Druckman, 2001) seems particularly relevant here.7
Also value pluralism models (Abelson, 1968; Tetlock, 1986), concerned

with how individuals negotiate competing interests, have great potential to
inform our research as we move from unidimensional models of decision
making to more realistic models where multiple issues and arguments are
present in light of complex, real-world fact patterns. After all, what is legal
decision making but a competition between competing values and interests?
As discussed in the chapter by Robbennolt, MacCoun, and Darley in this
volume (ch. 2), the psychological literature on how people negotiate value
conflict can help us move beyond the world of unidimensional theorizing
toward a more realistic understanding of how judges operate where more
than one issue is raised.
In a classic chapter on cognitive consistency, Abelson (1968) suggests
four distinct strategies for dealing with value conflict: bolstering, denial,
differentiation, and transcendence. Unfortunately not much has been done
to investigate when decision makers use particular conflict reducing techni-
ques since Ableson identified them. Kunda (1990, p. 480) points out that
although researchers have been quite interested in the scenarios that give rise
to value conflict, few have systematically explored specific strategies indivi-
duals use for dealing with it. Legal decision making provides fertile ground to
undertake this inquiry. Content analysis strategies may be devised to inves-
tigate where judges ignore (deny) or emphasize (bolster) arguments made by
litigants to get at these issues.
A final aspect of argumentation where we could benefit from psycholo-
gical approaches to decision making is, of course, related to the ultimate goal
of legal argument: persuasion. As there is an entire essay by Lawrence S.
Wrightsman on the topic in this volume (ch. 4), I will not spend much time
with it here, except to mention two points that seem especially relevant to the
subject of this essay. First, among the accepted tools of decision making (legal
text, statutory or constitutional intent, and precedent), there is what some
consider a hierarchy of authoritative forcetext is generally considered the
most definitive statement of what the law is, and arguably the source that is
least vulnerable to subjective interpretation (although admittedly text can be
wrought with ambiguity). The interpretation of precedent and intent are
considered by some to be somewhat more suspect because there is more
subjectivity inherent in the endeavor (Bork, 1990); still each is an accepted
source of legal analysis.
It would be interesting to know if the constraining force of one type of
authority is stronger than the others. For instance, is a U.S. statute that
declares a contested rule to apply in a certain class of cases more likely to
constrain decision makers who disagree with the stated rule, than a control-
ling Supreme Court precedent declaring the same rule? It wouldnt be hard to
create an experiment to test this. One might even find existing cases to
analyze; for instance, you could measure compliance with a court decision
in subsequent cases before and after it was codified by some legislative body.
Related to this question is how decision makers act in the presence of
competing types of authority. Does text trump conflicting precedent? Can

overwhelming evidence of intent convince decision makers that a specific
outcome should obtain notwithstanding contrary legal arguments? Howard
and Segal (2002) have a piece on the use of originalism in Supreme Court
jurisprudence that begins to investigate some of these interesting issues.
A final point related to persuasion and legal decision making is that
several studies have evidence suggesting individual differences in how deci-
sion makers respond to constraining authority related to attitudinal/ideolo-
gical traits of the decision maker (Braman, 2006a, findings suggesting
differences in responses to controlling precedent; Randazzo, Waterman, &
Fine, 2006, finding differences in responses to legislative authority). Until
now legal scholars and political scientists have all assumed law constrained or
(failed to constrain) uniformly across different types of decision makers. If
findings of individual differences persist we will need to take this into account
in our theorizing. Work in psychology and political psychology on individual
differences in motivated reasoning (Jost et al., 2003) and motivated skepti-
cism (Taber & Lodge, 2006) in evaluation of political arguments and evidence
might be particularly helpful in this regard.
Issues Related to Decision Rules

A second sort of normative constraint involves the specific application of
stylized rules of decision making in legal reasoning. Are individuals able to
comply with norms concerning how they should think in the context of
adversarial disputes? This is an especially interesting area of research as
judges and legal academics often suggest anecdotally that decision makers
are violating norms to reach decisions consistent with ideologically preferred
outcomes. We can now bring our empirical skills to bear on important
questions about the process of legal reasoning. How prevalent are violations
of decision norms? Under what conditions are they most likely to occur? Do
decision makers who are violating decision norms do so with self awareness?
Or are unconscious mechanisms of influence at play in the process of complex
decision making?
1. APPLICATION OF LEGAL PRESUMPTIONS. In terms of empirical investiga-

tions researchers have been almost obsessively focused on outcomes. This
ignores a potentially important source of constraint involving the ability of
decision makers to apply required presumptions in their legal analyses. The
ability to comply with rules about where reasoning should start from has
important consequences for the administration of justice in our judicial
system. One of the most important rules is, of course, the presumption of
innocence in criminal trials. It is this presumption that creates a heavy burden
of proof necessary for the state to take the freedom or lives of its citizens. If
decision makers come to the table with preconceptions about the guilt of
defendants the standard of evidence necessary for conviction is reduced. One
could imagine that decision makers are more able to comply with this
presumption in some types of cases (i.e., white collar vs. violent crime) or
more skeptically, with regard to some type of individuals (class or race based
differences in ability to apply the presumption). Experimental techniques
might help us understand if decision makers fall short of making required
assumptions and where they are especially likely to do so.
Prospect theory, especially as it relates to status quo bias, can help us
theorize about the application of presumptions in legal decision making.8
Psychological research reveals that the price at which people are willing to sell
things that already belong to them is typically higher than the price they
would be willing to pay to acquire them. This endowment effect (Knetch,
2000) seems especially important to the aspect of constraint requiring deci-
sion makers to apply presumptions in the context of thinking about legal
disputes. For instance, one of the most famous axioms in legal academia is the
Coase Theorem: in the absence of transaction costs it does not matter to
whom a right is initially assigned because the parties will be able to bargain
efficiently and it will wind up in the hands of whoever values it most (and is
therefore willing to pay to get it). Introducing endowment effects may
complicate matters significantly. If sellers attribute some unstated value to
simply possessing an objectbuyers who actually value that item more
may not be able or willing to meet their terms to acquire it. This endowment
effect is kind of a surcharge not adequately encompassed by the theory.
Moreover, it is not clear if the endowment effect for owning some objects (for
example, real property) is greater than the endowment effect for owning other
objects (for example, cups and penswhich is what much of the experi-
mental research on this topic has involved). One could hypothesize that the
endowment effect could be more of an obstacle for some types of Coasean
transactions than others.
Also, it is not clear whether there are third-party, observer, endowment
effects that might be relevant for judges adjudicating disputes concerning the
allocation of rights and resources. Are judges able to do the mental acrobatics
required of them if presumptions call for them to consider counterfactuals
that are at odds with the status quo? How, if at all, does who currently possesses
a right or resource influence their seemingly objective decision processes? For
instance, law students commonly study water rights in Property classes. In
western states, where water is scarce, a common presumption in this area of
the law is prior appropriation, the first person to use water from some
natural source acquires priority rights to that source against later users.
Often junior appropriators divert water from the beneficial use of land
owners with superior right to a particular source for several years before such
conflicts become the subject of litigation. How does the current state of affairs
influence the decisions of judges and/or juries in such matters?
Finally, presumptions are especially important in motion practice. As
discussed more fully in the essay by Rowland, Traficanti, and Vernon in this
volume (ch. 12), motions represent a highly significant aspect of legal
reasoning not often studied by empirical scholars. For instance, in deciding

motions for summary judgment, where parties argue there is no question of
material fact, judges are required to decide the case as a matter of law,
presuming all factual representations in the plaintiffs portrayal of the dispute
are true. We do not have a good idea of whether judges are able to do thisor
what role attitudinal factors play in deciding whether or not issues of material
fact actually exist. It would also be good to know the relationship between
factual and legal arguments in this sort of decision making. For example, if the
law favors the person who files the motion, are decision makers hostile to that
partys claim more likely to find some factual question precluding summary
judgment than decision makers who are not hostile to the claim? Or are all
decision makers constrained by interpretation of seemingly objective facts in
the same way?
2. ISSUE INDEPENDENCE AND ORDER OF OPERATIONS. One norm I have

gotten some leverage on with the use of experimental methods involves the
ability of legal decision makers to think independently about distinct issues in
litigation where multiple issues are raised (Braman, 2006a) Specifically, I did
this in the context of testing how decision makers decide a threshold ques-
tion in light of their views on policy questions related to the merits of a
politically charged case. I have used intuitions about separable preferences to
theorize about these concerns, but there might be other theories that are
equally (or more) appropriate.
There are certainly other aspects of legal decision making where the
violation of this particular norm can have important implications. For
instance, in bench (and jury) trials can decision makers separate their rea-
soning about liability (or guilt in criminal matters) from potential penalties
defendants are facing? If this is a problem, does bifurcation help? Or would we
do better to have separate decision makers at each stage? Recent work by
Wistrich, Guthrie, and Rachlinski (2005) suggests that magistrate judges have
trouble disregarding inadmissible information they may have been privy to in
preconference discussions. The authors recommend that different judges
handle distinct stages of litigation. Their study and recommendations
demonstrate how psychological research can suggest relatively easy changes
in our judicial system to make litigation more equitable.
Another very interesting question involving legal decision rules concerns
the order of operations legal decision makers are supposed to follow when
considering arguments and authority. Akin to mathematics, there is a logical
sequence decision makers are supposed to follow when reasoning through
cases where certain types of grounds are available for a decision. The rationale
behind these rules (which are sometimes referred to as the Ashwander rules
for a Supreme Court decision in 1936 that summarized them) is that judges
should cause as little disruption to existing democratic forms as possible when
deciding specific disputes. The most familiar rule is that decision makers
should not reach constitutional issues unless absolutely necessary. This
means if a matter can be determined on statutory (or other) grounds judges

are required to do so without considering the constitutional questions that
are potentially raised by the fact pattern.
Have judges been able to follow this norm? Do attitudinal factors come
into play where judges reach different conclusions about the necessity of
addressing constitutional concerns? An analysis of cases specifically citing
Ashwander (1936) could help us answer some of these questions. Is it cited
more often where judges choose to avoid or reach the constitutional issue?
Where judges do find it necessary to address constitutional matters does it
depend on the ideology of the judge(s) deciding the case in some respect?
A study much like Spaeths (1964) classic critiquing Justice Frankfurters
practice of legislative deference may be useful here. Looking at the actual
grounds of decision in cases where Ashwander is cited might shed significant
light on whether there are ideological factors at play in using this seemingly
neutral decision rule.
3. LATITUDE IN APPLICATION OF ACCEPTED DECISION RULES. Another issue

regarding decision rules concerns the level of discretion decision makers have
in the application of specific rules that are supposed to govern their analyses.
Perhaps the most commonly discussed issue in this vein is whether and to
what extent decision makers are constrained by the application of precedent
(Sunstein, 1993, 1996; Sherwin, 1999; Simon, 2004). Much of the legal work
on analogical reasoning in legal discourse borrows from disciplines like
philosophy or linguistics. Psychological findings and theories of analogical
reasoning (Holyoak & Thagard, 1995; Gentner, 1998) have been vastly under
utilized in this discourse. The clear exception to this rule is Simons work
(with Holyoak and on his own) on the importance of coherence in similarity
judgments (Simon, 2004, Holyoak & Simon, 1999); although others are
starting to think carefully about applications in legal research (Spellman,
this volume, ch. 10). Clearly there is much more to learn about how and
whether decision makers are meaningfully constrained by precedential
authority. Psychological research can greatly inform this important inquiry.
Similarly, research on more general categorization processes may help us
understand decisions about the applicability of statutes involving certain
types of litigants and/or categories of behavior in the minds of judges.
A final question having to do with the differential constraining authority
of decision rules has to do with the oft made distinctions between rules and
standards in legal discourse. Conventional wisdom holds that black-letter
rules confine decision makers to a greater extent than more flexible legal
standards (see, for instance, Schauer, 1991; Schlag, 1985; Sullivan, 1992).
This is actually an empirical question that has yet to be tested (Braman,
2006b). One could do this experimentally by giving decision makers an
identical factual scenario and asking them to make the same decision with
different decision criteria. One condition would involve a decision rule; the
other a legal standard. Decisions across conditions could then be compared
with respect to the degree of attitudinal influence and variance across decision
makers. If conventional wisdom is right, decision makers in the rule condi-
tion should be less influenced by their attitudes and there should be less
variance across decision makers within that condition.9 Regardless of what
the findings show, this sort of experiment would tell us something very
interesting about the confining influence of different decision criteria that
have been a subject of substantial interest in legal discourse.
Contextual Sources of Constraint
An essay on constraint in legal decision making would be remiss if it failed

to mention contextual or institutional constraint on the ability of judges to
decide cases as their preferences dictate (Gibson, 1980). Rather than go
through an exhaustive list of contextual sources of constraint I will mention
two where research from cognitive and social psychology seem particularly
relevant: the group nature of appellate decision making and the requirement
that judges justify and be held accountable for their decisions.
Issues Related to Group Decision Making

Political scientists generally attribute directional policy motivations to judges.
Whether researchers take a strategic or sincere view of the role of pre-
ferences in decision making, the presence of others in the decision process acts
as a constraint on the ability of judges to reach outcomes they most prefer.
Neither strategic nor sincere accounts of attitudinal influence portray judges
as open to reasoned arguments about why the decision should be different
than they believe at the outset of discussions. One could argue that this is at
odds with how decision makers themselves actually conceive of the process of
appellate decision making. The democratic justifications for having judges
decide cases in groups are to (1) avoid error and (2) reach the best resolu-
tion of legal questions by having several decision makers come to a compro-
mised agreement about what outcome should obtain in particularized
disputes. To the extent judges have internalized these justifications for
group decision making, portrayals of them posturing to achieve predeter-
mined policy goals in conference are taken as a personal affront.
Furthermore, much of the extant empirical research treats group pro-
cesses as the sum of individual preferences, measuring for instance, how
much traditionally liberal judges move toward conservative positions in
the process of appellate decision making (Epstein & Knight, 1998; Collins &
Martinek, 2007). This emphasis on discrete individuals in the group context is
due, in part, to path dependence in the way political science operationaliza-
tions have evolved. Ideology is an individual trait; separate case votes are easy
to count and analyze.
I would like to draw attention to another conception of group decision
making that has been ignored by political scientists who assume directional
policy motivations on behalf of judges, by raising a possibility more in line

with legalistic conceptions of appellate decision making and more consistent
with psychological research on group decision processes (see Morehead,
1998, for review; Gruenfeld & Hollingshead, 1993). This is the idea that
group decision making is more appropriately conceptualized as a collective
cognition process with its own unique properties. Research on sociocognition
in groups (Gruenfeld & Hollingshead, 1993) tends to demonstrate that group
outcomes are more than (and different from) the sum of individual views of
group members.
When legal decision makers act in groups they are arguing about more than
case votes; they are talking about doctrine and what the appropriate grounds of
a decision should be. In this process decision makers likely come to a collective
understanding of what the law requires in a particular case that is distinct from
what any individual thought at the outset. Indeed group discussion may itself
lead to insights and considerations that would not have come into play if several
decision makers had not come together to discuss a particular matter.
Perhaps where decision makers write separate opinions traditional
notions from political science are appropriate; there judges may refuse to
accommodate beliefs or opinions. But when judges are part of majorities, I
think political scientists should at least allow for the possibility that decisions
result from group processes more akin to sincere compromise and reasoned
accommodation than strategic interaction. Creative research minds might
even be able to come up with a way to test these competing conceptions of
appellate group dynamics. Simulations where researchers collect individual
data before group discussion, then have decision makers come to some
collective judgmentcarefully observing group dynamics and outcomes
seem like the most promising avenue to get at these processes. Law students
or retired judges who are familiar with stylized rules of decision making could
be used as participants in simulations.
Finally, there are many psychological studies on related topics like social
influence (Sherif, 1936; Asch, 1955), majority and minority influence
(Moscovici, 1980), group polarization (Moscovici & Zavalloni, 1969), and
groupthink (Janis, 1982) that could be relevant to appellate decision pro-
cesses. Gruenfeld has done some very interesting work looking at the inte-
grative complexity of majority versus separate opinions on the Supreme
Court (Gruenfeld, 1995). I reference this work here to acknowledge that I
have not dealt with the potential contribution of literature on group decision
processes as fully as I might have in this chapter. I strongly believe, however, it
is a literature we should be paying much more attention to in our theorizing.
Issues Related to Accountability
There is some excellent work on the psychology of accountability as a result of

a research agenda of Tetlock and colleagues that spans a period of over 20
years (see Lerner & Tetlock, 1999, for a cogent review of findings). Here I will
highlight findings that seem particularly relevant to the behavior of judicial

decision makers required to account for decisions via written opinions that
justify their choices in adversarial disputes. First, through a series of very
clever pre- versus postjudgment experimental manipulations, Tetlock
(1983b, 1985) and Tetlock, Skitka, and Boettger (1989) have shown that
accountability concerns influence not only how decision makers justify
their choices, but how they think about the decisions for which they may be
held accountable. This is important and perhaps underappreciated in the
judicial literature, where researchers sometimes talk as if judges are only
concerned with coming up with explanations to satisfy (or perhaps pacify)
external observers of court outputs (Epstein & Knight, 1998).
Although not all the effects of accountability on thought processes are
positive,10 generally decision makers who understand they will be accountable
for their choices tend to engage in more in-depth processing than decision
makers who do not expect to have to justify choices. Accountability tends to
stimulate open-minded, self-critical thinking and attenuate biases under spe-
cific conditions including those where an audience (1) is interested in accuracy,
(2) is reasonably well informed, and (3) has a legitimate reason for inquiring
into the reasons behind particular judgments (Lerner & Tetlock, 1999, p. 259).
Put another way, accountability tends to heighten accuracy motivations. When
we know others are watching we want to get things right and we also strive to
use appropriate decision criteria to avoid criticisms that may be raised down
the line. One could argue judicial decision makers operate under these condi-
tions and, as such, accountability concerns should heighten their motivation to
comply with norms of appropriate decision making.
A second, but equally relevant, point is that whom decision makers are
accounting to is important. Judges have multiple audiences including liti-
gants, attorneys, legal academics, other judges, and members of ideological
groups (Baum, 2006). A judges desire to court, appease, or satisfy any one of
these audiences can act as a real constraint on their decisional behavior.11
Here a second strand of the value pluralism literature mentioned previously
regarding self-categorization processes may help understand how judges
think about themselves in light of multiple possible identities (Tajfel &
Turner, 1985; see also, Brewer, 1991). Also, as we come to realize judicial
actors are human and potentially concerned with issues like self-presentation
(Baum, 2006), value pluralism models can help us understand when they are
likely to court certain constituencies and when their desire to do so may be a
real limit on their decisional behavior.
Conclusion
This is an exciting time in the development of our knowledge of legal decision
making and the constraints that act on judges engaged in the task. Part of this
excitement is because of the important implications these findings have for
democratic notions of expert legitimacy and for how we must accommodate

our theories to recognize what we know and have yet to learn about how
judicial actors decide cases in our constitutional system. More of the enthu-
siasm I feel is based on what we all have to gain from the endeavor. I hope I
have made a convincing argument that understanding constraint in decision
making will greatly contribute to our mutual understanding of legal decision
making behavior.
Notes
1. There are, of course, also contextual sources of constraint in our legal system
including the group nature of appellate review and accountability mechanisms.
I will deal with these more fully later in the essay.
2. See Braman and Nelson (2007) for a detailed critique of extant operationaliza-
tions of the constraining influence of precedent:
For a variety of reasons, the Supreme Court is perhaps the least likely
place to detect the powerful gravitational pull of cited authority.
Because the Supreme Court sits atop our judicial hierarchy, justices
are not subject to the constant threat of review that may serve to
heighten accuracy goals for lower court judges. Moreover, the
Supreme Court has the ability to overrule itself. Lower court judges
must follow binding authority without explicit license to change or
ignore higher court rulings. Many of the justices on the Supreme
Court see themselves as high-profile representatives of a specific
jurisprudential approach (Baum 2006), and are thus unlikely to back
away from public statements about how the law should be applied in
prior dissenting or concurring opinions. Thus, tests of constraint that
have been applied to the Supreme Court set an exceedingly high bar,
and may not be appropriate to assess the influence of precedent more
generally.
3. See Braman (2004, pp. 4552) Braman and Nelson (2007, p. 942) regarding top
down versus bottom up theories of attitudinal influence in legal decision
making on why this is important distinction. See also Bartels (this volume, ch.
3) Top-Down and Bottom-Up Models of Judicial Reasoning.
4. Gillman (2001) writes, [I]f some contemporary positivists are willing to make
empirical claims about the determinate influence of rules, then it is not unrea-
sonable for scholars such as Spaeth and Segal to ask them to formulate those
claims in ways that allow for hypothesis testing (Gillman 2001, p. 486). That is
what I attempt to do here.
5. I do not mean to imply that empirical studies have been blind to the influence of
legal factors in models of decision making (Segal, 1984 [search and seizure];
Songer & Haire, 1992 [obscenity cases]; Gryski, Main, & Dixon, 1986 [sex
discrimination]) but where such factors are treated as control variables that
improve our attitudinal predictions (i.e., Segal & Spaeth, 1993) their influence
tends to be obscured.
6. Significantly, altruism (as I have defined it) may be less of a motivating factor at
the Supreme Court, where researchers have actually looked at this question
especially if one assumes that litigants who get to our highest court are likely to
be relatively well represented.
7. See also, Nelson, Maruska, and Braman (2002) for an analysis of how alternative
argument frames can influence lay perceptions of important issues in litigation.
8. The challenge Kahneman and Tverskys (1979) theory of decision under risk
poses to purely rational models of decision making has been realized and
acknowledged across multiple disciplines. Legal scholars have shown through
their own experimental techniques that magistrate judges, like ordinary people,
can be subject to framing effects in choosing between settlement options
(Guthrie, Rachlinski, & Wistrich, 2001).
9. I have yet to come up with a psychological theory to express this expectation
that seems almost intuitive to legally trained individuals. It might be an example
of the kind of decision domain where there is no psychological theory that
neatly corresponds with conventional legal wisdom.
10. For instance, decision makers who will be held accountable tend to overem-
phasize the importance of irrelevant information; this is commonly referred to
as the dilution effect.
11. Moreover, where judges have a prior record of deciding cases in a certain way,
memorialized in written opinions accounting for prior decisions, consistency
motives (Abelson et al., 1968) are likely to kick in and act as a further constraint
on decisional behavior.
Part III
EVALUATING AND IMPROVING JUDGING
14
Evaluating Judges
Gregory Mitchell
Much of the interest in empirical studies of judges lies in the comparison of

actual to ideal behavior. When we ask what makes a good judge or whether a
judge rightly decided a case, we implicitly compare the judges decisions to a
normative standard. In some instances, the content of the normative standard
is uncontroversial and its application straightforward. Hence, a trial judge
who sentences African American defendants to longer terms of imprisonment
than white defendants, all other things being equal, departs from the norma-
tive ideal that application of criminal laws should be color-blind (e.g., Blair
et al., 2004; Pruitt & Wilson, 1983). Or the circuit court judge who votes to
uphold a state law banning abortions on grounds that Roe v. Wade was
wrongly decided departs from the normative ideal that inferior court judges
should follow the Supreme Courts constitutional decisions (for evidence on
how circuit judges view their obligations within the legal hierarchy, see Klein,
2002).
In many instances, however, the content of the normative standard
chosen for comparison purposes proves controversial or hard to delimit
and its application unclear. Is the judge who applies a rule of evidence
according to its plain language a better judge than one who applies the
rule to achieve its purpose when following the text would contravene the
purpose? Is the judge who deviates from the law to correct a disparate impact
that the law is having on minorities, say, in sentencing, a bad judge? Is the
circuit judge who narrowly construes the Supreme Courts abortion cases a
better judge than the circuit judge who broadly construes this precedent?
Does it matter if the former judge favors this minimalist approach because of
221
222 Evaluating and Improving Judging
pragmatic commitments rather than political preferences? Is there even a

neutral approach to construal of Supreme Court precedent that can be labeled
the way of the ideal judge?
We may decompose any empirical study that attempts to address the
competence of judges or the quality of judging into three basic components:
(1) the specification of a normative benchmark; (2) the conversion of the
benchmark into testable form and judicial behavior into measurable units;
(3) the interpretation of the results of any comparison to draw appropriate
conclusions about the descriptive-normative gap.1 In this chapter, I consider
complications at each stage in the comparison process, with illustrations from
existing studies of judicial competence and studies from psychology that
examine the gap between behavior and norms of rational judgment and
decision making.
I contend that we presently know very little about the degree to which
judges depart from normative ideals because political scientists focus too
much on the Supreme Court and often employ flawed benchmarks, psychol-
ogists focus too much on juries, and legal scholars focus too much on the
normative side of the descriptive-normative comparison. I discuss seven
different conceptual approaches to the evaluative study of judging: bench-
marking based on social, moral, legal, coherence (or internal consistency),
and efficacy (or correspondence to an external criterion) norms; evaluations
based on notions of good judicial character; and evaluations based on relative
performance in judicial tournaments. In the process, I consider the benefits
that may be achieved by embracing a more diverse set of benchmarks for
judicial evaluations. I predict that greater diversity in normative perspectives
will lead to a greater appreciation of the importance of legal structure relative
to personal characteristics in judicial behavior and lead to new views about
the contours of the descriptive-normative gap.2
Why Study Judicial Competence?

Knowledge about the descriptive-normative gap in judicial behavior is
important for what it may reveal about the lawfulness of particular cases
and the predictability of courts, but more generally for what the size and
contours of this gap may reveal about the legitimacy of the legal system and
the meaning of the rule of law. Describing judicial behavior in relation to
norms of good judging holds an important place within debates about the
proper role and power of the federal judiciary relative to other branches of
the federal government and relative to state governments. As Sherry (2005)
argues, the rise of the view of federal judges as just another political actor
within the government likely contributes to elite, and perhaps popular (Scheb
& Lyons, 2001), discontent with judicial review and provides cover for those
elites seeking to change the balance of power (see Sisk & Heise, 2005).
Characterizing judges as unaccountable yet political actors rather than
Evaluating Judges 223
experts in the administration of law takes the sting out of arguments to limit
judicial review and give greater power to the legislature and executive.
Likewise, to the extent that the legitimacy of the courts and their ability to
motivate voluntary compliance with the law derives from the popular view of
the courts as detached from ordinary political processes (Tyler & Mitchell,
1994), evidence that judges are just political actors in robes should undercut
the legitimacy of the courts. However, these arguments should lose steam if
existing studies into the descriptive-normative gap are far from compelling,
as I contend they are.
From a prescriptive perspective, reliable knowledge about the descriptive-
normative gap is needed to understand where judicial incompetence exists so
that we may attempt to correct it.3 If judicial behavior systematically deviates
from legal norms in some domains, then steps may need to be taken to rein
in this predictably unlawful behavior. If judicial behavior commonly, but
randomly, deviates from legal norms in other domains, then steps may need
to be taken to eliminate chance influences in these domains.4 If judicial
behavior proves unresponsive to prescriptive interventions in these domains,
then the legal norm, or the task to which the norm is to be applied, should be
reconsidered on grounds that ought implies can. If some types of judges
perform better than others under consensual norms, then we can evaluate
whether our preconceptions of the personal qualities and training associated
with competent judging are correct and gain valuable information about
whether the judicial selection process should be altered.
By suggesting the many different standards that may be used to examine
judicial competence and by suggesting the limits of the usual comparison
focused on law versus politics (between attitudinal-type models and simple
legal models of judicial behavior5), I hope to encourage the debate to move
beyond characterizations of judges as driven primarily by ideology or law (for
a more comprehensive effort in this regard, see Baum, 2006). At all levels of
the judiciary, judges political attitudes undoubtedly play some role in their
judgments and decisions, but many other aspects of a judges psychological
make-up and of the social-psychological setting in which a judge acts will be
influential as well, with some of these influences pushing the judge toward
idealized normative behavior and some pushing the judge away from judicial
ideals. Gaining a more sophisticated understanding of when and why judges
deviate from various judicial norms should provide a better understanding of
potential sources of judicial legitimacy or illegitimacy and aid in the important
task of developing prescriptive models of judicial behavior.
The Limits of Our Present Knowledge on the

Descriptive-Normative Gap
One need not endorse a skeptical view of the limits of social scientific knowledge
to endorse the view that we presently know little about the size and contours
of the descriptive-normative gap in judicial behavior. Political scientists, who

by far have conducted the greatest number of empirical studies of judicial
behavior, focus a disproportionate amount of their efforts on appellate courts
and the U.S. Supreme Court in particular (see Friedman, 2006; Klein, 2002;
Maveety, 2003). Even if the strongest claims of the attitudinalists and
quasi-attitudinalists (Baum, 2006, p. 7) are true about the role of personal
policy preferences and strategic preference-maximization behavior in shaping
the decisions of Supreme Court justices, there is little reason to believe that
any strong version of the attitudinal model generalizes to judges on inferior
courts (see Gerber & Park, 1997). Although lower courts are studied much
less commonly, existing empirical evidence indicates that legal norms
influence judges on lower courts to a greater extent than on the Supreme
Court (Cross, 2007; Klein, 2002; Rowland & Carp, 1996), and many social
psychologists would argue that the situation of the judge will often exert
greater influence on judicial behavior than preexisting preferences. From
this perspective, studies of supreme courts have limited external validity:
one cannot assume that judicial actors in very different situations (especially
those who do not control their docket in the way the Supreme Court does)
will act the same; equivalence must be empirically demonstrated rather than
assumed.
Psychologists, for their part, simply have not conducted many empirical
studies of judicial behavior. Whereas mock jury studies make up a high
percentage of all law and psychology empirical studies, studies of judicial
behavior are infrequent. Consider, for instance, that the chapter on
Psychology and Law in the most recent edition of the Handbook of Social
Psychology contains no discussion devoted specifically to studies of judicial
decision-making (Ellsworth & Mauro, 1998). Thus, while psychology may
have much to offer the study of judging (Wrightsman, 1999), psychologists
have had little to say about judges.
Legal scholars, for their part, spend enormous amounts of time telling the
courts how they should decide casesindeed, normativity seems almost a
prerequisite to publishing in law reviews (Rubin, 1997)but relatively little
time studying empirically what, beyond the reasons given in opinions, moti-
vates judges and whether courts can in fact follow the norms that legal
scholars recommend to them. When legal scholars do consider the descriptive
side of the descriptive-normative comparison, they tend to favor qualitative
studies that identify descriptive-normative gaps in particular cases (e.g.,
Barnett, 2005; Vermeule, 2006) or lines of cases (e.g., Barnett, 2004). While
qualitative studies can generate important hypotheses for future testing and
rich stories about how judges do or do not comply with legal norms, such
studies can provide only limited evidence about the systematicity of this
behavior and cannot test competing causal hypotheses due to sample restrictions
(Buthe, 2002; King et al., 1994).
A variation on the last point is worth emphasizing: beware of vivid
anecdotal evidence of alleged judicial misbehavior (of the kind found, for
instance, in Boot, 1998) or alleged theory-confirming judicial decisions. Segal

and Spaeth (2002) begin the revised edition of The Supreme Court and the
Attitudinal Model with discussion of the Supreme Courts decision in Bush v.
Gore, which they take to be a recent example of judicial policymaking. Baum
(1997), on the other hand, begins his review of judicial behavior research with
anecdotes about conservative Supreme Courts deciding Roe v. Wade and then
affirming Roe in Planned Parenthood v. Casey to suggest that judicial attitudes
cannot do the explanatory work often assigned to them.6 Neither book bases
its main conclusions on anecdotal evidence; rather, each understandably
uses anecdotes for narrative purposes. But the contrast illustrates the point:
empirical debates cannot be decided with anecdotal evidence, as each side
will be able to find a seemingly endless supply of supportive stories. That is
not to say that exemplary judges or exemplary cases cannot be important
sources of information about the characteristics and behavior of good (and
bad) judges and the institutional features that promote or inhibit good
judging. But reaching consensus on exemplary behavior returns us to the
original problems of identifying the norms to use for evaluating judges and
measuring the exemplary qualities.
Alternative Normative Perspectives on Judging

Normative standards derive from analytical frameworks that enjoy authoritative
status, either by general acceptance because the framework seems to work well
or because it has been justified or promulgated by a source to whom deference
is widely given (Baron, 2004). Some frameworks impose order, such as the
rules of arithmetic; other frameworks are supposed to promote desired goals
efficiently, reliably, or accurately, such as expected utility theory for the
maximization of utility or scientific principles for the development of accurate
and reliable empirical knowledge. Some normative systems, including
systems of social and legal norms, are not easily reducible to a single clear,
consensual goal or function (Elster, 1989), and some normative systems
remain quite controversial, as with the debate over the propriety of norms
of rationality for evaluating human judgment and choice behavior (e.g.,
Cohen, 1981). For the most part, empirical legal researchers take norms as
identified by others (usually philosophers, statisticians, game theorists, or
legal scholars) and apply them to actual judgment and decision-making
behavior without wading into the foundational normative disputes.
The potential norms for evaluating judicial behavior may be divided into
five categories: social, moral, legal, coherence, and efficacy norms.7 Social
norms, in this context, involve interjudge or intercourt comparisons: Did
Judge X behave as the median judge behaves in such a case? Are the judges of
the Ninth Circuit outliers in their treatment of Establishment Clause cases?
Although some judicial studies do discuss social norms within courts, few
treat deviation from judicial social norms as a reason in and of itself for
criticizing a judges behavior.8
Moral norms, as I use the term here, encompass any norm derived from a
theory of justice or ethical obligations. Legal scholars often use moral norms
to evaluate individual judicial decisions (see Posner, 1999), but these studies
typically dispute the normative status of the law by recourse to moral arguments
rather than describe judicial behavior. Several empirical studies examine the
degree to which parties or citizens in general perceive judges and the courts to
be acting justly or to be allocating outcomes justly (e.g., Tyler & Huo, 2002;
Tyler & Mitchell, 1994). Procedural justice norms are an important source for
evaluation, but they typically are used in connection with efficacy norms as
discussed below. Other than procedural and distributive justice norms, few
empirical studies of judging evaluate judges against moral norms as defined
here (except with respect to selected decisions).
Legal norms for present purposes are norms established by an author-
itative governmental body, including constitutional, statutory, and adminis-
trative law and judge-made common law, to bind public and private
institutions and the persons running and governed by these institutions.
The power of courts to clarify, establish, or alter legal norms complicates
matters from an empirical perspective, for it means that in some instances it is
legitimate for a court to ignore, alter, or invalidate one norm to resolve
conflict with another norm (as may occur through judicial review or
common law development) or to embellish a norm (as may occur through
legislative or constitutional interpretation). Uncertainty about what counts as
law or binding legal norms likewise complicates matters (e.g., Hart, 1961), as
we have seen with recent disputes about the role of foreign law in U.S. courts
(e.g., Rahdert, 2007; see generally Shapiro, 2007).
Accordingly, a secondary set of legal norms has developed to govern
interpretive practices and to establish a hierarchy of precedent and deference
to guide superior and inferior courts, future panels of the same court, and
branches of the government as they sort out the content of legal norms and
the obligatory status of these norms. For instance, the norm of rationality
review within constitutional law provides that a legislative act with some
conceivable purpose behind it should be upheld by the courts.9 It is often this
secondary set of interpretive and role norms that serve as the primary legal
norms in comparisons of actual to ideal judicial behavior (e.g., Segal &
Spaeth, 2002, operationalize the legal model of judicial behavior as
requiring fidelity to the plain meaning of statutory or constitutional text, to
the legislative or constitutional drafters intent, and to precedent).
Coherence norms, including norms of rationality, require logical consis-
tency and coherence in the formation and ordering of beliefs and preferences,
whatever those beliefs or preferences may be:
technical discussions of rationality generally adopt a logical
conception, in which an individuals beliefs and preferences are said
to be rational if they obey a set of formal rules such as

complementarity of probabilities, the sure-thing principle or
independence of irrelevant alternatives. In the laissez-faire spirit of
modern economics and decision theory, the content of beliefs and of
preferences is not a criterion of rationalityonly internal coherence
matters. (Kahneman, 1997, pp. 105106)
The goal of this coherence metatheory of judgment is to describe and explain
the process by which a persons judgments achieve logical, or mathematical,
or statistical rationality (Hammond, 2000a, p. 53).10
Rationality as coherence requires that judgments exhibit (1) resistance to
logically irrelevant features of the situation (e.g., post hoc judgments about
the reasonableness of a course of conduct should be immune to whether the
course of conduct caused no harm or severe harm, given that this information
was unavailable when the conduct was undertaken), and (2) appropriate
sensitivity to logically relevant features of the situation (e.g., base rate information
should be properly assimilated into a Bayesian calculation) (Kahneman,
1991). Furthermore, (3) judgments should lead to a coherent and consistent
web of beliefs (i.e., a change in one belief may require that other related beliefs
be altered) (Tversky & Kahneman, 1982).
Coherence norms evaluate behavior solely in terms of coherence and
consistency, with no necessary connection to real world success or empirical
accuracy:
It may come as a surprise to the reader that rationality does not
directly imply accuracy and vice versa, but brief reflection shows that
this is the case. Rationality always operates in a closed system; given
the premises, certain conclusions follow if a rational reasoning
process is followed. When the reasoning process satisfies a logical
test, the system is termed coherent, and that is all it is and all it claims
to be. (Hammond, 2000a, p. 53)
This does not mean that behavior within a coherence metatheory is
purposeless, nor that goals or beliefs are irrelevant to rationality in this
system. It means only that goals and beliefs are to be defined by the individual,
and then behavior is examined to determine whether it is logically consistent
and coherent in light of these goals and beliefs and whether beliefs are
appropriately updated in response to new evidence.
In contrast, what may be called efficacy norms test judgments and
decisions against empirical reality. The most common efficacy norm used
in psychological research on judgment and decision-making is a correspon-
dence norm: The goal of a correspondence metatheory is to describe and
explain the process by which a persons judgments achieve empirical accuracy
(Hammond, 2000a, p. 53). Thus, whereas psychologists applying a coherence
norm may examine whether causal attributions seem to follow norms for
unbiased testing of causal hypotheseswith no attempt being made to assess
the accuracy of these attributionspsychologists applying a correspondence

norm examine the circumstances under which causal attributions tend to be
more or less accurate (Funder, 1999). Correspondence norms, in theory,
provide the simplest source of comparison standards because they provide
an objective baseline for comparison and hold obvious attraction for evalu-
ating the decisions of judges. Of course, in many instances, reality is not easily
measured (e.g., truth is a correspondence norm for judging the outcome in
criminal trials, but it is often very hard to determine truth in criminal trials).
Efficacy norms focus on the outputs of a system to test whether the
system is functioning as it is supposed to according to a functional analysis
or a division of labor, whereas coherence norms focus on fidelity to the rules
prescribed to govern the system. Thus, if our perceptual system leads to
effective navigation in our environment, then the perceptual system receives
high marks under efficacy norms despite evidence of systematic perceptual
illusions (Funder, 1999; Hammond, 2000b).
A variety of possible efficacy norms for judicial evaluation exist, most
obviously norms of accuracy and error avoidance: how accurate are a judges
fact-based rulings and empirical predictions and how does a judge resolve
evidential ambiguities: does she err in the direction desired by the system
(e.g., significantly more Type II than Type I errors in verdicts in criminal cases
and slightly more Type II than Type I errors in verdicts in civil caseswhere
the null hypothesis is not guilty or not liableor substantially more Type II
errors on threshold rulings in civil cases, such as rulings on motions for failure
to state a claim)? Or we may ask whether judges succeed at obtaining
compliance with the law in their adjudications, however that goal is achieved.
Or the related question of whether parties are subjectively satisfied with a
judges rulings.
Note that these efficacy norms derive from analytical frameworks devel-
oped independent of what happens in actual cases. That is, the norms derive
from some theory or observable consensus about what makes judgment and
decision-making behavior good or bad or about what goals judges should
strive to achieve. Alternatively, one could reason backward from cases where,
say, litigants (or others) express low and high levels of satisfaction to search
for patterns of judicial behavior associated with those reactions and try to
discern what norms, if any, the judges tend to follow or what cognitive and
motivational qualities these judges tend to possess. Such an empirics-based
approach to norms of good judgment has not, to my knowledge, ever been
undertaken in a formal or systematic fashion with respect to judicial behavior,
but historical and anthropological studies of courts contain elements of this
approach (e.g., Philips, 1998).11
Many evaluative studies of judicial behavior begin with one of these
norms of good judgment and compare specific judicial behaviors to that
norm, but that is not the only way to approach the descriptive-normative
gap. One increasingly popular alternative is to conduct a tournament of
judges (see, e.g., Choi & Gulati, 2004a, 2004b), in which judges compete on
one or more measurable criterion variables and the tournament produces a

relative ranking of judicial performance on these variables. The impetus
behind such tournaments is, in part, recognition of the contested nature of
specific norms of good judging and the desire to emphasize objective mea-
sures of performance in judicial evaluations.12 Another popular alternative is
to proceed from a dispositional or character-based theory of good judging,
which posits that the possession of certain traits or qualities is sufficient for
judicial competence (e.g., Solum, 2003). According to these theories, judges
should be evaluated for good character rather than good decisions. Because
these last two alternatives present unique issues, I consider the particulars of
each after first considering complications that arise with more standard
empirical approaches to the descriptive-normative gap.
General Complications
The process of choosing a normative standard and applying it to judicial
behavior presents predictable complications, some more difficult than others.
First, the researcher must decide whether to take an internal or external
normative perspective. From an internal perspective, the researcher asks the
judge (or otherwise determines, perhaps through the judges opinions) what
goal he or she was trying to achieve, selects the proper evaluative norm given
that goal, and compares behavior to the norm to measure degree of success or
compliance with the norm.13 An external perspective asks what goal the judge
should be trying to achieve and applies norms proper to those goals; no effort
is made to gain judicial input on proper normative standards. An external
perspective may be justified from a metatheoretical stance about the proper
role of judges within the legal hierarchy or from the perspective of interested
parties subject to the authority or influence of the courts (e.g., litigants may
hope that judges decisions are legal, just, and accurate). Or the external
perspective may be chosen simply to test some theory about the degree to
which judicial behavior, whether intentionally or not, accords with particular
normative standards.
The choice of norms is crucial, because different norms will often point
to different conclusions about competence. A judge presented with inadequate
evidence at trial may irrationally convict the truly guilty defendant, while
another judge presented with this inadequate evidence may rationally acquit
the truly guilty defendant. The first judge did a good job under the corre-
spondence standard but a poor job under the coherence standard, and
conversely with respect to the competence of the second judge. Hence, the
choice of normative criteria controls the assessment of the judges competence
on particular tasks.
It may be that the judge was capable of providing the normative response
the researcher is testing for, but the judge had different values or concerns than
the researcher assumed or intended and these other concerns overwhelmed
the judges interest in applying the norm of the researchers choice. In such a
case, it is not that the judge has misunderstood her task or role nor that biases
drove the judges behavior, but rather that there is a mismatch between the
task or role as the researcher perceives it and as the judge perceives it. In some
instances, being perceived as logical, consistent, legally principled, or fair-
minded may be less important to the judge than achieving some other goal
(e.g., moral and public policy concerns may override base rate information in
determining liability14), and, in other instances, the researcher may have
mistaken beliefs about the judges goals or beliefs in light of the larger social
setting. In these latter cases, where the researcher simply misconstrues the
judges beliefs or goals but the judges behavior is logical in light of her true
goals or beliefs, then we cannot say that the behavior is irrational under a
coherence standard if that is our normative perspective of choice.
Unifunctionalist tunnel vision blinds the research community to empirical
and normative boundary conditions on basic effects. Inconvenient though it is,
people are multifunctional entities that demand cumbersomely complex expla-
nations (Tetlock, 2002, p. 469).15 Even within the confines of a theory of the role
of a judge, much room exists for different conceptions of the judges proper
functions. Thus, a narrow view of behavior may reveal departures from the
coherence definition of rationality, but a broader view of the same behavior may
reveal a larger purpose that makes this lack of coherence look instrumentally
rational and that confounds conclusions that the irrationality was due to
cognitive bias.16
Existing discussions of judicial competence typically employ externally
determined measures of competence, rather than internally determined mea-
sures based on the judges self-chosen goals or norms. Because we are usually
interested in how judges fulfill their roles within the larger legal-institutional
setting, it is fair to impose values and goals on the judge that he or she might
not endorse or that might even be contrary to the judges own goals or
valuesalthough sharp disparities between internal and external normative
perspectives should be rare if the judge deems her legal system legitimate
and the researcher accurately perceives the goals of the legal system.17
Nevertheless, a finding of external incompetence does not necessarily mean
that the judge was incompetent from an internal perspective; it only means
that the judge failed to meet an externally imposed normative standard.
Indeed, a judge may be judged externally incompetent precisely because the
judge is quite competent at achieving a goal that is internally desirable but
undesirable from an institutional, external standpoint.
Second, because there may be dispute about the goals that should be
served by certain tasks or because the judge may need to serve multiple goals,
the researcher may need to employ multiple, possibly conflicting normative
criteria or justify her choice of some normative criteria to the exclusion of
others (for more on this topic, see the chapter by Robbennolt, MacCoun, &
Darley in this volume).18 Otherwise, the research may be dismissed as irrelevant
or be disputed on grounds of incompleteness.19 Absent an agreed ranking of
values and norms, however, contending factions may exalt or impugn judicial
competence simply by invoking selectively those norms against which judges

fare better or worse.
Third, we come to what may be the single most difficult problem in
attempting to study compliance with legal norms (and which explains in part
the preference for the two alternative conceptions of judicial competence that
I discuss below): even if there is no dispute over the proper norms, there may
be dispute over, or uncertainty about, how to operationalize these norms for
testing purposes. If efficacy norms of accuracy and party satisfaction are
employed, for example, then the researcher must of course devise a measure
of accuracy and outcome satisfaction. Experiments provide a particularly
attractive setting for addressing this problem, because they allow the
researcher to vary the dimensions of hypothetical or staged cases to examine
whether fact-finder behavior conforms to different formulations of the norms
and allows the specification of true and false results (as, for example, in tests of
false-confession detection).
In some cases, it may be impossible to formulate workable tests of goals,
which may result in the selection of a second-best goal. Indeed, we might
prefer that truth be the ultimate touchstone in fact-finding, but we may have
to settle for a coherence measure of competence over a correspondence
measure because we lack a reliable measure of truth in many real cases. Of
course, while many of the rules of evidence can be seen as rules designed to
lead to accurate decision-making (or risk-of-error allocations, per Stein,
2005), other rules endorse values other than truth and prevent the introduc-
tion of evidence that might improve accuracy. We may thus favor a coherence
standard that asks whether judges follow the rules of evidence because we
believe that these rules are the most feasible and manageable means to truth
approximation in light of epistemic constraints and because of the other
values contrary to truth detection that the trial process must serve. Which
brings us to a more detailed discussion of legal norms.
Legal Norms
Many empirical studies of judicial competence employ perfect lawfulness
as the normative standard: Do extralegal factors affect judicial decisions that
ideally should turn solely on what the law and legally relevant facts of the
case dictate? The numerous studies by political scientists into the correla-
tion of Supreme Court justices political attitudes with the direction of case
votes on judicial decisions employ this normative standard (e.g., Unah &
Hancock, 2006). Most prominently, Segal and Spaeth (2002) contrast
their attitudinal model of Supreme Court judicial decision-making with
what they call the legal model, which is supposed to represent how a
law-bound ruling would be made (through fidelity to text, drafters intent,
or existing precedent), and they find that justices political attitudes relate
significantly to patterns of votes on cases.20 They conclude that the legal
model and its components serve only to rationalize the Courts decisions
and to cloak the reality of the Courts decision-making process (Segal &
Spaeth, 2002, p. 53).
There are considerable limitations to the use of perfect lawfulness as the
normative standard. First, the norm is typically just in the air and does not
serve as a source for direct comparison. In these studies, the normative
position of perfect lawfulness at best serves as the null hypothesis, and any
measured extralegal variable that is statistically related to case outcomes leads
to a rejection of the null hypothesis. In Segal and Spaeths (2002) main tests of
the attitudinal model of judicial decision making, there is no direct comparison
between judicial decisions and the predictions of any specific instantiation of
the legal model. Rather, Segal and Spaeth, after finding that ideology can
explain some significant percentage of variance in judicial decisions, simply
assume that this finding contradicts what a legal model would predict.
Of course, to the extent a legal model would predict absolutely no
influence of justices attitudes, then the legal model is contradicted, but I
am aware of no legal model that would predict that liberals and conservatives
would perfectly agree on the controlling legal norm in all Supreme Court
cases. Indeed, I am aware of no single legal model or set of coherent legal
norms that all conservative scholars and judges would agree on as the correct
approach, nor that liberal elites would similarly agree on, and I suspect that
many politically conservative and politically liberal elites would disagree over
the proper outcome in many Supreme Court cases from even a narrowly self-
interested, purely political perspective.21 In other words, Segal and Spaeths
support for the attitudinal model certainly casts into doubt perfect lawfulness
as an accurate description of judicial behavior, but the degree of deviation
from this vague standard goes unspecified, and, given competition over
what perfect lawfulness requires across cases, interpretations of a particular
outcome as reflecting inappropriate policy preferences or principled disagree-
ments between liberals and conservatives as to what the law requires in
particular cases will often be contestable.22
This last point is not a minor quibble with political science studies of the
Supreme Court. If a conservative judge can choose among legally principled
courses C1, C2, and C3 to reach the conservative result that he politically
prefers, but if the judge feels constrained by legal norms to choose C1 over the
other two options, then this constraint is significant if writing the opinion
pursuant to C1, as opposed to pursuant to C2 or C3, has effects on the lower
courts and later Supreme Court decisions (see Richards & Kritzer, 2002; see
also Friedman, 2006; Hansford & Spriggs, 2006).
Second, because these studies almost invariably test for an extralegal
influence rather than compliance with a legal norm, a finding that the variable
of interest is not significantly related to judicial behavior tells us little about
judicial competence other than that this variable, in this study, did not exert
any influence. It may be the case that judges were acting pursuant to the
normative standard, that the wrong extralegal influences were investigated, or
that some other norm or feature of the institution or situation other than the
particular law of the case dampened individual differences in voting inclina-

tions (as Fischman, 2008, argues with respect to the norm of consensus).23
Third, dependent variables (i.e., the judicial behavior of interest in
extralegal influence studies, such as votes to grant certiorari or votes on the
merits of a case) are often operationalized in categorical terms that may be too
crude to capture the influence of extralegal factors or, if a significant influence
is found, may obscure the real nature of this influence (as Maveety, 2003,
notes, this is a long-standing criticism of political scientists judicial studies).
For instance, judicial outcomes may reflect a wide range of political positions,
from very liberal to very conservative, but Segal and Spaeth (2002, ch. 8)
categorize Supreme Court decisions as liberal or conservative, and the decisions
these researchers make regarding what constitutes a liberal or conservative
decision reflect particularized versions of liberalism and conservatism (e.g.,
pro-federal-government rulings are generally deemed conservative, but of
course some varieties of conservatism would probably not uniformly endorse
such rulings). That is not to say that dichotomous dependent measures are
not appropriate for testing some deviations from legal or other normsjust
that one needs to be sensitive to the loss of sensitivity that accompanies such
choices.
Fourth, many legal norms are process norms as opposed to substantive,
or outcome, norms (e.g., rules of procedure and evidence at the trial level,
or interpretive norms for statutory and constitutional construction and
standards of review at the appellate level). Process norms specify the range of
permissible considerations and at best forbid some categories of outcomes
that could be reached only through reliance on certain sorts of impermissible
evidence or considerations, such as using lotteries or coin flips to assign
responsibility for torts; process norms narrow the range of permissible out-
comes but do not compel particular outcomes. Unless one takes the position
that it is impossible to test compliance with process norms (a proposition that
psychological studies of the rules of evidence have falsified, see, e.g., Wistrich
et al., 2005) or that process norms have no important constraining force (an
empirical claim itself), then the overwhelming focus of political science
studies on case outcomes as opposed to decision processes greatly limits the
utility of these studies of the descriptive-normative gap in judicial behavior
(Richards & Kritzer, 2002).24
A related problem is the requirement that a legal norm lead, a priori, to
definitive predictions for empirical testing purposes. This requirement is, of
course, reasonable from an empirical perspective, but its implementation has
been problematic. As Segal and Spaeth (2002, p. 59) correctly point out, a
legal norm that cannot tell a judge how to act in a particular case cannot be the
subject of empirical testing (nor can it have normative force) (see also Segal &
Spaeth, 1994), but a legal norm need not compel a particular decision on the
merits to have some constraining force (stated counterfactually, would the
same evidence have been considered, would the same rulings on all issues
have been made, or would the decision have been written in the same way had
the legal norm not existed?). How an opinion is written may have more
lasting influence on the behavior of other courts as well as that of lawyers and
parties than the particular outcome of a case (Hansford & Spriggs, 2006); so it
is no small matter if norms constrain opinions without altering outcomes
(Richards & Kritzer, 2002). Whereas legal scholars may give too much
explanatory weight to the reasons stated in opinions, the tendency to ignore
the reasons stated and focus on case outcomes and external motives greatly
restricts the usefulness and acceptance of empirical judicial studies by judges
and lawyers.
Stated another way, if one of the goals of empirical judicial studies is to
predict how judges will act and rule under specified conditions rather than
assess normative compliance (see Leiter, 1997), then disregard of the text of
statutes and cases constitutes a curious omission unless the prediction game is
confined only to how a particular courtmost notably the Supreme Court
will rule on a case once the judges hearing the case are known. Attorneys
cannot play the prediction game in that truncated manner because they must
often advise clients before the ideology of the whole series of judges who may
ultimately hear a case is known. Accordingly, attorneys must look to existing
case law and the text of statutes and regulations to predict how a generic judge
would rule and advise their clients accordingly. Until studies demonstrate
convincingly that looking to these documentary sources of the law is pointless
to predict judicial behavior outside the Supreme Court, attorneys will continue
to do so for lack of any better guidance on predicting how the law will be
applied. Furthermore, when the prediction game is really a multilevel predic-
tion game with tremendous uncertainty (trial judge A will rule this way, but
appellate judges X, Y, & Z are likely to rule that way assuming X, Y, & Z are
assigned to the case. . . ), it is unclear whether any rational attorney, ex ante,
should look more to judicial ideology than the law to advise her client.
Additional Distinctions for Legal Norms as Benchmarks

These concerns suggest that it may be wise to look for alternatives to treating
perfect lawfulness as the null hypothesis and then declaring the legal model
invalid when significant deviations from the null are found. In general, direct
tests for compliance with legal norms, rather than indirect tests using the
perfect lawfulness standard, will provide the information we are most
interested in from an evaluative perspective. The nature of the direct test
depends on the nature of the legal norm being tested. One way (and surely not
the only way) to classify legal norms for purposes of testing is with respect to
their injunctive content: (1) proscription norms: certain kinds of considera-
tions must not be taken into account or certain processes or outcomes cannot
be chosen under specified conditions; (2) permission norms: certain kinds of
considerations may be taken into account or certain processes or outcomes
may be chosen under specified conditions; (3) prescription norms: certain
kinds of considerations must be taken into account or certain processes or
outcomes must be chosen under specified conditions; (4) ordering norms: if

multiple considerations or outcomes are prescribed or permissible, prioritize
the considerations or outcomes in a particular way; (5) decision norms:
decide a case in a particular way in light of permissive and prescriptive
considerations.
Proscription norm tests will look very much like many of the extralegal
influence studies found within psychology that examine whether juries (and
occasionally judges) can disregard inadmissible evidence as required by the
rules of evidence (e.g., Wistrich et al., 2005). Proscription norms can be
found, of course, at the appellate level as well (e.g., rules about the record
on appeal). The difference from the generic perfect lawfulness test is that the
improper extralegal influence is specifically defined by the specific proscription
norm, and the norm typically specifies the conditions under which a particular
consideration/outcome/process is not appropriate.
Permission norms vest judges with discretion to decide the propriety of
certain outcomes or considerations with broad guidance as to the range of
acceptable discretion, and, as such, these norms pose difficult testing pro-
blems for they require an empirical measure of acceptable discretion. One
way to test for compliance with permission norms is to test whether discre-
tion is exercised in a biased fashion. Thus, if judges with sentencing discretion
impose higher sentences to black than white defendants, all else being equal,
then the judges exercise their discretion in a biased way. These tests require,
however, another norm to resolve whether the bias is itself nonnormative, as
with a proscription norm against taking race into account in sentencing. If
younger judges appointed by Democrats grant summary judgment more
frequently than older judges appointed by Republicans in sexual harassment
cases (e.g., Kulik et al., 2003), all else being equal, then it appears that one
set of judges is using its discretion in a biased way, but it is not clear which
direction constitutes the nonnormative response. Another approach to
tests of permission norms would be to have experts agree on the appropriate
range of permissive outcomes in a set of cases and test judges against these
benchmarks.
Prescription norms allow point-specific tests of normative compliance.
This category of tests includes testing for enforcement of bright-line rules,
particularly those triggered by objective conditions such as deadlines. Because
tests of prescription norms will typically yield clear results (assuming the
prescription norm can be operationalized and compliance with it measured
without much controversy), detection of deviation from these norms should
be straightforward.
Ordering norms also allow for straightforward tests in many instances.
Lower court obedience to Supreme Court precedent rather than contrary
circuit precedent and preference orderings for types of evidence in certain
types of disputes (e.g., statutory interpretation cases) will often yield clear
results (of course, there may be dispute about whether the proper ordering
norm is being applied). In other instances, such as implementation of the
Chevron norm of deference to administrative agency interpretations of sta-

tutes where the interpretation is reasonable, will prove more difficult because
implementation of the ordering rule is based on subjective determinations
about a case (on the politics of applying Chevron, see Miles & Sunstein, 2006).
In such instances, we must revert to looking for systematic disparities (tied to
types of cases/values at stake, types of parties, or other situational or judge-specific
factors) or deviations from expert orderings.
Decision norms specify particular outcomes when triggering conditions
are met. The more objective the trigger and the conditions for determining
whether the trigger is pulled, the easier the decision norm will be to test. Some
decision norms function as constraints on the authority of the judge, such as
minimalist or incrementalist norms about case law development or constitu-
tional interpretation, and the focus will be on testing whether judges abide by
such constraints. In tests of norms such as these, the key will be ranking
possible decision outcomes in terms of restraint to monitor for compliance
(or ranking for allegiance to text, intent, or purpose depending on the nature
of the decision norm being tested).
From a psychological perspective, we should expect the greatest deviation
to occur with permission norms (that assign discretion to judges and thus
invite individual differences in application), where triggering conditions for
norms are most ambiguous (and possibility of detection of norm violation
lowest),25 and where accountability is the least (e.g., with supreme courts
reviewing constitutional issues). If the deviation appears to be in one direc-
tion across judges, then we should look for problems with how the norm is
presented to judges (e.g., perhaps the text of a rule contains an unnoticed
confusion) or for some systematic situational influence driving behavior in
that direction. If the deviation appears bimodal (or some other systematic but
not unidirectional pattern), then it is likely that an individual difference
variable or some covarying situational influence is driving the deviations. If
the deviation appears to be random, then it may be that the norm is difficult
to apply in some subset of cases, this subset of cases is randomly distributed
across judges, and, for this subset, some irrelevant situational or individual
difference variable is having an influence that it does not outside this subset of
cases. Random deviation may also represent the operation of a norm that is
hard to implement for some reason (e.g., some complex rules of evidence may
be difficult to implement consistently under the time demands of making
real-time rulings on motions at trial).
Coherence Norms
There are, of course, many coherence norms other than legal norms that we
may want judges to follow. Axioms of rational choice, rules of logic and
probability, hypothesis-testing rules, and Bayes formula for updating beliefs
are the most common coherence norms employed in behavioral decision
theory studies.26 Many legal norms, such as interpretive rules and the rules of
evidence, clearly allow room for other coherence norms to operate. In general,
the more closely a judge follows norms of rationality, logic, and probability,
the more consistent and predictable her behavior should be and the more
likely the judge will maximize the values she deems most important.
In the abstract, the notion that judges and jurors should act rationally
when evaluating the law and evidence in a case is uncontroversial, and
therefore proof from psychology of apparently rampant irrationality under-
standably commands the attention of legal scholars (as Saks & Kidd, 1980,
noted many years ago). In some ways, this evidence should be more troubling
to positive political theorists and others who advocate rational choice models
of judicial decision-making as explanatory models of actual judicial behavior
(e.g., Epstein & Knight, 1998; Schanzenbach & Tiller, 2006) than to scholars
examining judicial behavior for the size of the descriptive-normative gap. For
there are a host of other norms that we may care equally about from an
external perspective that do not require strict compliance with norms of
rationality. Indeed, to the extent that systematic deviations from rationality
are caused by cognitive and motivational mechanisms that have evolved or
are used to serve alternative goals than those served within the closed rationality
system under observation (e.g., efficiency in processing at acceptable levels of
accuracy [Hammond, 2000b] or attributing greater causal responsibility to
persons than rationally justified to encourage care and deter excuse-making
[Tetlock, 2000]), then we should expect deviations from rationality to accom-
pany satisfactory performance under alternative normative standards.
Nevertheless, whether judges, given their experience, training, and the
decision aids available to them, comply with norms of causal hypothesis
testing, probability and rationality is of interest from an external normative
perspective because this information may lead to prescriptive models of
judging. We now know that several predictable judgmental biases can be
debiased (Heath et al., 1998; Larrick, 2004), and we know that the structure of
some environments ameliorates or exacerbates some biases (see Mitchell,
2002b, 2003b). We also know that there are individual differences in ration-
ality, a topic addressed in other chapters in this volume (see generally
Mitchell, 2003b; Stanovich, 1999). Identifying the norms that judges have
the greatest difficulty following and identifying the characteristics of more
and less rational judges may lead to feasible prescriptive models for legal
reform.
Tests of norms of rationality and probability will look like many of the
tests of legal norms, because all are coherence norms. Thus, if a norm of
rationality directs the judge to ignore outcomes and focus only on ex ante
information available to the parties to assess their behavior, we will test
whether the judge can ignore these improper influences (see Rachlinski,
1998). Where the norm directs that certain information should be consid-
ered, such as base rate evidence in assessing the probative value of evidence,
we will test whether the judge properly uses this information. Experimental
tests allow the cleanest tests of judges abilities to engage in analytical thought
and comply with norms of rationality, but testing judges outside their natural
environments may lead us to worry more than we should about the harms
associated with laboratory normative deviations (Gigerenzer, 2006). Thus,
experimental tests should be seen as providing an important first step toward
understanding the rationality of judges and identifying possible normative
deviations, but that step should be followed by studies assessing how institutions
and individual judges adapt, if at all, to correct for such deviations.
Efficacy Norms
If one is interested in judges primarily as one among several interdependent

actors within the governmental division of labor, then efficacy norms may be
of particular interest for evaluative purposes. For instance, if a key role of
judges within the larger institution of government is to induce voluntary
compliance with the law to ease pressure on executive enforcement of the law,
then testing judges for their ability to persuade parties to accept legal decisions
and voluntarily obey the law become important efficacy norms for evaluative
purposes (hence the popularity of measuring perceptions of procedural
justice and its relation to perceptions of the legitimacy of the legal system
and voluntary compliance with the law; Gibson, 2006; Tyler & Huo, 2002).
Likewise, the function posited for constitutional adjudication by the Supreme
Court will determine the efficacy norm one would choose to evaluate con-
stitutional decisions.
The empirical and theoretical importance of efficacy norms is that they
take the focus away from whether the law can constrain judges from turning
their own preferences into law and refocuses attention on how effective judges
are at achieving whatever goal is sought, by whatever means. Thus, a judge
who possesses preternatural lie detection skills (see OSullivan & Ekman,
2005) and employed those skills to achieve highly accurate results in trials,
while completely ignoring the rules of evidence, would be an excellent judge
under a correspondence norm. The results one obtains from an efficacy norm
study will only be as compelling as the argument one advances for the
propriety of that norm, but, given agreement on the norms, efficacy norms
arguably provide the most important information of any of the evaluative
comparisons that can be conducted because such tests directly link performance
to desired outcomes.
Once one satisfies oneself that the efficacy norm chosen is defensible
and worth studying, then one confronts the task of devising a criterion
measure. Once outside experimental settings, where the experimenter con-
trols how the comparison to the norm will be conducted and measured (e.g.,
asking judges to rate known true and false confessions for truthfulness; cf.
Kassin et al., 2005), a backward or known-outcomes approach for testing
compliance with correspondence norms in trial settings may be wisest. In
this approach one identifies cases where the values of the output variable of
interest are known (e.g., innocence or guilt as determined with a high level
of certainty based on DNA evidence or corroborated confessions by others

[e.g., Garrett, 2008], or the outcomes of predictions made previously by
judges), and then one examines the conditions under which judges made
accurate or otherwise efficacious decisions. One must take into account, of
course, that legal norms may constrain judges from giving certain types of
evidence, at certain stages of a case, the weight the evidence might deserve
under a strict correspondence test. Thus, such studies may reveal places
where correspondence and coherence norms work at cross-purposes.
Efficacy norms have been relatively neglected compared to coherence
norms and legal coherence norms in particular. Yet getting things empirically
right is what we often care most about in courts, and at times there will be
objective (or at least intersubjectively reliable) measures of efficacy. Appellate
judges often make predictions about the empirical consequences of choosing
one legal interpretation over another, and the most dire predictions are
typically found in dissents, which makes their predictions testable in some
instances (we should expect dissenters whose parades of horribles never
materialize to claim that conditions changed, making predictions no longer
applicable; Tetlock, 2005). For instance, Albiston and Nielsen (2007) tested
the Supreme Courts empirical prediction that rejection of the catalyst theory
of fee awards in civil rights cases would not discourage the filing of suits by
surveying public interest organizations, and their findings cast doubt on the
accuracy of the Courts prediction. Or we can treat judicial decisions as
natural experiments and examine the effects of these decisions on other
courts, other institutions, and the pattern of lawsuit filings (e.g., examining
the effects changes in the standards for admitting expert evidence; see Vickers,
2005). The tournament approach to judicial competence, with its emphasis
on objective comparisons, may be particularly well suited to testing compliance
with efficacy norms.
Judicial Tournaments and Judicial Character

Recently, researchers have shown interest in going beyond examinations in
which correlations between party affiliation and voting patterns play a central
role, as is so often the case in political science studies of judicial behavior. The
two leading alternative approaches, already mentioned above, involve judicial
tournaments that emphasize multiple performance measures and theories
of good judging that emphasize judicial character rather than judicial perfor-
mance in particular cases. These two very different approaches present their
own difficulties, but each has the potential to push research in new and
important directions.
Choi and Gulati (2004a, 2004b; Gulati & Choi, 2007) recently popularized
the judicial tournament approach as part of a proposal to reform the Supreme
Court appointment process (for discussion of variants on the judicial tournament
idea, see Goldberg, 2005; WERL, 2004). Traditional judicial studies, which
examine justices for ideological voting and typically contrast behavior to perfect
legality (or rather a norm of no ideological voting), often have data that can be
reported in tournament format (e.g., Segal & Spaeth, 2002, report the relative
influence of values on different justices votes), but the emphasis is on testing for
deviations from a specified norm or finding confirmation that some extralegal
factor influences judicial behavior. The emphasis in judicial tournaments is on the
relative ranking itself and, at least in Choi and Gulatis judicial tournaments, on
relative rankings derived from objectively measurable variables. This emphasis on
relative performance gives rise to a fundamental criticism of the tournament
approach: a relative ranking of judges fails to provide compelling evidence of
judicial competence absent some independent normative benchmark (e.g.,
Solum, 2005a): the winner may simply be the best judge in a tournament of
bad judges.
The skeptical response to this criticism is that relative standing on weakly
diagnostic objective measures of performance is the best we can do given the
contested nature of the concept of good judging and given the difficulties of
translating judicial votes and opinions into data. Under this view, any eva-
luative study of judges yields at best a relative ranking of judges according to
the researchers flawed theory or methodology, and hence we avoid at least
the problems of subjective judgment by employing objectively determined
rankings.
The ecumenical response is that relative standing yields useful information
about judicial competence along dimensions that may hold relevance for a
variety of conceptions of good judging, and perhaps provides definitive
information for some norms. A well-theorized selection of objective measures
can provide good, if imperfect, proxies for testing abstract norms of judicial
competence (e.g., some of Solums [2003, 2005a] thin judicial virtues may
be captured roughly by objective measures of ethics compliance, opinion
clarity, educational achievement, and productivity), and a number of objective
measures may approximate widely accepted conceptions of good judging
(e.g., most theorists would place a judge who writes her own opinions
above one who does not; Choi & Gulati, 2005). When the tourneys results
converge across multiple measures designed to get at the same qualities, then
we should have confidence that the tournament meaningfully and reliably
assesses qualities related to differential effort or ability at judging.
Furthermore, understanding how judges compare on objective measures
moves debate away from the mechanics of operationalizing independent and
dependent variables to a debate about the relevance of these and other
measures to notions of judicial competence (Choi and Gulatis tournaments
certainly motivated much fruitful debate in this regard; see Gey & Rossi,
2005). When objective measures identify sets of judges as high versus low
achievers, and alternative approaches to judicial competence yield different
outcomes, then understanding why the divergence occurs can only improve
theorizing about what constitutes good judging and how to identify it
(e.g., Farber, 2005; Goldberg, 2005; WERL, 2004). For instance, on what
objective dimensions do Democrats and Republicans agree, and is such
agreement consistent with the attitudinal model of judges? Along what
objective dimensions other than voting behavior are liberal and conservative
judges predicted to act in an ideological fashion, and do they behave as
predicted (for one example along these lines, see Law, 2005)? And of
course, the judicial tournament need not involve only objective measures.
Many states systematically collect data on judicial performance, including
attorney and juror surveys of judicial demeanor and performance, and this
wealth of data, if made available to empiricists, could be readily adapted for
academic research and judicial tournaments in particular (see Kourlis &
Singer, 2007, on judicial performance evaluation programs).
The relative ranking approach to judicial competence may also proceed
in an atheoretical fashion to explore courts and areas of law for unexpected
patterns. For instance, are there consistent sets of characteristics associated
with judges on particular courts who perform well versus poorly on mundane
measures of productivity, influence, or reversal rates? Judicial tournaments
can identify the high and low achievers on such measures, who can then be
systematically studied for similarities and dissimilarities in their backgrounds
and situational settings that may account for these differences. Areas of law
that provide objectively determinable outcomes allow for the use of judicial
tournaments based on a correspondence norm: for instance, courts that
randomly assign criminal cases, and hence provide a natural control on
factual differences across cases, and that collect data on probation and
parole violations provide a setting for a judicial tournament involving the
ability of judges to discriminate between good and bad risks.
The judicial tournament, in short, presents a host of research opportunities
and should be seen as a necessary, but not sufficient, part of the judicial
competence research repertoire. The relative ease of conducting tournaments
based on objective measures compared to labor-intensive and debatable
coding of opinions and outcomes makes this tool a particularly valuable
one as judicial researchers seek to expand knowledge about judicial competence
beyond the U.S. Supreme Court.
The second popular alternative to traditional political science studies of
judicial competence is to focus on the possession of certain traits or qualities
rather than particular behaviors to determine judicial competence (e.g.,
Farber & Sherry, this volume, ch. 18). This dispositional approach risks
converting intuition into unproven tautology: wise judges will issue wise
decisions (because the desirable trait of wisdom means acting wisely). To
move beyond tautology, some justification needs to be provided for giving a
dispositional theory normative weight. Normative authority could arise from
agreement, as with a thin theory of judicial virtue, in which only uncon-
troversial traits that are desired in all judges are chosen (Solum, 2003). Or its
normative status could arise from showing that the dispositional theory of
judging derives from another theory that already enjoys normative status
(e.g., Solum derives his virtue-centered theory of judging from virtue ethics),
or from empirical studies showing that other types of judges with particular
characteristics perform better and worse on various normative tests (for a
discussion of the concept of expertise and its application to judicial studies,
see the chapter in this volume by Shanteau & Dalgleish, ch. 16). Or the theory
could be justified through empirical observation (which appears to be the
basis for Sherrys [2005, 2006] views on the traits and experiences that
appellate judges should have or should not have; namely, in this last respect,
experience as an academic): if empirical observation reveals that judges who
possess qualities X, Y, and Z reliably produce excellent decisions, then these
qualities should be favored. The problem with this last approach is that it
simply moves the normative question to another point in the equation: we
will need justified normative standards for deciding whether the decisions of
these judges were indeed excellent.27
One interesting feature of character-based approaches to evaluating
judges is that, if one accepts the character traits chosen, then the evaluative
focus shifts to whether judges possess these traits and away from perfor-
mance. The task becomes measuring the traits of existing and potential judges
to evaluate them for good and bad judicial character.
From an empirical perspective, the measurement problems associated
with dispositional approaches to judicial evaluation are severe for judicial
traits of the kind Solum advances (abstract virtues and vices),28 but much less
so for judicial traits of the kind Sherry advances (at least for the experience-
based factors she argues for as proxies for the more abstract character traits
she ultimately favors). Consider a popular judicial trait within some disposi-
tional approaches: good judgment, which may go by a variety of labels, such
as wisdom or pragmatism (see Sherry, 2003).29 There is no single theory of
judicial pragmatism or judicial wisdom, and certainly no well-validated
measure of pragmatism or judicial wisdom. It may be difficult to distinguish
a pragmatic from a textualist outcome in a particular case (because, as Posner
[2005] notes, at times the pragmatic judge will adopt the textualist outcome),
but it will be near to impossible to distinguish a pragmatist from a textualist
from biographical material about judges (though Solum [2005b] disputes this
with respect to his proposed virtues and vices) (and, in any event, clinical
assessments of character are considerably less reliable than those based on
validated assessment tools, none of which exist for pragmatism). If a judges
opinions serve as the source for making determinations about the judges
pragmatic character, then we return to our tautology problem.30
Another interesting feature of these theories, and the most problematic
feature from a social-psychological perspective, is the assumption that judges
with ideal judicial character are more likely to follow the law strictly, wisely
interpret and apply the law, or, perhaps, do justice (whatever those terms
imply) than those without this character. While it is certainly the case that
individuals differ in important ways that may relate to judicial behavior (e.g.,
Tetlock et al., 1985), it is also the case that these individual difference variables
interact with situational variables (such as being in the majority or minority;

Gruenfeld, 1995) to produce judicial outcomes. Neither a strong personality-
based nor a strong situationist view of judging (or of behavior in general)
can be supported by the evidence from the long debate about the primacy of
the situation versus the person; rather, interactionism is the proper view (see
Funder, 1999, 2001). A salutary by-product of empirical tests of dispositional
theories of good judging might well be a greater appreciation for the role of
interactions between personality and situational variables in achieving nor-
mative compliance.
Thus, while I am optimistic that a dispositional approach to evaluating
judges will lead to new insights because researchers taking this approach will
likely ask different questions than those posed in standard norm-based
evaluative approaches (or at least employ new and different norms), I am
not optimistic that, ultimately, dispositional approaches to good judging will
avoid the usual problems encountered in attempts to study the descriptive-
normative gap. For if tests of judicial character remain unmoored from actual
decisions, critics will eventually dismiss the dispositional claims as irrelevant
or invalid. In the end, we evaluate judges to learn how well they are doing their
jobs, however doing their jobs may be defined, and so we must focus on
what judges actually do in relation to what we wish they would do.
Conclusion
Empirical evaluative studies of judges have historically been dominated by a
focus on the U.S. Supreme Court and a narrow debate about the degree to
which Supreme Court justices allow their personal preferences or political
values to influence their decisions. I have argued that these studies fail to
provide compelling evidence on the law versus politics debate as applied to
the Supreme Court, much less other courts, and these studies distract attention
from a host of other possible normative comparisons that judicial scholars
should care about, and from alternative empirical approaches that may be
more easily, or at least less controversially, implemented.
A diverse and robust literature on the normative-descriptive gap in
judgment and decision-making exists within psychology, and this literature
contains much healthy debate about the size, contours, and meaning of this
gap (see, e.g., Krueger & Funder, 2004). The insights and methods of this
literature are just starting to filter into empirical studies of judicial compe-
tence, and by explicating some of the norms utilized in this literature and
some of the methodological issues that evaluative studies present, this chapter
is meant to accelerate the rate of interdisciplinary exchange and encourage
diversification in evaluative studies. Trite though it now is to suggest that
judicial researchers should let a thousand empirical flowers bloom (given the
long history of invocation of this dictum within judicial studies; Maveety,
2003), this recommendation may have no more apt application within the
field than to studies of judicial competence. If the attitudinalist orientation to

judicial studies continues to dominate empirical research in this field, then
outside interest in the field will likely remain limited and a host of important
dimensions of judicial competence with possible prescriptive implications
will continue to be unexplored.
Notes
1. I use the term competence to refer to a judges (degree of) compliance with a
normative standard (see Stein, 1996, for a similar usage). The term is meant to
have the same meaning here as the terms rational or irrational do in
psychological studies of the concordance of actual judgment and choice with
norms of rationality.
2. My primary and modest goal is to provide a framework for organizing empirical
questions about judicial competence, but I do venture some predictions and offer
some critical assessments of existing evaluative studies. For an extended discus-
sion of the difficulty in operationalizing good judgment for assessing the quality
of real world judgments, see chapter 1 in Tetlock (2005).
3. Within psychology, prescriptive models are distinguished from normative and
descriptive models of judgment and decision making. Normative models
supply the standards or rules for evaluating judgments; descriptive models
compare judgments to the normative models; prescriptive models specify the
conditions under which judgments can be brought more into line with the
normative models (see Baron, 2004).
4. These random deviations constitute mere performance errors, whereas sys-
tematic deviations would suggest an underlying incompetence (Stein, 1996).
Both may need attention, but likely different sorts of attention.
5. Of course, not all political scientists embrace the strict attitudinalist view and
very few legal scholars embrace a strict legalistic view. Segal and Spaeths
influential framing of the debate in this way, however, surely has framed the
larger debate, for better or worse, and their strongly worded skeptical view of
the power of the law to constrain (e.g., Segal & Spaeth, 2002), or the modified
strategic view in which other political institutions but not the law constrain
judges (e.g., Epstein & Knight, 1998), now regularly serves as the starting
position within empirical quantitative studies of judicial behavior.
6. Segal and Spaeth (2002, pp. 289292) assimilate Casey to their attitudinal model
and thus do not treat it as dissonant evidence.
7. Others may parse the norms differently; this particular categorization is used for
organizational purposes only and holds no special importance in the chapter.
8. The few works that do evaluate judges in relation to social norms usually involve
norms of secrecy and civility (e.g., Gaffney, 1994). Baum (2006) argues for greater
attention to collegiality as an important judicial motive, which might lead to a
greater emphasis on social norms and the degree to which judges comply with
them. Fischman (2008) presents interesting evidence that a norm of consensus
may significantly moderate individual judges policy preferences. For further
discussion of collegiality as an important influence on appellate judges behavior,
see the chapter by Martinek in this volume (ch. 5).
9. See, e.g., FCC v. Beach Communications, Inc., 508 U.S. 307, 315 (1993): On
rational-basis review, a classification in a statute such as the Cable Act comes to
us bearing a strong presumption of validity, and those attacking the rationality
of the legislative classification have the burden to negative every conceivable
basis which might support it. Moreover, because we never require a legislature
to articulate its reasons for enacting a statute, it is entirely irrelevant for
constitutional purposes whether the conceived reason for the challenged dis-
tinction actually motivated the legislature (citations omitted). Whether
rationality review actually operates in this minimal, deferential form is the
subject of debate (see Farrell, 1999), but the standard does serve as a normative
baseline.
10. Technically, legal norms are a subcategory of coherence norms because they
constitute a closed system of rules to be followed without regard to some
independent measure of the quality of system outputs. The goals are deter-
mined by the body that promulgates the legal norms, and the legal norms are
designed to produce outputs in line with the desired goals. Legal norms are thus
analogous to scientific norms: scientific principles for conducting empirical
research, if followed, should lead to reliable empirical knowledge; the scientific
status of research is determined by its fidelity to scientific principles rather than
the reliability or truth of the knowledge produced (Kitcher, 1993). Likewise,
fidelity to legal norms should lead to the kinds of outcomes lawmakers desire. I
treat legal norms separately from other coherence norms because of their
special importance for judicial evaluations.
11. Alexander George (1980) and Irving Janis (1982) developed their important
models of good political decision-making through historical studies of flawed
and successful executive decision making.
12. In addition to moving selection away from political litmus tests and irresolvable
debates about merit, Choi and Gulati (2004b) also hypothesize that tourna-
ments with implications for selections for higher courts may provide positive
incentives to judges.
13. We may question the sincerity of judges stated goals, and the possibility of
insincerity does complicate the interpretation of results obtained from an
internal perspective: if empirical tests reveal a large gap between actual and
ideal behavior under the judge-chosen norm, then the gap may be due to
insincerity or incompetence.
14. Windschitl and Wells make this argument:
Consider . . . the general implication of a system of justice that could
find people liable, based on the balance of probability that derives
merely from membership in a group or class. Would it be good social
or legal policy, for instance, to consider a man to have harassed female
coworkers merely because he is a member of a corporate organization
in which 80% of the males have harassed female coworkers. Our point
is that it would be inappropriate to use such base rates for these types
of judgments. (Windschitl & Wells, 1996, p. 41)
15. As an alternative way to understand the possible problem of confounding
between seemingly biased behavior and underlying values and beliefs, consider
how the interplay between cognitive bias and different aversions to the types of
verdict errors may play out at trial. Imagine a criminal case with Judge 1 who
considers conviction of an innocent person to be a mortal error that must be

avoided at all costs and Judge 2 who has aversions of equal magnitude to the
errors of convicting an innocent person and failing to convict a guilty person.
Imagine further that Judge 1 rationally processes all evidence but, at the end of
the trial, Judge 1s judgment of the probability that the defendant committed
the crime does not reach the high level of subjective certainty needed by this
judge to convict. Judge 2, on the other hand, irrationally processes the evidence
and acquits due to this biased processing of evidence. A focus only on the end
resulttwo votes for not guiltywill obscure the very different thought pro-
cesses and error aversions or values that led to the same votes.
16. With surprising frequency within behavioral decision theory, researchers simply
assume that study participants would endorse a particular goal with a particular
set of normative standards, without justifying this external normative perspec-
tive. Which may lead to premature judgments of incompetence:
Even when people appear to be making systematically biased
judgments or irrational decisions, it is likely that they are trying to
solve some problem or achieve some goal to the best of their abilities.
The behavioral researcher is well advised to look carefully at his or her
research participants behavior, beliefs, and goals to discern the
method in the apparent madness. (Hastie, 2001, p. 659)
17. If the judge takes, that is, something like Harts (1961) internal point of view
and feels obliged to obey and uphold the law. That does not mean the judge will
do so with complete success, given psychological limitations as well as eviden-
tial and normative ambiguities, but the judge internalizes the legal systems
rules and tries to follow them.
18. Or it may be that different legal systems embrace different goals, and it may be
impracticable to consider all of the different goals within a single study if each
goal requires a variation in design to test for competence.
19. For instance, Feigenson (2003) criticized the Sunstein et al. (2002) study of
punitive damage awards (mock jurors served as the primary subjects, but one
study in the collection examined judges) for its emphasis on optimal deterrence
as the goal of punitive damages to the exclusion of a retribution as a goal.
20. Given the public policy issues often at stake in appellate cases, particularly
constitutional cases, the emphasis within political science studies of judging on
personal policy preferences as a likely motivating factor in how cases are
decided is quite understandable. But political ideology or attitudinal orienta-
tion, as presently measured by political scientists (typically from the party of the
president or the senators in whose state the judicial vacancy is being filled; see
Sisk & Heise, 2005) is unlikely to capture other important, enduring extralegal
influences. Basic assumptions about human behavior and the role of internal
and external forces in determining behavior, about the ability of the market to
correct inequalities and errors of judgment and choice, and about the harm of
different types of errors may exert much more pervasive influence on judicial
decisions than policy attitudes (see Tetlock & Mitchell, Forthcoming), though
no doubt there will be a correlation between conservatism-liberalism and
various assumptions about human behavior and the power of markets. From
a political psychology standpoint, we should focus not simply on political
values but also causal beliefs as determinants of public policy positions, not to
mention a host of more mundane motives and influences (see Baum, 2006;
Schauer, 2000). Of course, how to measure these other extralegal factors can
create large new problems.
21. To the extent that ideology and normative perspectives are confounded (e.g., if
conservatism is positively correlated with acceptance of textualism as a legal
process norm and liberalism is positively correlated with purposivism as a legal
process norm), the correlation between outcomes and ideology does not compel
the conclusion that legal norms fail to constrain decisions. Correlation cannot
establish causation: the conservative may conveniently favor textualism to
advance desired policy ends, or the conservative may endorse textualism as part
of a sincere set of beliefs about the proper role of the judiciary; partial correlations
can suggest the proper causal model but cannot establish that ideology causes a
strategic choice of normative perspectives.
22. A direct comparison would compare actual votes to the votes predicted by a
legal norm. Very few studies undertake this direct comparison, perhaps because
it is very difficult to operationalize a quantitative empirical test with legal
variables (Cross, 2004, p. 31). For instance, Spaeth and Segal (1999) have only
directly tested compliance with the follow precedent norm within their
simplified legal model.
23. As I discuss below, when testing legal proscription norms, which forbid con-
sideration of some specific factor(s), a finding of no influence from these factors
may vindicate the judges conduct relative to the legal norm. But tests for
extralegal influences will often not provide a direct test of compliance with
other types of norms, as I discuss below.
24. In some domains, process norms dominate legal scholars normative discus-
sions. For instance, process norms dominate discussions of constitutional
interpretation (i.e., norms for how to make constitutional decisions rather
than norms about the specific decisions that should be made), making their
relative neglect within empirical studies all the more unfortunate given the great
interest in constitutional cases within judicial behavior studies.
25. Arguably the central mediating variable in studies of extralegal influences is the
degree of discretion or subjectivity permitted or perceived with respect to any
given judgment or decision, yet this variable typically goes unmeasured in
judicial studies.
26. Although probability and hypothesis-testing rules ultimately serve the purpose of
making accurate and reliable judgments, they are still types of coherence rather
than correspondence norms, just like the rules of logic or rationality, because they
do not guarantee correspondence to reality. If one complies with rules of prob-
ability and hypothesis testing, one is more likely to reach reliable conclusions
about probabilistic and causal relations, but compliance is measured by fidelity to
the rules rather than the correspondence of judgments with reality.
27. Or we might ask what kinds of judges seem to act in ways that are contrary to
existing normative theories, using the character of these judges as negative
evidence. For instance, the attitudinal model research would suggest that
judges who hold less extreme ideological views should rule very differently
than ideologues, and, if moderates are more constrained by the law (i.e., if
their decisions are less predicted by ideology), then political moderation should
be favored to the extent legal constraint is desired.
28. Solum accepts that virtues of the kind he advocates are not easily quantified
(Solum, 2005a). He offers templates of the judicial qualities to look for (Solum,
1988, 2003) and offers some specific ideas on how to screen out those with
judicial vices (Solum, 2005a). But on the more abstract judicial virtues, Solum
offers little specific guidance on how exactly to identify these qualities in
particular judicial candidates and is largely content to rely on recommendations
of persons who know the judges, with those recommendations proceeding from
vague standards.
29. Posner (1995b) provides one of the more explicit defenses of pragmatism, even
if the content remains somewhat vague, though Posner seems to argue less that
some are predisposed to be pragmatic judges and more that all judges should
strive to be pragmatists.
30. The point here is not that looking to opinions to find pragmatic judges is
wrongheaded; the point is that the character of pragmatism is, at this point,
doing no independent work except serving as a label for how this judge ruled. It
will be impossible to identify pragmatic judges and predict their rulings if we
can only identify pragmatic judges from their rulings. In this approach, prag-
matism may be seen as a proper norm for judging judicial acts, but it is now just
another of our normative standards rather than part of some character-based
theory of good judging. Of course, this may be the sense most advocates of
pragmatic judging intend; few scholars advance as explicitly as Solum does a
dispositional theory of good judging.
15
Defining Good Judging
Andrew J. Wistrich
When Supreme Court Justice Oliver Wendell Holmes Jr. was quite old, he
took a trip by train. During the trip, a conductor asked him for his ticket.
After unsuccessfully searching through one pocket after another, Holmes
became flustered. The conductor recognized the famous jurist and tried to
reassure him: You dont need to find your ticket now, Mr. Holmes. Im sure
you bought one. Just send it to the railroad office when it turns up. Holmes
replied with impatience: The problem is not that I cant find my ticket, the
problem is I dont know where Im going!1
Some have suggested that we must define good judging before we begin
studying, evaluating, and reforming judicial decision making.2 That plausible
suggestion flows from the same intuition as the one expressed in the story about
Holmes; that is, before we embark on a journey, it would be nice to know where
we are going. While I agree that it might be nice, I am not sure that it is possible,
or necessary, or even that it would be helpful. Therefore, when we are thinking
about what research should be done concerning judges and judicial decision
making during the next decade, and what improvements to our justice system
might result from that research, perhaps what seems like the most logical place
to start is actually not the best place to start. In fact, beginning by trying to define
good judging might delay our departure and lead us in the wrong direction.
Attempts to Define Good Judging

What is good judging? Attempts to formulate a definition have been pursued
for millennia. Since ancient times, philosophers, theologians, judges, lawyers,
249
politicians, social scientists, law professors, bar associations, and judicial

nominating commissions have wrestled with the question of what makes a
judge good. Many of their definitions have been intriguing, and some have
been enlightening, but none has provided a fully satisfactory answer. Trying
to formulate a meaningful definition of good judging is like trying to find the
end of a rainbow; for all of our efforts, we never seem to get there. Reviewing
just a few of the many attempts illustrates the point.
Early efforts are a convenient place to start. In about 1500 B.C., Pharaoh
Thutmose III of Egypt issued the following instructions to Chief Justice
Rekhmire:
Mayest thou see to it for thyself, to do everything after that which is

in accordance with law; to do everything according to the right
thereof. . . .
It is an abomination of the god to show partiality. This is the
teaching: thou shalt do the like, shalt regard him who is known to
thee like him who is unknown to thee, and him who is near. . . like
him who is far. . . .
Be not enraged toward a man unjustly, but be thou enraged
concerning that about which one should be enraged. (quoted in
Breasted, 2001, p. 269)
The Old Testament, dating back to the twelfth century B.C., teaches
judges, Thou shalt not wrest judgment; thou shalt not respect persons,
neither take a gift: for a gift doth blind the eyes of the wise, and pervert the
words of the righteous (Deuteronomy 16:19).
Several hundred years later, the Greek philosopher Socrates said, Four
things belong to a judge: To hear courteously, to answer wisely, to consider
soberly and to decide impartially(Sampson, 2004). Writing at about the
same time, and in the same place, Plato argued that the good judge must
not be young but old (Plato, 2006, bk. 2 div. 409).
St. Augustine, who served not only as Bishop of Hippo, but also as a civil
and ecclesiastical judge during the 4th and 5th centuries A.D. (Schaff, 1954),
said that [a] good judge does nothing according to his private opinion but
pronounces sentence according to the law and the right.3
Not surprisingly, many definitions of good judging have been authored
by judges. Often their definitions consist of a mixture of personal qualities a
judge should possess and ways in which a judge should act or behave while
doing his job.
In 1612, Sir Francis Bacon, philosopher, statesman, and Lord Chancellor
of England, wrote: Judges ought to be more learned than witty, more
reverend than plausible, and more advised than confident. Above all things,
integrity is their portion and proper virtue. . . . Let judges also remember that
Solomons throne was supported by lions on both sides: let them be lions, but
yet lions under the throne (Montagu, 1850, pp. 5859).
Defining Good Judging 251
An especially notable effort by a judge dates from the mid-1600s. It was

authored by Sir Matthew Hale, Lord Chief Justice of England, and is remark-
able for its detail and down-to-earth practicality:
LORD HALES RULES FOR HIS JUDICIAL GUIDANCE:
Things Necessary To Be Continually Had In Remembrance
...
4. That in the execution of justice I carefully lay aside my own
passions, and not give way to them, however provoked.
5. That I be wholly intent upon the business I am about, remitting
all other cares and thoughts as unseasonable, and interruptions.
6. That I suffer not myself to be prepossessed with any judgment
at all, till the whole business and both parties be heard.
7. That I never engage myself in the beginning of a cause, but
reserve myself unprejudiced till the whole be heard.
8. That in business capital, though my nature prompt me to pity,
yet to consider there is also pity due to the country.
9. That I be not too rigid in matters purely conscientious, where
all the harm is diversity of judgment.
10. That I be not biased with compassion to the poor or favor to
the rich, in point of justice.
11. That popular or court applause, or distaste have no influence
upon any thing I do, in point of distribution of justice.
12. Not to be solicitous what men will say or think, so long as I
keep myself exactly according to the rules of justice.
13. If in criminals it be measuring cast, to incline to mercy and
acquittal.
14. In criminals, that consist merely in words where no more
harm ensues, moderation is no injustice.
15. In criminals of blood, if the fact be evident, severity is justice.
16. To abhor all private solicitations, of what kind whatsoever,
and by whomever, in matters depending.
...
18. To be short and sparing at meals, that I may be fitter for
business.4
In more recent times, several judges have compiled shorter lists of the
qualities judges should attempt to display while doing their jobs. One
thoughtful attempt offered ten commandments for new judges: (1) be
kind; (2) be patient; (3) be dignified; (4) dont take yourself too
seriously; (5) a lazy judge is a poor judge; (6) dont fear reversal; (7)
there are no unimportant cases; (8) be prompt; (9) common sense; and
(10) pray for divine guidance (Devitt, 1979).
Twenty years later, another judge offered his own ten-part definition of
what it means to be a good judge:
1. Being a judge means having self-confidence without conceit,

decisiveness without arrogance, and passion without
pretension.
2. Being a judge means using common sense as well as the law to
handle problems.
3. Being a judge means being temperate under circumstances
that at times would try the patience of the most serene.
4. Being a judge means being civil to those who are uncivil.
5. Being a judge means using awesome judicial power sparingly
and with restraint.
6. Being a judge means accepting criticism, justified or
unjustified, without always being able to respond.
7. Being a judge means not being afraid to make mistakes.
8. Being a judge means having confidence in the system, that
mistakes can be corrected, and that justice is attainable.
9. Being a judge means recognizing that the great principles
upon which this country was founded and endures apply not
just to the best of us, not just to the worst of us, but to all of us.
10. Being a judge means accepting the responsibility to represent
the justice system at your very best to exhibit patience,
tolerance, and understanding (Nadeau, 2000, p. 35).
One judge approached the question from a different perspective. He
essentially asked: What would a judge like to have written on her tombstone?
His answer was: An appropriate epitaph for a trial judge is this: this judge
exercised common sense, was slow to anger, rich in courtesy and under-
standing; unquestionably fair but firm; and steadfastly encouraged the search
for truth (Weber, 1979, p. 11).
Many of these lists of the qualities a good judge should possess or exhibit
are aspirational. They sound more like the Boy Scout oath than a set of criteria
that actually would be useful to reform-minded judges, law professors, and
social scientists. They also set an unrealistically high bar against which to
measure judicial performance.5 No wonder one judge observed that [w]e
shall find the ideal judge if, as and when there is combined in a single person
the strength of Samson, the patience of Job, the wisdom of Solomon, the
statesmanship of Marshall, the economic foresight of Mansfield, the political
acumen of Machiavelli, the trenchant pen of Holmes and the nervous system
of William J. Bryan (Dobbie, 1951, p. 477).
Not surprisingly, appellate judges trying to define good judging have
focused on the qualities or behaviors good appellate judges should possess or
display. One offered the following thoughts about the characteristics that
distinguish good appellate judges from bad or mediocre ones: [A]n ideal
appellate judge should possess the following six characteristics: the quality
of being fair, just and impartial; the twin qualities of devotion and decisive-
ness; the quality of clarity of thought and expression; the quality of being
professionally literate; the quality of institutional fidelity; and the quality of

political responsibility (Aldisert, 1982, p. 29).
Commenting on the especially weighty responsibility borne by Supreme
Court justices relative to appellate court judges, Justice Felix Frankfurter said
that in choosing the former, what is essential is that you get men who bring to
their task, first and foremost, humility and an understanding of the range of the
problems and of their own inadequacy in dealing with them; disinterestedness,
allegiance to nothing except the search, amid tangled words, amid limited
insights, loyalty and allegiance to nothing except the effort to find their path
through precedent, through policy, through history, through their own gifts of
insight to the best judgment that poor fallible creatures can arrive at in that
most difficult of all tasks, the adjudication between man and man, between
man and state, through reason called law (Frankfurter, 1953, p. 95).
Academics, especially law professors, also have tried their hand at the
task. One law professor made the following perceptive observations:
Most of the cases, and thus most of the judicial work, involve trials;
either criminal prosecutions or civil disputes. There are a few
important skills which are required of the judges dealing with such
trials. First, and often overlooked, they must be good triers of fact.
They must know how to conduct the trial efficiently but with respect
and courtesy, so that the parties will have a sense that justice is being
done. They must have the legal competence to identify the laws
relevant to the facts and the basic ways of legal reasoning and
interpretation. They must have a basic sense of justice and a measure
of common sense which will check the applications of these norms.
They must act, and seem to be acting, impartially and without bias or
prejudice to either party. They must also act, and seem to be acting,
in obedience to the laws of the land. They must have the ability to
make decisions with caution and good judgment but without delay,
so that legal justice will be as speedy as it can be, on the one hand, but
that the parties will not feel that the judge came to the trial with a
closed mind about the case, on the other hand. Finally, judges must
know how to justify their decisions in a way that will be persuasive
both to the parties and to the legal community. (Gavison, 1988, pp.
16231624)
Another law professor, now also a federal appellate judge, has suggested
that good judging may best be defined as conformity to widely endorsed
conceptions of the judicial role. In his view,
To regard oneself and be regarded by others, especially ones peers as a
good judge requires conformity to the accepted norms of judging.
One cannot be regarded as a good judge if one takes bribes, decides
cases by flipping a coin, falls asleep in the courtroom, ignores legal
doctrine, cannot make up ones mind, bases decisions on the personal
attractiveness or unattractiveness of the litigants or their lawyers, or

decides cases on the basis of politics. (Posner, 2008, p. 61)
Some scholars have attempted to begin by defining bad judging, and then
back into a definition of good judging. Professor Lawrence Solum (2003) has
employed this strategy. He lists as judicial vices: (1) corruption; (2) civic
cowardice; (3) bad temper; (4) incompetence; and (5) foolishness. He lists as
judicial virtues: (1) judicial temperance; (2) judicial courage; (3) judicial
temperament; (4) judicial intelligence; and (5) judicial wisdom. He then
adds a sixth judicial virtue, justice, of which he identifies three aspects:
(1) judicial impartiality; (2) judicial integrity; and (3) legal vision.6
Two law professors approached the problem from an unconventional
point of view. Eschewing many of the typical criteria of good judging, they
instead chose to emphasize the human relationships constructed and
implicated by the act of judgment (Minow & Spelman, 1989, p. 59). Their
definition contains the following elements:
(1) The judge should try to take the perspective of all parties before the
court prior to reaching a decision.
(2) The judge should try to remain open to the newness of each case,
even if it resembles previous ones, while also subjecting new
understandings to scrutiny through comparison with past
experiences.
(3) The judge should not disguise how he or she actually reached the
decision, and should explain the decision not only through post-
hoc justifications but also with reference to intuitions and reasons
for selecting one principled justification over other possible ones.
(4) The judge should confront the difficulty of rejecting the arguments
of a party by trying to develop reasons that would persuade that
party or explain the result in terms that party would concede are fair.
(5) The judge should acknowledge what it feels like to have power over
the lives of others in the act of judgment, and, if the judge does not
experience such power, the judge should reflect on why, despite the
actual effects of the decisions he or she will make, he or she
experiences not having much power.
Codes of judicial ethics are a rich source of guidance about what con-
stitutes good, or perhaps more accurately, bad, judging (e.g., Model Code of
Jud. Conduct, 2007; Miller, 2004). Here is one early example from the ABA
canons of Judicial Ethics:
In every particular his conduct should be above reproach. He should
be conscientious, studious, thorough, courteous, patient, punctual,
just, impartial, fearless of public clamor, regardless of public praise,
and indifferent to private political or partisan influences; he should
administer justice according to law, and deal with his appointments
as a public trust; he should not allow other affairs or his private
interests to interfere with the prompt and proper performance of his

judicial duties, nor should he administer the office for the purpose of
advancing his personal ambitions or increasing his popularity.
(Dobie, 1951, p. 477)
More recently, the American Bar Association (2005) published a set of
guidelines by which the performance of judges could be evaluated. Those
guidelines, which are reproduced in the Appendix at the end of this chapter,
are remarkable for their thoroughness. They identify 31 separate dimensions
on which judicial performance might be measured. While portions of the
ABA guidelines focus on adjudication and overlap with some of the earlier
definitions described or quoted above, they also include several elements that
earlier definitions do not. These include the precepts that a good judge
promote[s] public understanding of . . . the courts, utilizes hiring and
promotion practices that will ensure a diverse pool of applicants for court
employment, and ensure[s] that disabilities and linguistic and cultural
differences do not limit access to the justice system (sections 5-4.6, 5-5.9,
5-5.10). The ABA guidelines, then, extend the definition of good judging into
an important but often neglected realm, albeit one that is only tangentially
related to the judges disposition of cases or behavior in court, namely,
judicial administration. On the other hand, for all of their detail, even the
ABA guidelines are somewhat vague, sometimes exhorting judges to be
dignified, courteous, or patient, and often stating only that a judge should
perform a particular task or pursue a particular goal appropriately.
One scholar tried an empirical approach to defining a good judge
(Rosenberg, 1966). He asked 144 trial judges of varying experience and back-
grounds to answer the following question: What qualities best equip a lawyer
to become a trial judge? He gave each judge a list of 23 possible attributes
compiled from the preexisting literature on judicial qualifications and a variety
of other sources, and then asked them to rate their importance. Those the
judges rated as highest in importance were: (1) moral courage; (2) decisive-
ness; (3) reputation for fairness and uprightness; (4) patience; (5) good health,
physical and mental; (6) consideration for others. Those the judges rated as
lowest in importance were: (23) past honorable partisan political activity;
(22) higher earnings in practice than as a judge; (21) active in civic and com-
munity affairs; (20) experience in supervision of subordinates; (19) well above
average law school record; (18) active in professional associations and work.
Modern empirical researchers arguably are more sophisticated. They use
computerized citation analysis to determine which appellate judges opinions
are cited most frequently and by which courts or by which scholars (Choi &
Gulati, 2004a, 2004b; Gulati & Sanchez, 2002). This taps into a market
reflecting how judges are viewed by their peers or by those who study judges.
Good judges are defined as those whose opinions are cited most often or are
included in the leading law school casebooks. In addition to measuring the
quality of a judges opinions, other criteria, such as the amount of effort a
judge exerts in approaching her job and the willingness to decide cases
impartially and independent of political ideology, are considered (Choi &
Gulati, 2004b).
Whatever its value with respect to appellate judges, however, citation
analysis is of little use with respect to trial judges. In most states, trial court
decisions are not published. In the few states where they are published, or in
the federal courts, published decisions by trial courts are usually of negligible
importance as precedent (except as to issues that are rarely litigated on
appeal), and most are seldom cited.
The results of these laudable efforts to define good judging make inter-
esting reading. They also possess some practical value. Reviewing them occa-
sionally can remind judges of the abstract ideals to which they should aspire.
Some of the definitions conceivably could be used as a crude measuring stick to
assess the quality of judicial performance. As the basis for scholarly research or
judicial reform, however, they are inadequate (Frankel, 1976).
While a handful of traits have persisted over the centuries and appear
to be universally accepted as desirable, no clear consensus has emerged
regarding the criteria that should be employed to determine whether
someone is a good judge or whether a particular decision embodies good
judging (Fennell, 1999). Moreover, widely accepted criteria such as integrity,
impartiality, dignity, and patience are subjective or immeasurable, making it
difficult to determine whether a judge did or did not display them in making a
particular decision (Minow & Spelman, 1989, p. 50).
Modern definitions of good judging are more detailed than their pre-
decessors, and tend to emphasize more practical judicial qualities or goals.
They also acknowledge that judges do more than simply decide cases, and that
judges ancillary responsibilities should be reflected in the definition of good
judging as well. Although those newer definitions may be slightly more
realistic when compared to the lofty and often superhuman standards of
good judging emphasized in the past, they still contain many terms that are
vague or subjective (Schaefer, 1960).
The failure of so many sincere efforts by so many capable people leads me
to believe that defining good judging in the abstract is infeasible. Others have
reached the same conclusion (e.g., Posner, 2008, p. 12; Champagne, 1986, pp.
104, 110; Farthing-Capowich, 1985, p. 23; Dubois, 1980, p. 17). Gregory
Mitchells understandable failure to offer such a definition in his fine essay
on evaluating judges in this volume (ch. 14), coupled with his ingenious and
exhaustive account of the difficulty and complexity of doing so, only confirms
that attempting to define good judging in the abstract is a fools errand.
Conceptual Difficulties
In addition to the difficulty of crafting a meaningful definition of good
judging, attempting to do so creates conceptual problems. One such problem
is that beginning with a definition of a good judge or good judging causes us

to focus too much on the qualities or characteristics of judges, and not
enough on the nature of the context or environment in which they make
decisions. People systematically tend to give too much weight to the indivi-
dual, and too little weight to the situation, when assessing causation or
responsibility. Psychologists call this widespread phenomenon the funda-
mental attribution error (Plous, 1993, pp. 180181). Attempts to define
good judging exacerbate this tendency by encouraging us to give even more
weight to judges as individuals, and even less weight to the situations in which
judges act, than we ordinarily would. In fact, the emphasis on defining what
makes a judge good may itself be a product of cognitive bias.
The context in which judges decide matters. As an example, courts
depend on the adversary system. When it works well, the adversary system
can help judges avoid mistakes and ensure that they are confronted with both
sides of the story before issuing a ruling. But it is not always effective.
Sometimes one lawyer misses something important. Occasionally both law-
yers do a poor job, and an overworked judge fails to detect their mistake.
When the lawyers slip, it is not just the litigants, but also the judge, who falls
(Lasky, 1965, p. 680; Brandeis, 1916, p. 470). An important difference
between a good judge and a bad judge might be that the former benefited
from having better lawyers appear before him than did the latter. This
insightthat actors operate within a system, and that errors often are the
result of a confluence of factors of which the quality of the actors personal
performance is only oneis crucial to achieving meaningful reform (Bates &
Gawande, 2000).
A second problem is that any single definitionassuming that a work-
able definition could ever be formulatedinevitably would foster the mis-
guided notion that there can be a one size fits all standard of good judging.
Not all judges do the same work. To take the simplest example, the role of
appellate court judges is very different from the role of trial court judges. As
Supreme Court Justice Stephen Breyer has observed, theyre simply different
jobs.7 Even within trial courts, the variety of judicial tasks is enormous. Some
judges are generalists, while others are specialists assigned to hear only bank-
ruptcy matters or family law cases. Some spend most of their time presiding
over civil or criminal trials, while others devote the bulk of their attention to
pretrial motions or case management. Still others serve in nontraditional
collaborative or therapeutic courts (see Judicial Council of California, 2007;
Williams, 2007), in which they perform functions foreign to the classical
conception of the umpireal judge (Frankel, 1975), or even to the more
modern conception of the managerial judge (Resnik, 1982). To adequately
define good judging, then, we would have to come up with a different
definition for trial court judges than for appellate court judges (Mott, 1948,
p. 265), and we likely would have to distinguish among several different types
of trial court assignments as well (Gavison, 1988, p. 1623). This would
complicate the task enormously, but failing to draw those distinctions
would be misleading. In todays courts, there is no one size fits all descrip-
tion of what judges do, and therefore no single, all-encompassing definition
of what they should do.
The role of trial court judges in performing nontraditional tasks such as
mediation and case management creates especially challenging problems of
quality evaluation or measurement. What counts as bad judging in a media-
tion setting? Is it any behavior that either coerces or discourages settlement in
the particular context presented? Or is it conducting a settlement conference
in a manner that is more evaluative than transformative regardless of the
outcome? Similarly, how should an error in case management be defined? Is it
setting a schedule that is too strict, so that the facts cannot be developed
adequately or the lawyers are severely stressed? Or is it setting a schedule that
is too lax, so that a case languishes rather than being disposed of promptly and
inexpensively? Might it also consist of sequencing motions in a way that gives
one party an unfair tactical advantage over the other? Defining and measuring
quality in these dimensions of judicial performance would be difficult indeed.
Perhaps for that reason, few attempts have been made. Even the ABAs
judicial performance guidelines, which go further than just about any defini-
tion of good judging in accounting for the varying roles that judges play, fail
to address these sorts of questions in a meaningful way. Yet these sorts of
decisions and actions comprise a large chunk of what many trial court judges
do (see, e.g., Otis & Reiter, 2006; Parness, 2006).
The nature of judging in collaborative or nontraditional trial court
contexts makes it especially difficult to evaluate quality in those contexts.
For example, consider child custody determinations, the appropriate
response to juvenile delinquency, or public law cases involving structural
relief. As one scholar has noted, in these types of assignments,
what is important is not to mete justice, but to help the parties design
an arrangement which will be least destructive and vindictive,
especially when children are involved. The prime commitment of the
decision maker should not be justice or the application of (legal)
norms, but a viable resolution of the dispute. For this commitment,
legal education is neither necessary nor clearly desirable.
Similarly, for cases in which intervention is needed to supervise
a complex administrative agency structure, such as a company or a
prison system, judges must have the skills to help them identify the
right and the feasible arrangement, and design the supervision
mechanism which might be most effective and least demanding of
constant review. I know of no law school which seeks to provide its
students with such skills, nor of a [judicial] selection criterion which
reflects them. (Gavison, 1988, p. 1625)
A third problem is that attempting to formulate a definition of good
judging encourages the misperception that the judicial role is fixed. Judges
jobs, however, evolve over time. The mix of tasks judges perform has changed
during the past few decades (Parness, 2006, pp. 18921898), as has our
perception of the relative importance of those tasks. For example, as the
role of case management has waxed and the role of presiding over trials has
waned (Galanter, 2004), our assessment of the skills judges should possess, as
well as our vision of good judging, has begun to shift. As one law professor has
pointed out,
the claim that the more dispositions, the better raises difficult
valuation tasks; decisionmaking must be assessed not only
quantitatively, but also qualitatively. On any given day, are four
judges who speak with parties to sixteen lawsuits and report that
twelve of those cases ended without trial more productive than
four judges who preside at four trials? (Resnik, 1982, p. 422)
Questions such as this, which no one would have thought to ask a few decades
ago, must be addressed today.
Today, excellence at presiding over jury trials, which remains a valuable
skill, arguably has become less important relative to excellence at the now
far more common tasks of mediating, scheduling, and ruling on pretrial
motions. While the personal qualities and skill sets that judges need to excel
at the wide range of tasks they are assigned undoubtedly overlap, they are not
congruent. Historical wish-lists of judicial qualities and guidelines for how
judges should do their jobs need to be revised to take this into account. Any
definition of good judging we develop today will have to be updated as judges
jobs continue to evolve. Thinking that we can settle on oneor even
severaldefinitions of good judging that are capable of providing a stable
platform for research over a decade or more is unrealistic.
Proceeding Without Definitions

Can we improve the justice system without first defining good judging? I
believe we can. One need look no farther than the field of medicine for a
striking example of what rigorous interdisciplinary analysis might do for law.
General anesthesia is risky. Some patients, who otherwise would have
survived their surgeries, are killed or seriously harmed by the anesthesia itself.
Not surprisingly, death from anesthesia initially was common. By the 1940s,
however, the situation had improved to the point that there was only 1 death
from anesthesia in every 2,500 operations. By the 1960s, the rate of death from
anesthesia had dropped even lower, to merely 1 in every 5,000 operations.
Considering the substantial risks posed by anesthesia, that death rate
seems relatively low.8 Some anesthesiologists, however, regarded it as unac-
ceptably high. They asked a young engineer named Jeffrey Cooper for help. In
an exhaustive study of 359 anesthesia-related errors, Cooper and several
colleagues researched the procedures followed by anesthesiologists, analyzed
the design of anesthesia machines, and, in meticulous detail, reconstructed
from interviews of the nurses and physicians involved exactly how and why
each mishap occurred (Cooper et al., 1978; see also Cooper et al., 1984).
Based on Coopers work, the American Society of Anesthesiologists
brought together anesthesiologists, surgeons, nurses, human factors engi-
neers, equipment manufacturers, and others at an international conference.
The combination of a willingness to acknowledge mistakes, careful empirical
study, and the consideration of interdisciplinary and cross-cultural perspec-
tives proved fruitful. Sources of error were identified, and strategies to
mitigate or eliminate them were implemented. Hours on duty for anesthe-
siology residents were shortened so that fewer slips or lapses would be caused
by fatigue. Manufacturers redesigned their machines, standardizing dials to
turn in a uniform direction, adding a mechanism to prevent the accidental
administration of more than one type of anesthetic gas simultaneously, and
changing controls so that oxygen delivery could not be turned off. Where an
error could not be eliminated, methods for detecting it before irreparable
harm was done to the patient were developed.
These efforts paid off. Within a decade, anesthesia-related deaths fell
from 1 in every 5,000 operations to just 1 in every 200,000 operations.9
What does this inspiring story about anesthesiology have to do with
judges? A lot. It suggests how judicial decision making might be improved
if the justice system adopted a similar approach.
Law is different from medicine in a variety of respects, of course, so it may
not be possible to replicate in the justice system the remarkable success
achieved in anesthesiology. It may be easier to identify errors in medicine
than in law, where the standards of correctness may be more subjective
(Posner, 2008, p. 3). Also, physicians may be more accountable for poor
performance than judges, because they can be sued for malpractice while
judges generally are immune from civil liability or criminal prosecution for
judicial acts.10 There already is at least one example, however, of the use of
empirical research in psychology to improve the quality of decisions in trial
courts.
Psychologists have known for decades that the traditional way of con-
ducting lineups in which an eyewitness is shown six photographs or six
people simultaneously under the supervision of a law enforcement officer
involved in investigation is likely to result in mistaken identifications
(Behrman & Davey, 2001; Steblay et al., 2001). There is a much better
technique that significantly reduces false positive identifications without
significantly increasing false negative identifications, and without signifi-
cantly increasing the cost of conducting lineups. That method is to show
the eyewitness one person or photograph at a time, and to ask the eyewitness
to make a decision about that individual before moving on to the next person
or photograph. If the eyewitness wants to view a particular person or photo-
graph again, he is allowed to do so, but only after he has first seen all of the
individuals separately, and only after the order in which the individuals are
presented has been reshuffled. In addition, the live or photographic lineup
should be conducted by a law enforcement officer who does not know which
individual is the suspect and which individuals are not.
These two changes improve the accuracy of identifications because
research has shown that when confronted by a simultaneous array, witnesses
tend to assume that the perpetrator is included, and then to compare the
individuals in the array to determine which one best matches their recollec-
tion of the perpetrator. Also, having the lineup supervised by an officer who is
not involved in the investigation reduces the danger that something might tip
off the eyewitness as to who the suspect is.
Although the advantages of this new method of conducting lineups have
been known for a long time, law enforcement agencies have been slow to
adopt it. Recently, however, New Jersey decided to require that this new
method of conducting lineups be used throughout its criminal justice
system (Kolata & Peterson, 2001). Other jurisdictions are following suit.11
This is a good example of how social science research can be used to improve
the accuracy of the justice system by strengthening the quality of the evidence
available to judges and juries. It also demonstrates that not all clearly bene-
ficial changes to our justice system necessarily require expensive, unpalatable,
or dramatic changes to existing procedures.
Improving Judging
Assuming that we want to follow in the footsteps of anesthesiologists, and
skip or defer the step of defining good judging, how should we proceed?
Probably by continuing some of the research we already are conducting, and
by trying a few new tacks as well. One approach would be to begin with a few
existing criteria of good judging we can all agree on (e.g., Klein, 2005; Farber
& Sherry, this volume, ch. 18). For example, no one would argue that the
length of a federal criminal sentence should be based on the race or gender of
the defendant rather than the relevant factors specified by statute. Nor is that
something that is likely to change over time, or even to vary depending on
whether the sentence is a conventional one imposed by a traditional criminal
court or a disposition implemented by an innovative collaborative court. One
thing we could do without first defining good judging would be to determine
whether judges have been imposing racially biased sentences by carefully
analyzing past decisions. Several scholars already have done this. I discuss
just one example, notable for its thoroughness and perceptiveness.
David Mustard (2001) examined the sentences imposed on 77,236 fed-
eral offenders during 19911994. His analysis revealed large disparities in the
length of sentences based on race, gender, education, income, and citizenship.
On average, white offenders received shorter sentences than black or Hispanic
offenders, and female offenders received shorter sentences than male offen-
ders. After controlling for a myriad of factors, the difference in sentence
length amounted to about six months in both instances. Significantly, most
of the difference occurred when the sentencing judge departed from the
guidelines promulgated by the United States Sentencing Commission.
Mustard started with a criterion of good judging we can all agree on
judges should not determine sentences based on statutorily forbidden
characteristics like the defendants race or genderand simply tested how
well judges were doing. His research revealed that judicial discretion for
individualizingwhich seems desirable in the abstractcontinued to create
the risk of forbidden discrimination even under the then tighter control of the
guidelines.12 This suggests that, whatever their other shortcomings, tight
sentencing guidelines may not be so bad after all.
Another approach is the one that Jeffrey Rachlinski, Chris Guthrie, and I
have taken in our work.13 We have tackled the problem experimentally rather
than archivally. We began by examining the psychology literature to identify
mistakes people commonly commit when making decisions. Assuming that
judges should avoid cognitive error while ruling on cases, we then attempted
to discern which of those mistakes plausibly might occur in situations in
which judges work. Our next step was to test judges, eliciting from them
responses to hypothetical cases or other sorts of problems to determine if,
when faced with choices similar in content and structure to those they
confront when doing their jobs, they would be vulnerable to the same sorts
of cognitive biases or illusions that cause ordinary people to make mistakes in
experiments conducted by psychologists.
This approach possesses some advantages. There is an extensive literature
concerning flaws in judgment and decision making on which one can draw,
rather than starting from scratch. That made it possible for us to start
identifying nonobvious potential sources of error immediately. If judges are
susceptible to committing particular types of errors, then we can think about
the situations in which judges act and try to identify solutions. It is not
necessary to have a comprehensive definition of good judging to know that
judges should not consider evidence they themselves have correctly decided is
inadmissible under the rules of evidence, be wildly influenced by an arbitrary
number mentioned by a litigant, make rulings based on a litigants attrac-
tiveness rather than the merits of the case, and so on. If it turns out that judges
are not susceptible to a particular type of error across a range of situations,
then we can cross that type of error off the list, at least provisionally, and move
on to others on the list.
Of course, this approach is not perfect. There may be questions about its
external validity,14 and it may be that some of the problems in human
decision making that have the most serious consequences for judges have
not been identified or systematically studied by psychologists. Cognitive
biases, however, are things we can identify, and in many instances, do some-
thing about, right now.
Careful study of reversals in the aggregate would be another way of
identifying potential sources of error. There are many studies of reversals,
of course, but most seem to focus on reversal rates, either overall or in
particular categories of cases (e.g., Guthrie & George, 2005). Others explore
the impact of reversals on subsequent trial court decisions (e.g., Gellis, 1985;
Liberato & Rutter, 2003; Smith, 2005). There do not seem to be any, however,
that attempt to determineas Cooper did in his study of anesthesia mishaps
the types of errors lower courts are making. That is something that can, and
should, be tried.15
Another criterion of performance could be developed from studying the
appellate records of individual lower court judges. It is well known that the
judgments of some judges are more frequently appealed than are judgments
of others. It is likewise recognized that some lower courts conform more
closely to the law as interpreted by the higher courts than do other lower
courts. The extent to which a trial judge should attempt to conform to the
opinions of her appellate court superiors may be debatable in some cases, but
regardless of the side one takes in that debate, both the frequency of appeal
and the frequency of reversal are objective criteria of judicial performance of
considerable significance (Mott, 1948, p. 276).
Even if an appellate court finds that the trial court erred, it will only
reverse the decision if it concludes that the error was not harmless. We should
study cases in which error was found to be harmless very carefully. Not only
may the error have affected the outcomedespite the appellate courts
determination to the contrarybut if it did not that may have been a fortuity.
Like the situation in which a doctor makes a mistake, but subsequently
notices it and fixes it before it harms the patient, the error occurred and
might have had injurious consequences in a different case, so we ought to
figure out how to prevent it. These near misses could be a valuable source of
information about needed reforms that may be overlooked by focusing only
on reversals.
We also could go a step further and review a sample of cases that were not
appealed to determine whether errors were made that escaped appellate
review entirely. For example, a confidential peer review process could be
implemented, in which respected judges would periodically visit other
courts and study a few randomly selected cases in which no appeal was
taken. The judges under study could be given feedback on their performance
to help them improve. This would help us to determine whether the errors
that are appealed are representative of those that are not appealed. It also
would help us to ascertain whether errors that are likely to be harmful are
escaping appellate review.
Of course, reversals provide noisy signals of error. Appeals are infrequent,
particularly in civil cases, and reversals are rare (Guthrie & George, 2005, pp.
359363). Some interim decisions are isolated from appellate review, and
final judgments are reviewed deferentially. Further, some reversals result from
disagreements over policy, or occur in cases that are extremely close. These
arguably cannot be characterized as errors in any meaningful sense. Despite
these shortcomings, however, studying the causes of reversal might be
fruitful.
An even better place to begin might be exonerations based on postcon-

viction determinations that a criminal defendant actually was innocent. We
know that such wrongful convictions occur. By carefully studying cases in
which that happened, we might be able to discern why our usually good
justice went bad in those instances, and what might be done to prevent such
miscarriages of justice in the future. Some have pursued this line of inquiry
(Borchard, 1932; Christianson, 2004), but more remains to be done.
Conclusion
Like the talented and hard-working anesthesiologists, judges (and others,
such as scholars and legislators) possess both the abilityand, I would
argue, the responsibilityto try to reduce the frequency and the magnitude
of error in trial court decision making. What we need is the dedication and
the courage to do what anesthesiologists did: openly pursue a painstaking
interdisciplinary inquiry into the question of how best to structure the
process of judicial decision making, and then implement reforms based on
what we learn. And we need to get started right away.
It is significant, I think, that anesthesiologists did not begin by trying to
define a good anesthesiologist or good anesthesiology. Instead, they
rolled up their sleeves and collected data. Then, viewing the delivery of
medical care as a system of which anesthesia and anesthesiologists were
merely a part, they attacked the most common and serious problems, leaving
the philosophizing for later. That is exactly what judicial reformers should do.
If, on the other hand, defining good judging is made a prerequisite to
empirical research concerning judicial decision making, then I doubt we will
progress very far anytime soon. We will have missed the opportunity to make
rapid progress like the anesthesiologists did, and that would be a terrible shame.
My respectful disagreement with those who contend that we must begin
our efforts to achieve judicial reform by defining good judging reminds me of
a story about two clerics who were debating the finer points of theology. After
they had been at it for awhile, one put his arm around the others shoulders
and said: Brother, why do we argue? After all, we both serve God. You in your
way, and I in His.
Appendix
American Bar Association Committee On Judicial Performance Evaluation,
Black Letter Guidelines for the Evaluation of Judicial Performance with
Commentary, 5-1 to 5-6 (2005)
Guideline 5-1. A judge should be evaluated on his or her legal ability,
including the following criteria:
1.1. Legal reasoning ability.

1.2. Knowledge of substantive law.
1.3. Knowledge of rules of procedure and evidence.
1.4. Keeping current on developments in law, procedure, and evidence.
Guideline 5-2. A judge should be evaluated on his or her integrity and
impartiality, including the following criteria:
2.1. Avoidance of impropriety and the appearance of impropriety.
2.2. Treating all people with dignity and respect.
2.3. Absence of favor or disfavor toward anyone, including but not limited
to favor or disfavor based upon race, sex, religion, national origin,
disability, age, sexual orientation, or socioeconomic status.
2.4. Acting fairly by giving people individual consideration.
2.5. Consideration of both sides of an argument before rendering a
decision.
2.6. Basing decisions on the law and the facts without regard to the
identity of the parties or counsel, and with an open mind in con-
sidering all issues.
2.7. Ability to make difficult or unpopular decisions.
Guideline 5-3. A judge should be evaluated on his or her communication
skills, including the following criteria:
3.1. Clear and logical oral communication while in court.
3.2. Clear and logical written decisions.
Guideline 5-4. A judge should be evaluated on his or her professionalism and
temperament, including the following criteria:
4.1. Acting in a dignified manner.
4.2. Treating people with courtesy.
4.3. Acting with patience and self-control.
4.4. Dealing with pro se litigants and litigation fairly and effectively.
4.5. Participating and providing leadership to an appropriate degree
in professional development activities and in jurisdiction-wide and
statewide court improvement and judicial education activities.
4.6. Promoting public understanding of and confidence in the courts.
Guideline 5-5. A judge should be evaluated on his or her administrative
capacity, including the following criteria:
5.1. Punctuality and preparation for court.
5.2. Maintaining control over the courtroom.
5.3. Appropriate enforcement of court rules, orders, and deadlines.
5.4. Making decisions and rulings in a prompt, timely manner.
5.5. Managing his or her calendar efficiently.
5.6. Using settlement conferences and alternative dispute resolution
mechanisms as appropriate.
5.7. Demonstrating appropriate innovation in using technology to

improve the administration of justice.
5.8. Fostering a productive work environment with other judges and
court staff.
5.9. Utilizing recruitment, hiring, and promotion policies and practices
to ensure that the pool of qualified applicants for court employment
is broad and diverse.
5.10. Acting to ensure that disabilities and linguistic and cultural differ-
ences do not limit access to the justice system.
Guideline 5-6. Additional criteria should be developed reflective of jurisdic-
tion (specialized versus general) and level of court (trial versus appellate).
6.1. A specialized court judge should be evaluated according to whether he
or she demonstrates the knowledge and skills necessary.
6.2. An appellate court judge should be evaluated on the quality of his or
her preparation for and participation in oral argument and on his or
her effectiveness in working with other judges of the court.
Notes
1. Differing versions of this story have been recounted. See, for example, Reed
(2005).
2. See Gregory Mitchells essay in this volume and Schauer (1988b). See generally
Dworkin (1986).
3. Quoted in St. Thomas Aquinas, Summa Theologica, pt. 2, question 67, art. 2,
objection 4.
4. Reprinted in In re Code of Judicial Conduct (1994), 103738 n.2. See also
Campbell (2005).
5. For a fine example of this, consider the following:
I venture to believe that it is as important to a judge called upon to
pass on a question of constitutional law, to have at least a bowing
acquaintance with Acton and Maitland, with Thucydides, Gibbon
and Carlyle, with Homer, Dante, Shakespeare and Milton, with
Machiavelli, Montaigne and Rabelais, with Plato, Bacon, Hume and
Kant, as with the books which have been specifically written on the
subject. For in such matters everything turns upon the spirit in which
he approaches the questions before him. (Hand, 1930, pp. 1213)
6. See also Solum (2005a).
7. Stephen J. Breyer, quoted in Toward Active Liberty: A Supreme Court Justice
Offers a View from the Top, Harvard Law Bulletin, 57:14, 17 (Spring 2006).
8. To put these anesthesia-related death rates into perspective, consider the error
rate in adjudication. Perhaps the best measurealbeit an imperfect oneof
whether a decision is mistaken is whether it is reversed on direct appeal or set
aside on collateral review. In the federal courts, reversal rates in appeals

terminated on the merits typically hover around 9 percent, or about 450 in
every 5,000 appeals from civil and criminal trial court judgments. See 2006 Fed.
Judiciary Ann. Rep. Director, Table B-5, available at http://www.uscourts.gov/
judbususc/judbus.html. Arguably more relevant is the ratio of trial court
decisions reversed on the merits to all trial court judgments whether appealed
or unappealed. That would yield a reversal rate of approximately .7 percent, or
about 35 in every 5,000 trial court judgments. (See Tables B-5, C, D.) Thus, the
error rate appears to be higher for law than for anesthesia.
Death penalty cases offer another basis for comparison. There are about 300
death sentences imposed annually in the United States. Most are subjected to
both direct appeal and collateral review, often in both state and federal courts.
Approximately 68 percent of death sentences are overturned or materially
altered because either the conviction or the sentence was wrong (Liebman,
2000). Therefore, the error rate in capital cases from 1973 to 1995, appears to be
higher than the error rate in anesthesia during the 1960s.
9. This account of what anesthesiologists accomplished is based on the summary
presented in Gawande (2000).
10. See, for example, Mireles v. Waco, 502 U.S. 9, 11-12 (1991) (per curiam). See
generally Olowofoyeku (1993).
11. In 2007, eyewitness identification reform legislation was signed into law by the
governors of Georgia, Maryland, North Carolina, Vermont, and West Virginia.
Press Release, California Commission on the Fair Administration of Justice,
Commission Chair John Van de Kamp Responds to Governors Vetoes of
Critical Criminal Justice Reform Bills Passed By California State Legislature
(Oct. 18, 2007), available at http://ccfaj.org/news.html.
12. Initially, the federal sentencing guidelines were mandatory, but after the period
Mustard studied, they were relaxed somewhat to allow for greater judicial
discretion (Stith, 2008).
13. Guthrie, Rachlinski, and Wistrich (2001; 2002; 2007a; 2007b); Wistrich,
Guthrie, and Rachlinski (2005); Rachlinski, Guthrie, and Wistrich (2009);
Rachlinski, Johnson, Wistrich, and Guthrie (2009). Of course, we are neither
the first nor the only researchers to have adopted this approach. See, for
example, Landsman and Rakos (1994); and Solan, Rosenblatt, and Osherson
(2008).
14. We have attempted to address these concerns in Guthrie, Rachlinski, and
Wistrich (2001, pp. 819821).
15. Of the studies of reversal, Gelliss (1985) article comes the nearest to doing this.
16
Expertise of Court Judges
James Shanteau and Len Dalgleish
Even a casual glance at papers on the expertise of court judges reveals wide-
spread disagreement about the quality of judicial decisions. Consider the
following recent articles:
Playing Dice with Criminal Sentences, by Englich, Mussweiler, and
Strack (2006)
Judicial Decision Thresholds for Violence Risk Management, by
Monahan and Silver (2003)
Conditional Bail Decision Making in the Magistrates Court by
Dhami (2004)
The first article implies a deficiency in judicial decision making, whereas the
second implies competence. The position taken here is consistent with the
third article: the expertise of court judges is tied to the particular domain in
which judges are asked to make factual or probabilistic assessments in the face
of uncertainty or ambiguity.
We believe that expert performance is not best understood and assessed
in terms of general characteristics applying across all domains. Rather, it
should be analyzed in the context of the specific tasks that experts in a
particular domain are called on to accomplish. Drawing on a substantial
line of research from psychology, we will argue that considerable empirical
evidence supports this domain-specific view of expert competence, that there
are reasons to believe this view holds with respect to court judges, and that, to
be as fair and useful as possible, assessments of judicial performance should
incorporate this view.
269
The paper is organized into five sections as follows. First, there is a review
of the literature on expertise in general. Second, the implicit theory behind
expert decision making (judges are either good or bad decision makers) is
described. Third, the role that domain differences play in disagreement
between experts across a wide variety of domains is considered. Fourth, an
alternate theory of expertise is proposed that posits that decision quality
depends on domain differences. Finally, the paper concludes with implica-
tions for judicial decision making and future research directions.
Background
Since the start of systematic analyses of skilled performance in the 1950s,
investigators have expressed surprise and dismay at the extent to which
experts make poor or inconsistent decisions. For example, if we ask two
court judges to assess a court case, the expectation of most analysts is that
they should come to the same conclusion. If they arrive at different conclu-
sions, then we wonder whether they are as skilled as they claim to be.
In a seminal paper, Einhorn (1974) argued that consensus or between-
expert reliability is a necessary condition for expertise.1 He found, however,
significant differences in diagnoses made by three expert medical pathologists.
The average between-expert correlation (r) was .55 (where .0 is chance and 1.0
is perfect). In comparison, weather forecasters were reported to have high
consensus values, r .95, for short-term predictions (Stewart, Roebber, &
Bosart, 1997). Thus there is widespread variability in the agreement of experts.
It is also possible to examine internal consistency, the extent to which an
expert makes the same decision for the same case if repeated. For pathologists,
Einhorn (1974) reported an average within-expert consistency r of .50. For
weather forecasters, on the other hand, the average internal consistency is
near perfect, r .98 (Stewart, Roebber, & Bosart, 1997).
In a study of livestock judges, four professionals were asked to evaluate
overall breeding quality of swine (Phelps, 1977). Despite a high level of
internal consistency (average r .96), the consensus agreement was much
lower, r .50. Apparently, livestock experts have internally consistent strate-
gies, but they disagree with each other about what those strategies should be.
Comparable results have been reported for other types of experts. For
instance, grain inspectors were found to have a consensus value between
judges of r .60, with internal consistency of r .62 (Trumbo, Adams,
Milner, & Schipper, 1962).
In other domains, the correlations are often lower. For example,
Hoffman, Slovic, and Rorer (1968) and Goldberg and Werts (1966) reported
consensus values of less than .40 for judgments by professional stockbrokers
and clinical psychologists. The internal consistency values were slightly
higher, with correlations of just over .40.
Expertise of Court Judges 271
Several studies have explored whether between-expert consensus

increases with experience. Ettenson, Shanteau, and Krogstad (1987) found
that between-auditor correlations increased from .66 to .76 to .83, for stu-
dents, audit seniors (mid-level professionals), and full partners, respectively.
Messier (1983) reported similar resultsaudit partners with more than
15 years experience had greater consensus than partners with less experience.
These results lead to three conclusions: First, experts in a variety of domains
often disagree with each other; the consensus correlations range from .40 to .60.
Second, some experts, such as weather forecasters, do show high levels of
agreement, with r values up to .95. Finally, in nearly all domains, the internal
consistency values are higher than the between-expert consensus values, that is,
experts usually agree with themselves, even if they disagree with others.
Experts-Should-Be-Perfect Argument
The less-than-impressive results from most studies of expertise have led many
analysts to question the abilities of experts in general. Following Einhorns
logic, these investigators assumed that agreement on the correct answer is a
necessary condition for expertise. The lack of agreement, therefore, suggested
that experts are no damn good (Gettys, personal communication, 1980).
This interpretation apparently derives from an implicit five-part argument
about experts:
(1) In expert domains, there is an assumed gold standard or unique
ground truth. If this truth is readily accessible, then we dont
need expertsanyone can obtain it. In most cases, however, it
requires an expert, such as court judges, to access truth. In these
cases, the answer is outside the realm of common knowledge of
most people. That is why we need experts.
(2) Because of their special skills and experience, the experts job is to
tell the rest of us about this ground truth. That is because experts
should be able to access answers that others cannot access. Put
another way, experts should have a unique ability to know the
truth.
(3) Since by definition there can be only one ground truth, all
experts should arrive at the same (single) correct answer. The
special abilities of experts thus allow them to arrive at the ground
truth correct answer.
(4) If experts disagree, then someone is wrong. Since there is one and
only one correct answer, some (or all) of them must not be real
experts. Thus, disagreement about the correct answer is a
reflection of incompetence within the group of would-be experts.
(5) Since lay people do not know which of the so-called experts is
correct, the only safe course of action is to distrust all (or at least
most) of the claimants. Thus, disagreement between experts

implies that we should be suspicious of any claim of special
abilities.
This argument, of course, is not a formal chain of logic. However, it is
implicit in the way that many analysts have interpreted the meaning of dis-
agreements between domain experts about correct answers. The bottom line:
disagreements are a sign of likely incompetence.
Domain Differences
It is common knowledge that experts in different domains perform different
tasks. Yet, many decision analysts persist in treating all fields of expertise alike,
that is, the term expert is used generically. For instance, Kahneman (1991)
concludes, there is much evidence that experts (in general) are not immune
to the cognitive illusions that affect other people (parenthesis in the ori-
ginal). Of course, this may be true for some, or even many, domains. But there
are domains, such as weather forecasting, where experts show little sign of
biases or cognitive illusions (Lichtenstein, Fischhoff, & Phillips, 1982;
Murphy & Winkler, 1977; Stewart, Roebber, & Bosart, 1997). Thus, despite
the broad generalizations of incompetence drawn about experts, there are a
number of exceptions to the rule.
In an effort to account for these domain differences, Shanteau (1992a,
1992b) constructed a table to differentiate those domains where experts do
well from those where experts do not. Table 16.1 is an updated version based
on a continuum from high to low competence. In the left column are those
domains where experts make aided decisions using Decision Support Systems
(DSS) or other computerized tools, for example, weather forecasters. The
Table 16.1 Progression of Domains from High to Low Performance
Highest Levels of Performance Lowest Levels of Performance
Aided Decisions Competent Restricted Random
Weather Livestock Judges Clinical Stockbrokers

Forecasters Psychologists
Auditors Grain Inspectors Pathologists Polygraphers
Test Pilots Chess Masters Parole Officers Office
Managers
Insurance Photo Student Parole Officers
Analysts Interpreters Admissions
Physicists Soil Judges Intelligence Court Judges
Analysts
Note: Each domain was classified based on analyses of expert performance reported in the literature.
next column contains domains where experts make skilled, but largely
unaided decisions, for example, livestock judges. The third column lists
domains where experts show limited competence, for example, clinical psy-
chologists. The behavior of experts in the last column is close to random, for
example, stockbrokers and, unfortunately, court judges.
It should be noted that assignment of domains within the table was based
on our review of the conclusions of the authors of each study. That is, the level
of competence is categorized from the assessments of the researchers who
studied each domain.
There are many ways to describe the differences in this table (Shanteau,
1992a, 1992b). For present purposes, it is sufficient to observe that domains to
the left (more competent) side involve stable (static) properties. That is, the
stimuli and the problem hold still for experts to evaluate. The domains to
the right (less competent) side involve more changeable (dynamic) proper-
ties. Thus, the stimuli and problem are less stable, more difficult to specify
more like moving targets.
Another way to view this distinction is to note that most domains to the
left side of the table involve physical or natural properties, whereas most
domains to the right involve human or social behavior. This may reflect the
fact that natural sciences in left-side domains have a longer history of devel-
opment. In contrast, the domains to the right have relatively young social
sciences. Given this difference, it makes sense that, as a general rule, expert
competence will tend to be higher to the left side and lower to the right side.
The question remains whether this distinction between domains is pre-
dictive of different levels of agreement. To examine this proposition, Table 16.2
summarizes consensus values reported by the author(s) for a variety of domains.
Two specialties are listed under each category, with the between-expert agree-
ment (consensus) given as average correlations. As can be seen, the mean
consensus r value for weather forecasters is .95, whereas the average value for
livestock judges is .50. The consensus values for clinical psychologists, and
stock forecasters are .40, and .32, respectively. Comparable results for other
domains appear in the second line. The trend supports the prediction outlined
Table 16.2 Between-Expert (Consensus) Reliability Values
Weather Forecasters Livestock Judges Clinical Psychologists Stockbrokers

r = .95 r = .50 r = .40 r = .32
r = .76 r = .60 r = .55 r = .33
Note: The values cited in this table were drawn from the following studies (from left to right): Stewart,
Roebber and Bosart (1997); Phelps and Shanteau (1978); Goldberg and Werts (1966); Slovic (1969); Kida
(1980); Trumbo, Adams, Milner, and Schipper (1962); Einhorn (1974); and Lykken (1979).
Table 16.3 Within-Expert (Internal Consistency) Reliability Values
Weather Forecasters Livestock Judges Clinical Psychologists Stockbrokers

r = .98 r = .96 r = .44 r = <.40
r = .90 r = .62 r = .50 r = .91
Note: The values cited in this table were drawn from the following studies (from left to right): Stewart,
Roebber and Bosart (1997); Phelps and Shanteau (1978); Goldberg and Werts (1966); Slovic (1969); Kida
(1980); Trumbo, Adams, Milner, and Schipper (1962); Einhorn (1974); and Raskin and Podlesny (1979).
abovemore-structured natural-science domains lead to high consensus values

whereas less-structured social-science domains leading to lower consensus.
For comparison, the average within-expert reliability (consistency) cor-
relations for these same domains are listed in Table 16.3. The trends are
similar, with better-structured domains leading to higher internal consistency.
As expected from psychometric theory (Nunnally, 1967), the consistency
values are higher than the corresponding consensus values in Table 16.2
(except for pathologists). In two domains (livestock judges and polygraphers),
there are notable discrepancies between consensus and consistency correla-
tions. These apparently reflect the existence of different schools of thought
as to how experts should do their jobs in these fields, that is, there are several
approaches to evaluating polygraphs that reflect the training of the operators.
Analysts View of Experts

Investigators in artificial intelligence, expert system design, cognitive science,
systems analysis, and computer science have all concluded that experts are
superior decision makers. That is why knowledge engineers design computer
simulations around what experts know. Similarly, most domain-specific
researchers (such as in medicine and weather forecasting) view experts as
possessing unique information essential for making good decisions. In short,
investigators in these disciplines see human expertise as an ultimate goal to be
emulated.
In contrast, decision analysts have concluded that experts are typically
flawed and prone to making simple errors (e.g., Kahneman, 1991). Moreover,
experts and novices are viewed as sharing the same shortcomings. For
instance, Tversky (quoted in Gardner, 1985, p. 360) states, whenever there
is a simple error that most laymen fall for, there is always a slightly more
sophisticated version of the same problem that experts fall for.
Decision analysts have apparently overlooked the fact that there are
important domain differences between areas of expertise. For instance, both
weather forecasters and livestock judges are skilled professionals. Yet, there is
a major difference for these two fields in the level of disagreement between
experts. The former is based on a well-developed science, whereas the latter is
based more on informed judgment. It should not be surprising, therefore,
to find livestock experts disagreeing more with each other about their judg-
ments than weather forecasters.
Nonetheless, judgment and decision researchers continue to view experts
as extensions of the flawed decision making typically seen in undergraduates
(Kahneman, 1991). That is, the characteristics of flawed decision making
associated with the biases and heuristics tradition are generalized to skilled
professionals (Gigerenzer, 1993; Klein, Orasanu, Calderwood, & Zsambok,
1993; Yates, 1990).
Importance of Domains
The position taken here is that previous analysts have unknowingly adopted
the experts-should-be-perfect view of expertise. As argued above, this view
implies that disagreement between experts is taken as a sign that something is
wrong. That in turn leads to the conclusion that experts are not as skilled or as
competent as they claim. In this section, we propose an alternative perspective
based on a domain-dependent view of expert performance. This perspective is
based on a five-part argument:
(1) The primary job of an expert is not to make decisions but to help
others reach a broadly defined target state. For example, the goal may
be to help policy makers design better strategies to reduce recidivism
or to increase the efficiency (i.e., reduce the cost) of the court system.
These goals do not involve single answers, but instead require
something more elaborate from the expert, such as strategic plans.
(2) To reach these goal states requires dealing with multiple,
constantly changing, and dynamic factors. As noted by Klein et al.
(1993), the situations faced by experts are different and more
complex than the simplified situations considered by most
analysts. Thus, experts work on problems that are much more
complex than those studied in idealized settings with
undergraduates making decisions.
(3) Using their knowledge and experience, the role of the expert is to
recognize patterns and find consistencies in a dynamic problem
space. The experts job is to clarify the issues and to identify
alternative approaches. In other words, the challenge for an expert
is to make sense out of chaos. This is certainly a description of
what many court judges have to do.
(4) Based on their experience and insights into the nature of problems,
experts try to help others (such as jury members) clarify their
thinking. Typically, an expert will identify various alternate paths to

a desired goal state (arriving at a verdict). The experts role is to lay
out the options and the consequences in a clear and comprehensible
fashion so that others may make the final decision.
(5) In the end, it is others, for example, a jury, not the expert judge who
makes most decisions. The judge offers insights and observations, but
it is up to others to make the final choice(s). In a jury trial, the ultimate
responsibility for the decision rests on the jury, not the judge.
The view is nicely summarized by the management consultant Golde
(1969: We seem to expect too much and the wrong things of our experts.
That is, experts generally act more like knowledgeable consultants. Rarely do
they function as the all-knowing, single-answer decision makers envisioned
by many analysts (who often see a parallel to economic/statistical theory).
When experts disagree, therefore, it is because they often see alternative
paths to the goal state. In court settings, there are almost always multiple
perspectives to any problem. Thus, disagreements between expert judges are
not only expected, they are a necessary consequence of the situation in which
judges find themselves.
Implications for Research

By relying on an inappropriate argument for expert decision making, many
analysts have unknowingly adopted a distorted view that leads them to expect
that experts should always agree on the correct decision. The next section
looks further at the research implications arising from this distortion.
By drawing a parallel to economic/statistical theory, decision analysts
have adopted a single correct answer approach to assessing expertise. In
most quantitative assessments, we expect to find one and only one answer to a
question. When an expert (or anyone else) gives an answer different from the
correct answer, he/she is said to have a bias (Tversky & Kahneman, 1974).
Moreover, when two or more experts give different answers, their claim of
special competence is questioned (Einhorn, 1974).
The position here is that these analysts have relied on an inappropriate
view of how experts function. For instance, the environment in which experts
work is much different from that reflected in the idealized world envisioned
by these analysts. The complex, changeable environment that experts actually
operate in is considerably more complicated. In reality, problems rarely are
simple enough to lead to single correct answers. Instead, there are almost
always multiple answers (or at least multiple routes to answers). If so, it
should not be surprising to find that experts, especially in domains involving
human behavior, often take different approaches to finding solutions.
The underlying problem is that analysts misunderstand what experts do
and what is expected of them. These investigators seem to think that experts
see the world as they dowith simplifying assumptions and single-answer

solutions. However, experts generally have a different worldview, with many
complexities and contingencies, but with few optimal solutions. In addition,
experts have a flexible approach to adaptation that helps them manage
uncertainty (Shanteau, 1992a).
From the present perspective, therefore, disagreements between experts
are expected. Although analysts view disagreements as evidence of incompe-
tence, our view is that experts see disagreements as a more-or-less inevitable
part of their job.
One explanation for these errors in judgment was offered by Gaeth
and Shanteau (1984). They noted that irrelevant materials (e.g., excessive
moisture) significantly impacted the decisions of soil judges. They also found
that training to compensate for irrelevancies was successful in reducing the
impact of these inappropriate factors. Another approach to improving expert
judgment was developed for weather forecasting; Murphy and Winkler
(1977) found that precipitation forecasts could be improved using a feedback
system based on Brier scores (a quadratic lossfunction). Since the introduc-
tion of Brier scores, the accuracy of weather forecasts has increased drama-
tically (Stewart et al., 1997).
Thus, it appears that experts can perform reasonably well, although there
is clearly room for improvement, for example, experts judgments are not
perfectly consistent. Moreover, there are sizable domain differences, that is,
there is greater agreement between experts in some areas than others. Experts
working in domains with better-defined properties, such as tax law, would be
expected to show sizable agreement. In contrast, experts in domains with
more subjective standards, such as sexual harassment, would be expected to
perform less well.
Conclusion
Dawes (personal communication, 1987) offered an insightful observation
about an earlier version of Table 16.1. The performance standards expected
by clients are different for the left and right sides. Weather forecasters are
allowed to make occasional mistakes. However, court judges and managers
are expected to be correct almost all the time. That is, in the less predictable
(right-side) domains, experts are held to higher standards of performance.
This is important for court judges, since with increased computerization
and media coverage, there has been a shift in skills needed. Traditionally, a
judge needed to be a jack-of-all-trades, with general skills in many areas.
Today, modern courtrooms place more demand on cognitive (thinking)
abilities. For example, with the trend toward more complex legal cases,
there is a greater need for judges with management abilities. These are
precisely the psychological skills associated with right side of the table (as
opposed to the more technical skills on the left side).
As argued here, disagreements are a natural product of the various

domains that experts work in. If disagreement between experts is not a critical
issue, then what should be? Let us suggest four useful goals for future research
on expertise. First, as argued above, the superiority of experts depends on their
ability to distinguish between relevant and irrelevant information (Ettenson,
Shanteau, & Krogstad, 1987; Jacavone & Dostal, 1992; Mosier, 1997; Schwartz
& Griffin, 1986; Shanteau, 1992b). One goal in future research should be to
discover how experts make these discriminations and to find ways to enhance
the process (also see Wistrich, Guthrie, & Rachlinkski, 2005).
A second goal should be to understand the kinds of intellectual and
physical tools expert judges use to enhance their decisions. Experts seldom, if
ever, make unaided judgments of the sort emphasized in laboratory research. In
fact, analysts regularly use the very tools denied their subjects. The experi-
menters themselves, using tools and expertise, are able to perform (laboratory)
tasks rather well (Edwards, 1983, p. 511). The type of tools used by experts
needs to be better understood. Perhaps it will be possible to borrow some of the
tools used by left-side experts to assist those making decisions on the right side.
The third goal should be to develop insights into domain differences. As
Edwards (1983, p. 512) argues, we have no choice but to develop a taxonomy
of intellectual tasks themselves. Only with the aid of such a taxonomy can we
think with reasonable sophistication about how to identify among the myriad
types of experts and the myriad types of tasks . . . just exactly what kinds of
people and tasks deserve our attention. The analyses in Tables 16.1, 16.2, and
16.3 offer an initial start in building a taxonomy.
The final goal is to find better methods for assessing the expertise of
individual experts in particular situations. As demonstrated here, expert
performance varies both between individuals in a given situation and
within individuals across various domains. The consensus and consistency
correlations presented above capture some of this variation. What is needed,
however, are measures that reflect both the difficulty of the domain and the
skill of the judge. Dalgleish and Weiss (2007) are working on this measure-
ment problem with hopeful progress.
Research on such goals will help broaden our understanding of expertise.
In contrast, concern about the supposed incompetence of experts based on
disagreements about correct answers offers little opportunity for expanding
our insights about expertise in the court room. Instead, we should focus our
efforts on analyses of relevance/irrelevance, tool usage, domain differences,
and measurement methods.
Note
1. Preparation of this chapter was supported by funds provided by the Scottish
Funding Council to HealthQWest; A research consortium for the West of
Scotland.
17
Cognitive Style and Judging
Gregory Mitchell and Philip E. Tetlock
An important dimension of judicial decisions much discussed by legal scho-

lars (e.g., Breyer, 2005; Molot, 2004; Peters, 2000; Sunstein, 1999, 2005), but
little studied by social scientists, is the scope of a courts opinion. Traditional
political science studies of judicial decisions emphasize the political valence of
judicial opinionsas endorsing a liberal or conservative outcomebut do
not address the scope with which commands or proclamations are issued. The
valence of an opinion is, however, theoretically orthogonal to the scope of an
opinion. The judge who writes opinions that curtail the use of affirmative
action programs in education (e.g., Justice OConnor in Grutter v. Bollinger
(2003)) and the judge who writes opinions that would widely bar the use of
affirmative action programs in education (e.g., Justice Thomas in Grutter)
may both take politically conservative positions, but they do so in ways that
may have significantly different implications for future cases and the devel-
opment of the law (see Richards & Kritzer, 2002, on the importance of
jurisprudential regimes). Thus, focusing only on the valence of decisions
ignores an important dimension of judicial opinions.1
Case details and other situational features surely affect the nature and
scope of judicial opinions, but differences in the manner in which opinions
are written may also reflect deep differences in the ways that individual judges
understand the causes of human behavior and their own ability to predict
behavior and influence it through court action. These deep differences
involve what psychologists broadly refer to as variations in cognitive style,
a psychological construct that has been operationalized in a variety of ways
to refer to individuals consistent ways of perceiving stimuli, processing
279
information, and thinking about problems.2 From a cognitive styles perspec-

tive, judges will naturally vary in the ways that they process case-relevant
information and solve problems, and these stable processes of thought are
likely to have important implications for the content of judges thoughts and
ultimately the content of their judicial opinions. For instance, integratively
complex judgesjudges who differentiate among many facets of a problem
but also look for connections across these dimensions that can provide
integrative solutionsare more likely to write opinions that take into
account multiple sources of evidence and multiple, potentially conflicting
legal rules and interests, whereas integratively simple judges are more likely to
write opinions that focus on key pieces of evidence, see some arguments as
clearly right and others as clearly wrong, and avoid trade-offs (see Tetlock et
al., 1985; Gruenfeld, 1995).
One instantiation of the cognitive style construct that we believe holds
particular relevance to judicial behavior is what Tetlock (2005) labels, after
Berlin (1953), the hedgehog-fox cognitive-style continuum: persons toward
the hedgehog pole on this continuum hold relatively simple theories about
human nature, which they believe apply across many domains, and they have
relatively high confidence in the long-run predictive success of these theories;
persons toward the fox pole on this continuum exhibit skepticism about
grand theories of human nature and deductive application of such theories,
believing instead that behavior and outcomes are contextually dependent and
subject to many random perturbations, and they have relatively low confi-
dence in their forecasting abilities.3 In a 15-year study of expert prediction of
political and business events, Tetlock (2005) found that a seven-item measure
of the hedgehog-fox cognitive style distinguished among his sample of experts
in several important ways: foxes were better calibrated and more accurate
in their predictions, particularly within their areas of expertise, were more
rational in the updating of their beliefs following failed predictions, and were
less likely to exhibit defensive maneuvers after predictive failures and knew-
it-all-along effects after outcomes were known.
The results from Tetlocks (2005) prognostication tournament suggest
that judicial foxes will more accurately predict future conditions, and be more
humble in their concomitant predictions and ambitions, when acting as social
engineers who must predict how their rulings will interact with future societal
conditions (see, e.g., Albiston & Nielsen, 2007). Judicial hedgehogs, on the
other hand, should be much more likely to see the world in deterministic and
predictable terms and thus should issue more expansive and less contingent
proclamations that will tend more often to be wrong factually. Extension of
this distinction to judging finds support in legal discussions of judicial
philosophy, particularly the work of Cass Sunstein on constitutional deci-
sion-making. Indeed, Gelman (2001) notes the striking similarity between
Berlins dichotomization of historical thinkers into hedgehogs and foxes and
Sunsteins (1999) dichotomization of appellate judges into maximalists and
minimalists. Judicial maximalists (i.e., hedgehogs) favor broad rulings that
Cognitive Style and Judging 281
implement abstract theories about individual rights and institutional compe-

tencies, whereas judicial minimalists (i.e., foxes) do not apply broad theories
deductively to particular cases but instead proceed incrementally, through
analogy, and issue narrow rulings (see Sunstein, 1999). Cognitive style rese-
arch suggests that the consistent differences that Sunstein identifies reflect
judges broad orientations to the world and their information-processing
tendencies more than consciously chosen judicial philosophies (i.e., cognitive
style likely leads to a preference for minimalist or maximalist approaches to
judging).
The potential relevance of cognitive style research to judicial behavior
extends beyond the formulation of opinions. Judicial foxes, who see much
greater causal murkiness in the world, should exhibit greater humility in their
choices to end cases without the aid of trial, where evidence can be more fully
developed. Thus, at the appellate level, they should be more likely to remand
cases for new trials rather than resolve the cases through review of the
appellate record, and, at the trial level, they should be less likely to grant
summary judgment motions and Rule 12(b)(6) motions on grounds of
implausibility, per Bell Atlantic v. Twombly (2007).
Further, Tetlock (2005)s finding that hedgehogs tend to be less corrigible
and more protective of their preexisting views than foxes suggests also that the
opinions of judicial hedgehogs should show less change over time in both
their content and direction because of the greater rigidity of these judges
views, and their opinions should thus be more predictable and internally
consistent than those of judicial foxes. And, given their certitude, judicial
hedgehogs should be less prone to compromise in response to pleas from
members of the same court, particularly pleas from judicial foxes who see
issues so differently, and more prone to write separate opinions when the
majority opinion is authored by a judicial hedgehog of a different ideology or
by a judicial fox of any political type.
Tetlock (2005) found hedgehogs on both the left and the right in his
study of experts, and legal scholarship suggests the same may be true of
judicial hedgehogs (cf. Sunstein, 1999). If so, we should find liberal and
conservative judicial hedgehogs looking more ideological and less sensitive
to legal and case-specific variationsand thus more predictablewithin
political science studies of the role of attitudes in judicial decision-making
(e.g., Segal & Spaeth, 2002), as compared to judicial foxes of either political
bent. This hypothesis runs counter to theories of political ideology founded
on a rigidity-of-the-right thesis, which associates conservatism with right-
wing authoritarianism and a need for closure and associates liberalism with
open-mindedness (e.g., Jost et al., 2003), but our hypothesis is consistent with
the view that judicial activism is not the primary domain of liberal judges
(Cross & Lindquist, 2007; Sunstein, 2005) but rather of ideologues generally
(see Tetlock, 1986).
While Tetlocks (2005) results with respect to forecasting accuracy and
learning, and our predictions about the relation of judicial hedgehoggishness
to judicial activism, might suggest that judicial foxes should be favored over
judicial hedgehogs (e.g., Farber & Sherry, this volume, ch. 18, advocate giving
preference in judicial selections to foxes), the attractiveness of hedgehog-like
or fox-like tendencies will depend on the interests of the audience. The judge
who appears to some to be a wise judicial fox, deciding only what is necessary
to resolve the case at hand, will appear to others as an unprincipled, timid
judge who fails to provide necessary guidance to future legal actors
(cf. Tetlock et al., 1993; Waldron, 2007).
Fox-like tendencies are likely to be especially valued in public law
domains, which preoccupy many legal scholars, where the issues presented
in cases often implicate many contentious trade-offs and implicate debates
about the relative competency of courts versus legislatures versus executives
to make these trade-offs. Thus, a humble, incremental approach will often be
seen as the least likely to lock-in-place bad guesses about future conditions
and unpopular resolutions of trade-offs (see, e.g., Sunstein, 2007). In the
domain of private law, however, where parties can often contract around
undesirable legal rules or choose from competing legal regimes, we suspect
that judicial hedgehogs, with their more sweeping, definitive, and predictable
approaches to cases, will fare much better in the eyes of the key attendant
audiencesbusinesses and business lawyers.
If we are right that the hedgehog-fox cognitive style dimension can
illuminate how judges go about their business, then researchers should
naturally be concerned with issues of measurement and testing of our
hypotheses on judges, for it is perilous to assume that findings from
Tetlocks research on forecasting experts holds with respect to judges, who
inhabit unique institutional positions and generally have similar educational
backgrounds that may greatly affect the expression of individual differences in
cognitive style. The most straightforward approach would be to recruit judges
to complete Tetlocks (2005) hedgehog-fox scale and then examine these
judges with respect to their opinion-writing and other behaviors consistent
with the different cognitive styles.4 Alternatively, observers may score a
sample of judges opinions for hedgehog- versus fox-like perspectives and
examine the consistency of these tendencies across cases and the ability of
these scores to predict other judicial behaviors. Unfortunately, no content-
analytic protocol for the placement of writers on the hedgehog-fox conti-
nuum presently exists.
However, given the potential gains from cognitive styles research on
judges, we believe that investments in the development of empirical tools to
measure judicial cognitive styles would be well-rewarded. Because interest in
the hedgehog-fox variable supplements rather than supplants traditional
political-science interest in the judicial ideology variable, and offers a way
to explain more of the non-law-related variance in judicial behavior, political
scientists should find this research perspective congenial to prevailing attitu-
dinal theories of judicial behavior. Because the hedgehog-fox distinction
mirrors existing legal discussions of judicial philosophy, and offers a way to
Cognitive Style and Judging 283
explain the bases for these philosophies and to focus on elements of judicial
behavior largely ignored by political scientists, legal scholars should find this
research perspective congenial to prevailing theories of legal interpretation
and the development of the common law. We thus see cognitive style research
as offering an empirical bridge between the political science and legal litera-
tures. Most provocatively, cognitive style research may show that traditional,
competing positions of the political scientist and the legal scholar are both
correct, but only for subsets of judges: judicial hedgehogs may allow their
personal beliefs and values to drive their legal decisions, thus supporting
many political scientists views of judging, but judicial foxes may be much
more sensitive to the facts and law of particular cases and much less willing to
inject their own beliefs and values into cases, thus supporting many legal
scholars views of judging.
Notes
1. Scope itself can be seen as multidimensional. For instance, Sunstein (1999)
distinguishes between the width and depth of an opinion, with the former
referring to the reach or breadth of the legal ruling announced and the latter
referring to the nature of the doctrinal or theoretical foundations for the
opinion.
2. In this brief discussion, we focus only on one area of cognitive style research.
For reviews of this large body of research, see Kozhevnikov (2007), Suedfeld
(2000), and Suedfeld & Tetlock (2001). We focus here on individual differ-
ences in cognitive style because of their potential for differentiating among
judges and their opinions. There are, however, situational differences in
cognitive style as well, and these situational differences may cause convergence
across judges who otherwise would exhibit different individual styles (see
Tetlock, 2005).
3. This distinction follows from Isaiah Berlins now classic essay, The Hedgehog and
the Fox, in which he describes the distinction as one of the deepest differences
which divide writers and thinkers, and it may be, human beings in general:
For there exists a great chasm between those, on one side, who relate
everything to a single central vision, one system less or more coherent
or articulate, in terms of which they understand, think and feela
single, universal, organizing principle in terms of which alone all that
they are and say has significanceand, on the other side, those who
pursue many ends, often unrelated and even contradictory,
connected, if at all, only in some de facto way, for some psychological
or physiological cause, related by no moral or aesthetic principle;
these last lead lives, perform acts, and entertain ideas that are
centrifugal rather than centripetal, their thought is scattered or
diffused, moving on many levels, seizing upon the essence of a vast
variety of experiences and objects for what they are in themselves,
without, consciously or unconsciously, seeking to fit them into, or
exclude them from, any one unchanging, all-embracing, sometimes
self-contradictory and incomplete, at times fanatical, unitary inner

vision. The first kind of intellectual and artistic personality belongs to
the hedgehogs, the second to the foxes; and without insisting on a
rigid classification, we may, without too much fear of contradiction,
say that, in this sense,. . . Dante belongs to the first category and
Shakespeare to the second; Plato, Lucretius, Pascal, Hegel,
Dostoyevsky, Nietzsche, Ibsen, and Proust are, in varying degrees,
hedgehogs; Herodotus, Aristotle, Montaigne, Erasmus, Moliere,
Goethe, Pushkin, Balzac, and Joyce are foxes (Berlin, 1953, pp. 12).
4. A few items in this scale will likely need substitution or revision for a judicial
context, however (see Tetlock, 2005, p. 268).
18
Building a Better Judiciary
Daniel A. Farber and Suzanna Sherry
We have spent much of our academic careers arguing that judicial decision
makingeven in constitutional casesis a specialized craft, not merely an
exercise in politics (Farber & Sherry, 2002; Farber, 1992, 1995; Farber &
Adams, 1999; Sherry, 2003, 2005, 2007). We have suggested that good judging
requires both expertise and a certain set of dispositional traits, and that it can
be enhanced or hindered by both personal traits and situational character-
istics. In a recent book, we describe and defend our vision of the process of
constitutional adjudication, provide examples of good and bad judicial opi-
nions, and identify existing and proposed structural supports conducive to
good constitutional decision making. (Farber & Sherry, 2008).
Our task in this essay is to translate our theorizing into concrete sugges-
tions for further research. Anyone who has read this far in the current volume
is probably persuaded that the task is hopeless. In constitutional cases, there
seems to be no hope of reaching agreement on an appropriate normative
standard. Certainly there is no consensus on outcome-based norms: If
everyone agreed about the ultimate meaning of constitutional provisions,
the Supreme Court would have no docket. Decisional norms fare no better,
insofar as judges and scholars disagree about both the appropriate method of
constitutional interpretation and the role of the judiciary in our constitu-
tional democracy. As for our claim that certain dispositional traits can lead to
better judging, the empirical testing of such a claim would prove difficult if
not impossible.
Being legal academics as well as pragmatists, however, we remain
undaunted by philosophical dilemmas and empirical gaps. The reader will
285
have to judge whether our optimism is warranted. Mindful of the fate of Gaul,
we nevertheless divide this essay into three parts: In the first part, we describe
what judges do when they decide constitutional questions, concluding that
they are primarily exercising the same legal expertise that judges and lawyers
utilize in all of their professional decisions. In the second part we focus briefly
on the personal and contextual characteristics that have been shown to
produce or interfere with expert decision making in general. Finally, we
turn to our main focus: the legal structures that might enhance the positive
characteristics and minimize the negative ones.
Deciding Constitutional Questions

Constitutional decision making, like all legal decision making, is a process of
constrained discretion. While there are rarely unequivocally correct answers,
some answers are better than othersand some methods are more or less
likely to produce better answers. Consider such run-of-the mill legal ques-
tions as how the parties to a contract would interpret an ambiguous provi-
sion, or whether a particular piece of evidence is likely to be more prejudicial
than probative to a jury, or whether the efficacy of a safety regulation out-
weighs its cost. In each case, there is room for legitimate disagreement about
the correct outcome. In other words, there is no formula for reaching the
correct answer, and discretion is inevitable although not unlimited.
We suggest that the same standard can be used to evaluate judicial
decisions. We can ask, essentially, whether the judge considered all of the
appropriate legal and factual factors (and only those factors) and whether the
conclusion is reasonable rather than arbitrary. This move does not eliminate
discretion or provide certainty, since there will often be more than one
reasonable outcome. And ultimately, knowing what is reasonable is a
matter of judgment. But this standard does rule out both some arguments
and some answers. It requires attention to precedent, constitutional history,
and public values, and demands that judicial justification of the outcome
meet a fairly high level of plausibility and coherence.
No formula can tell us what it means to provide due process or to
afford everyone equal protection of the laws, nor what constitutes an
establishment of religion or probable cause for a search. (Easy constitu-
tional questions do exist, but rarely reach the courts: No one is ever going to
litigate the meaning of the requirement that one must be at least 35 years old
to become president.) While some judges and scholars have tried to devise
such formulas, their theories raise more questions than they answer, and do
not in fact cabin discretion. Originalism, the most commonly invoked for-
mulaic theory of interpretation, is a case in point. It merely transposes the
question from the meaning of the ambiguous text into the equally unans-
werable historical query of what the founding generation meant by that
Building a Better Judiciary 287
ambiguous text (Farber & Sherry 2002, pp. 1028). Given that professional
historians themselves disagree about such issues, and that judges are not
trained in historical analysis, originalism cannot effectively constrain judicial
discretion. Indeed, it is likely to prove less effective in constraining discretion
than the common law method, with which judges have years of professional
training and experience to guide them (Farber & Sherry, 2002, pp. 153154).
So how do judges make decisions in constitutional and nonconstitutional
cases? What distinguishes a reasonable answer from an unreasonable one, a
legitimate decision from an illegitimate one? As we argue in our book, we can
look to administrative law to help us answer these questions. Legislatures
often delegate broad but not unlimited discretion to administrative agencies,
and courts are frequently called on to review agency action to determine
whether it is a legitimate exercise of the agencys discretion. In doing so,
courts focus on two aspects of the agencys decision: They ask whether the
agency considered all of the relevant factors and only the relevant factors, and
they ask whether the decision was arbitrary and capricious.1
Administrative decision making also helps us see how sound judgment
and expertise can play a key role in decision making even in situations where
politics and ideology matter. Clearly, it matters whether the administrator
of EPA is appointed by a Republican or a Democrat. But science, law, and
economic analysis also matter. Two good professional EPA administrators
appointed by presidents of different parties (say, William Ruckelshaus and
Carol Browner) have more in common with each other, and their decisions
have more similarity, than either has with a political ideologue like Anne
Burford. The exercise of discretion leaves room for a variety of factors, but not
all of those factors are political. Good administrative decisions are also shaped
by legal directives, prior practice, and rigorous analysis. So it is with consti-
tutional interpretation.
When good judging is described in this way, it should be apparent that it
is not very different from good decision making in other contexts. Expert
decision making, judicial and otherwise, is the ability to identify and take into
account all of the relevant information and then draw reasonable conclusions.
Lawyers rely on this sort of expertise all this time: when advising a client about
whether a proposed course of action is likely to lead to legal liability, when
drafting a contract designed to have a certain legal effect, or when deciding
whether a particular legal argument might persuade a court. Of course,
lawyers do not always agree about the right answer, and these disagreements
may relate in part to broader differences in personal experience and perspec-
tive, but there is a core of shared skills in these tasks. Interpreting the
Constitution may have more political salience than interpreting a statute, a
contract, or a body of precedents, but it is fundamentally analogous. Judicial
expertise is simply legal expertise in a different context.
Once we conclude that judicial decision-making is the exercise of legal
expertise, we can turn to the field of psychology and expert decision making
more broadly to help us understand what personal and situational factors
enhance decision making. In particular, we can look for characteristics that

incline decision makers toward the careful consideration of all relevant
factors, that increase the likelihood of reasonable and coherent decisions,
and that allow individuals to overcome the cognitive biases to which humans
are so prone.
Personal and Contextual Characteristics

As suggested by several essays in this section, especially that by Mitchell and
Tetlock (ch. 17), numerous studies have pointed to a cluster of personal
characteristics that tend to improve decision making by increasing the ability
to overcome common cognitive biases. By reducing the effect of these biases,
such characteristics are also likely to make judges better able to identify and
rely on relevant factors and to process the information correctly.
The primary finding seems to be that foxes make better decisions than
do hedgehogs. The more broadly a person is inclined to search, the better
for both the reasoning process and the ultimate decision (Tetlock, 2005).
This should be true for judges as well as for other decision makers.
Moreover, to the extent that foxiness incorporates specific characteristics
such as integrative complexity and open belief systems, it is especially
important that judges be foxes rather than hedgehogs. Recall that our
standard for good judging requires a willingness to consider carefully all
relevant information. For the reasons identified at the outset of this essay,
we cannot measure directly whether a judge has done so (except to the
extent that the opinion reflects the judges decision process) or whether a
failure to do so has produced a worse decision. Instead, we must rely on a
judges dispositional preference for multiple perspectives over dogmatism.
Focusing on these individual characteristics is most helpful at the stage of
judicial selection.
Turning to situational characteristics, two seem to make the most
difference: collegial decision making in a heterogeneous group (Sunstein,
2000), and accountabilityespecially accountability to persons whose
views are unknown. This insight can be of use both in selecting judges and
in designing the structure of the legal regime generally and the judiciary in
particular.
Legal Characteristics
With these personal and situational characteristics in mind, we can now
examine which aspects of the American legal landscape are likely to be
conducive to good judicial decision-making, which might be in need of
improvement, and what changes we might want to make.
Let us begin at the beginning, with legal education. A traditional legal

education seems quite well suited to fostering a judicial disposition. Most law
schools use the Socratic method: Teachers call on individual students and ask
sequential questions that add new or different information and make the
students think ever more deeply about an issue. This method forces students
to confront arguments on all sides of an issue, and to consider an increasingly
complex web of information. They will often be called on to identify the weak
points in their own arguments or to make the strongest arguments on the
other side. All of this helps students develop the professional judgment that
they will need as lawyers and the fox-like disposition that makes good
judgesespecially since the discussion is usually focused on existing prece-
dent, which does not allow students to run freely with any preconceived big
ideas. Throw in the traditional emphasis on the incremental development
of the common law and the careful attention to texts, and American legal
education has the capacity to foster exactly the expertise that facilitates good
judgment.
There are some flaws, of course. Most lie with individual teachers and
scholars. The critical legal studies scholars and their heirs (including the
popular constitutionalists, as we argue in our book) teach students that con-
stitutional law is simply politics by another name, denying the possibility of
expert judgment. Many conservative constitutional scholars want judges to be
hedgehogs, urging a single overarching method of constitutional interpreta-
tion. Both of these schools of thought, to the extent that they seep into the
classroom, undermine the efficacy of a legal education in producing good
judgment as we have defined it. The problem is exacerbated by the lack of
curricular requirements after the first year of law school: Students in their last
two years of law school are free to choose the courses (and teachers) they prefer,
and can thus avoidinadvertently or deliberatelyconfronting uncomfor-
table intellectual challenges. The contemporary pedagogical philosophy that
students are consumers and that education ought to be entertaining and
relevant also makes it easier for students to escape some of the experiences
that might make them better lawyers and judges. Despite these problems,
however, legal education provides a good basis for development of legal judg-
ment and legal expertise. But perhaps legal educators should be more knowl-
edgeable about the ingredients of good judgment and more conscious of their
role in fostering it.
Clinical education, properly conceived, can also cultivate important
habits of thought. Good lawyering requires an ability to imagine opposing
arguments and alternative perspectives on a case (Farber, 1994). It also
requires an ability to think broadly about the interests and goals of a client,
which may require consideration of a range of institutional and interpersonal
factors as well as deliberation about the ethical factors that define the lawyers
role. Although clinical education can sometimes be no more than basic
training in the rudiments of legal practice, it is also capable of inculcating
important elements of good judgment.
What legal education begins, a life in law practiceespecially private

practicehones. Professional judgment makes the difference between a suc-
cessful lawyer and an unsuccessful one. Everything a lawyer does depends on
his ability to accurately predict legal consequences. Like experts in all fields,
then, good lawyers must develop the dispositional characteristics that are
most conducive to good decision-making. The context of private practice also
fosters good decision-making. Clients make sure that lawyers are accountable.
And the practice of law is a collaborative enterprise, often (especially for
young lawyers) with little or no choice of collaborators. The heterogeneous
working groups that result further contribute to good decision making.
Two groups of lawyers, however, work in environments that may not
provide such incentives to the exercise of good judgment. Long-time lawyers
for a particular cause or clientwhether it is the NRA, the ACLU, the
prosecutors office, or the public defenderare less often forced to confront
multiple perspectives that differ from their own. While they do have to
anticipate opposing arguments, the homogeneous environment tends to
reinforce their own perspectives and denigrate the validity of any counter-
arguments. They are also less accountable, because their clients often lack the
choice, control, or incentives of private paying clients. We might therefore
expect that lawyers with only this type of legal experience are likely to make
poorer judges than those who have a more diverse background. This will not
always be true, particularly if the individual has played multiple roles, such as
working as a prosecutor and later as a defense lawyer.
Legal academics are even less constrained by the realities of private
practice, and indeed are often subject to exactly the wrong incentives
(Farber & Sherry, 2008; George, 2001). Legal scholars generally work alone,
and are (in their scholarship, at least) accountable to no one (at least after
tenure!). Their own biases are often reinforced as they seek out like-minded
colleagues. Unlike practicing lawyers, academics are free to ignore the legal
world around them, and legal expertise can sometimes seem almost irrele-
vant, particularly for those whose work is the most theoretical. Moreover,
the most highly valued scholarship is the most innovative; legal scholars are
thus encouraged to flout common sense and exhibit poor legal judgment.
Innovation is obviously a desirable trait, but what makes it treacherous in
legal scholarship (as opposed to more empirically oriented social sciences) is
the absence of a strong reality check.
All of these factorsas well as others we canvass in Judgment Calls
suggest that academics are among the least likely to internalize the disposi-
tional traits conducive to good judicial decision making. We should note,
however, that this depressing picture of legal academia is of relatively recent
origin, and may be moderating somewhatin addition, there are obviously a
large number of exceptions to this generalization. (For instance, we ourselves
have excellent judgment. Identifying the cognitive biases in the preceding
sentence is left as an exercise to the reader.) Also, many legal academics have
had other kinds of experiences, either prior to teaching or in the thankless
tasks of academic administration, which may be more conducive to devel-

oping balanced judgment.
Identifying which professional experiences are most likely to produce
good judges leads us to the next, and most problematic, topic: the judicial
selection process. What we have already discussed suggests that we should
prefer judges with private practice experience, and avoid most academics. But
the selection process for federal judges is in need of even more improvement.
Once upon a time, presidents generally nominated (and the Senate confirmed)
experienced and competent moderatesusually, although not always, from
the presidents party. While a judges overall political perspective was usually
aligned with that of the nominating president, the difficult cases that reached
appellate courts were resolved largely through legal expertise and good judg-
ment. Being good lawyers rather than political animals, judges exercised
independent judgment that sometimes surprised their nominating presidents.
President Eisenhower, for example, was famously disappointed in Chief
Justice Warren and Justice Brennan.
In recent years, however, some presidents have been much more deter-
mined to influence judicial decision making and avoid future surprises. They
have substituted ideological litmus tests for professional competence in
nominating judges. And in trying to ensure ideological purity and fidelity,
they have deliberately sought nominees with an unyielding record of ideolo-
gical commitment. Offices within the executive branch are now charged with
the task of identifying true believers, and the slightest hint of ideological
impurity will sink a candidate. Of course, some presidents are more com-
mitted to this strategy than others, but the trend is worrisome.
Whether or not this contemporary selection process means that federal
judges decide cases based on political considerations (and we believe that most
judges still do not, at least not consciously), it does diminish the likelihood of
appointing foxes rather than hedgehogs. The stronger the commitment to a
particular ideology, the less open a judge will be to other perspectives. Instead
of focusing on ideological commitment, then, presidents and senators should
be looking for evidence of the dispositional traits that have been shown to
enhance judgment and good decision making (Sherry, 2003). We should be
seeking an openness to other perspectives, a willingness to revise ones views in
the face of new information, and a refusal to adopt a single approach to
decision making. Hedgehogs should be shunned: A nominee who espouses
an overarching grand theory of constitutional interpretation, of whatever
ideology, should be immediately disqualified.
The selection of state judges is even worse. Most state judges are popu-
larly elected or subject to popular reappointment or recall, and judicial
electionsespecially at the appellate levelare becoming ever more politi-
cized. The same campaign strategies that are lamented in legislative races are
now infecting judicial elections. Our constitutional commitment to free
speech means that states cannot effectively limit judicial electioneering or
organized smear campaigns. Here the cure is obvious: State judges, like
federal judges, should be appointed rather than elected. The public has no way
to evaluate a candidates judgment or expertise, so it is no surprise that elected
judges are only coincidentally good decision makers.
One further factor should guide the selection process. Since heteroge-
neity has been shown to improve collegial decision making, we should try to
maintain a diversity of viewpoints on the bench. This often happens naturally,
as presidents and senates with different perspectives have a chance to nomi-
nate and confirm new judges. But long periods of single-party rule, random
fluctuations in the judicial vacancy rate, and particular commitments to
ideological nominees can sometimes produce a largely homogeneous judi-
ciary. When this seems likely, senators of both parties should take seriously
our constitutional ideal of an independent judiciary, and insist on greater
diversity. (We can dream, cant we?) To the extent that presidents focus on
expertise and open-mindedness rather than ideology, of course, deliberately
creating diversity will be less necessary.
Moving from individual character traits to the institutional context, a
number of existing structural features enhance judicial decision making. The
two most important are the collegial and hierarchical structure of the judi-
ciary, and the tradition of issuing written opinions to explain judicial deci-
sions. A less important but still significant factor is the influence of multiple
nonjudicial actors on judicial decision making.
All American appellate courts make their decisions collegially, on
multimember panels. Since virtually all important constitutional questions
get to the appellate level, we need not be overly concerned about constitu-
tional adjudication by single judges at the trial level. Additionally, all judges
save the nine on the Supreme Court are subject to oversight by other
judges. Thus all judges must persuade others in order to prevail, and all
except the nine justices are directly or indirectly accountable to some other
court. Both collegiality and accountability, the two most important situa-
tional characteristics for good decision making, are therefore already
present.
The heterogeneous nature of the judiciary contributes two further ben-
efits. First, collegial decision making by a heterogeneous group injects mul-
tiple perspectives and limits the problem of self-reinforcement within the
members of the group. Second, because judges often know neither the
identities nor the perspectives of those who will be reviewing their decisions,
the beneficial effect of accountability is increased. This will always be true for
decisions by trial court judges. It can also be true for judges on three-member
federal appellate panels, whose decisions might be scrutinized by other
members of their court via a petition for en banc review, by the Supreme
Court, or by both. These judges may also be concerned about how their
decisions will be received by colleagues in different circuits, who are not
bound by their decisions and are free to criticize them.
Even more important is the fact that in the United States (as in most
countries whose legal systems are primarily derived from the English
common law regime), judges routinely issue written opinions explaining and
justifying their decisions. Judicial decision making is therefore relatively
transparent. This transparency enhances decision making by subjecting deci-
sions to public scrutiny and thereby increasing accountability. It also forces
judges to confront counterevidence and counterarguments. Many judges
recount the experience of having reached a particular decision, only to have
a change of heart after finding that the opinion wont write. A judicial
career spent deliberating with colleagues and writing opinions is also likely to
lead to greater open-mindedness and critical thinking over time, further
improving the quality of decision making in life-tenured judges. It is well-
known that judges can grow on the bench, and the psychological literature
on decision making may offer a partial explanation.
Unfortunately, opinion-writing is declining in the federal appellate
courts, as more and more cases are decided by perfunctory unpublished
decisions. The recent adoption of Fed. R. App. Proc. 32.1, which bars courts
from prohibiting the citation of unpublished opinions, should help alleviate
this problem; since most unpublished opinions are easily located through
electronic databases, it was only the courts prior ban on citation that allowed
such decisions to escape scrutiny. Now that judges know that all their
opinions are subject to citation, they are more likely to take them all seriously.
But the underlying problem is one of time and resources: As long as the
federal appellate docket is overloaded, judges will not be able to devote
sufficient time to each case. Increasing the number of courts or judges
might help, but is not a panacea, because it could create more inter- and
intracircuit conflicts. In the end, we have to decide which cases are most in
need of federal judicial resources, and cut back on the others. Our own view is
that Congress has been acting irresponsibly by increasing federal jurisdiction
in diversity cases (which do not need federal adjudication in the first place)
and by enacting too many federal criminal statutes that simply criminalize
behavior that is already punishable under state law. In part, this trend is
driven by a sense that federal judges are more capable of handling complex,
major cases, but a better solution would be federal funding to help state courts
develop this capacity. In any case, we cannot expect good decisions from
judges whose dockets are bloated with cases unworthy of their attention.
Finally, although judges are ultimately responsible for their own deci-
sions, they are subject to multiple influences and thus to multiple perspec-
tives. The adversary system ensures that all interested parties will have a say.
Most appellate judges have law clerks, whose short tenure and youthful
perspective can complement the judges experience. And to the extent that
judicial opinions are subject to critical scrutiny by law professors, journalists,
politicians, and the public, judges may have to refine their views over time.
We also have some improvements to suggest, especially at the Supreme
Court level. First, the problem of insufficient attention to each case is not
limited to the lower courts, although the problem is somewhat different in the
Supreme Court. It is not that they have too many casesindeed, we think
they decide too fewbut that they place unnecessary artificial limits on the
amount of time for each case. Two small changes could have a great effect:
Routinely allow more than one hour for oral argument, and permit cases to
carry over into the next term if some number of justices think more time
should be devoted to deliberation or to opinion writing.
Another problem is that the Courts focus on constitutional and other
politically salient cases has a tendency to make its job seem more exceptional,
and less like ordinary legal decision making. To the extent that the justices
themselves believe this, it might interfere with their judgment. Again, two
changes could help alleviate the problem by exposing the justices to a greater
number of more routine (albeit legally difficult) cases. The Supreme Courts
mandatory jurisdiction might be expanded, so that it is required to decide
more cases: for example, cases in which the circuit courts disagreein mid-
twentieth century the Court tried to resolve almost all circuit splits, but now it
ignores most of them. Another possibility would be to reinstate the justices
circuit-riding duties, requiring them to sit regularly with courts of appeals
and hear the gamut of cases (Sherry, 2005). Some appellate judges, including
Richard Posner on the Seventh Circuit and the late Chief Justice William
Rehnquist, have voluntarily chosen to sit as trial judges, and we suspect the
experience was good for them.
We also think that it is useful for judges to be exposed to other view-
points. The problem is that most people with whom they associate have every
reason to cater to their viewsfew lawyers will tell a sitting judge that his
views are wrong, and lower courts judges are equally unlikely to provide
challenges to the intellectual viewpoints of their reviewing courts. It is useful
for high court judges to meet with their peers: state supreme court judges with
colleagues in other states; U.S. Supreme Court justices with colleagues from
foreign constitutional courts. To encourage this, Congress should provide
funding to assist in meetings of state court judges and to finance travel to
foreign conferences, as well as funding to hold conferences with foreign jurists
on U.S. soil.
A final proposal would be to encourage peer review, especially at the
Supreme Court level. A panel of eminent retired judges, lawyers, and academics
might be created, on the understanding that subgroups of those with relevant
expertise would be selected to give feedback on Supreme Court opinions before
they are issued. A particularly noncontroversial form of this review would be to
require the circulation of opinions to retired justices for comment. Playing this
peer review role might make retirement more attractive to some justices,
increasing turnover and therefore potentially diversity on the Court. Retired
solicitors general and attorneys general might also be recruited with minimal
controversy. A more modest, and less controversial, program would be to
appoint a nonpartisan commission with the charge of issuing assessments of
the Courts performance on a regular basis.
Considered individually, it is possible that none of these proposals would
have a substantial effect. They might, however, have a significant cumulative
impact. They might also have some effect on the recruitment and retention of
judges, making the job more attractive to the kind of people we wish to have
as judges: those who enjoy being exposed to different viewpoints and who like
engaging in legal analysis rather than political rhetoric.
None of these existing or proposed structural features, of course, guar-
antees good decision making. A judge who is dogmatic and closed-minded or
whose hedgehog-like devotion to certainty or to an abstract theory is unshak-
able will not be swayed by colleagues or critics nor dissuaded by the impos-
sibility of writing a coherent opinion. In the end, then, it is the character traits
that are most important. Former Attorney General Nicholas Katzenbach put
it most eloquently when he told the Senate Judiciary Committee deliberating
on the nomination of Robert Bork: Were I in your position . . . the central
question I would be asking is this. Is Judge Bork a man of judgment? . . . Is he a
wise person? (Lewis, 1987). It is our hope that the project begun with this
book will eventually help us answer that question about future nominees.
Note
1. The leading case is Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402
(1971).
References
Abelson, R. P. (1968). Psychological implications. In R. Abelson, E. Aronson.

W. McGuire, T. Newcob, M. Rosenberg, & P. Tannenbaum (Eds.), Theories of
Cognitive Consistency: A Sourcebook. Chicago: Rand McNally.
Abelson, R. P., & Rosenberg, M. (1958). Symbolic psycho-logic: A model of attitudinal
cognition. Behavioral Science, 3, 18.
Abrams, D., Wetherill, M., Cochrane, S., Hogg, M. A., & Turner, J. C. (1990).
Knowing what to think by knowing who you are: Self-categorization and the
nature of norm formation, conformity and group polarization. British Journal of
Social Psychology, 29, 91119.
Ackerman, B. (1999). Taxation and the Constitution. Columbia Law Review, 99, 1.
Adams v. The New Jersey Steamboat Company, 151 N.Y. 163, 45 N.E. 369 (N.Y. Ct. of
Appeals, 1896).
Ajzen, I., & Fishbein, M. (1980). Understanding Attitudes and Predicting Social
Behavior. Englewood Cliffs, NJ: Prentice-Hall.
Albiston, C. R., & Nielsen, L. B. (2007). The procedural attack on civil rights: The
empirical reality of Buckhannon for the private attorney general. UCLA Law
Review, 54, 10871134.
Aldisert, R. J. (1982). What makes a good judge? Pennsylvania Lawyer, 4, 22.
Alexander, J. C. (1994). Judges self-interest and procedural rules: Comment on
Macey. Journal of Legal Studies, 23, 647.
Alexander, L. (1989). Constrained by precedent. Southern California Law Review, 63, 3.
Alexander, L. (1991). The gap. Harvard Journal of Law and Public Policy, 14, 695.
Alexander, L. (1996). Bad beginnings. University of Pennsylvania Law Review, 145, 5787.
Alexander, L., & Sherwin, E. (2001). The Rule of Rules: Morality, Rules, and the
Dilemmas of Law. Durham, NC: Duke University Press.
Alexander, L., & Sherwin, E. (2008). Demystifying Legal Reasoning. Cambridge, UK:
Cambridge University Press.
297
298 References
Aliotta, J. M. (1988). Social backgrounds, social motives, and participation on the U.S.
Supreme Court. Political Behavior, 10, 267284.
Alpert, L., Atkins, B. M., & Ziller, R. C. (1979). Becoming a judge: The transition from
advocate to arbiter. Judicature, 62, 325335.
Altman, S. (1990). Beyond candor. Michigan Law Review, 89, 296.
American Bar Association Committee on Judicial Performance Evaluation. (2005).
Black Letter Guidelines for the Evaluation of Judicial Performance with Commentary.
Anderson, A. A., Lindsay, J. L., and Bushman, B. J. (1999). Research in the
psychological laboratory: Truth or triviality? Current Directions in Psychological
Science, 8, 39
Anderson, M. C., & MacCoun, R. J. (1999). Goal conflict in juror assessments of
compensatory and punitive damages. Law & Human Behavior, 23, 313.
Arlen, J. (1998). The future of behavioral economic analysis of law. Vanderbilt Law
Review, 51, 17651788.
Arlen, J. H., & Talley, E. L. (in press). Introduction. In J. H. Arlen & E. L. Talley (Eds.),
Experimental Law and Economics, pp. 156. Cheltenham, UK: Edward Elgar
Publishing.
Aronson, E., Wilson, T. D., & Brewer, M. B. (1998). Experimentation in social
psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The Handbook of
Social Psychology, Vol. 1 (4th ed.). New York: McGraw-Hill.
Arrow, H., McGrath, J. E., & Berdahl, H. L. (2000). Small Groups as Complex Systems:
Formation, Coordination, Development, and Adaptation. Thousand Oaks, CA: Sage.
Asch, S. (1955). Opinions and social pressure. Scientific American, 193(5), 3135.
Asch, S. (1956). Studies of independence and conformity: A minority of one against a
unanimous majority. Psychological Monographs, 70(9), Whole Number 416.
Asch, S. E. (1940). Studies in the principles of judgments and attitudes: II.
Determination of judgments by group and by ego standards. The Journal of
Ashwander v. Tennessee Valley Authority, 297 U.S. 288 (1936).
Atkins, B. M., & Zavoina, W. (1974). Judicial leadership on the court of appeals: A
probability analysis of panel assignment in race relations cases on the Fifth
Circuit. American Journal of Political Science, 18, 701711.
Austin, J. (2006). Do the Supreme Court Justices Questions during Oral Arguments
Predict their Decisions? Honors Thesis, Department of Psychology, University of
Kansas, Lawrence.
Bainbridge, S. M., & Gulati, G. M. (2002). How do judges maximize? (The same way
everyone else doesboundedly): Rules of thumb in securities fraud opinions.
Emory Law Journal, 51, 83151.
Bales, R. F. (1950). Interaction Process Analysis. Cambridge: Harvard University Press.
Bandstra, R. A. (2005). Looking toward Lansing: Could you be a lawyer-legislator?
Michigan Bar Journal, 89, 2829.
Barber, J. D. (1965). The Lawmakers: Recruitment and Adaptation to Legislative Life.
New Haven, CT: Yale University Press.
Barberis, N., & Thaler, R. (2003). A survey of behavioral finance. In G. M.
Constantinides, M. Harris, & R. M. Stulz (Eds.), Handbook of the Economics of
Finance. Amsterdam: Elvesier.
Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of being.
American Psychologist, 54, 462479.
References 299
Bargh, J. A., & Ferguson, M. L. (2000). Beyond behaviorism: On the automaticity of

higher mental processes. Psychological Bulletin, 126, 925945.
Bargh, J. A., Gollwitzer, P. M., Lee-Chai, A., Barndollar, K., & Trotschel, R. (2001).
The automated will: Nonconscious activation and pursuit of behavioral goals.
Journal of Personality and Social Psychology, 81, 10141027.
Barnett, R. E. (2004). Restoring the lost Constitution: The presumption of liberty.
Princeton, NJ: Princeton University Press.
Barnett, R. E. (2005). Foreword: Limiting Raich. Lewis & Clark Law Review, 9, 743750.
Barnett, S. M., & Ceci, S. J. (2002). When and where do we apply what we learn? A
taxonomy for far transfer. Psychological Bulletin, 128(4), 612637.
Baron, J. (2004). Normative models of judgment and decision making. In D. J.
Koehler & N. Harvey (Eds.), Blackwell Handbook of Judgment and Decision
Making. Malden, MA: Blackwell.
Baron, R. E., Byrne, D., & Branscomb, N. R. (2006). Social Psychology (11th ed.). New
York: Pearson.
Bartels, B. L. (2005). Heterogeneity in Supreme Court Decision-Making: How Case-Level
Factors Alter Preference-Based Behavior. Paper presented at the annual meeting of
the Midwest Political Science Association, Chicago.
Bartels, B. L. (2006). Heterogeneity in Supreme Court Decision Making: How Situational
Factors Shape Preference-Based Behavior. Unpublished doctoral dissertation,
Ohio State University, Columbus.
Bartlett, K. (1990). Feminist legal methods. Harvard Law Review, 103, 829888.
Bates, D. W., & Gawande, A. A. (2000). Error in medicine: what have we learned?
Annals of Internal Medicine, 132, 763.
Baum, L. (1990). American Courts: Process and Policy (2nd ed.). Boston: Houghton-
Mifflin.
Baum, L. (1994). What judges want: Judges goals and judicial behavior. Political
Research Quarterly, 47, 749768.
Baum, L. (1997). The Puzzle of Judicial Behavior. Ann Arbor: University of Michigan
Press.
Baum, L. (2006). Judges and Their Audiences. Princeton, NJ: Princeton University Press.
Baum, L. (2007). The Supreme Court (9th ed.). Washington, DC: CQ Press.
Baumeister, R. F., & Newman, L. S. (1994). Self-regulation of cognitive inference and
decision processes. Personality and Social Psychology Bulletin, 20, 319.
Beck, R. C. (2000). Motivation: Theories and Principles. Upper Saddle River, NJ:
Prentice-Hall.
Becker, G. (1969). Crime and punishment: An economic analysis. Journal of Political
Economy, 79, 169.
BedRock Limited v. United States, 124 S.Ct. 1587 (2004).
Beebe, B. (2006). An empirical study of the multifactor tests for trademark
infringement. California Law Review, 94, 15811654.
Behrman, B. W. & Davey, S. L. (2001). Eyewitness identification in actual criminal
cases: An archival analysis. Law & Human Behavior, 25, 475.
Bell Atlantic Corp. v. Twombly, 27 S. Ct. 1955 (2007).
Bergara, M., Richman, B., & Spiller, P. T. (2003). Modeling Supreme Court strategic
decision making: The congressional constraint. Legislative Studies Quarterly, 28,
247280.
Berger, J., Fisek, M. H., Norman, R. Z., & Zelditch, M. Jr. (1977). Status Characteristics
in Social Interaction: An Expectation-States Approach. New York: Elsevier.
300 References
Berger, J, Wagner, D. G., & Zelditch, M. Jr. (1985). Expectation states theory. In
J.M. Berger & M. Zelditch (Eds.), Status, Rewards, and Influence. San Francisco:
Jossey-Bass.
Berlin, I. (1953). The Hedgehog and the Fox: An Essay on Tolstoys View of History.
Chicago, IL: Ivan R. Dee.
Bertrand, M., & Mullainathan, S. (2004). Are Emily and Greg more employable than
Lakisha and Jamal? A field experiment on labor market discrimination. The
American Economic Review, 94, 9911013.
Bickel, A. (1986). The Least Dangerous Branch. New Haven, CT: Yale University Press.
Biskupic, J. (2004, June 8). June is often the time of compromise in court. USA Today,
p. 2A.
Biskupic, J. (2005). Sandra Day OConnor. New York: HarperCollins.
Black, H. C. (1912). Handbook on the Law of Judicial Precedents. St. Paul, MN: West
Publishing Co.
Blair, I. V., Judd, C. M., & Chalpeau, K. M. (2004). The influence of Afrocentric facial
features in criminal sentencing. Psychological Science, 15, 674679.
Blanchette, I., & Dunbar, K. N. (2001a). The inVivo/inVitro approach to cognition:
The case of analogy. Trends in Cognitive Science, 5, 334339.
Blanchette, I., & Dunbar, K. N. (2001b). Analogy use in naturalistic settings: The
influence of audience, emotion, and goals. Memory & Cognition, 29, 730735.
Blanchflower v. Blanchflower, 834 A.2d 1010 (N.H. 2003).
Blanck, P. D. (1991). What empirical research tells us: Studying judges and juries
behavior. American University Law Review, 40, 775, 777.
Blanck, P. D., Rosenthal, R., & Cordell, L. H. (1985). The appearance of justice:
Judges verbal and nonverbal behavior in criminal jury trials. Stanford Law
Review, 38, 89.
Boot, M. (1998). Out of Order. New York: Basic Books.
Borchard, E. M. (1932). Convicting the Innocent: Sixty-five Actual Errors of Criminal
Justice. New Haven, CT: Yale University Press.
Bork, R. (1990). The Tempting of America. New York: Touchtone.
Bowen, T., & Scheb, J. M. (1993). Freshman opinion writing on the U.S. Supreme
Court, 19211991. Judicature, 79, 239243.
Bowers v. Hardwick, 478 U.S. 186 (1986).
Brace, P. R., & Hall, M. G. (1997). The interplay of preferences, case facts, context, and
rules in the politics of judicial choice. Journal of Politics, 59, 1206.
Braman, E. (2004). Motivated Reasoning in Legal Decision Making. Doctoral
dissertation, Ohio State University, Columbus.
Braman, E. (2006a). Reasoning on the threshold: Testing the separability of
preferences in legal decision making. Journal of Politics, 68, 308321.
Braman, E. (2006b, February 15). Exploring discretion in legal decision making: An
experimental approach. Paper prepared for Presentation at Law & Society
Workshop, University of Indiana University Law School.
Braman, E., & Nelson, T. E. (2007). Mechanism of motivated reasoning?: Analogical
perception in discrimination disputes. American Journal of Political Science,
51(4), 940956.
Brandeis, L. D. (1916). The living law. Illinois Law Review, 10, 461.
Breasted, J. H. (2001). Ancient Records of Egypt: Volume Two. The Eighteenth Dynasty.
Urbana-Champaign: University of Illinois Press.
References 301
Brehm, S. S., & Brehm, J. W. (1981). Psychological Reactance: A Theory of Freedom and
Control. New York: Academic Press.
Brenner, S., & Hagle, T. M. (1996). Opinion writing and acclimation effect. Political
Behavior, 18, 235261.
Breyer, S. (2005). Active Liberty. New York: Vintage Books.
Brenner, S., & Spaeth, H. J. (1995). Stare Indecisis. Cambridge, UK: Cambridge
University Press.
Brewer, M. (1979). In-group bias in the minimal intergroup situation: A cognitive-
motivational analysis. Psychological Bulletin, 86, 307324.
Brewer, M. (1991). The social self: On being the same and different at the same time.
Personality and Social Psychology Bulletin, 17, 475482.
Brewer, S. (1996). Exemplary reasoning: Semantics, pragmatics, and the rational force
of legal argument by analogy. Harvard Law Review, 109, 9251028.
Bright, S. B., & Keenan, P. J. (1995). Judges and the politics of death: Deciding between
the bill of rights and the next election in capital cases. Boston University Law
Review, 75, 759.
Brooks, R. R. W., & Raphael, S. (2003). Life terms or death sentences: The uneasy
relationship between judicial elections and capital punishment. Journal of
Criminal Law and Criminology, 92, 609.
Brudney, J. J., & Ditslear, C. (2001). Designated diffidence: District court judges on the
courts of appeals. Law & Society Review, 35, 565606.
Brust, R. (2005, September). Balancing act. American Bar Association Journal, 37.
Buchman, J. (2007). The effects of ideology on federal district judges decisions to
admit expert testimony. American Political Science Review, 35, 671.
Burnet v. Coronado Oil & Gas Co., 285 U.S. 393 (1932).
Burton, S. (2005). An Introduction to Law and Legal Reasoning (3d ed.). Frederick,
MD: Aspen Publishers.
Bush v. Vera, 517 U.S. 952 (1996).
Bushman, B. J. (1995). Moderating role of trait aggressiveness in the effects of
violent media on aggression. Journal of Personality and Social Psychology, 69, 950960.
Buthe, T. (2002). Taking temporality seriously: Modeling history and the use of
narratives as evidence. American Political Science Review, 96, 481494.
Byrne, M. D. (1995). The convergence of explanatory coherence and the story model:
A case study in juror decisions. In J. D. Moore & J. F. Lehman (Eds.), Proceedings
of the 17th Annual Conference of the Cognitive Science Society. Mahwah, NJ:
Lawrence Erlbaum Associates.
Caldeira, G. A. (1977). Judicial incentives: Some evidence from urban trial courts.
Iusticia, 4(2), 128.
Caldeira, G. A., & Wright, J. R. (1990). The discuss list: Agenda building in the
Supreme Court. Law & Society Review, 24, 807836.
California Supreme Court. (n.d.) Internal Operating Practices and Procedures of the
California Supreme Court. Retrieved August 27, 2007, from http://
www.courtinfo.ca.gov/courts/supreme/documents/sc082504.pdf
Callins v. Collins, 510 U.S. 1141 (1994).
Caminker, E. H. (1994). Why must inferior courts obey superior court precedents?
Stanford Law Review, 46, 817873.
Campbell, J. (2005). The Lives of the Chief Justices of England: From the Norman
Conquest Till the Death of Lord Mansfield. London: Elibron Classics.
302 References
Cardozo, B. N. (1921). The Nature of the Judicial Process. New Haven, CT: Yale
University Press.
Carp, R. A., & Rowland, C. K. (1983). Policymaking and Politics in the Federal District
Courts. Knoxville: University of Tennessee Press.
Cartwright, D. (1968) The nature of group cohesiveness. In D. Cartwright & A.
Zander (Eds.), Group Dynamics: Research and Theory(3d ed.). New York: Row,
Peterson, & Co.
Cartwright, D., & Zander, A. (1968). Origins of group dynamics. In D. Cartwright &
A. Zander (Eds.), Group Dynamics: Research and Theory(3d ed.). New York: Row,
Peterson, & Co.
Catrambone, R., & Holyoak, K. J. (1989). Overcoming contextual limitations on
problem-solving transfer. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 15, 11471156.
Cattell, R. B. (1963). Theory of fluid and crystallized intelligence: A critical
experiment. Journal of Education Psychology, 54, 122.
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of
source versus message cues in persuasion. Journal of Personality and Social
Psychology, 39, 752766.
Chaiken, S., Giner-Sorolla, R., & Chen, S. (1996). Beyond accuracy: Defense and
impression motives in heuristic and systematic information processing. In
P. M. Gollwitzer & J. A. Bargh (Eds.), The Psychology of Action: Linking
Cognition and Motivation to Behavior. New York: Guilford Press.
Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic
information processing within and beyond the persuasion context. In J. S.
Uleman & J. A. Bargh (Eds.), Unintended Thought. New York: Guilford Press.
Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-Process Theories in Social Psychology.
New York: Guilford Press.
Champagne, A. (1986). The selection and retention of judges in Texas. Southwestern
Law Journal, 40, 53.
Champagne, A. (2005). Tort reform and judicial selection. Loyola of Los Angeles Law
Review, 38, 1483.
Chapman, G. B., & Bornstein, B. H. (1996). The more you ask for, the more you get:
Anchoring in personal injury verdicts. Applied Cognitive Psychology, 10, 519, 527, 531.
Chapper, J. A., & Hanson, R. A. (1990). Intermediate Appellate Courts: Improving Case
Processing. Williamsburg, VA: National Center for State Courts.
Chen, S., & Chaiken, S. (1999). The heuristic-systematic model in its broader context.
In S. Chaiken & Y. Trope (Eds.), Dual-Process Theories in Social Psychology. New
York: Guilford Press.
Chi, M. T. H. (2006). Two approaches to the study of experts characteristics. In
K. A. Ericsson, N. Charness, P. J. Feltovich, & R. R. Hoffman (Eds.), The
Cambridge Handbook of Expertise and Expert Performance. Cambridge, UK:
Chi, M. T. H., Farr, M. J., & Glaser, R. (Eds.). (1988). The Nature of Expertise. Hillsdale,
NJ: Lawrence Erlbaum.
Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981). Categorization and representation
of physics problems by experts and novices. Cognitive Science, 5, 121152.
Choi, S. J., & Gulati, G. M. (2004a). Choosing the next supreme court justice: An
empirical ranking of judge performance. Southern California Law Review, 78,
23117.
References 303
Choi, S. J, & Gulati, G. M. (2004b). A tournament of judges? California Law Review,

92, 299322.
Christianson, S. (2004). Innocent: Inside Wrongful Conviction Cases. New York: New
York University Press.
Christie, G. C. (1969). Objectivity in law. Yale Law Journal, 78, 13111342.
Chutkow, D. M. (2007). The Executive Chief Justice: Appointments to the Judicial
Conference committees, May 31, 2007. Available from Social Science Research
Network. Web site: http://ssrn.com/abstract=990167
Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402 (1971).
Clark, H. R. (1995). Justice Brennan: The Great Conciliator. New York: Birch Lane Press.
Clayton, C. W. (1999). The Supreme Court and political jurisprudence: New and old
institutionalism. In C. W. Clayton & H. Gillman (Eds.), Supreme Court Decision-
Making. Chicago: University of Chicago Press.
Clemen, R. T. (1999). Does decision analysis work? Unpublished manuscript.
Cohen, F. (1935). Transcendental nonsense and the functional approach. Columbia
Law Review, 35, 809845.
Cohen, J. M. (2002). Inside Appellate Courts: The Impact of Court Organization on
Judicial Decision Making in the United States Courts of Appeals. Ann Arbor:
University of Michigan Press.
Cohen, L. J. (1981). Can human irrationality be experimentally demonstrated?
Behavioral and Brain Sciences, 4, 317370.
Cohen, M. A. (1991). Explaining judicial behavior, or whats unconstitutional about
the Sentencing Commission? Journal of Law, Economics, and Organization, 7, 183.
Coke, E. (1985). Commentary upon Littleton (18th ed., Charles Butler, Ed.) Legal
Classics Library. Delran, NJ: Gryphon Books. (Original work published 1628).
Collins, P., and Martinek, W. L. (2007). The small group context: Designated district
court judges on the United States Courts of Appeals. Paper presented at the
meeting of the Southern Political Science Association.
Cooksey, R. W. (1996). Judgment Analysis: Theory, Methods and Applications. San
Diego, CA: Academic Press.
Coombs, C. H., Dawes, R. M., & Tversky, A. (1970). Mathematical Psychology: An
Elementary Introduction. Englewood Cliffs, NJ: Prentice-Hall.
Cooper, J. B., Newbower, R. S., & Kitz., R. J. (1984). An analysis of major errors and
equipment failures in anesthesia management: Considerations for prevention
and detection. Anesthesiology, 60, 34.
Cooper, J. B., Newbower, R. S., Long, C. D., & McPeek, B. (1978). Preventable
anesthesia mishaps: A study of human factors. Anesthesiology, 49, 399.
Cooper, P. J. (1995). Battles of the Bench: Conflict inside the Supreme Court. Lawrence:
University of Kansas Press.
Cooter, R. D., & Ulen, T. S. (2007). Law and Economics (5th ed.). Boston: Pearson/
Addison Wesley.
The Courts uncompromising libertarian. (1975, November 24). Time Magazine, p. 69.
Craig v. Boren, 429 U.S. 190 (1976).
Croskerry, P. (2002). Achieving quality in clinical decision making: Cognitive
strategies and detection of bias. Academic Emergency Medicine, 9, 11841204.
Cross, F. B. (1997). Political science and the new legal realism: A case of unfortunate
interdisciplinary ignorance. Northwestern University Law Review, 92, 251.
Cross, F. B. (2003). Decisionmaking in the U.S. Courts of Appeals. California Law
Review, 91, 14571515.
304 References
Cross, F. B. (2004). Explaining U.S. Circuit Court Decision Making. Judicature, 88,
3135.
Cross, F. B. (2007). Decision Making in the U.S. Courts of Appeals. Stanford, CA:
Stanford University Press.
Cross, F., & Tiller, E. (1998). Judicial partisanship and obedience to legal doctrine:
Whistleblowing on the Federal Courts of Appeals. Yale Law Journal, 107,
21552176.
Cross, F. B., & Nelson, B. J. (2001). Strategic institutional effects on Supreme Court
decisionmaking. Northwestern University Law Review, 95, 14371493.
Cross, F. B., & Lindquist, S. A. (2007). The scientific study of judicial activism.
Minnesota Law Review, 91, 17521784.
Cross, R., & Harris, J. W. (1990). Precedent in English Law (4th ed.). Oxford:
Clarendon Press.
Cushman, B. (2000). Formalism and realism in Commerce Clause jurisprudence.
University of Chicago Law Review, 67, 1089.
Dagan, H. (2007). The realist conception of law. University of Toronto Law Journal, 57,
607660.
Dalgleish, L. I. (2003). Risk, needs and consequences. In M.C. Calder (Ed.),
Assessments in Child Care: A comprehensive Guide to Frameworks and Their Use.
(pp. 8699). Dorset, UK: Russell House Publishing.
Dalgleish, L., & Weiss, D. (2007). Evidence-based expertise: A theoretical framework and
some data. Paper presented at the annual meeting of the Brunswik Society, Long
Beach, CA.
Dancy, J. (1993). Moral Reasons. Oxford: Blackwell.
Danelski, D. J. (1960). The influence of the Chief Justice in the decisional process. In
W. M. Murphy & C. H. Pritchett (Eds.), Courts, Judges, and Politics. New York:
Random House.
Daniels, N. (1979). Wide reflective equilibrium and theory acceptance. Journal of
Philosophy, 76, 256.
Darley, J. M. (2001). The dynamics of authority influence in organization and the
unintended action consequences. In J. M. Darley, D. M. Messick, & T. R. Tyler
(Eds.), Social Influences on Ethical Behavior in Organizations. Mahwah, NJ:
L. Erlbaum.
Darley, J. M., Carlsmith, K. M., & Robinson, P. H. (2000). Incapacitation and
just deserts as motives for punishment. Law and Human Behavior, 24,
659683.
Darley, J. M., & Pittman, T. S. (2003). The psychology of compensatory and
retributive justice. Personality & Social Psychology Review, 7, 324336.
Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579 (1993).
Dekay, M. L. (1996). The difference between Blackstone-like error ratios and
probabilistic standards of proof. Law & Social Inquiry, 21(1), 95132.
Denniston, L. (2005, November 27). Court to rule on military tribunals. Web site:
http://www.scotusblog.com/movabletype/archives/2005/
Denzau, A., Riker, W., & Shepsle, K. (1985). Farquharson and Fenno: Sophisticated
voting and home style. American Political Science Review, 79, 117134.
Deutsch, M. (1975). Equity, equality, and need: What determines which value will be
used as the basis of distributive justice. Journal of Social Issues, 31, 137.
Devins, N. (2004). The judicial safeguards of federalism. Northwestern University Law
Review, 99, 131.
References 305
Devins, N., & Fisher, L. (2004). The Democratic Constitution. New York: Oxford
University Press.
Devins, N., & Meese, A. (2005). Judicial review and non-generalizable cases. Florida
State University Law Review, 32, 323.
Devitt, E. J. (1979). Ten commandments for the new judge. American Bar Association
Journal, 65, 574.
Dhami, M. K. (2003). Psychological models of professional decision making.
Psychological Science, 14, 175180.
Dhami, M. K. (2004). Conditional bail decision making in the magistrates court.
Howard Journal of Criminal Justice, 43(1), 2746.
Dhami, M. K. (2005) From discretion to disagreement: Explaining disparities in
judges pretrial decisions. Behavioral Sciences & the Law, 23(3), 367386.
Dobie, A. M. (1951). A judge judges judges. Washington University Law Quarterly,
1951, 471.
Dorff, R. H., & Steiner, J. (1981). Political decision making in face-to-face groups:
Theory, methods, and an empirical application in Switzerland. American Political
Science Review, 75, 368380.
Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A. (1997). On the
nature of prejudice: Automatic and controlled processes. Journal of Experimental
Downs, A. (1957). An Economic Theory of Democracy. New York: Harper & Row.
Drahozal, C. R. (1998). Judicial incentives and the appeals process. Southern Methodist
Law Review, 51, 469503.
Druckman, J. (2001). The limits of framing effects: Who can frame. Journal of Politics,
63(4), 104166.
Dubois, P. L. (1980). From Ballot to Bench: Judicial Elections and the Quest for
Accountability. Austin: University of Texas Press.
Duckworth, K. L., Bargh, J. A., Garcia, M., & Chaikin, S. (2002). The automatic
evaluation of novel stimuli. Psychological Science, 13, 513519.
Duncker, K. (1945). On problem-solving (L. S. Lees, Trans.). Psychological
Monographs, 58(5), Whole No. 270.
Duxbury, N. (1995). Patterns of American Jurisprudence. New York: Oxford University
Press.
Dworkin, R. (1986). Laws Empire. Cambridge, MA: Harvard University Press.
Eagly, A. H., & Chaiken, S. (1993). The Psychology of Attitudes. Fort Worth, TX:
Harcourt Brace Jovanovich.
Edwards, H. T. (1985). Public misperceptions concerning the politics of judging:
Dispelling some myths about the D.C. Circuit. University of Colorado Law Review,
56, 619.
Edwards, H. T. (1998). Collegiality and decision making on the D.C. Circuit. Virginia
Law Review, 83, 13351370.
Edwards, H. T. (2003). The effects of collegiality on judicial decision making.
University of Pennsylvania Law Review, 151, 16391689.
Edwards, W. (1983). Human cognitive capacities, representativeness, and ground
rules for research. In P. Humphreys, O. Svenson, & A. Vari (Eds.), Analyzing
and Aiding Decision Processes. Budapest, Hungary: Akademiai Kiado.
Edwards, W., & Newman, J. R. (1982). Multiattribute Evaluation. Beverly Hills, CA: Sage.
Einhorn, J. (1974). Expert judgment: Some necessary conditions and an example.
Journal of Applied Psychology, 59, 562571.
306 References
Eisenberg, M. A. (1988). The Nature of the Common Law. Cambridge, MA: Harvard
University Press.
Eisenberg, T., Garvey, S. P., & Wells, M. T. (2001). Forecasting life and death: Juror
race, religion, and attitude toward the death penalty. The Journal of Legal Studies,
30, 277311.
Eisenberg, T., Hannaford-Agor, P. L., Hans, V. P., Waters, N. L.; Munsterman, G. T.,
Schwab, S. J., et al. (2005). Judge-jury agreement in criminal cases: A partial
replication of Kalven and Zeisels The American Jury. Journal of Empirical Legal
Studies, 2(1), 171207.
Eisenberg, T., LaFountain, N., Ostrom, B., Rottman, D., & Wells, M. T. (2002). Juries,
judges, and punitive damages: An empirical study. Cornell Law Review, 87, 743,
747749.
Eisenstadt v. Baird, 405 U.S. 438 (1972).
Eisler, K. I. (1993). A Justice for All: William J. Brennan, Jr., and the Decisions That
Transformed America. New York: Simon & Schuster.
Ellsworth, P. C., & Mauro, R. (1998). Psychology and law. In D. T. Gilbert, S. T. Fiske
& G. Lindzey (Eds.), Handbook of Social Psychology (4th ed., pp. 684732). New
York: McGraw-Hill.
Elster, J. (1989). The Cement of Society. Cambridge, UK: Cambridge University Press.
Enelow, J. M., & Hinch, M. (1990). Advances in the Spatial Theory of Voting. New York:
Englich, B., Mussweiler, T., & Strack, F. (2006). Playing dice with criminal sentences:
The influence of irrelevant anchors on experts judicial decision making.
Personality and Social Psychology Bulletin, 32(2), 188200.
Epstein, L., & Jacobi, T. (2008). Super medians. Stanford Law Review, 61, 37.
Epstein, L., & Knight, J. (1998). The Choices Justices Make. Washington, DC: CQ Press.
Epstein, L., & Knight, J. (2000). Toward a strategic revolution in judicial politics: A
look back, a look ahead. Political Research Quarterly, 53, 625661.
Epstein, L., Knight, J., & Martin, A. (2003). The political (science) context of judging.
St. Louis University Law Journal, 47, 783.
Epstein, L., Segal, J. A., & Johnson, T. (1996). The claim of issue creation on the U.S.
Supreme Court. American Political Science Review, 90(4), 845852.
Ericsson, K. A. (2006). The influence of experience and deliberate practice on the
development of superior expert performance. In K. A. Ericsson, N. Charness,
P. J. Feltovich, & R. R. Hoffman (Eds.), The Cambridge Handbook of Expertise and
Expert Performance. New York: Cambridge University Press.
Ericsson, K. A., Charness, N., Feltovich, P. J., & Hoffman, R. R. (Eds.). (2006). The
Cambridge Handbook of Expertise and Expert Performance. New York: Cambridge
University Press.
Ericsson, K. A., & Ward, P. (2007). Capturing the naturally occurring superior
performance of experts in the laboratory: Toward a science of expert and
exceptional performance. Current Directions in Psychological Science, 16, 346350.
Eskridge, W. N., Jr. (1991). Reneging on history? Playing the court/congress/president
civil rights game. California Law Review, 79, 613684.
Estrich, S. R., & Sullivan, K. M. (1989). Abortion politics: Writing for an audience of
one. University of Pennsylvania Law Review, 138, 119.
Ettenson, R., Shanteau, J., & Krogstad, J. (1987). Expert judgment: Is more
information better? Psychological Reports, 60, 227238.
References 307
Evans, G., Heath, A., & Lalljee, M. (1996). Measuring left-right and libertarian-
authoritarian values in the British electorate. British Journal of Sociology,
47(1), 93112.
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating
the use of exploratory factor analysis in psychological research. Psychological
Methods, 4, 27299.
Farber, D. A. (1992). The inevitability of practical reason: Statutes, formalism, and the
rule of law. Vanderbilt Law Review, 45, 533559.
Farber, D. A. (1994). Missing the play of intelligence. William & Mary Law Review,
36, 147171.
Farber, D. A. (1995). Reinventing Brandeis: Legal pragmatism for the 21st century.
University of Illinois Law Review, 1995, 163190.
Farber, D. A. (2005). Supreme Court selection and measures of past judicial
performance. Florida State University Law Review, 32, 117596.
Farber, D. A, & Adams, E. (1999). Beyond the formalism debate: Expert reasoning,
fuzzy logic, and complex statutes. Vanderbilt Law Review, 52, 12431340.
Farber, D. A., & Sherry, S. (2002). Desperately Seeking Certainty: The Misguided Quest
for Constitutional Foundations. Chicago: University of Chicago Press.
Farber, D. A., & Sherry, S. (2008). Judgment Calls: Principle and Politics in
Constitutional Law. New York: Oxford University Press.
Farrell, R. C. (1999). Successful rational basis claims in the Supreme Court from the
1971 term through Romer v. Evans. Indiana Law Journal, 32, 357.
Farthing-Capowich, D. P. (1985). Designing programs to evaluate judicial
performance: Participation is key to success. Supreme Court Journal, 9, 22.
Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino & E. T.
Higgins (Eds.), Handbook of Motivation and Cognition. New York: Guilford Press.
Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: The MODE
model as integrative framework. In M. P. Zanna (Ed.), Advances in Experimental
Social Psychology. San Diego, CA: Academic Press.
Fazio, R. H. (1995). Attitudes as object-evaluation associations: determinants,
consequences, and correlates of attitude accessibility. In R. E. Petty & J. A. Krosnick
(Eds.), Attitude Strength: Antecedents and Consequences. Hillsdale, NJ: Erlbaum.
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1982). Attitude
accessibility, attitude-behavior consistency, and the strength of the object-
evaluation association. Journal of Experimental Social Psychology, 18, 339357.
Fazio, R. H., & Towles-Schwen, T. (1999). The MODE model of attitude-behavior
processes. In S. Chaiken & Y. Trope (Eds.), Dual Process Theories in Social
Psychology. New York: Guilford Press.
Fazio, R. H., & Williams, C. J. (1986). Attitude accessibility as a moderator of the
attitude-perception and attitude-behavior relations: An investigation of the 1984
presidential election. Journal of Personality and Social Psychology, 51, 505514.
FCC v. Beach Communications, Inc., 508 U.S. 307, 315 (1993).
Feigenson, N. R. (2003). Can tort juries punish competently? Chicago-Kent Law
Review, 78, 239.
Feltovich, P. J., Prietula, M. J., & Ericsson, K. A. (2006). Studies of expertise from
psychological perspectives. In K. A. Ericsson, N. Charness, P. J. Feltovich, & R. R.
Hoffman (Eds.), The Cambridge Handbook of Expertise and Expert Performance.
New York: Cambridge University Press.
308 References
Fennell, L. A. (1999). Between monster and machine: Rethinking the judicial function.
Southern California Law Review, 51, 183.
Fenno, R. M. (1973). Congressmen in Committees. Boston: Little, Brown.
Festinger, L. (1957). A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson.
Festinger, L. (1968). Informal social communication. In D. Cartwright & A. Zander
(Eds.), Group Dynamics: Research and Theory (3d ed.). New York: Row,
Peterson, & Co.
Festinger, L., Schacter, S., & Back, K. (1968). Operation of group standards. In D.
Cartwright & A. Zander (Eds.), Group Dynamics: Research and Theory (3d ed.).
New York: Row, Peterson, & Co.
Fiorina, M. P., & Plott, C. R. (1978). Committee decisions under majority rule: An
experimental study. American Political Science Review, 72, 575598.
Fischman, J. B. (2008). Decision-making under a norm of consensus: A structural
analysis of three-judge panels. Unpublished manuscript.
Fishburn, P. C. (1991). Decision theory: The next 100 years? Economic Journal, 101, 27.
Fisher, W. W., III, Horwitz, M. J., & Reed, T.A. (Eds.). (1993). American Legal Realism.
New York: Oxford University Press.
Fiske, S. T. (2003). Five core social motives, plus or minus five. In S. Spencer, S. Fein,
M. P. Zanna, & J. M. Olson, Motivated Social Perception. Mahwah, NJ: Lawrence
Erlbaum Associates.
Fiske, S. T., and Taylor, S. E. (1991). Social Cognition. New York: McGraw-Hill.
Fong, G. T., Krantz, D. H., & Nisbett, R. E. (1986). The effects of statistical training on
thinking about everyday problems. Cognitive Psychology, 18, 253.
Forgas, J. P. (Ed.). (2000). Feeling and Thinking: The Role of Affect in Social Cognition.
Foroni, F. & Mayr, U. (2005). The power of a story: New, automatic associations from
a single reading of a short scenario. Psychonomic Bulletin & Review, 12, 139144.
Forsyth, D. R. (1999). Group Dynamics (3d ed.). Belmont, CA: Wadsworth.
Foucha v. Louisiana, 504 U.S. 71 (1992).
Frank, J. (1930). Law and the Modern Mind. New York: Brentanos.
Frankel, M. E. (1975). The search for truth: An umpireal view. University of
Pennsylvania Law Review, 123, 1031.
Frankel, M. E. (1976). The adversary judge. Texas Law Review, 54, 465.
Frankfurter, F. (1953). Chief justices I have known. Virginia Law Review, 39, 83.
Frey, D. (1986). Recent research on selective exposure to information. In L. Berkowitz
(Ed.), Advances in Experimental Social Psychology (vol. 19). New York: Academic
Press.
Frickey, P. (1996). The fool on the hill: Congressional findings, constitutional
adjudication, and United States v Lopez. Case Western Law Review, 46, 695.
Fried, C. (1981). The artificial reason of the law or: What lawyers know. Texas Law
Review, 60, 3558.
Friedman, B. (2006). Taking law seriously. Perspectives on Politics, 4, 261276.
Friedman, L. (2002). The Rehnquist court: Some more or less historical comments. In
M. Belky (Ed.), The Rehnquist Court: A Retrospective. New York: Oxford
University Press.
Fumerton, R. A. (1990). Reason and Morality. Ithaca, NY: Cornell University Press.
Funder, D. C. (1999). Personality Judgment: A Realistic Approach to Person Perception.
San Diego, CA: Academic Press.
Funder, D. C. (2001). The Personality Puzzle. New York: W.W. Norton & Company.
References 309
Furgeson, J. R., Babcock, L., & Shane, P. M. (2008a). Behind the mask of method:
Political orientation and constitutional interpretive preferences. Law and Human
Behavior, 32, 502510.
Furgeson, J. R., Babcock, L., & Shane, P. M. (2008b). Do a laws policy implications
affect beliefs about its constitutionality? An experimental test. Law and Human
Behavior, 32, 21927.
Gaeth, G. J., & Shanteau, J. (1984). Reducing the influence of irrelevant information
on experienced decision makers. Organizational Behavior and Human
Performance, 33, 263282.
Gaffney, E. M., Jr. (1994). The importance of dissent and the imperative of judicial
civility. Valparaiso University Law Review, 28, 583646.
Galanter, M. (2004). The vanishing trial: An examination of trials and related matters
in federal and state courts. Journal of Empirical Legal Studies, 1, 459.
Gardner, H. (1985). The Minds New Science: The History of the Cognitive Revolution.
New York: Basic Books.
Garrett, B. (2008). Judging innocence. Columbia Law Review, 108, 55142.
Gatowski, S. I., Dobbin, S. A., Richardson, J. T., Ginsburg, G. P., Merlino, M. L., & Dahir,
V. (2001). Asking the gatekeepers: A national survey of judges on judging expert
evidence in a post-Daubert world. Law and Human Behavior, 25, 433, 442.
Gavison, R. (1988). The implications of jurisprudential theories for judicial election,
selection, and accountability. Southern California Law Review, 61, 1617.
Gawande, A. (2000). When doctors make mistakes. In J. Gleick & J. Cohen (Eds.), The
Best American Science Writing 2000. New York: HarperCollins.
Gellis, S. L. (1985). Reasons for case reversal in Texas. St. Marys Law Journal, 16, 299.
Gelman, S. (2001). The hedgehog, the fox, and the minimalist. Georgetown Law
Journal, 89, 22972350.
Gentner, D. (1983). Structure-mapping: A theoretical framework. Cognitive Science, 7,
155170.
Gentner, D. (1989). The mechanisms of analogical reasoning. In S. Vosniadou & A.
Ortony (Eds.), Similarity and Analogical Reasoning. London: Cambridge
University Press.
Gentner, D. (1998). Analogy. In W. Bechtel and G. Graham (Eds.), Companion to
Cognitive Science. Malden, MA: Blackwell.
Gentner, D., & Kurtz, K. (2006). Relations, objects, and the composition of analogies.
Cognitive Science, 30, 609642.
Gentner, D., & Kurtz, K. (2007). Relational categories. In W. Ahn, R. L. Goldstone, B.
C. Love, A. B. Markman, & P. Wolff (Eds.), Categorization Inside and Outside the
Laboratory: Essays in Honor of Douglas L. Medin. Washington, DC: American
Psychological Association.
Gentner, D., & Loewenstein, J. (2002). Relational language and relational thought. In E.
Amsel & J. P. Byrnes (Eds.), Language, Literacy, and Cognitive Development: The
Development and Consequences of Symbolic Communication. Mahwah, NJ: Erlbaum.
Gentner, D., & Markman, A. B. (1994). Structural alignment in comparison: No
difference without similarity. Psychological Science, 5, 152158.
Gentner, D., Rattermann, M. J., & Forbus, K. D. (1993). The roles of similarity in
transfer: Separating retrievability from inferential soundness. Cognitive
George, T. E. (1998). Developing a positive theory of decisionmaking on the U.S.
Courts of Appeals. Ohio State Law Journal, 58, 1635.
310 References
George, T. E. (1999). The dynamics and determinants of the decision to grant en banc
review. Washington University Law Review, 74, 1635.
George, T. E., & Yooh, A. H. (n.d.). Chief judges: The limits of attitudinal theory and
possible paradox of managerial judging. Available at Social Science Research
Network: http://ssrn.com/abstract=1001247.
George, A. (1980). Presidential Decisionmaking in Foreign Policy: The Effective Use of
Information and Advice. Boulder, CO: Westview Press.
George, T. E. (2001). Court fixing. Arizona Law Review, 43, 962.
George, T. E., & Epstein, L. (1992). On the nature of Supreme Court decision making.
American Political Science Review 86, 323337.
Gerber, S. D., & Park, K. (1997). The quixotic search for consensus on the U.S.
Supreme Court: A cross-judicial empirical analysis of the Rehnquist Court
justices. American Political Science Review, 91, 390408.
Gey, S. G., & Rossi, J. (2005). Empirical measures of judicial performance: An
introduction to the symposium. Florida State University Law Review, 32, 10011014.
Gibson, J. L. (1980). Environmental constraints on the behavior of judges: A representational
model of judicial decision making. Law and Society Review, 14, 343370.
Gibson, J. L. (1991). Decision making in appellate courts. In J. B. Gates & C. A. Johnson
(Eds.), The American Courts: A Critical Assessment. Washington, DC: CQ Press.
Gibson, J. L. (2006). The Legitimacy of the United States Supreme Court in a Polarized
Polity. Working paper.
Gick, M. L., & Holyoak, K. J. (1980). Analogical problem solving. Cognitive Psychology,
12, 306355.
Gick, M. L., & Holyoak, K. J. (1983). Schema induction and analogical transfer.
Cognitive Psychology, 15, 138.
Gigerenzer, G. (1993). The superego, the ego, and the id in statistical reasoning. In G.
Keren & C. Lewis (Eds.), A Handbook for Data Analysis in the Behavioral Sciences:
Methodological issues. Hillsdale, NJ: Erlbaum.
Gigerenzer, G. (2000). Adaptive Thinking: Rationality in the Real World. New York:
Oxford University Press.
Gigerenzer, G. (2006). Heuristics. In G. Gigerenzer & C. Engel (Eds.), Heuristics and
the Law. Cambridge, MA: The MIT Press.
Gigerenzer, G., & Selten, R. (Eds.). (2001). Bounded Rationality: The Adaptive Toolbox.
Cambridge, MA: MIT Press.
Gillman, H. (1993). The Constitution Besieged. Durham, NC: Duke University Press.
Gillman, H. (1997). Placing judicial motives in context: A response to Lee Epstein and
Jack Knight. Law and Courts, 7 (Spring), 1013.
Gillman, H. (1999). The Court as an idea. In C. W. Clayton & H. Gillman (Eds.),
Supreme Court Decision-Making. Chicago: University of Chicago Press.
Gillman, H. (2001). Whats law got to do with it? Judicial behavioralists test the legal
model of judicial decision making. Law & Social Inquiry, 26, 465504.
Gilovich, T. (1981). Seeing the past in the present: The effect of associations to familiar
events on judgments and decisions. Journal of Personality and Social Psychology,
40, 797808.
Gilovich, T., Griffin, D. W., & Kahneman, D. (Eds.). (2002). Heuristics and Biases:
The Psychology of Intuitive Judgment. Cambridge, UK: Cambridge University Press.
Ginossar, Z., & Trope, Y. (1987). Problem solving in judgment under uncertainty.
References 311
Glaser, R., & Chi, M. T. H. (1998). Overview. In M. T. H. Chi, R. Glaser, & M. J. Farr
(Eds.), The Nature of Expertise. Hillsdale, NJ: Erlbaum.
Glockner, A. (2007, August). Solving complex legal cases using the intuitive system.
Paper presented at the Subjective Probability, Utility, and Decision Making
Conference, Warsaw, Poland.
Glockner, A., Betsch, T., & Schindler, N. (n.d.). Construction of probabilistic inferences
by constraint satisfaction. Unpublished manuscript.
Goffman, E. (1959). The Presentation of Self in Everyday Life. New York: Doubleday.
Goldberg, L. R., & Werts, C. E. (1966). The reliability of clinicians judgments: A
multitrait-multimethod approach. Journal of Clinical Psychology, 30, 199206.
Goldberg, S. (2005). Federal judges and the Heisman trophy. Florida State University
Law Review, 32, 12371244.
Golde, R. A. (1969). Can You Be Sure of Your Experts? New York: Award Books.
Golding, M. P. (1963). Principled decision-making and the Supreme Court. Columbia
Law Review, 63, 3558.
Gollwitzer, P. M., & Bargh, J. A. (2005) Automaticity in goal pursuit. In A. J. Elliot &
C. S. Dweck (Eds.), Handbook of Competence and Perception. New York: Guilford
Press.
Graber, M. A. (2006). Dred Scott and the Problem of Constitutional Evil. New York:
Graber, M. L., Franklin, N., & Gordon, R. (2005). Diagnostic error in internal
medicine. Archives of Internal Medicine, 165, 14931499.
Granberg, D., & Bartels, B. (2005). On being a lone dissenter. Journal of Applied Social
Greenawalt, K. (1978). The enduring significance of neutral principles. Columbia Law
Review, 78, 9821021.
Greenburg, J. C. (2007). Supreme Conflict: The Inside Story of the Struggle for Control of
the United States Supreme Court. New York: Penguin Press.
Greenhouse, L. (2002, April 18). U.S. joins inmate in prison discipline case. New York
Times, p. A23.
Greenhouse, L. (2004, June 2). Teenagers murder conviction is reinstated after
challenge. New York Times, p. A17.
Greenhouse, L. (2005). Becoming Justice Blackmun: Harry Blackmuns Supreme Court
Journey. New York: Times Books.
Griswold v. Connecticut, 381 U.S. 479 (1965).
Groopman, J. (2007). How Doctors Think. New York: Houghton Mifflin.
Groscup, J. L. (2004). Judicial decision making about expert testimony in the aftermath
of Daubert and Kumho. Journal of Forensic Psychology Practice, 4(2), 5766.
Gruenfeld, D. H. (1995). Status, ideology, and integrative complexity on the U.S.
Supreme Court: Rethinking the politics of political decision making. Journal of
Personality and Social Psychology, 68, 520.
Grutter v. Bollinger, 539 U.S. 306 (2003).
Gruenfeld, D., & Hollingshead, A. B. (1993). Sociocognition in work groups: The
evolution of integrative complexity and its relation to task performance. Small
Groups Research, 24, 383405.
Gruenfeld, D. H., & Preston, J. (2000). Upending the status quo: Cognitive complexity
in Supreme Court justices who overturn legal precedent. Personality and Social
Psychology Bulletin, 26, 10131022.
312 References
Gryski, G. S., Main, E. C., & Dixon, W. J. (1986). Models of high court decision
making in sex discrimination cases. Journal of Politics, 48(1), 143155.
Gulati, M., & Choi, S. J. (2005). Which judges write their opinions (and should we
care?). Florida State University Law Review, 32, 10771122.
Gulati, M., & Choi, S. J. (2007). Ranking judges according to citation bias. Notre Dame
Law Review, 82, 12791309.
Gulati, M., & Sanchez, V. (2002). Giants in a world of pygmies? Testing the
superstar hypothesis with judicial opinions in casebooks. Iowa Law Review, 87, 1141.
Guthrie, C., & George, T. E. (2005). The futility of appeal: Disciplinary insights into
the affirmance effect on the United States Courts of Appeals. Florida State
University Law Review, 32, 357.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2001). Inside the judicial mind. Cornell
Law Review, 86, 777830.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2002). Judging by heuristic: Cognitive
illusions in judicial decision making. Judicature, 86, 44.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2007a). Blinking on the bench: How
judges decide cases. Cornell Law Review, 93, 1.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2007b). Inside the bankruptcy judges
mind. Boston University Law Review, 86, 1227.
Guthrie, C., Rachlinski, J. J., & Wistrich, A. J. (2009). The hidden judiciary: An
empirical examination of executive branch justice. Duke Law Journal, 58,
14771530. Hagerstrom, A. (1953). Inquiries into the Nature of Law and Morals
(C. D. Broad, Trans.). Stockholm: Almquist & Wiksell.
Hagle, T. M. (1993). Freshman effects for supreme court justices. American Journal of
Political Science 37, 11421157.
Hall v. Brooklands Auto Racing Club, K.B. 205 (1933).
Hall v. Brooklands Auto-Racing Club, 1 KB 205 (1933).
Hall, M. G. (1992). Electoral politics and strategic voting in state supreme courts.
Journal of Politics, 54, 427446.
Hall, M. G. (1995). Justices as representatives: Elections and judicial politics in the
American states. American Politics Quarterly, 23, 485503.
Hall, R. L. (1996). Participation in Congress. New Haven, CT: Yale University Press.
Hamdan v. Rumsfeld, 126 S.Ct. 2749 (2006).
Hammond K. R. (1996). Human Judgement and Social Policy: Irreducible Uncertainty,
Inevitable Error and Unavoidable Injustice. Oxford: Oxford University Press.
Hammond, K. R. (2000a). Coherence and correspondence theories in judgment and
decision-making. In T. Connolly, H. R. Arkes, & K. R. Hammond (Eds.),
Judgment and Decision Making: An Interdisciplinary Reader. Cambridge, UK:
Hammond, K. R. (2000b). Judgments Under Stress. Oxford: Oxford University Press.
Hammond, T. H., Bonneau, C. W, & Sheehan, R. S. (2005). Strategic Behavior and
Policy Choice on the U.S. Supreme Court. Stanford, CA: Stanford University Press.
Hand, L. (1930). Sources of tolerance. University of Pennsylvania Law Review, 79, 1.
Hansford, T. G., & Spriggs, J. F., II. (2006). The Politics of Precedent on the U.S.
Supreme Court. Princeton, NJ: Princeton University Press.
Hanson, J., & Yosifon, D. (2004). The situational character: A critical realist
perspective on the human animal. Georgetown Law Journal, 93, 1177.
Hare, A. P., Borgatta, E. F., & Bales, R. F. (Eds.) (1965). Small Groups: Studies in Social
Interaction, Revised Edition. New York: Alfred A. Knopf.
References 313
Harman, G., & Kulkarni, S. R. (2006). The problem of induction. Philosophy and
Phenomenological Research, 72, 559575.
Hart, H. L. A. (1961). The Concept of Law. Oxford: Clarendon Press.
Haslett, D. W. (1987). What is wrong with reflective equilibria? Philosophical
Quarterly, 37, 305.
Hastie, R. (2001). Problems for judgment and decision making. Annual Review of
Hastie, R., Schkade, D. A., & Payne, J. W. (1999). Juror judgments in civil cases: Effects
of plaintiffs requests and plaintiffs identity on punitive damage awards. Law &
Human Behavior, 23, 445.
Hausegger, L., & Baum, L. (1999). Inviting congressional action: A study of Supreme
Court motivations in statutory interpretation. American Journal of Political
Science, 43,162185.
Hausegger, L., & Haynie, S. (2003). Judicial decisionmaking and the use of panels in
the Canadian Supreme Court and the South African Appellate Division. Law &
Society Review, 37, 635657.
Haynie, S. L. (1992). Leadership and consensus on the U.S. Supreme Court. Journal of
Politics, 54, 11581169.
Heath, C., Larrick, R. P., & Klayman, J. (1998). Cognitive repairs: How organizational
practices can compensate for individual shortcomings. Research in Organizational
Behavior, 20, 137.
Heider, F. (1946). Attitudes and cognitive organization. Journal of Psychology, 21,
107111.
Heider, F. (1958). The Psychology of Interpersonal Relations. New York: Wiley.
Helmke, G. (2005). Courts under Constraints: Judges, Generals, and Presidents in
Argentina. New York: Cambridge University Press.
Henderson, L. (1987). Legality and empathy. Michigan Law Review, 85, 1574.
Hersch, J., and Viscusi, W. K. (2004). Punitive damages: How judges and juries
perform. Journal of Legal Studies, 33, 1.
Hettinger, V. A., Lindquist, S. A., & Martinek, W. L. (2003). Acclimation effects on the
United States Courts of Appeals. Social Science Quarterly, 84, 792810.
Hettinger, V. A., Lindquist, S. A., & Martinek, W. L. (2006). Judging on a Collegial
Court: Influences on Federal Appellate Decision Making. Charlottesville: University
of Virginia Press.
Higgins, R. W., & Rubin, P. H. (1980). Judicial discretion. Journal of Legal Studies, 9,
129138.
Hinckley, B. (1979). Twenty-one variables beyond the size of winning coalitions.
Journal of Politics, 41,192212.
Hirsch, H. N. (1981). The Enigma of Felix Frankfurter. New York: Basic Books.
Ho, D. E., & Quinn, K. M. (2007). Assessing political positions of media. Stanford Law
and Economics Olin Working Paper No. 343.
Hoffman, P., Slovic, P., & Rorer, L. (1968). An analysis of variance model for the
assessment of configural cue utilization in clinical judgment. Psychological
Bulletin, 69, 338349.
Hofstadter, D. R., & Mitchell, M. (1994). An overview of the copycat project. In K. J.
Holyoak & J. A. Barnden (Eds.), Advances in Connectionist and Neural
Computation Theory: Vol. 2. Analogical Connections. Norwood, NJ: Ablex.
Hollander, E. P. (1960). Competence and conformity in the acceptance of influence.
Journal of Abnormal and Social Psychology, 61, 365369.
314 References
Holmes, O. W. (1881). The Common Law. Boston: Little Brown.

Holmes, O. W. (1897). The path of the law. Harvard Law Review, 10, 457.
Holyoak, K. J. (1991). Symbolic connectionism: Toward third generation theories of
expertise. In K. A. Ericsson & J. Smith (Eds.), Toward a General Theory of
Expertise: Prospects and Limits. New York: Cambridge University Press.
Holyoak, K. J. (2005). Analogy. In K. J. Holyoak & R. G. Morison (Eds.), The
Cambridge Handbook of Thinking and Reasoning. Cambridge, UK: Cambridge
University Press.
Holyoak, K. J., & Koh, K. (1987). Surface and structural similarity in analogical
transfer. Memory & Cognition, 15, 332340.
Holyoak, K. J., & Simon, D. (1999). Bidirectional reasoning in decision making
by constraint satisfaction. Journal of Experimental Psychology: General, 128,
331.
Holyoak, K. J., & Thagard, P. (1989). Analogical mapping by constraint satisfaction.
Cognitive Science: A Multidisciplinary Journal, 13, 295355.
Holyoak, K. J., & Thagard, P. (1995). Mental Leaps: Analogy in Creative Thought.
Cambridge MA: MIT Press.
Hope v. Pelzer, 122 S.Ct. 2508 (2002).
Horn, J., & Masunaga, H. (2006). A merging theory of expertise and intelligence. In K.
A. Ericsson, N. Charness, P. J. Feltovich, & R. R. Hoffman (Eds.), The Cambridge
Handbook of Expertise and Expert Performance. New York: Cambridge University
Press.
Howard, J. W., Jr. (1981). Courts of Appeals in the Federal Judicial System: A Study of
the Second, Fifth, and District of Columbia Circuits. Princeton: Princeton
University Press.
Howard, R. M. and J. A. Segal. 2002. An original look at originalism. Law & Society
Review, 36(1), 113138.
Hsee, C. K. (1996). Elastic justification: How unjustifiable factors influence
judgments. Organizational Behavior and Human Decision Processes, 66, 122129.
Huber, G. A., & Gordon, S. C. (2004). Accountability and coercion: Is justice blind
when it runs for office? American Journal of Political Science, 48, 247.
Hudson v. McMillian, 503 U.S. 1 (1992).
Hughes, H. D. (1917). An interesting corn seed experiment. The Iowa Agriculturalist,
17, 424425.
Hummel, J. E., & Holyoak, K. J. (2003). A symbolic-connectionist theory of relational
inference and generalization. Psychological Review, 110, 220264.
Hunter, D. (2001). Reason is too large: Analogy and precedent in law. Emory Law
Journal, 50, 11971264.
Hurd, H. M. (1991). Challenging authority. Yale Law Journal, 100, 16111677
Hurd, H. M. (1992). Justifiably punishing the justified. Michigan Law Review, 90, 203.
Hurd, H. M. (1999). Moral Combat. Cambridge, UK: Cambridge University Press.
Hurwitz, M.S., & Stefko, J.V. (2004). Acclination and attitudes: Newcomer justices
and precedent conformance on the Supreme Court. Political Research Quarterly,
57, 121129
Hutcheson, J. C., Jr. (1929). The judgment intuitive: The role of the hunch in
judicial decision. Cornell Law Quarterly, 14, 274307.
In re Code of Judicial Conduct, 643 So. 2d 1037, (Fla. 1994).
Isikoff, M. (2006, April 3). Detainees rights: Scalia speaks his mind. Newsweek,
147(14), 6.
References 315
Jacavone, J., & Dostal, M. (1992). A descriptive study of nursing judgment in assessment
and management of cardiac pain. Advances in Nursing Science, 15, 5463.
Jacob, H. (1983). Trial courts in the United States: The travails of explanation. Law
and Society Review, 17, 407423.
Janis, I. L. (1982). Groupthink (2nd ed.). Boston: Houghton Mifflin.
Johnson, T. R. (2004). Oral Arguments and Decision Making on the United States
Supreme Court. Albany: State University of New York Press.
Johnson, T. R., Spriggs, J. F., II, & Wahlbeck, P. F. (2007). Supreme Court oral
advocacy: Does it affect the justices decisions? Washington University Law
Review, 85, 457.
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political conservatism
as motivated social cognition. Psychological Bulletin, 129, 339375.
Judicial Council of California. (2007). The Strategic Plan for Californias Judicial
Branch 20062012. San Francisco: Judicial Council of California, Administrative
Office of the Courts.
Kahneman, D. (1991). Judgment and decision making: A personal view. Psychological
Science, 2, 142145.
Kahneman, D. (1997). New challenges to the rationality assumption. Legal Theory, 3, 105.
Kahneman, D., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment Under Uncertainty:
Heuristics and Biases. New York: Cambridge University Press.
Kahneman, D., & Tversky, A. (1979) Prospect theory: An analysis of decision under
risk. Econometrica, 47, 263291.
Kahneman, D., & Tversky, A. (1981). The framing of decisions and the psychology of
choice. Science, 211, 453458
Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American
psychologist, 39, 341350.
Kahneman, D., & Tversky, A. (Eds.). (2000). Choices, Values, and Frames. New York:
Kalman, L. (1986). Legal Realism at Yale, 192760. Chapel Hill: University of North
Carolina Press.
Kalven, H,. & Zeisel, H. (1966). The American Jury. Boston: Little, Brown.
Kansas v. Colorado, 121 S.Ct. 2023 (2001).
Kassin, S. M., Meissner, C. A., & Norwick, R. J.(2005). Id know a false confession if I
saw one: A comparative study of college students and police investigators. Law
and Human Behavior, 29, 211227.
Kennedy, D. (1986). Freedom and constraint in adjudication: A critical
phenomenology. Journal of Legal Education, 36, 518547
Kerlinger, F. N. (1967). Social attitudes and their criterial referents: A structural
theory. Psychological Review, 74, 110122.
Kerr, N. & Tindale, R. S. (2004). Group performance and decision making. Annual
Review of Psychology, 55, 623.
Kerr, N. L., Atkin, R. S., Stasser, G., Meek, D., Holt, R. W., & Davis, J. H. (1976). Guilt
beyond a reasonable doubt: Effects of concept definition and assigneddecisionrule on
judgments of mock jurors. Journal of Personality and Social Psychology, 34, 282294.
Kida, T. (1980). An investigation into auditors continuity and related qualification
judgments. Journal of Accounting Research, 8, 506523.
Kim, N. S., & Ahn, W. (2002). Clinical psychologists theory-based representations of
mental disorders predict their diagnostic reasoning and memory. Journal of
Experimental Psychological: General, 131, 451476.
316 References
King, G., Keohane, R. O., & Verba, S. (1994). Designing Social Inquiry: Scientific
Inference in Qualitative Research. Princeton: Princeton University Press.
Kintsch, W. (1988). The role of knowledge in discourse comprehension: A construction-
integration model. Psychological Review, 95, 163182.
Kirkpatrick, S. A. (1976). Introduction to the special issue on small group theory.
American Behavioral Scientist, 20, 59.
Kitcher, P. (1993). The Advancement of Science. Oxford: Oxford University Press.
Klayman, J. (1995). Varieties of confirmation bias. The Psychology of Learning and
Motivation, 32, 385418.
Klein, D. E. (2002). Making Law in the United States Courts of Appeals. New York:
Klein, D. E. (2005). Unspoken questions in the Rule 32.1 debate: Precedent and
psychology in judging. Washington & Lee Law Review, 62, 1709.
Klein, D. E. & Morrisroe, D. (1999). The prestige and influence of individual judges on
the U.S. Courts of Appeals. Journal of Legal Studies, 28, 371391.
Klein, G. A., Orasanu, J., Calderwood, R., & Zsambok, C. E. (1993). Decision Making in
Action: Models and Methods. Norwood, NJ: Ablex Publishing Corp.
Knetch, J. C. (2000). The endowment effect and evidence of irreversible indifference
curves. In D. Kahneman and A. Tversky (Eds.), Choices Values and Frames. New
York: Cambridge University Press.
Koehler, D. J., Brenner, L., & Griffin, D. (2002). The calibration of expert judgment:
Heuristics and biases beyond the laboratory. In T. Gilovich, D. Griffin, & D.
Kahneman (Eds.), Heuristics and Biases: The Psychology of Intuitive Judgment.
Kolata, G., & Peterson, I. (2001, July 21). New Jersey is trying new way for witnesses to
say, Its Him. New York Times, p. A1.
Konecni, V. J., & Ebbesen, E. B. (1986). Courtroom testimony by psychologists on
eyewitness identification issues: Critical notes and reflections. Law and Human
Behavior, 10, 117126.
Kornhauser, L. A., & Sager, L. G. (1993).The one and the many: Adjudication in
collegial courts. California Law Review, 81, 159.
Kourlis, R. L., & Singer, J. M. (2007). Using judicial performance evaluations to
promote judicial accountability. Judicature, 90, 200207.
Kovera, M. B., & McAuliff, B. D. (2000). The effects of peer review and evidence
quality on judge evaluations of psychological evidence: Are judges effective
gatekeepers? Journal of Applied Psychology, 85, 574.
Kozhevnikov, M. (2007). Cognitive styles in the context of modern psychology:
Toward an integrated framework of cognitive style. Psychological Bulletin, 133,
464481.
Kozinski, A. (1993). What I ate for breakfast and other mysteries of judicial decision
making. Loyola Los Angeles Law Review, 26, 993.
Kritzer, H. M., & Richards, M. J. (2003). Jurisprudential regimes and Supreme Court
decisionmaking: The Lemon regime and Establishment Clause cases. Law &
Society Review, 37, 827840.
Kritzer, H. M., & Richards, M. J. (2005). The influence of law in the Supreme Courts
search-and-seizure jurisprudence. American Politics Research, 33, 3355.
Krueger, J. I., & Funder, D. C. (2004). Toward a balanced social psychology: Causes,
consequences, and cures for the problem-seeking approach to social behavior
and cognition. Behavioral and Brain Sciences, 27, 313376.
References 317
Kruglanski, A. W. (1996). Motivated social cognition: Principles of the interface. In E.

T. Higgins & A. W. Kruglanski (Eds.), Social Psychology: Handbook of Basic
Principles. New York: Guilford Press.
Kruglanski, A. W., & Webster, D. M. (1996). Motivated closing of the mind: Seizing
and freezing. Psychological Review, 103, 263283.
Kruglanski, A. W., Shah, J. Y., Fishbach, A., Friedman, R., Chun, W. Y., & Sleeth-
Keppler, D. (2002). A theory of goal systems. Advances in Experimental Social
Kulik, C. T., Perry, E. L., & Pepper, M. B. (2003). Here comes the judge: The influence
of judge personal characteristics on federal sexual harassment case outcomes.
Law and Human Behavior, 27, 6997.
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108, 480498.
Landeo, C. M. (n.d.). Tort reform and disputes under endogenous beliefs. Unpublished
manuscript.
Landes, W. M., & Posner, R. A. (1975). The independent judiciary in an interest-group
perspective. Journal of Law and Economics, 18, 875901.
Landes, W. M., & Posner, R. A. (1979). Adjudication as a public good. Journal of Legal
Studies, 8, 235284.
Landsman, S., & Rakos, R. F. (1994). A preliminary inquiry into the effect of
potentially biasing information on judges and jurors in civil litigation.
Behavioral Sciences and the Law, 12, 113.
Langer, L. (2002). Judicial Review in State Supreme Courts: A Comparative Study.
Albany: State University of New York Press.
Larrick, R. P. (2004). Debiasing. In D. J. Koehler & N. Harvey (Eds.), Blackwell
Handbook on Judgment and Decision Making. Malden, MA: Blackwell Publishing.
Lasky, M. (1965). A return to the observatory below the bench. Southwestern Law
Journal, 19, 679.
Law, D. S. (2005). Strategic judicial lawmaking: Ideology, publication, and asylum law
in the ninth circuit. University of Cincinnati Law Review, 73, 817866.
Lawrence, J. A. (1988). Expertise on the bench: Modeling magistrates judicial
decision-making. In M. T. H. Chi, R. Glaser, & M. J. Farr (Eds.), The Nature of
Expertise. Hillsdale, NJ: Erlbaum.
Lazarus, E. P. (1998). Closed Chambers. New York: Times Books.
Lazarus, E. P. (2000, July 9). Its all about OConnor. Los Angeles Times, p. 2.
Leary, M. R. (1996). Self Presentation: Impression Management and Interpersonal
Behavior. Boulder, CO: Westview Press.
Leblanc, V. R., Brooks, L. R., & Norman, G. R. (2002, October Supplement). Believing
is seeing: The influence of a diagnostic hypothesis on the interpretation of clinical
features. Academic Medicine, 77(10).
Lehman, D. R., Lempert, R. O., & Nisbett, R. E. (1988). The effects of graduate training
on reasoning: Formal discipline and thinking about everyday-life events.
American Psychologist, 43, 431442.
Leiter, B. (1997). Rethinking legal realism: Toward a naturalized jurisprudence. Texas
Law Review, 76, 267315
Leiter, B. (2004). American legal realism. In M. P. Golding & W. A. Edmundson
(Eds.), The Blackwell Guide to the Philosophy of Law and Legal Theory. Oxford:
Blackwell.
Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of accountability.
Psychological Bulletin, 125, 255275.
318 References
Lerner, J. S., Goldberg, J. H., & Tetlock, P. E. (1998). Sober second thought: The effects
of accountability, anger, and authoritarianism on attributions of responsibility.
Personality and Social Psychology Bulletin, 24, 563574.
Leuchtenburg, W. E. (1995). The Supreme Court Reborn: The Constitutional Revolution
in the Age of Roosevelt. New York: Oxford University Press.
Levi, E. (1948). An Introduction to Legal Reasoning. Chicago: University of Chicago Press.
Levine, J. M., & Moreland, R. L. (1994). Group socialization: Theory and research. In
W. Stroebe & M. Hewstone (Eds.), European Review of Social Psychology.
Chichester: Wiley.
Lewin, K. (1935). A Dynamic Theory of Personality. New York: McGraw Hill.
Lewis, A. (1987, September 27). Abroad at home: Question of judgment. New York
Times, p. D23.
Liberato, L., & Rutter, K. (2003). Reasons for reversal in the texas courts of appeals.
South Texas Law Review, 44, 431.
Lichtenstein, S., Fischhoff, B., & Phillips, L. D. (1982). Calibration of probabilities:
The state of the art to 1980. In D. Kahneman, P. Slovic, & A. Tversky (Eds.),
Judgment Under Uncertainty: Heuristics and Biases. Cambridge, UK: Cambridge
University Press.
Liebman, J. S. (2000). The overproduction of death. Columbia Law Review, 100, 2030.
Linder, D. O. (1985). How judges judge: A study of disagreement on the United States
Court of Appeals for the Eighth Circuit. Arkansas Law Review, 38, 479560.
Lindquist, S., & Klein, D. (2006). The influence of jurisprudential considerations on
Supreme Court decisionmaking: A study of conflict cases. Law and Society
Review, 40, 135161.
Lithwick, D. (2004). A high court of one. In N. Devins & D. M. Douglas (Eds.), A Year
at the Supreme Court. Durham, NC: Duke University Press
Llewellyn, K. N. (1930). The Bramble Bush: Law and its Study. New York: Columbia
Law School.
Llewellyn, K. N. (1950). Remarks on the theory of appellate decision and the
rules or canons about how statutes are to be construed. Vanderbilt Law Review,
3, 395.
Llewellyn, K. N. (1960). The Common Law Tradition: Deciding Appeals. Boston: Little,
Brown.
Lodge, M., & Taber, C. (2000). Three steps toward a theory of motivated political
reasoning. In A. Lupia, M. D. McCubbins, & S. L. Popkin (Eds.), Elements of
Reason: Cognition, Choice, and the Bounds of Rationality. New York: Cambridge
University Press.
Loewenstein, G., & Elster, J. (Eds.). (1992). Choice Over Time. New York: Russell Sage
Foundation.
Loewenstein, J., Thompson, L., & Gentner, D. (1999). Analogical encoding facilitates
knowledge transfer in negotiation. Psychonomic Bulletin & Review, 6(4),
586597.
Lord, C. G., Ross, L., & Lepper, M. (1979). Biased assimilation and attitude
polarization: The effects of prior theories on subsequently considered evidence.
Journal of Personality & Social Psychology, 37, 2098.
Louis, M. R. (1980). Surprise and sense making: What newcomers experience in
entering unfamiliar organizational settings. Administrative Science Quarterly,
25, 226251.
References 319
Lundberg, C. G. (2004). Modeling and predicting emerging inference-based decisions

in complex and ambiguous legal settings. European Journal of Operational
Research, 153, 417432.
Lundberg, C. G. (2007). Models of emerging contexts in risky and complex decision
settings. European Journal of Operational Research, 177, 13631374.
Lykken, D. T. (1979). The detection of deception. Psychological Bulletin, 80, 4753.
MacCormick, N. (2005). Rhetoric and the Rule of Law: A Theory of Legal Reasoning.
Oxford: Oxford University Press.
MacCoun, R. (1998). Biases in the interpretation and use of research results. Annual
Review of Psychology, 49, 259287.
MacCoun, R. J. (2005). Voice, control, and belonging: The double-edged sword of
procedural fairness. Annual Review of Law and Social Science, 1, 171201.
MacCoun, R. J., & Tyler, T. R. (1988). The basis of citizens perceptions of the criminal
jury: Procedural fairness, accuracy, and efficiency. Law and Human Behavior, 12,
333352.
Macey, J. R. (1994). Judicial preferences, public choice, and the rules of procedure.
Journal of Legal Studies, 23, 627.
Maltzman, F., & Wahlbeck, P. J. (1996). May it please the Chief? Opinion assignments
in the Rehnquist Court. American Journal of Political Science, 40, 421443.
Maltzman, F., Spriggs, J. F., II, & Wahlbeck, P. J. (2000). Crafting Law on the Supreme
Court: The Collegial Game. New York: Cambridge University Press.
Mandler, J. M., & Orlich, F. (1993). Analogical transfer: The roles of schema
abstraction and awareness. Bulletin of the Psychonomic Society, 31, 485487.
Manning, J. (2003). The absurdity doctrine. Harvard Law Review, 116, 23872486
Markman, A. B., & Gentner, D. (1993). Structural alignment during similarity
comparisons. Cognitive Psychology, 25, 431467.
Martin, A. D., & Quinn, K. M. (2002). Dynamic ideal point estimation via Markov
chain Monte Carlo for the U.S. Supreme Court, 19531999. Political Analysis, 10,
134.
Martin, A. D., Quinn, K. M., & Epstein, L. (2005). The median justice on the United
States Supreme Court. North Carolina Law Review, 83, 1275.
Martinek, W. L. (2006). Small group theory and the United States Courts of Appeals.
Paper presented at the annual meeting of the Midwest Political Science
Association.
Mather, M., Shafir, E.; & Johnson, M. K. (2000). Misremembrance of options past:
Source monitoring and choice. Psychological Science, 11, 132.
Mauro, T. (2005). Fluent in tea leaves. American Lawyer, 27(5), 7576.
Maveety, N. (1996). Justice Sandra Day OConnor: Strategist on the Supreme Court.
Lanham, MD: Rowman & Littlefield.
Maveety, N. (2003). The study of judicial behavior and the discipline of political
science. In N. Maveety (Ed.), The Pioneers of Judicial Behavior. Ann Arbor:
Mayhew, D. R. (1974). Congress: The Electoral Connection. New Haven, CT: Yale
University Press.
McAtee, A., & McGuire, K. T. (2007). Lawyers, justices, and issue salience: When and how
do legal arguments affect the U.S. Supreme Court? Law & Society Review, 41, 259278.
McCarty, N., Poole, K. T., & Rosenthal, H. (2006). Polarized America: The Dance of
Ideology and Unequal Riches. Cambridge, MA: MIT Press.
320 References
McClelland, J. L., & Rumelhart, D. E. (1981). An interactive activation model of

context effects in letter perception: I. An account of basic findings.
Psychological Review, 88, 375407.
McGuire, K. T. (1996, November). Explaining executive success in the United States
Supreme Court. Paper presented at the conference on Scientific Study of Judicial
Politics, St. Louis.
McGuire, K. T., & Palmer, B. (1995). Issue fluidity on the United States Supreme
Court. American Political Science Review, 89, 691702.
McGuire, K. T., & Palmer, B. (1996). Issues, agendas, and decision making on the
Supreme Court. American Political Science Review, 90, 85365.
McNicol, D. (2004) A Primer of Signal Detection Theory. (2nd ed.). London: Routledge.
Medin, D. L., Goldstone, R. L., & Gentner, D. (1990). Similarity involving attributes
and relations: Judgments of similarity and difference are not inverses.
Psychological Science, 1, 6469.
Mehrabian, A. (1996). Relations among political attitudes, personality, and
psychopathology assessed with new measures of libertarianism and
conservatism. Basic and Applied Social Psychology, 18, 469491.
Merida, K., & Fletcher, M. (2007). Supreme Discomfort: The Divided Soul of Clarence
Thomas. New York: Doubleday.
Merrill, T. W. (2003). The making of the second Rehnquist Court. St. Louis University
Law Journal, 47, 569.
Messier, W. F. (1983). The effect of experience and firm type on materiality/disclosure
judgments. Journal of Accounting Research, 21, 611618.
Miceli, T. J., & Cosgel, M. M. (1994). Reputation and judicial decision-making.
Journal of Economic Behavior and Organization, 23, 3151.
Michael H. v. Gerald D., 491 U.S. 110 (1989).
Miles, T. J., & Sunstein, C. R. (2006). Do judges make regulatory policy? An empirical
investigation of Chevron. University of Chicago Law Review, 73, 823881.
Milgram, S. (1974). Obedience to Authority. New York: Harper & Row.
Miller, G. P. (2004). Bad judges. Texas Law Review, 83, 431.
Minow, M. L., & Spelman, E. V. (1989). Passion for justice. Cardozo Law Review, 10, 37.
Minow, M. L., & Spelman, E. V. (1990). In context. Southern California Law Review,
63, 15971653.
Mireles v. Waco, 502 U.S. 9, 11-12 (1991).
Mishler, W., & Sheehan, R. S. (1996). Public opinion, the attitudinal model, and
Supreme Court decision making. Journal of Politics, 58, 169.
Mitchell, G. (2002a). Taking behavioralism too seriously? The unwarranted pessimism of
the new behavioral analysis of law. William and Mary Law Review, 43, 19072021.
Mitchell, G. (2002b).Why law and economics perfect rationality should not be traded
for behavioral law and economics equal incompetence. Georgetown Law Journal,
9, 67167.
Mitchell, G. (2003a). Tendencies versus boundaries: Levels of generality in behavioral
law and economics. Vanderbilt Law Review, 56, 17811812.
Mitchell, G. (2003b). Mapping evidence law. Michigan State Law Review, 2003, 1065
1148.
Mitchell, M. L., & Jolley, J. M. (2007). Research Design Explained (6th ed.). Belmont,
CA: Wadsworth/Thompson Learning.
Molot, J. T. (1998). How changes in the legal profession reflect changes in civil
procedure. Virginia Law Review, 84, 955.
References 321
Molot, J. T. (2004). Principled minimalism: Restriking the balance between judicial

minimalism and neutral principles. Virginia Law Review, 90, 17531847.
Monahan, J., & Silver, E. (2003). Judicial decision thresholds for violence risk
management. International Journal of Forensic Mental Health, 2(1), 16.
Montagu, B. (ed.). (1850). The Works of Francis Bacon, Lord Chancellor of England.
London: A. Hart.
Moors, A., & De Houwer, J. (2006). Automaticity: A theoretical and conceptual
analysis. Psychological Bulletin, 132, 297326.
Morehead, T. (1998). Small groups. In D. T Gilbert & G. Lindzey (Eds.), The Handbook
of Social Psychology. New York: McGraw Hill.
Moscovici, S. (1980). Toward a Theory of Conversational Behavior. In L. Berkowitz
(Ed.), Advances in Experimental Social Psychology (vol. 13). New York: Academic
Press.
Moscovici, S., & Zavalloni, M. (1969). The group as a polarizer of attitudes. Journal of
Moseley v. V Secret Catalogue, Inc, 537 U.S. 418 (2003).
Mosier, K. L. (1997). Myths of expert decision making and automated decision aids. In C.
Zsambok & G. Klein, (Eds.), Naturalistic Decision Making. Hillsdale, NJ: Erlbaum.
Mott, R. L. (1948). Measurement of judicial personnel. New York University Law
Quarterly Review, 23, 262.
Murphy, A. H., & Winkler, R. L. (1977). Can weather forecasters formulate reliable
forecasts of precipitation and temperature? National Weather Digest, 2, 29.
Murphy, B. A. (2003). Wild Bill: The Legend and Life of William O. Douglas. New York:
Random House.
Murphy, W. (1964). Elements of Judicial Strategy. Chicago: University of Chicago Press.
Murphy, W. F. (1966). Courts as small groups. Harvard Law Review, 79, 15651572.
Mustard, D. B. (2001). Racial, ethnic, and gender disparities in sentencing: Evidence
from the U.S. federal courts. Journal of Law and Economics, 44, 285.
Nadeau, J. P. (2000). What it means to be a judge. Judges Journal, 39, 34.
Nagel, R. F. (1994). Judicial Power and American Character: Censoring Ourselves in an
Anxious Age. New York: Oxford University Press.
Nagel, R. F. (2006, April 17). Bowing to precedent. The Weekly Standard, p. 24.
Nelson, T. E., and Kinder, D. R. (1996). Issue framing and group centerism in
American public opinion. Journal of Politics, 58(4), 10551078.
Nelson, T. E., Oxley, Z. M., & Clawson, R A. (1997). Toward a psychology of framing
effects. Political Behavior, 19, 221246.
Nelson, T. E., Maruska, S., & Braman, E. (2002). What is the issue? Legal and media
constructions and political attitudes. Paper presented at the Annual Conference of
the American Political Science Association.
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many
guises. Review of General Psychology, 2, 175220.
Nisbett, R. E., & Wilson, T. D. (1977). The halo effect: Evidence for unconscious
alteration of judgments. Journal of Personality and Social Psychology, 35, 250256.
Norton, M. R. (2003). Texts and manuscripts of the Old Testament. In F. F. Bruce, J. I.
Packer, P. Comfort, & C. F. H. Henry (Eds.), The Origin of the Bible. Wheaton, IL:
Tyndale House Publishers.
Novick, L. (1988). Analogical transfer, problem similarity, and expertise. Journal of
Experimental Psychology: Learning, Memory and Cognition, 14, 510520.
Nunnally, J. C. (1967). Psychometric Theory. New York: McGraw-Hill Book Company.
322 References
OBrien, D. M. (2005). A diminished plenary docket. Judicature, 89, 134137, 182.

Oklahoma Tax Commission v. Jefferson Lines, Inc., 514 U.S. 175 (1995).
Oliphant, H. (1928). A return to stare decisis. American Bar Association Journal, 14,
159163.
Olivecrona, K. (1971). Law as Fact (2nd ed.). London: Stevens & Sons.
Olowofoyeku, A. (1993). Suing Judges: A Study of Judicial Immunity. New York:
Oxford University Press.
Operario, D., & Fiske, S. T. (1999). Social cognition permeates social psychology:
Motivated mental processes guide the study of human social behavior. Asian
Journal of Social Psychology, 2, 6378.
OSullivan, M., & Ekman, P. (2005). The wizards of deception detection. In P. A.
Granhag & L. Stromwell (Eds.), The Detection of Deception in Forensic Contexts.
Cambridge, UK: Cambridge University Press.
Otis, L., & Reiter, E. H. (2006). Mediation by judges: A new phenomenon in the
transformation of justice. Pepperdine Dispute Resolution Law Journal, 6, 351.
Parness, J. A. (2006). Improving judicial settlement conferences. U.C. Davis Law
Review, 39, 1891.
Payne, J. L., & Woshinsky, O. H. (1972). Incentives for political participation. World
Politics, 24, 518546.
Pennington, N., & Hastie, R. (1986). Evidence evaluation in complex decision making.
Pennington, N., & Hastie, R. (1988). Explanation-based decision making: Effects of
memory structure on judgment. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 14, 521.
Pennington, N., & Hastie, R. (1992). Explaining the evidence: Tests of the story model
for juror decision making. Journal of Personality and Social Psychology, 62, 189.
Pennington, N., & Hastie, R. (1993). The story model for juror decision making. In R.
Hastie (Ed.), Inside the Juror: The Psychology of Juror Decision Making. New York:
Perry, H. W., Jr. (1991). Deciding to Decide: Agenda Setting in the United States
Supreme Court. Cambridge, MA: Harvard University Press.
Peters, C. J. (2000). Assessing the new judicial minimalism. Columbia Law Review,
100, 14541537.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and Persuasion: Central and
Peripheral Routes of Attitude Change. New York, NY: Springer-Verlag.
Petty, R. E., & Wegener, D. T. (1993). Flexible correction processes in social judgment:
Correcting for context induced contrast. Journal of Experimental Social
Pfennig v. The Queen, 182 CLR 461 at 487488 (1995).
Phelps, R. H. (1977). Expert Livestock Judgment: A Descriptive Analysis of the
Development of Expertise. Unpublished doctoral dissertation, Kansas State
University, Manhattan.
Phelps, R. H., & Shanteau, J. (1978). Livestock judges: How much information
can an expert use? Organizational Behavior and Human Performance, 21, 209219.
Philips, S. U. (1998). Ideology in the Language of Judges: How Judges Practice Law,
Politics, and Courtroom Control. New York: Oxford University Press.
Phillips, F. (2002). The distortion of criteria after decision-making. Organizational
Behavior and Human Decision Processes, 88, 769784.
References 323
Pines, J. M. (2005). Profiles in patient safety: Confirmation bias in emergency

medicine. Academic Emergency Medicine, 13, 9094.
Pittman, T. S. (1998). Motivation. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.),
The Handbook of Social Psychology (4th ed.). Boston: McGraw-Hill.
Planned Parenthood v. Casey, 505 U.S. 833 (1992).
Plato. (2006). The Republic (R. E. Allen, Trans.). New Haven, CT: Yale University Press.
Plous, S. (1993). The Psychology of Judgment and Decision Making. New York:
McGraw-Hill.
Plous, S., & Williams, T. (1995). Racial stereotypes from the days of american slavery:
A continuing legacy. Journal of Applied Social Psychology, 25, 796817.
Pollock, P. H., III, Lilie, S.A., & Vittes, M. E. (1993). Hard issues, core values, and
vertical constraint: The case of nuclear power. British Journal of Political Science,
23, 2950.
Posner, R. A. (1992). Legal reasoning from the top down and from the bottom up: The
question of unenumerated constitutional rights. University of Chicago Law
Review, 59, 43350.
Posner, R. A. (1993). What do judges and justices maximize? (The same thing
everybody else does). Supreme Court Economic Review, 3, 1.
Posner, R. A. (1995a). Judges writing styles (and do they matter?). University of
Chicago Law Review, 62, 14211436.
Posner, R. A. (1995b). Overcoming Law. Cambridge, MA: Harvard University Press.
Posner, R. A. (1999). The Problematics of Moral and Legal Theory. Cambridge, MA:
Harvard University Press.
Posner, R. A. (2005). Foreword: A political court. Harvard Law Review, 119, 31102.
Posner, R. A. (2006). Reasoning by analogy. Cornell Law Review, 91, 761774.
Posner, R. A. (2008). How Judges Think. Cambridge, MA: Harvard University Press.
Postema, G. M. (1982). Coordination and convention at the foundation of law.
Journal of Legal Studies, 11, 165.
Pritchett, C. H. (1948). The Roosevelt Court: A Study in Judicial Politics and Values,
19371947. New York: Macmillan Company.
Pritchett, C. H. (1954). Civil Liberties and the Vinson Court. Chicago: University of
Chicago Press.
Pritchett, C. H. (1969). The development of judicial research. In J. B. Grossman and J.
Tanenhaus (Eds.), Frontiers of Judicial Research. New York: John Wiley and Sons.
Pruitt, C. R., & Wilson, J. Q. (1983). A longitudinal study of the effect of race on
sentencing. Law & Society Review, 17, 613636.
Pyszczynski, T., & Greenberg, J. (1987). Toward and integration of cognitive and
motivational perspectives on social inference: A biased hypothesis-testing model.
In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (vol. 20). New
York: Academic Press.
Rachlinski, J. J. (1998). A positive psychological theory of judging in hindsight.
University of Chicago Law Review, 65, 571.
Rachlinski, J. J. (2006). Bottom-up and top-down decisionmaking. University of
Chicago Law Review, 73, 933.
Rachlinski, J. J., Guthrie, C., & Wistrich, A. J. (2006). Inside the bankruptcy judges
mind. Boston University Law Review, 86, 1227.
Rachlinski, J. J., Johnson, S. L., Wistrich, A. J., & Guthrie, C. (2009). Does unconscious
racial bias affect trial judges? Notre Dame Law Review, 84(3), 11951246.
324 References
Radin, M. (1925). The theory of judicial decision: Or how judges think. American Bar
Association Journal, 11, 357362.
Rahdert, M. C. (2007). Comparative constitutional advocacy. American University
Randazzo, K. A., Waterman, R. W., & Fine, J. A. (2006). Checking the federal courts:
The impact of congressional statutes on judicial behavior. Journal of Politics,
68(4), 10061017.
Raskin, D. C., & Podlesny, J. A. (1979). Truth and deception: A reply to Lykken.
Psychological Bulletin, 86, 5459.
Ratterman, M. J., & Gentner, D. (1998). More evidence for a relational shift in the
development of analogy: Childrens performance on a causal-mapping task.
Cognitive Development, 13, 453478.
Ratzlaf v. United States, 510 U.S. 135 (1994).
Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Belknap Press.
Raz, J. (1979). The Authority of Law: Essays on Law and Morality. Oxford: Clarendon Press.
Raz, J. (1986). The Morality of Freedom. Oxford: Oxford University Press.
Raz, J. (1994). Ethics in the Public Domain. Oxford: Oxford University Press.
Read, S. J., & Marcus-Newhall, A. (1993). Explanatory coherence in social
explanations: A parallel distributed processing account. Journal of Personality
and Social Psychology, 65, 429.
Read, S. J., & Miller, L. C. (1998). On the dynamic construction of meaning: An
interactive activation and competition model of social perception. In S. J. Read &
L. C. Miller (Eds.), Connectionist Models of Social Reasoning and Social Behavior.
Mahwah, NJ: Lawrence Erlbaum.
Read, S. J., Vanman, E. J., & Miller, L. C. (1997). Connectionism, parallel constraint
satisfaction processes, and Gestalt principles: (Re)introducing cognitive
dynamics to social psychology. Personality and Social Psychology Review, 1,
2653.
Redding, R. E., & Reppucci, N. D. (1999). Effects of lawyers socio-political attitudes
on their judgments of social science in legal decision making. Law and Human
Behavior, 23, 31.
Redelmeier, D. A. (2005). The cognitive psychology of missed diagnoses. Annals of
Internal Medicine, 142, 115120.
Reed, J. W. (2005). Theyre playing a tango. Michigan Bar Journal, 84, 18.
Reeve, J. (2005). Understanding Motivation and Emotion. New York: John Wiley & Sons.
Resnik., J. (1982). Managerial judges. Harvard Law Review, 96, 374.
Resnik, J. (2002). Mediating preferences: Litigant preferences for process and judicial
preferences for settlement. Journal of Dispute Resolution, 2002, 155.
Richards, M. J., & Kritzer, H. M. (2002). Jurisprudential regimes in Supreme Court
decision making. American Political Science Review, 96, 305320.
Rickard, A. (1998). Juror Judgment and Decision Making: Prior Convictions Evidence,
Instructions to Disregard and Different Standards of Proof. Unpublished honors
thesis, School of Psychology, University of Queensland.
Ridgeway, C. L. (1978). Conformity, group-oriented motivation, and status
attainment in small groups. Social Psychology, 41, 175188.
Ridgeway, C. L., & Berger, J. (1986). Expectations, legitimation, and dominance
behavior in task groups. American Sociological Review, 51, 603617.
Ridgeway, C. L., Johnson, C., & Diekema, D. (1994). External status, legitimacy, and
compliance in male and female groups. Social Forces, 72, 10511077.
References 325
Ringhand, L. A. (2007). Judicial activism: An empirical examination of voting behavior

on the Rehnquist natural court. Constitutional Commentary, 24, 43102.
Robbennolt, J. K. (2002). Determining punitive damages: Empirical insights and
implications for reform. Buffalo Law Review, 50, 103.
Robbennolt, J. K. (2005). Evaluating juries by comparison to judges: A benchmark for
judging? Florida State University Law Review, 32(2), 469509.
Robbennolt, J. K., Darley, J. M., & MacCoun, R. J. (2003). Symbolism and
incommensurability in civil sanctioning: Decision makers as goal managers.
Brooklyn Law Review, 68, 1121.
Roberts, J. G., Jr. (2005). Oral advocacy and the re-emergence of a Supreme Court bar.
Journal of Supreme Court History, 30(1), 6881.
Robinson, P. H., & Spellman, B. A. (2005). Sentencing decisions: Matching the
decisionmaker to the decision nature. Columbia Law Review, 105, 11241161.
Roe v. Wade, 410 U.S. 113 (1973).
Rogers v. Tennessee, 532 U.S. 451 (2001).
Rohde, D. W. (1972). Policy goals and opinion coalitions in the Supreme Court.
Midwest Journal of Political Science, 16, 208224.
Rohde, D. W., & Spaeth, H. J. (1976). Supreme Court Decision Making. San Francisco:
W.H. Freeman.
Rombeck, T. (2002, October 30). Justice takes time for Q and A. Lawrence Journal-
World, p. 5B.
Rosen, J. (2007a). The Supreme Court: The Personalities and Rivalries that Defined
America. New York: Times Books.
Rosen, J. (2007b, June 18). Supreme leader. The New Republic, p. 16.
Rosenberg, M. (1966). The qualities of justiceAre they strainable? Texas Law Review,
44, 1063.
Rosenberg, M. J., & Abelson, R. P. (1960). An analysis of cognitive balancing. In M. J.
Rosenberg, C. I. Hovland, W. J. McGuire, R. P. Abelson, & J. W. Brehm (Eds.),
Attitude Organization and Change: An Analysis of Consistency Among Attitude
Components. New Haven, CT: Yale University Press.
Ross, A. (1958). On Law and Justice. Berkeley, CA: University of California Press.
Ross, L., & Nisbett, R. E. (1991). The Person and the Situation: Perspectives of Social
Psychology. New York: McGraw-Hill.
Rowland, C. K., & Carp, R. A. (1996). Politics and Judgment in Federal District Courts.
Lawrence: University Press of Kansas.
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and
contagion. Personality and Social Psychology Review, 5, 296320.
Rubin, E. (1997). Law and the methodology of law. Wisconsin Law Review, 1997, 521.
Ruger, T. W., Kim, P. T., Martin, A. D., & Quinn, K. M. (2004). The Supreme Court
Forecasting Project: Legal and political-science approaches to Supreme Court
decision making. Columbia Law Review, 104, 11501209.
Rumble, W. E. (1968). American Legal Realism: Skepticism, Reform, and the Judicial
Process. Ithaca, NY: Cornell University Press.
Ryder, P., Pike, R., & Dalgleish, L. (1974). What is the signal in signal detection?
Perception and Psychophysics, 15, 479482.
Saks, M. J., & Kidd, R. F. (1980). Human information processing and adjudication:
Trial by heuristics. Law & Society Review, 15, 123160.
Sala, B. R., & Spriggs, J. F., II. (2004). Designing tests of the Supreme Court and the
separation of powers. Political Research Quarterly, 57, 197208.
326 References
Salthouse, T. A. (2005). Effects of aging on reasoning. In K. J. Holyoak & R. G.

Morrison (Eds.), The Cambridge Handbook of Thinking and Reasoning. New
Sampson, K. M., (Ed.). (2004). Handbook for Judges: An Anthology of Inspirational and
Educational Readings. Chicago: American Judicature Society.
Sapiro, V. (1983). The Political Integration of Women: Roles, Socialization, and Politics.
Urbana: University of Illinois Press.
Sarat, A. (1977). Judging in trial courts: An exploratory study. Journal of Politics, 39,
368398.
Scalia, A. (1994). The dissenting opinion. Journal of Supreme Court History, 1994,
3344.
Schaefer, W. V. (1960). Good judges, better judges, best judges. American Judicature
Society, 44, 22.
Schaff, P. (1954). The Life and Labors of St. Augustine (T. C. Porter, Trans.). New York:
J.C. Riker.
Schanzenbach, M. M., & Tiller, E. H. (2006). Strategic judging under the U.S.
sentencing guidelines: Positive political theory and evidence. Journal of Law,
Economics, & Organization, 23, 2456.
Schauer, F. (1987). Precedent. Stanford Law Review, 39, 571605.
Schauer, F. (1988a). Formalism. Yale Law Journal, 97, 509548.
Schauer, F. (1988b). Judging in the corner of the law. Southern California Law Review,
61, 17171733.
Schauer, F. (1991). Playing By the Rules: A Philosophical Examination of Rule-Based
Decision-Making in Law and in Life. Oxford: Clarendon Press.
Schauer, F. (1992). The practice and problems of plain meaning. Vanderbilt Law
Review, 45, 715741
Schauer, F. (2000). Incentives, reputation, and the inglorious determinants of judicial
behavior. University of Cincinnati Law Review, 68, 615636.
Schauer, F. (2008). Why precedent in law (and elsewhere) is not totally (or even
substantially) about analogy. Perspectives on Psychological Science, 3, 454460.
Schauer, F. (2009). Thinking Like a Lawyer: A New Introduction to Legal Reasoning.
Cambridge: Harvard University Press.
Schauer, F. (2003). Profiles, Probabilities, and Stereotypes. Cambridge, MA: Harvard
University Press.
Schauer, F. (2004a). The generality of law. West Virginia Law Review, 107, 217238
Schauer, F. (2004b). Judicial supremacy and the modest constitution. California Law
Review, 92, 10451067
Schauer, F. (2004c). The limited domain of the law. Virginia Law Review, 90, 19091956
Schauer, F. (2006a). Do cases make bad law? University of Chicago Law Review, 73,
883918.
Schauer, F. (2006b). On the supposed jury-dependence of evidence law. University of
Pennsylvania Law Review, 155, 165202.
Scheb, J. M., II, & Lyons, W. (2001). Judicial behavior and public opinion: Popular
expectations regarding the factors that influence Supreme Court decisions.
Political Behavior, 23, 181194.
Schick, M. (1970). Learned Hands Court. Baltimore: John Hopkins Press.
Schlag, P. (1985). Rules and Standards. University of California Law Review, 33, 379430.
Schlenker, B. R., & Pontari, B. A. (2000). The strategic control of information:
Impression management and self-presentation in daily life. In A. Tesser,
References 327
R. B. Felson, & J. M. Sulls (Eds.), Psychological Perspectives on Self and Identity.

Washington, D.C.: American Psychological Association.
Schroeder, T. (1918). Psychologic study of judicial opinions. California Law Review, 6,
89113.
Schubert, G. (1962). Policy without law: An extension of the certiorari game. Stanford
Law Review, 14, 284327.
Schubert, G. (1964). The power of organized minorities in a small group.
Administrative Science Quarterly, 9, 133153.
Schubert, G. (1965). The Judicial Mind: Attitudes and Ideologies of Supreme Court
Justices 19461963. Evanston, IL: Northwestern University Press.
Schuette, R. A., and Fazio, R. H. (1995). Attitude accessibility and motivation as
determinants of biased processing: A test of the MODE model. Personality and
Social Psychology Bulletin, 21, 704710.
Schwartz, S., & Griffin, T. (1986). Medical Thinking: The Psychology of Medical
Judgment and Decision Making. New York: Springer-Verlag.
Sears, D. O. (1986). College sophomores in the laboratory: Influences of a narrow data
base on social psychologys view of human nature. Journal of Personality and
Sears, D. O., Huddy, L., & Jervis, R. (Eds.). (2003). Oxford Handbook of Political
Psychology. New York: Oxford University Press.
Segal, J. A. (1984). Predicting Supreme Court cases probabilistically: The search and
seizure cases. American Political Science Review, 78, 891900.
Segal, J. A. (1986). Supreme Court justices as human decision makers: An
individual-level analysis of the search and seizure cases. Journal of Politics, 48,
938955.
Segal, J. A. (1997). Separation-of-powers games in the positive theory of law and
courts. American Political Science Review, 91, 2844.
Segal, J. A., & A. D. Cover. (1989). Ideological values and the votes of U.S. Supreme
Court justices. American Political Science Review 83, 557565.
Segal, J. A., & Spaeth, H. J. (1993). The Supreme Court and the Attitudinal Model. New
Segal, J. A., & Spaeth, H. J. (1994). The authors respond. Law and Courts, 4, 1012.
Segal, J. A., & Spaeth, H. J. (1996). The influence of stare decisis on the votes of
United States Supreme Court justices. American Journal of Political Science, 40,
9711003.
Segal, J. A., & Spaeth, H. J. (2002). The Supreme Court and the Attitudinal Model
Revisited. New York: Cambridge University Press.
Shah, J. Y. (2005). The automatic pursuit and management of goals. Current Directions
in Psychological Science, 14, 1013.
Shah, J. Y., & Kruglanski, A. W. (2000). The structure and substance of intrinsic
motivation. In C. Sansone & J. M. Karackiewicz (Eds.), Intrinsic and Extrinsic
Motivation: The Search for Optimal Motivation and Performance. San Diego:
Academic Press.
Shanteau, J. (1992a). Competence in experts: The role of task characteristics.
Organizational Behavior and Human Decision Processes, 53, 252266.
Shanteau, J. (1992b). How much information does an expert use? Is it relevant? Acta
Psychologica, 81, 7586.
Shapiro, S. (1998). The difference that rules make. In B. Bix (Ed.), Analyzing Law: New
Essays in Legal Theory. Oxford: Clarendon Press.
328 References
Shapiro, S. (2007). The Hart-Dworkin Debate: A Short Guide for the Perplexed.
University of Michigan Public Law Working Paper No. 77.
Sherif, M. (1936). The Psychology of Social Norms. New York: Harper and Row.
Sherry, S. (2003). Judges of character. Wake Forest Law Review, 38, 793812.
Sherry, S. (2005). Politics and judgment. Missouri Law Review, 70, 973987.
Sherry, S. (2006). Logic without experience: The problem of federal appellate courts.
Notre Dame Law Review, 82, 97154.
Sherry, S. (2007). Democracy and the death of knowledge. University of Cincinnati
Law Review, 75, 10531069.
Sherwin, E. (1999). A defense of analogical reasoning in law. University of Chicago Law
Review, 66, 11791197.
Shoemaker, P. J. H. (1982). The expected utility model: Its variants, purposes,
evidence, and limitations. Journal of Economic Literature, 20, 529.
Shullman, S. L. (2004). The illusion of devils advocacy: How the justices of the
Supreme Court foreshadow their decisions during oral argument. Journal of
Appellate Practice and Process, 6, 271293.
Simon, D. (1998). A psychological model of judicial decision making. Rutgers Law
Journal, 30, 1142.
Simon, D. (2002). Freedom and constraint in adjudication: A look through the lens of
cognitive psychology. Brooklyn Law Review, 67, 10971139.
Simon, D. (2004). A third view of the black box: Cognitive coherence in legal decision
making. University of Chicago Law Review, 71, 511586.
Simon, D., & Holyoak, K. J. (2002). Structural dynamics of cognition: From
consistency theories to constraint satisfaction. Personality and Social Psychology
Review, 6, 283294.
Simon, D., Krawczyk, D. C., & Holyoak, K. J. (2004a). Construction of preferences by
constraint satisfaction. Psychological Science, 15, 331336.
Simon, D., Krawczyk, D. C., Bleicher, A., & Holyoak, K. J. (2008). The transience of
constructed preferences. Journal of Behavioral Decision Making, 21, 114.
Simon, D., Pham, L. B., Le, Q. A., & Holyoak, K. J. (2001). The emergence of
coherence over the course of decision making. Journal of Experimental
Psychology: Learning, Memory, & Cognition, 27, 12501260.
Simon, D., Snow, C. J., & Read, S. J. (2004b). The redux of cognitive consistency
theories: Evidence judgments by constraint satisfaction. Journal of Personality and
Simon, J. F. (1980). Independent Journey: The Life of William O. Douglas. New York:
Harper & Row.
Simon, R., & Mahan, L. (1971). Quantifying burdens of proof: A view from the bench,
the jury, and the classroom. Law and Society Review, 5(3), 319330.
Simonson, I., & Staw, B. M. (1992). Deescalation strategies: A comparison of
techniques for reducing commitment to losing courses of action. Journal of
Applied Psychology, 77, 419426.
Sisk, G. C., & Heise, M. (2005). Judges and ideology: Public and academic
debates about statistical measures. Northwestern University Law Review, 99,
743803.
Slotnick, E. E. (1979). Judicial career patterns and majority opinion assignment on the
Supreme Court. Journal of Politics, 41, 640648.
Slovic, P. (1969). Analyzing the expert judge: A descriptive study of a stockbrokers
decision processes. Journal of Applied Psychology, 53, 255263.
References 329
Smith, J. L. (2005). Patterns and consequences of judicial reversals:

Theoretical considerations and data from a district court. Justice System
Journal, 27, 28.
Snow, R. E., Kyllonen, P. C., & Marshalek, B. (1984). The topography of ability and
learning correlations. In R. J. Sternberg (Ed.), Advances in the Psychology of
Human Intelligence (vol. 2). Hillsdale, NJ: Erlbaum.
Snyder, E. C. (1958). The Supreme Court as a small group. Social Forces, 36, 232238.
Solan, L., Rosenblatt, T., & Osherson, D. (2008). False consensus bias in contract
interpretation. Columbia Law Review, 108, 1268.
Solum, L. B. (1988). The virtues and vices of a judge: An Aristotelian guide to judicial
selection. Southern California Law Review, 61, 17351756.
Solum, L. B. (2003). Virtue jurisprudence: A virtue-centered theory of judging.
Metaphilosophy, 34, 178213.
Solum, L. B. (2005a). A tournament of virtue. Florida State University Law Review, 32,
13651400.
Solum, L. B. (2005b). Judicial selection: Ideology versus character. Cardozo Law
Review, 26, 659689.
Songer, D. R. & Haire, S. (1992). Integrating alternative approaches to the study of
judicial voting: Obscenity cases in the U.S. Courts of Appeals. American Journal of
Political Science, 36(4), 963982.
Spaeth, H. J. (1964). The judicial restraint of Mr. Justice FrankfuterMyth or reality?
Midwest Journal of Political Science 8(1), 2238.
Spaeth, H. J. (1979). Supreme Court Policy Making: Explanation and Prediction. San
Francisco: W. H. Freeman.
Spaeth, H. J., and Segal, J. A. (1996). The influence of stare decisis on the votes of
United States Supreme Court justices. American Journal of Political Science, 40,
9711003.
Spaeth, H. J., & Segal, J. A. (1999). Majority Rule or Minority Will: Adherence to
Precedent on the U.S. Supreme Court. New York: Cambridge University Press.
Spellman, B. A. (2004). Reflections of a recovering lawyer: How becoming a cognitive
psychologistand (in particular) studying analogical and causal reasoning
changed my views about the field of psychology and law. Chicago-Kent Law
Review, 79, 11871214.
Spellman, B. A. (2007). On the supposed expertise of judges in evaluating evidence.
University of Pennsylvania Law Review PENNumbra, 156, 19.
Spellman, B. A., & Holyoak, K. J. (1992). If Saddam is Hitler then who is George Bush?
Analogical mapping between systems of social roles. Journal of Personality and
Spellman, B. A., & Holyoak, K. (1996). Pragmatics in analogical mapping. Cognitive
Psychology, 31, 307346
Spiller, P. T., & Tiller, E. H. (1996). Invitations to override: Congressional reversals of
Supreme Court decisions. International Review of Law and Economics, 16, 503
521.
Stangor, C. (2004).Social Groups in Action and Interaction. New York: Psychology
Press.
Stainslaw, H., & Todorov, N. (1999). Calculation of signal detection theory measures.
Behaviour Research Methods, Instruments, and Computers, 31, 137149.
Stanovich, K. E. (1999). Who is Rational? Individual Differences in Reasoning. Mahwah,
NJ: Lawrence Erlbaum Associates.
330 References
Starr, K. W. (2002). First Among Equals: The Supreme Court in American Life. New
York: Warner Books.
Staudt, N., Friedman, B., & Epstein, L. (2008). On the role of ideological homogeneity
in generating consequential decisions. University of Pennsylvania Journal of
Constitutional Law, 10, 361.
Steblay, N., Dysart, J., Fulero, S., & Lindsay, R. C. L. (2001). Eyewitness accuracy rates
in sequential and simultaneous lineup presentations: A meta-analytic
comparison. Law & Human Behavior, 25, 459.
Stein, A. (2005). Foundations of Evidence Law. Oxford: Oxford University Press.
Stein, E. (1996). Without Good Reason: The Rationality Debate in Philosophy and
Cognitive Science. Oxford: Clarendon Press.
Sternberg, R. J. (1999). The theory of successful intelligence. Review of General
Stewart, T. R., Roebber, P. J., & Bosart, L. F. (1997). The importance of the task in
analyzing expert judgment. Organizational Behavior and Human Decision
Processes, 69, 205219.
Stith, K. (2008). The arc of the pendulum: Judges, prosecutors, and the exercise of
discretion. Yale Law Journal, 117, 1420.
Stone Sweet, A. (2002). Path dependence, precedent, and judicial power. In I. Shapiro
& A. Stone Sweet (Eds.), On Law Politics and Judicialization. New York: Oxford
University Press.
Suedfeld, P. (2000). Cognitive styles: Personality. In A. E. Kazdin (Ed.), Encyclopedia of
Psychology (Vol. 2, pp. 166169). Washington, DC and New York: American
Psychological Association and Oxford University Press.
Suedfeld, P., & Tetlock, P. E. (2001). Individual differences in information processing.
In A. Tesser & N. Schwarz (Eds.), Blackwell International Handbook of Social
Psychology (pp. 284304). Boston, MA: Blackwell Publishers.
Sullivan, K. M. (1992). The Supreme Court 1991 Term: The Justice of Rules and
Standards. Harvard Law Review, 106, 22123.
Sunstein, C. R. (1993). On analogy in legal reasoning. Harvard Law Review, 106, 741789.
Sunstein, C. R. (1996). On the expressive function of law. University of Pennsylvania
Sunstein, C. R. (1996). Legal Reasoning and Political Conflict. New York: Oxford
University Press.
Sunstein, C. R. (1999). One Case at a Time: Judicial Minimalism on the Supreme Court.
Cambridge, MA; Harvard University Press.
Sunstein, C. R. (2000). Deliberative trouble? Why groups go to extremes. Yale Law
Journal, 110, 71119.
Sunstein, C. R. (2003). Why Societies Need Dissent. Cambridge: Harvard University
Press.
Sunstein, C. R. (2005). Radicals in Robes: Why Extreme Right-Wing Courts Are Wrong
for America. Cambridge, MA: Basic Books.
Sunstein, C. R. (2007). If people would be outraged by their ruling, should judges care?
Stanford Law Review, 60, 155212.
Sunstein, C. R., Hastie, R., Payne, J. W., Schkade, D. A., & Viscusi, W. K.
(2002). Punitive Damages: How Juries Decide. Chicago: University of Chicago Press.
Sunstein, C. R., Schkade, D., Ellman, L. M., & Sawicki, A. (2006). Are Judges Political:
An Empirical Analysis of the Federal Judiciary. Washington, DC: Brookings
Institution Press.
References 331
Sunstein, C., & Ullman-Margalit, E. (1999). Second-order decisions. Ethics, 110,

526
Symposium. (2002). Expedited appeals in selected state appellate courts. Journal of
Appellate Practice and Process, 4, 191302.
Symposium. (2005). The behavioral analysis of legal institutions: possibilities,
limitations, and new directions. Florida State Law Review, 32, 315.
Taber, C., & Lodge, M. (2006). Motivated skepticism in political information
processing. American Journal of Political Science, 50, 755769.
Taha, A. (2004). Publish or Paris? Evidence of how judges allocate their time.
American Law and Economics Review, 6, 1.
Tajfel, H., & Turner, J. (1985). The social identity theory of intergroup behavior. In S.
Worchel and W. Austin (Eds.), Psychology of Intergroup Relations. Chicago:
Nelson-Hall.
Tate, C. N. (1981). Personal attribute models of the voting behavior of U.S. Supreme
Court justices: Liberalism in civil liberties and economic decisions, 19461978.
American Political Science Review, 75, 355367.
Taylor, S. (2005, September 12). The Roberts court will decide. Legal Times, pp. 6061.
Tetlock, P. E. (1983a). Cognitive styles and political ideology. Journal of Personality
and Social Psychology, 45, 118126.
Tetlock, P. E. (1983b). Accountability and the complexity of thought. Journal of
Tetlock, P. E. (1985). Accountability: A social check on the fundamental attribution
error. Social Psychology Quarterly, 48, 227236.
Tetlock, P. E. (1986). A value pluralism model of ideological reasoning. Journal of
Personality and Social Psychology, 50(4), 81927.
Tetlock, P. E. (2000). Cognitive biases and organizational correctives: Do both disease
and cure depend on the politics of the beholder? Administrative Science Quarterly,
45, 293.
Tetlock, P. E. (2002). Social functionalist frameworks for judgment and choice:
Intuitive politicians, theologians, and prosecutors. Psychological Review,
109, 451.
Tetlock, P. E. (2005). Expert Political Judgment: How Good is It? How Can We Know?
Tetlock, P. E., Bernzweig, J., & Gallant, J. L. (1985). Supreme Court decision making:
Cognitive style as a predictor of ideological consistency of voting. Journal of
Tetlock, P. E., & Mitchell, P. G. (1993). Liberal and conservative approaches to justice:
Conflicting psychological portraits. In B. Mellers & J. Baron (Eds.), Psychological
Perspectives on Justice. Cambridge, UK: Cambridge University Press.
Tetlock, P.E., & Mitchell, G. (Forthcoming). Implicit Bias and Accountability Systems:
What Must Organizations Do to Prevent Discrimination? Research in
Organizational Behavior, 28.
Tetlock, P. E., Skitka, L., & Boettger, R. (1989). Social and cognitive strategies for
coping with accountability conformity, complexity, and bolstering. Journal of
Thagard, P. (1989). Explanatory coherence. Behavioral and Brain Sciences, 12, 435.
Thagard, P. (1992). Conceptual Revolutions. Princeton: Princeton University Press.
Thagard, P. (2000). Coherence in Thought and Action. Cambridge, MA: MIT Press.
Thagard, P. (2006). Hot Thought. Cambridge, MA: MIT Press.
332 References
Thagard, P., & Kunda, Z. (1998). Making sense of people: Coherence mechanisms. In
S. J. Read & L. C. Miller (Eds.), Connectionist Models of Social Reasoning and
Social Behavior. Hillsdale, NJ: Lawrence Erlbaum.
Thagard, P., & Millgram, E. (1995). Inference to the best plan: A coherence theory of
decision. In A. Ram & D. B. Leake (Eds.), Goal-Driven Learning. Cambridge, MA:
Bradford Books.
Thaler, R. H. (1991). Quasi Rational Economics. New York: Russell Sage.
Thompson, W. C., Cowan, C. L., Ellsworth, P. C., & Harrington, J. C. (1984). Death
penalty attitudes and conviction proneness: The translation of attitudes into
verdicts. Law and Human Behavior, 8, 95113.
Tindale, R. S., Meisenhelder, H. M., Dykema-Engblade, A. A., & Hogg, M.A.
Shared cognition in small groups. In M. A. Hogg & R. S. Tindale (Eds.),
Blackwell Handbook of Social Psychology: Group Processes. Malden, MA:
Wiley-Blackwell.
Time Magazine. (1975, November 24). The Courts uncompromising libertarian,
p. 69.
Tinsdale, R. S., Meisenhelder, H. M., Dykema-Engblade, A. A., & Hogg, M. A. (2000).
Shared cognition in small groups. In M. A. Hogg & R. S. Tinsdale, (Eds.),
Blackwell Handbook on Psychology: Group Practices. Malden, MA: Blackwell.
Toobin, J. (2007). The Nine: Inside the Secret World of the Supreme Court. New York:
Doubleday.
Toward Active Liberty: A Supreme Court Justice Offers a View from the Top. (2006,
Spring). Harvard Law Bulletin, 57(14), 17.
Trumbo, D., Adams, C., Milner, M., & Schipper, L. (1962). Reliability and accuracy in
the inspection of hard red winter wheat. Cereal Science Today, 7, 6271.
Tushnet, M. (1992). Thurgood Marshall and the brethren. Georgetown Law Journal,
80, 21092130.
Tushnet, M. (2005). A Court Divided: The Rehnquist Court and the Future of
Constitutional Law. New York: Norton.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and
biases. Science, 185, 11241131.
Tversky, A., & Kahneman, D. (1982). Availability: A heuristic for judging frequency
and probability. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment
under Uncertainty: Heuristics and Biases. Cambridge, UK: Cambridge
University Press.
Twining, W. (1973). Karl Llewellyn and the Realist Tradition. London: Weidenfeld &
Nicolson.
Tyler, T. R., & Huo, Y. J. (2002). Trust in the Law. New York: Russell Sage Foundation.
Tyler, T. R., & Mitchell, G. (1994). Legitimacy and empowerment of discretionary
legal authority: The United States Supreme Court and abortion rights. Duke Law
Journal, 43, 703815.
Ulman, J. N. (1933). A Judge Takes the Stand. New York: Alfred A. Knopf.
Ulmer, S. S. (1971). Courts as Small and Not So Small Groups. New York: Basic Books.
Unah, I., & Hancock, A. M. (2006). U.S. Supreme Court decision making, case
salience, and the attitudinal model. Law & Policy, 28, 295320.
United States v. Locke, 471 U.S. 84 (1985).
United States v. Terry, 392 U.S. 1 (1968).
Vanberg, G. (2005). The Politics of Constitutional Review in Germany. New York:
References 333
Vandevelde, K. J. (1996). Thinking Like a Lawyer: An Introduction to Legal Reasoning.

Boulder, CO: Westview Press.
Vermeule, A. (2006). Judging Under Uncertainty: An Institutional Theory of Legal
Interpretation. Cambridge, MA: Harvard University Press.
Vickers, A. L. (2005). Daubert, critique, and interpretation: What empirical studies tell
us about the application of Daubert. University of San Francisco Law Review, 40,
109147.
Volokh, A. (1997). N guilty men. University of Pennsylvania Law Review, 146(1), 173
216.
von Neumann, J., & Morgenstern, O. (1944). Theory of Games and Economic Behavior.
Wachovia Bank, N. A. v. Schmidt, 126 S.Ct. 941 (2006).
Wahrman, R., & Pugh, M. D. (1972). Competence and conformity: Another look at
Hollanders study. Sociometry, 35, 376386.
Waldron, J. (2007). Temperamental justice. The New York Review of Books, 54(8),
1517.
Walker, L., & Monahan, J. (2007). Sampling evidence at the crossroads. Southern
California Law Review, 80, 969.
Walker, T. G. (1973). Behavioral tendencies in the three-judge district court. American
Journal of Political Science, 17, 407413.
Walker, T. G. (1974). The decision-making superiority of groups: A research note.
Small Group Research, 5, 121128.
Walker, T. G. (1976). Leader selection and behavior in small political groups. Small
Group Research, 7, 363368.
Wallsten, T. (1981). Physician and medical student bias in evaluating diagnostic
information. Medical Decision Making, 1, 145164.
Wambaugh, E. (1894). The Study of Cases (2nd ed.). Boston: Little, Brown.
Ward, A., and Weiden, D. L. (2006). Sorcerers Apprentices: 100 years of Law Clerks at
the United States Supreme Court. New York: New York University Press.
Ware, S. J. (1999). Money, politics, and judicial decisions: A case study of arbitration
law in Alabama. Journal of Law and Politics, 15, 645.
Washington v. Glucksberg, 521 U.S. 702 (1997).
Wasserstrom, R. (1961). The Judicial Decision: Toward a Theory of Legal Justification.
Stanford, CA: Stanford University Press.
Wazereud-Din v. Goodwill Homes and Mission, Inc., 737 A.2d 683 (N.J. Super. A.D.,
1999).
Weber, H. J. (1979). The perfect judge. Court Review, 17, 11.
Wegener, D. T., & Petty, R.E. (1995). Flexible correction processes in social judgment:
The role of nave theories in corrections for perceived bias. Journal of Personality
and Social Psychology 68, 3651.
Weinreb, L. (2005). Legal Reason: The Use of Analogy in Legal Argument. New York:
Wells, G. L. (1992). Naked statistical evidence of liability: Is subjective probability
enough? Journal of Personality and Social Psychology, 62, 739.
WERL (Workshop on Empirical Research in the Law). (2004). On tournaments for
appointing great justices to the U.S. Supreme Court. Southern California Law
Review, 78, 15779.
Wertheimer, W. (1923/1938). Laws in organization of perceptual forms. In W. D. Ellis
(Ed.), A Source Book of Gestalt Psychology. London: Routledge & Kegan Paul.
334 References
West, A., Desdevises, Y., Fenet, A., Gaurier, D., & Heussaff, M. (1992). The French
Legal System. London: Fourmat Publishing.
Whittington, K. E. (2000.) Once more unto the breach: Post-behavioralist approaches
to judicial politics. Law & Social Inquiry, 25, 601634.
Wickard v. Filburn, 317 U.S. 111 (1942).
Wilkerson, J. D. (1990). Reelection and representation in conflict: The case of agenda
manipulation. Legislative Studies Quarterly, 15, 263282.
Williams, F. V., III. (2007). Reinventing the courts: The frontiers of judicial activism in
the state courts. Campbell Law Review, 29, 591.
Wilson, T. D., & Brekke, N. (1994). Mental contamination and mental correction:
Unwanted influences on judgments and evaluations. Psychological Bulletin, 116,
117142.
Windschitl, P. D., & Wells, G. L. (1996). Base rates do not constrain nonprobability
judgments. Behavioral & Brain Sciences, 19, 40.
Winter, D. G. (2002). Motivation and political leadership. In L. O. Valenty & O.
Feldman (Eds.), Political Leadership for the New Century: Personality and Behavior
among American Leaders. Westport, CT: Praeger.
Winter, D. G. (2003a). Measuring the motives of political actors at a distance. In J. M.
Post (Ed.), The Psychological Assessment of Political Leaders. Ann Arbor:
Winter, D. G. (2003b). Personality and political behavior. In D. O. Sears, L. Huddy, &
R. Jervis (Eds.), Oxford Handbook of Political Psychology. New York: Oxford
University Press.
Wissler, R. L., Hart, A. J., & Saks, M. J. (1999). Decisionmaking about general
damages: A comparison of jurors, judges, and lawyers. Michigan Law Review,
98, 751.
Wistrich, A. J., Guthrie, C., & Rachlinski, J. J. (2005). Can judges ignore inadmissible
information? The difficulty of deliberately disregarding. University of Pensylvania
Wold, J. T. (1978). Going through the motions: The monotony of appellate court
cecisionmaking. Judicature, 62, 5865.
Wood, S. L. Keith, L. C., Lanier, D. N., & Ogundele, A. (1998). Acclimation effects
for Supreme Court justices: A cross-validation, 18881940. American Journal of
Political Science, 42, 690697.
Wright, D. B., & Hall, M. (2006). How a reasonable doubt instruction affects
decisions of guilt. Basic and Applied Social Psychology, 29(1), 8592.
Wrightsman, L. S. (1999). Judicial Decision Making: Is Psychology Relevant? New York:
Kluwer Academic.
Wrightsman, L. S. (2006). The Psychology of the Supreme Court. New York: Oxford
University Press.
Wrightsman, L. S. (2008). Oral Arguments Before the Supreme Court: An Empirical
Approach. New York: Oxford University Press.
Yarborough v. Alvarado, 124 S.Ct. 2140 (2004).
Yates, J. F. (1990). Judgment and Decision Making. Englewood Cliffs, NJ: Prentice-Hall.
Yuille, J. C., & Cutshall, J. L. (1986). A case study of eyewitness memory of a crime.
Journal of Applied Psychology, 71, 291301.
Index
Note: The letter n following locators refers to notes.
accountability, 16, 25, 45, 48, 50, Blackmun, Harry, 11, 5961, 75, 84 n13
5254, 205, 21416, 236, 288, Blackstone Ratio, 168, 17380
29293 Brennan, William, 15, 60, 6869, 75,
Alito, Samuel, 70 291
analogical, analogical reasoning, xii, Burger, Warren, 69, 75, 84 n13
107, 11819, 147, 14950, 152,
15461, 163, 205, 212 cognitive illusion(s), 187, 19395, 198,
analogy, 104, 111, 11720, 14952, 272
15455, 15762, 281 cognitive style, xii, xiii, 27983
attitudinal model, 48, 10, 2830, 42, coherence based reasoning, 134, 138,
46, 50, 52, 54, 6162, 7677, 8384, 14042, 146
8687, 150, 158, 163, 189, 204, 209, competence, 7981, 119, 22223,
21113, 216, 22325, 23132, 241, 22932, 23941, 24346, 25354,
244, 24647, 282 269, 27173, 27678, 291
audience(s), 13, 18, 2022, 25, 3637, constraint(s), 27, 2930, 3339, 46, 52,
88, 9192, 96, 99, 124, 187, 192, 76, 92, 105, 112, 117, 11920, 131,
19597, 199200, 215, 282 139, 142, 147, 184, 186, 19193, 195,
203207, 20910, 213, 21517,
bias, 31, 4348, 5055, 58, 72, 82, 23132, 236, 247
104, 113, 12224, 12630, 133, 136, construct validity, 13435, 14345,
145, 154, 184, 187, 19294, 210, 215, 147
227, 230, 235, 237, 24546, 251,
253, 257, 26162, 272, 27576, descriptive-normative gap, 22224,
288, 290 22829, 233, 237
335
336 Index
discretion, 14, 17, 27, 30, 52, 80, 84, 142, 4748, 5052, 5455, 5868, 70,
166, 184, 18693, 195, 197, 200, 206, 7477, 80, 8587, 9096, 9899,
212, 23536, 247, 262, 267, 28687 146, 184, 192, 200, 204205, 207,
dogmatism, 7172, 288 209, 21213, 215, 223, 232, 234,
Douglas, William, 15, 68, 69, 24041, 24647, 256, 28182, 287,
119 n31 29192
impression management, 18, 88,
election(s), 15, 18, 20, 24, 32, 187, 9193, 98
291 incompetence, 223, 230, 24446, 254,
emotion(s), xv, 60, 122, 12425, 27172, 27778
14445 internal validity, 134, 14445, 147
expectation states theory, 8081
expertise, xii, 29, 67, 80, 82, 11314, judicial review, 32, 22223, 226
122, 13536, 14345, 149, 15255, judicial selection, 187, 223, 258, 282,
157, 162, 242, 26972, 27476, 278, 288, 291
280, 28587, 28992, 294 juries, 31, 39, 104, 210, 222, 235, 261
expert(s), xii, 31, 50, 6667, 11314, juror(s), 31, 39, 104, 114, 131, 133, 137,
125, 135, 145, 149, 15254, 156, 158, 141, 16569, 17172, 175, 179,
162, 186, 189, 200201, 216, 223, 18385, 18890, 192, 19798, 237,
23536, 239, 26978, 28082, 241, 246
28687, 28990
external validity, 13238, 141, 14344, Kennedy, Anthony, 71, 87, 9394,
146, 224, 262 9899
extralegal factors, 183, 18586,
19192, 198, 23133, 235, 240, leadership, 7980, 91, 265
24647 legal considerations, 45, 7, 25, 41, 43,
4748, 5153, 203204
fox(es), 72, 28084, 28889, 291 legal model(s), 4, 7, 6162, 76, 204, 223,
Frankfurter, Felix, 68, 212, 253 226, 23132, 234, 247
freshmen effects, 54, 81 legal realism, legal realists, 45, 28, 76,
84, 10910, 113, 11819, 129, 163,
Ginsburg, Ruth Bader, 70, 75 204
group decision making, 78, 8182, 89,
205, 21314 majority coalition, 85, 8797, 100
group formation, 8586, 88, 9091, 94, managerial model(s), 28, 30
9798, 100 Marshall, Thurgood, 75
Martin-Quinn ideology scores, 2829,
hedgehog-fox, 280, 282 38 n2
hedgehog(s), 72, 28084, 28889, 291, median Justice, 8587, 90, 95, 100
295 MODE attitude-behavior model,
heuristic(s), 14, 1721, 23, 31, 43, 163, 4445, 46, 48
192, 194, 275 Multiple Constraint Satisfaction, 27
heuristic-systematic model, 1720
New Deal Court, 86, 90, 9497
ideological cases, 19, 5868
ideology, ideological, xii, 6, 8, 10, 12, OConnor, Sandra Day, 11, 69, 71, 93,
14, 1920, 24, 2829, 39, 4143, 9899
Index 337
parallel constraint satisfaction, 3336, Warren, Earl, 68, 291

3839, 147
particularism, 111, 115, 12427 Table of Cases
persuasion, 42, 46, 5760, 62,
6569, 7172, 118, 144, 205, Adams v. New Jersey Steamboat
208209 Company, 152
preference-behavior relationship, Ashwander, 211, 212
5254
BedRoc Limited v. United
reflective equilibrium, 122, 130 States, 59
Roberts, John, 6365, 70 Bell Atlantic v. Twombly, 281
Blanchflower v. Blanchflower, 119
Scalia, Antonin, 18, 6061, 66, 70, 75, Bowers v. Hardwick, 160
99 Brown v. Board of Education, 68
second-order decision making, 110, Burnet v. Coronado Oil & Gas.
121, 146 Co., 118
second-order reasoning, 107, 109, Bush v. Gore, 147, 225
11112, 11415, 118, 12324, Bush v. Vera, 99
13233, 13536
selection process, 189, 223, Callins v. Collins, 60
29192 Chevron, 236
Signal Detection Theory, 16769, Citizens to Preserve Overton Park
17173, 175 v. Volpe, 295
small group behavior, xii, 74
small group(s), xii, 7384, 132 FCC v. Beach Communications, Inc.,
socialization, 20, 80, 193 245
Social Judgment Theory, 89, Foucha v. Louisiana, 70
192
Solicitor General, 54, 63, 65, 67 Grutter v. Bollinger, 279
Souter, David, 93
Stevens, John Paul, 70 Hall v. Brooklands Racing Club,
strategic behavior, 67, 2223, 29, 117
7374 Hamdan v. Rumsfeld, 60
strategic model(s), 611, 18, 22, Hudson v. McMillian, 70
2829, 86 Hunter v. Norman, 109
sufficiency principle, 18
Kansas v. Colorado, 58
Thomas, Clarence, 66, 7071,
9394, 99 Michael H. v. Gerald D., 60
threshold placement, 16769, Mireles v. Waco, 267
17173
threshold(s), 37, 137, 16573, 175, Pfennig v. The Queen, 16970
17981, 211, 228, 269 Planned Parenthood v. Casey, 93, 225,
thresholds for action, 16567, 169, 244
18081
tournament(s), 222, 22829, 23941, Ratzlaf v. United States, 147
245, 280 Raymond v. Carman, Jr., 83
338 Index
Roe v. Wade, 93, 98, 221, 225 Wachovia Bank v. Schmidt, 58

Rogers v. Tennessee, 147 Washington v. Glucksberg, 99
Wazereud-Din v. Goodwill
United States v. Locke, 109 Homes and Missions, 161
U.S. v. Terry, 181 Wickard v. Filburn, 9697

The Psychology of Judicial Decision Making

Uploaded by

Copyright:

Available Formats

The Psychology of Judicial Decision Making

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Psychology of Judicial Decision Making

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics are covered in the book?

What topics are covered in the book?

The Psychology of Judicial Decision Making

American Psychology-Law Society Series

Books in the Series

Oxford New York

Copyright 2010 by Oxford University Press.

Published by Oxford University Press, Inc.

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America

This book series is sponsored by the American Psychology-Law Society

Part I. Judges and Human Behavior

1 Motivation and Judicial Behavior: Expanding the Scope of Inquiry 3

2 Multiple Constraint Satisfaction in Judging 27

3 Top-Down and Bottom-Up Models of Judicial Reasoning 41

4 Persuasion in the Decision Making of U.S. Supreme Court Justices 57

5 Judges as Members of Small Groups 73

6 The Supreme Court, Social Psychology, and Group Formation 85

Part II. Judging as Specialized Activity

7 Is There a Psychology of Judging? 103

8 Features of Judicial Reasoning 121

9 In Praise of Pedantic Eclecticism: Pitfalls and Opportunities in the

10 Judges, Expertise, and Analogy 149

11 Thresholds for Action in Judicial Decisions 165

12 Every Jury Trial Is a Bench Trial: Judicial Engineering of Jury

13 Searching for Constraint in Legal Decision Making 203

Part III. Evaluating and Improving Judging

14 Evaluating Judges 221

15 Defining Good Judging 249

16 Expertise of Court Judges 269

17 Cognitive Style and Judging 279

18 Building a Better Judiciary 285

Brandon L. Bartels David Klein

James Shanteau Philip E. Tetlock

Over the years, psychologists have devoted uncountable hours to learning

reasoning, judgment and decision-making, and the psychology of small

different conditions. It also teaches us to expect variation across individual

Students of judicial behavior have taken only limited steps to incorporate

The Study of Judicial Behavior

Models of Judicial Behavior

Legal and SemiLegal Models

labeled historical institutionalism. Adherents to this school, such as Howard

Legal realism was reflected in the quantitative studies of Supreme Court

information in the justices papers to analyze strategy at several stages of

Probing the Models

Second, in their treatment of judges motivations, the dominant models

Indeed, the evidence of differences among judges on the same court is

The Psychology of Judicial Motivation

Goals and Motives

Judges interest in making good legal policy might be fueled by each of

The Energizing Dimension

Variation by Judge and Situation

differences among legislators. It seems highly unlikely that judges are

judicial behavior, because that research emphasizes motivation so heavily.

Motivation and Cognition

Referring to motivated reasoning, Segal and Spaeth (2002, p. 433) argue