THE U.S. SUPREME COURT’S SELECTION OF
PETITIONS IN FORMA PAUPERIS
DISSERTATION

Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University

By

Wendy L. Watson, M.P.P., J.D.
*****
The Ohio State University
2004

Dissertation Committee:
Approved By
Professor Lawrence Baum, Adviser
Professor Elliot Slotnick
Professor Dean Lacy

____________________________________
Adviser
Political Science Graduate Program

Copyright by
Wendy Lyn Watson
2004

ABSTRACT

This dissertation explores the question of whether, how, and how much the U.S.
Supreme Court’s selection of paid petitions differs from its selection of petitions in forma
pauperis, or IFP petitions -- those petitions filed by individuals who are financially
unable to pay the Court’s filing fee. In addition to expanding upon existing scholarship
on the U.S. Supreme Court’s agenda-setting function, the study also examines the
accessibility of the Court to the poor and the Court’s attention to issues affecting lowincome people.
In examining the Court’s agenda-setting function, existing studies have identified
certain case characteristics, called “cue characteristics,” that increase the likelihood that
the Supreme Court will select a petition for review: reference to a conflict between
circuits or state courts of last resort; an allegation of a novel or important legal issue; a
dissent on the court below; a reversal in the case’s procedural history; filing of a sua
sponte responsive brief, as opposed to the absence of a responsive brief or a responsive
brief filed only at the request of the Court; and cert-stage participation by amici curiae.
Those studies, however, have generally limited their analysis to paid petitions to the
exclusion of those petitions filed by indigent litigants, the IFP petitions; even those
studies that have included IFP cases in their analysis have failed to compare the Court’s
agenda-setting process across the two dockets.
ii

Analyzing an original dataset based on a sample of paid and unpaid petitions
disposed of during the 1976 through 1985 Terms of the Court, I find that IFP petitions are
disproportionately criminal and prisoner civil rights cases, as opposed to civil cases, and
they are less likely to present most indicia of certworthiness that have been identified in
past studies. Moreover, I find that IFP petitions are more likely to be filed by pro se
litigants (litigants representing themselves) rather than by attorneys. Finally, I find that
certain issues--specifically minority rights issues, family law issues, and issues involving
access to welfare benefits--are disproportionately represented on the IFP docket. These
findings suggest that the cases on the IFP docket are at a disadvantage during the Court’s
agenda-setting process and that this disadvantage may have implications for the Court’s
attention to certain classes of legal issues.
Dynamic analysis of the Court’s attention to IFP petitions over time indicates that
the proportion of the Court’s plenary docket that is devoted to IFP petitions varies based
on the Court’s ideological composition and the attention devoted to IFP petitions in the
term immediately preceding. These findings suggest that the Court’s IFP docket is
distinct from its paid docket, making comparison between the two dockets meaningful.
Case-level multivariate analysis of the Court’s selection of paid and unpaid
petitions indicates that IFP cases are less likely to be accepted for review even when
analysis controls for the presence or absence of cue characteristics identified in prior
studies.
iii

Moreover, the effects of such case characteristics vary between the paid and
unpaid dockets, with the case characteristics having stronger effects in the context of the
docket on which they are less prevalent. This finding indicates that the case
characteristics identified as correlates of plenary review in past studies do, in fact,
provide information to the Court during the agenda-setting process and are not merely
correlates of some underlying quality of certworthiness.
Ultimately, the analysis indicates that IFP petitions are at a disadvantage relative
to the paid petitions during the agenda-setting process. That disadvantage takes two
forms. First, the cue characteristics are more important to the selection of the IFP
petitions and yet they are also less common; this interaction effect between the cue
characteristics and IFP status is quite significant. Second, IFP status has an independent
negative effect on the probability of the Court granting review; although controlling for
the interaction effects causes that independent effect to lose statistical significance, the
effect is nevertheless impressive when compared with the effect of other known cue
characteristics.

iv

Dedicated to Judge Donovan W. Frank,
for his tireless commitment to equality and justice

v

ACKNOWLEDGMENTS

I wish to thank my adviser, Lawrence Baum, for his unflagging support and
patience. Mere words cannot express my indebtedness to him.
I further wish to thank Elliot Slotnick for being there for me when the going got
tough; the fortune cookie was right. Dean Lacy talked me down off of more than one
methodological ledge, and his sense of humor was grounding during some fairly dark
moments.
Many people provided insight on sampling and statistical issues, including Jan
Box-Steffensmeier, Herb Weisberg, Kevin Sweeney, and Brandon Bartels.
The arduous task of data collection was made infinitely easier by the wonderful
staff of the National Archives and Records Administration, especially archivist Robert
Ellis; the reference staff at the Ohio State University Law Library; and Nancee Halling in
the St. Paul branch of the Eighth Circuit Law Library.
My family and my in-laws have all provided emotional support throughout this
long process. My husband, Peter Hawkinson, made many sacrifices along the way and
has been my inspiration. There never was a sweeter kismet.
This research was supported by a grant from the National Science Foundation. I
am also grateful for the financial support provided by a fellowship from the American
Association of University Women.
vi

VITA

January 19, 1970……………………Born - Zanesville, Ohio
1990…………………………………B.A. English, Psychology, University of Michigan
1994…………………………………M.P.P., J.D., College of William and Mary
1994-1996…………………………...Law Clerk,
The Honorables Donovan W. Frank,
Gary Pagliaccetti, and Robert Berger
Sixth Judicial District, Virginia, Minnesota
1996-1999…………………………...Graduate Research and Teaching Associate,
The Ohio State University
1999-2003…………………………...Law Clerk,
The Honorable Donovan W. Frank
U.S. District Court, District of Minnesota

PUBLICATIONS
1.
Neal Devins and Wendy L. Watson, eds. 1995. Federal Abortion Politics: A
Documentary History. New York, N.Y.: Garland Publishing.

FIELDS OF STUDY
Major Field: Political Science

vii

TABLE OF CONTENTS

Abstract……………………………………………………………………………………ii
Dedication………………………………………………………………………………....v
Acknowledgments………………………………………………………………..………vi
Vita………………………………………………………………………………...…….vii
List of Tables……………………………………………………………………………..xi
List of Figures…………………………………………………………………………...xiii
Chapters:
1.

The U.S. Supreme Court’s Unpaid Docket…………...……………………..……1
Introduction………………………………..……………………………………...1
Background on IFP Status……………………...…………………………………9
Eligibility for IFP Status…………………………...…………………………….10
Benefits of IFP Status…………………………………...……………………….15
History of the IFP Docket…………………………………...…………………...16

2.

The Court’s Case Selection Process……………………………………………..23
Introduction……………………………………………..…………………….….23
Goals in Supreme Court Agenda-Setting…………………..……………….……24
Constraints in Supreme Court Agenda-Setting………………..…………….…...28
Cognitive Task……………………………………………………..……….……30
Potential Cues to Certworthiness……………………………………..…….……37
Ideology of Decision Below………………………………………..……37
Lower Court Behavior…………………………………………………...39
The Parties: Attorney Representation and Solicitor General
Involvement……………………………………………………………...41
Amicus Briefs……………………………………………………………46
Conclusion……………………………………………………………………….47

3.

Hypotheses and Data……………………………………………………………..49
Introduction……………………………………………………………...……….49
Cues in the Context of the IFP Docket: Hypotheses………………………..…...49
Data………………………………………………………………………………55
viii

Population………………………………………………………………..55
Sampling…………………………………………………………………57
Coding……………………………………………………………………60
Identifying Information…………………………………………..60
Case Disposition (Dependent Variable)…...……………………..62
Case Characteristics……………………………………………...64
Petitioner Characteristics………………………………………...65
Responding Party………………………………………………...66
Other Parties and Participants…………………………................67
Decision Below…………………………………………………..69
Rule 10 Factors…………………………………………………..71
Conclusion……………………………………………………………………….72
4.

A Descriptive Analysis of the Court’s IFP Docket………………………………75
Introduction…………………………...………………………………………….75
Types of Cases on the IFP Docket………...……………………………………..77
Representation………………………………...………………………………….84
Participation by Other Parties……………………...…………………………….85
Lower Court Behavior…………………………………...………………………88
Rule 10 Considerations……………………………………...…………………...89
Conclusion………………………………………………………...……………..95

5.

The Supreme Court’s Selection of IFP Petitions………………………………...99
Introduction………………………………………………...…………………….99
Selection of IFP Cases in the Aggregate………………………...……………….99
Dynamic Analysis with Proportion of Cases Filed…………….....…….101
Dynamic Analysis with Proportion of Cases on Docket…………….....107
Implications…………………………………………………………….110
Case-Level Analysis of the Selection of IFP Petitions…………...…………….112
Independent Effect of IFP Status on Probability of Review……………112
Interaction Between IFP Status and Other Cue Characteristics………...120
Conclusion……………………………………………………………………...135

6.

Conclusions and Future Research………………………………………………137
Summary of Findings………………………………………………………...…137
Normative and Theoretical Implications…………………………………….....142
Directions for Future Research…………………………………………...…….146
ix

Appendices:
A.

Rule 39, Rules of the Supreme Court of the United States (1999)……………..150

B.

Example of Petition to Proceed In Forma Pauperis and Affidavit
of Indigency…………………………………………………………………….152

List of References………………………………………………………………………155

x

LIST OF TABLES

Table

Page

3.1

Sample Sizes for Strata in Study…………………………………………………58

3.2

Sampling Scheme for the 1976 through 1980 October Terms…………………..59

3.3

Sampling Scheme for the 1981 through 1985 October Terms…………………..59

4.1

Paid and IFP Cases by Type of Case…………………………………………….77

4.2

Paid and IFP Civil Cases by Court of Origin…………………………………….79

4.3

Incidence of Family Law and Custody Cases on the Paid and IFP Dockets…….82

4.4

Welfare Benefits Cases on the Paid and IFP Dockets…………………………...82

4.5

Representation of Petitioners on the Paid and Unpaid Dockets…………………84

4.6

Percent of Cases in Which Responsive Brief Was Filed………………………...87

4.7

Ideological Direction of Lower Court Decision, Paid and Unpaid Dockets…….88

4.8

Reference to Rule 10 Factors in Attorney-Prepared Petitions…………………...91

4.9

Reference to Rule 10 Factors in Pro Se Petitions………………………………..93

5.1

Predictors of Aggregate IFP Petition Selection (Fileprop)……………………..104

5.2

Predictors of Aggregate IFP Petition Selection (Docprop)……………………..108

5.3

Logit Analysis of All Cases, GVRs and Summary Dispositions Included…......114

5.4

Logit Analysis of All Cases, GVRs and Summary Dispositions Excluded…….117

5.5

Crosstab of IFP Status and Case Disposition, Weighted Data………………….118
xi

5.6

Crosstab of Solicitor General Position and Case Disposition,
Weighted Data………………………………………………………………….118

5.7

Logit Analysis of IFP Petitions Only………...…………………………………121

5.8

Logit Analysis of Paid Petitions Only………………………………………….122

5.9

Changes in Probability Associated with Cues, Conservative Lower Court
Decision………………………………………………………………………...124

5.10

Changes in Probability Associated with Cues, Liberal Lower Court
Decision………………………………………………………………………...125

5.11

Factor Change Associated with Cue Characteristics,
Paid and IFP Cases……………………………………………………………...127

5.12

Logit Analysis of All Cases, Including Interaction Terms……………………..130

5.13

Changes in Probability Associated with Cues, Conservative Lower Court
Decision………………………………………………………………………...131

5.14

Factor Change Associated with Cue Characteristics,
Paid and IFP Cases……………………………………………………………...133

xii

LIST OF FIGURES

Figure

Page

1.1

Paid and Unpaid Filings with the U.S. Supreme Court by Year…………………17

1.2

Case Filings and the Court’s Acceptance of IFP Petitions………………………18

4.1

Proportion of Paid and IFP Dockets Comprised by Different Case Types………78

4.2

Paid and IFP Civil Cases by Court of Origin…………………………………….79

4.3

Civil Cases by Issue Area………………………………………………………..81

4.4

Representation of Petitioners on the Paid and Unpaid Dockets…………………85

4.5

Rule 10 Factors in Attorney-Prepared Petitions…………………………………92

4.6

Rule 10 Factors in Pro Se Petitions………………………………………………94

5.1

Partial Autocorrelation Function of Residuals of Basic Regression……………102

5.2

PACF of Residuals from Model Employing Lagged Endogenous Variable…...105

5.3

IFP Filings and IFP Acceptances, as Percents of Totals, Smoothed…………...106

5.4

PACF of Residuals from Model Employing Lagged Endogenous Variable…...109

xiii

CHAPTER 1
THE U.S. SUPREME COURT’S UNPAID DOCKET

INTRODUCTION
Since the passage of the Judiciary Act of 1925, the Supreme Court has enjoyed
almost complete control over its docket by virtue of its ability to grant or deny petitions
for writs of certiorari.1 Some consider the Court’s control over its workload to be an
appropriate and valuable mechanism for allowing the Court to play a more overtly
policy-oriented role in the political process. Chief Justice William Howard Taft, for
example, considered the potential policy agenda of the Court to be the very purpose
behind granting the Court control over its docket:
The sound theory of [the Judiciary Acts of 1891 and 1925] is that litigants
have their rights sufficiently protected by a hearing or trial in the courts of
first instance, and by one review in an intermediate appellate Federal
court. The function of the Supreme Court is conceived to be, not the
remedying of a particular litigant’s wrong, but the consideration of cases
whose decision involves principles, the application of which are of wide
public or governmental interest, and which should be authoritatively
declared by the final court.

1

As the Court’s workload has ballooned, the Court has begun to treat even its original jurisdiction
as discretionary: “We construe 28 U.S. C. § 1251(a)(1), as we do Art. III, 2, cl. 2, to honor our original
jurisdiction but to make it obligatory only in appropriate cases. And the question of what is appropriate
concerns, of course, the seriousness and dignity of the claim; yet beyond that it necessarily involves the
availability of another forum where there is jurisdiction over the named parties, where the issues tendered
may be litigated, and where appropriate relief may be had. We incline to a sparing use of our original
jurisdiction so that our increasing duties with the appellate docket will not suffer” (Illinois v. City of
Milwaukee, 406 U.S. 91, 93-94 (1972)).

1

(Taft, 1925) Others, however, have noted that the certiorari process has become an end
in and of itself, distracting the Court from its true job of deciding cases and issuing
substantive opinions (Frankfurter and Landis, 1928).
Whether the Court’s gate-keeping authority is good or bad, it is certainly an
authority that the Court wields with fervor, and the certiorari process has become, for
better or worse, a substantial component of the Court’s workload (Reimann, 1999).
During the Court’s 2002-2003 Term, 8225 cases (including appeals, original jurisdiction
cases, and extraordinary writs) were brought before the Supreme Court, yet the Court
granted full review to only 92 (U.S. Law Week, July 16, 2002: 3080). The Court’s
selectivity itself raises interesting and potentially quite important questions: What makes
that handful of cases stand out from the crowd? What determines whether a particular
case gets heard by the Court? As discussed more fully in Chapter Two, scholars of the
judicial system have focused considerable effort on ascertaining the objectives underlying
the Court’s agenda-setting process (Epstein and Knight, 1998: 22-55; McGuire and
Caldeira, 1993; Songer, 1979; Provine, 1981; Armstrong and Johnson, 1982) and
identifying case characteristics that increase the probability of the Court granting review
(Tanenhaus et al, 1963; McGuire and Caldeira, 1993; McGuire, 1998; George and
Solimine, 2001).2
It has been suggested, however, that scholarship on the certiorari process should
begin to consider how specific categories of needs, issues, and petitioners are treated by
2

Surprisingly, the scholarship does relatively little to create explicit linkages between the
motivations of the Justices and the correlates of case selection; in other words, there has been little effort
made to create a unified theory of agenda-setting. The notable exception is the literature on strategic
agenda-setting, which posits a connection between the ideological direction of the lower court decision and
the Court’s “error correcting” objective (Armstrong and Johnson, 1982; Boucher and Segal, 1995; Brenner
and Krol, 1989; Schubert, 1958; Schubert, 1962; Songer, 1979; Ulmer, 1972).

2

the Court (Perry, 1991b). Already, some scholars have begun the process of parsing out
individual issue areas--such as business cases (Reimann, 1999), bankruptcy cases
(Lawless and Murray, 1997), and obscenity cases (McGuire and Caldeira, 1993)--for
consideration. Some of these studies use single issue areas as a control to allow for more
refined analysis of the certiorari process as a whole; for example, McGuire and Caldeira
(1993) focused exclusively on obscenity cases to allow for specific consideration of the
effect of organized interests and high profile attorneys on the Court’s certiorari
decisions. Other studies have compared the Court’s selection of cases addressing one
issue area to the Court’s overall agenda-setting process to determine whether particular
issue areas are relatively disadvantaged in that process; for example, Reimann (1999)
addressed critics of the Court who suggested that the Court overlooks important, complex
business cases in favor of higher-profile, more “glamorous” constitutional cases.
The effect of the certiorari process on a particular class of litigants, as opposed to
a particular class of issues, remains relatively uncharted territory. The one notable
exception is Kevin Smith’s 1999 study of pro se petitioners to the Supreme Court, those
individuals who represent themselves when seeking review of their cases (Smith, 1999).
That study considers whether the relative lack of success of pro se petitioners to the
Court derives from the quality of the petitions or from a bias against individuals who file
without the assistance of counsel. Smith ultimately concludes that the virtual absence of
pro se civil litigants granted review by the Supreme Court is most likely a result of the
legal insignificance of their petitions.
Interestingly, Smith justifies his inquiry by questioning whether “the Supreme
Court is denying pro se petitioners their last opportunity for legal and political justice in a
3

system in which justice often is expensive. Indeed, the essentially complete denial of pro
se petitions for certiorari presents a prima facie case of bias toward the poor . . .” (Smith,
1999: 384-385). Despite his apparent concern for equal access to justice for the poor,
Smith limits his analysis to pro se petitioners who are able to pay the $300 Supreme
Court filing fee.
While it is true that pro se status may be strongly correlated with poverty, the
ability of the poor to obtain access to the Court may be better measured by studying the
fate of petitions filed by those individuals who seek to file as indigents, those individuals
who seek leave to proceed in forma pauperis.3
Not only has scholarship on the U.S. Supreme Court failed to consider whether in
forma pauperis or “IFP” petitions are treated differently during the agenda-setting
process, but many empirical studies of the Court’s agenda-setting process look
exclusively at the Court’s acceptance of paid petitions (Caldeira and Wright, 1988;
Caldeira and Wright, 1990; Teger and Kosinski, 1980).4 The exclusion of IFP petitions
from these studies makes a great deal of practical sense: the costs of collecting
information on these petitions is very high and the relatively small percentage of the

3

Rule 39 of the Rules of the Supreme Court of the United States (1999) governs in forma pauperis
procedure; Rule 39 is set forth in Appendix A.
4
There are some exceptions to this general rule of exclusion. George and Solimine’s (2001)
examination of the relationship between en banc review and the Supreme Court’s plenary review drew a
sample of en banc and panel decisions by the Second, Fourth, and Eighth Circuits; thus their sample
undoubtedly included some IFP petitions. Hellman’s (2001) study of the resolution of intercircuit conflicts
drew a limited sample of IFP petitions denied review by the Supreme Court, but Hellman actually argues
that he would have done better to ignore them altogether and urges other researchers not to waste their time
with the IFP docket. Tanenhaus et al.’s (1963: 119) exploration of cue theory excluded “[e]ntries on the
Miscellaneous Docket [IFP petitions] other than petitions for certiorari carrying lower court citations”; it is
unclear how many IFP petitions were thus included in the sample. Jucewicz and Baum (1990) considered
the effect of workload on aggregate acceptance rates, which were based on both paid and IFP filings; their
focus, however, was not on individual case characteristics, nor did they compare the relative acceptance
rates of paid and IFP petitions.

4

petitions accepted dramatically skew statistical results (specifically, the application of
binary variables to rare events data results in biased coefficients, with the models
generating estimated event probabilities that are too small) (King and Zeng, 2001).
Scholars rationalize the exclusion of IFP petitions from analyses of the Court’s
agenda-setting function on the grounds that these petitions are, by and large, frivolous
and unworthy of the Court’s review. The categorical dismissal of IFP petitions by
academicians reflects a prevailing bias, justified or not, within the legal community as a
whole. One former Supreme Court law clerk described the IFP petitions as “sometimes
handwritten, occasionally illegible, and often inscrutable”5 (Lazarus, 1998: 30) Even
Justice William Brennan, whose liberal ideological stance and activist perception of the
Court’s role likely would have rendered him particularly receptive to the petitions of
indigent petitioners, considered the overwhelming majority of IFP petitions to be
unworthy of full Court review (Brown v. Herald Co., 194 S.Ct. 331, 333 (1983)
(Brennan, J., dissenting)). The perception that virtually all IFP petitions are without
merit leads Court commentators to be rather nonchalant about the disproportionately
small percentage of IFP petitions accepted for review. For example, this perception led
Hellman (1985: 961) to state as obvious that the Court’s apparent disregard of the IFP
petitions “is not surprising, nor is it the product of discrimination against the poor.”
Yet, as Chief Justice Harlan Fiske Stone observed, the IFP petitions “are mostly
chaff, but occasionally we find some grains of wheat in the chaff and those cases we
assign counsel, pay expense of printing the papers, and hear the case. This has
5

Pursuant to Rule 39.3 of the Rules of the Supreme Court of the United States (1999), the Clerk of
Court will refuse to accept illegible petitions (Stern et. al, 2002: 502). It is unclear, however, how often the
Clerk of Court actually invokes this provision.

5

occasionally resulted in unearthing grave abuses in trial courts which deprived the
petitioner of his constitutional rights” (Mason, 1956: 639).
Indeed, some of the Supreme Court’s landmark cases have been filed by indigent
petitioners. The most frequently cited example is Gideon v. Wainwright, 372 U.S. 335
(1963), which articulated the constitutional right to counsel in state criminal cases;
Clarence Gideon, a Florida prison inmate, hand-wrote his own Supreme Court petition in
pencil (O’Brien, 1993: 202). Yet other landmark Supreme Court decisions have
originated on the IFP docket, including Batson v. Kentucky, 476 U.S. 79 (1986), which
held that prosecutors may not exercise peremptory challenges to eliminate specific jurors
on the basis of race; Coker v. Georgia, 433 U.S. 583 (1977), which held that the
imposition of the death penalty for the crime of rape violates the Eighth Amendment;
Tison v. Arizona, 481 U.S. 137 (1987), which helped to define the degree of criminal
culpability necessary to justify imposition of the death penalty; Penry v. Lynaugh, 492
U.S. 302 (1989), which held that a criminal defendant’s mental retardation may mitigate
his culpability but that execution of the mentally retarded did not, per se, violate the
Eighth Amendment; and Atkins v. Virginia, 536 U.S. 304 (2002), which held that the
execution of the mentally retarded does, in fact, violate the Eighth Amendment.
Moreover, the composition and fate of the IFP docket has important implications
for equality of access to the courts and the countermajoritarian role that the courts play in
our system of government. To the extent that certain issues arise primarily in the context
of unpaid petitions and to the extent that the interests of certain groups--particularly the
poor and disenfranchised--are affected by the Court’s treatment of these petitions,
understanding the Court’s selection of IFP petitions is profoundly relevant.
6

The fate of IFP petitions should be of interest to scholars for a variety of reasons.
First, civil litigants and “prisoners have a constitutional right of access to the courts”
(Bounds v. Smith, 430 U.S. 817, 821 (1977)). That access cannot be merely technical, but
it must be meaningful. “[D]ifferences in access to the instruments needed to vindicate
legal rights, when based upon the financial situation of the defendant, are repugnant to
the Constitution” (Roberts v. LaValle, 389 U.S. 40, 42 (1967)). The constitutional
dimension of the right to access certainly requires fair and equal consideration of
petitions for certiorari without reference to the financial means of the petitioners.
Second, because the Supreme Court’s treatment of IFP petitions implicates the
procedural justice of the Court’s certiorari process, it has implications for the legitimacy
of the Court as a democratic institution. When people assess the fairness of a particular
institution or procedure, the perceived consistency in treatment across persons plays a
prominent role in their evaluations (Fondacaro, 1995). If there is a systematic bias
against IFP petitions, or perhaps even simply a systematic difference in the Court’s
treatment of IFP and paid petitions, the fundamental fairness of the Court’s procedures
becomes suspect. As Justice Marshall noted, “with each barrier that it places in the way
of indigent litigants, . . . the Court can only reinforce in the hearts and minds of our
society’s less fortunate members the unsettling message that their pleas are not welcome
here” (In re Demos, 500 U.S. 16, 19 (1991) (Marshall, J., dissenting, joined by Blackmun
and Stevens, JJ.)). A challenge to the fairness of any democratic institution would
present a threat to that institution’s legitimacy, but that threat is particularly acute for an
institution like the Court which lacks other political sources of legitimacy and thus

7

depends to heavily on the goodwill and support of the populace (Provine, 1980; Caldeira
and Gibson, 1992).6
Third, the relative fate of IFP petitions has implications for the Court’s role as a
countermajoritarian institution. For years, the legal and political science communities
have been troubled by the so-called “countermajoritarian difficulty,” the concern that an
ostensibly non-majoritarian or even anti-majoritarian institution should wield
considerable authority within a majoritarian democracy. Although scholars have
questioned whether the Court really is out of step with majoritarian interests (Dahl, 1957;
Friedman, 1993; Mishler and Sheehan, 1996; Epstein et al., 2001), the concern over the
countermajoritarian difficulty rages on (Issacharoff, 2001; Yoo, 2001). In the face of this
countermajoritarian crisis, one overriding justification for judicial review is that the Court
acts as a protector of individual rights against the potential tyranny of the majority (U.S.
v. Carolene Products, 304 U.S. 144, 152 n.4 (1938); Ely, 1980; Yingling, 1999). If the
justification for the Court’s authority is that the Court protects the interests of those who
would not otherwise have a voice in majoritarian politics, we should expect the Court to
pay particular attention to the interests and issues of the politically powerless. If
however, the bar is set higher for indigent petitioners to the Supreme Court, this liberal
justification for judicial review will be called into question.
Finally, on a more theoretical and less normative note, a study of the Court’s
certiorari process with respect to IFP petitions may cast further light on the Court’s
certiorari decisions more generally. Specifically, as discussed more fully in Chapter
6

Note, however, that Gibson (1989) found that institutional legitimacy was distinct from procedural
justice; while institutional legitimacy has an effect on the willingness of individuals to accept and comply
with unpopular decisions, perceptions of procedural justice have no such effect.

8

Two, social science researchers have identified a number of case characteristics which
appear to be correlated with an increased probability that a case will be reviewed by the
Supreme Court. Investigating the relationship between these “cue” or “signal”
characteristics and the decision to grant certiorari in the context of IFP petitions--in
which these characteristics may occur with less frequency or in which these
characteristics may be otherwise obscured--may illuminate the specific mechanism by
which these characteristic influence the certiorari process. Examining the way in which
these “cue” or “signal” characteristics affect the behavior of Supreme Court Justices in
the relatively low information / high noise environment of the IFP docket may provide
insight into the micro-level information processing task in which the Supreme Court
Justices engage when making certiorari decisions.

BACKGROUND ON IFP STATUS
The notion that the poor, who are otherwise frequently marginalized in social and
political life, should at a minimum enjoy equal access to the courts of justice is not new.
England’s Magna Carta makes provisions for the poor in court, and U.S. state
governments began making such provisions early in our nation’s history (Maguire, 1923).
In 1892, Congress passed the first federal statute designed to provide individuals
without financial means access to the federal court system (27 Stat. 252 (1892)). The
1892 statute allows individuals who are too poor to pay court costs or security to proceed
in forma pauperis (or “IFP”), literally “in the manner of a pauper.” The House Report
accompanying the legislation presents the underlying concern in stark terms: “Will the
Government allow its courts to be practically closed to its own citizens, who are
9

conceded to have valid and just rights, because they happen to be without the money to
advance pay to the tribunals of justice?” (H.R. Rep. No 1079, 52nd Cong., 1st Sess. 2
(1892)).
Despite the formal mechanism available to provide the poor access to the courts,
equality of access remains problematic. Although the federal IFP statute, currently
codified at 28 U.S.C. 1915 (2000), has changed little over the past 100 years, the history
of IFP status in the federal courts, particularly the U.S. Supreme Court, has varied
considerably, and the true equality of access to the courts remains an unanswered
question. This chapter provides further explanation of the IFP procedure and the history
of the U.S. Supreme Court’s administration of IFP petitions.

ELIGIBILITY FOR IFP STATUS
Generally, criminal defendants who are provided with counsel in the lower federal
courts are automatically granted pauper status, but other criminal defendants and civil
litigants may request leave to file as paupers (Rule 39.1; Baum, 1998: 108-109). Both
petitioners and respondents may proceed IFP, and, while leave to proceed IFP may be
sought at any point during litigation or appeal, it is generally sought at the outset of a
party’s involvement in the court. A party seeking leave to proceed in forma pauperis
must file a motion to do so along with an affidavit or declaration of indigency; these
documents are filed simultaneously with the substantive petition for review or appeal
(Rule 39.1; Stern et al., 2002: 501-504). Appendix B provides an example of both a
motion for leave to proceed in forma pauperis and an affidavit of indigency, both filed by
a pro se litigant.
10

Interestingly, debtors proceeding in federal bankruptcy court are not automatically
considered indigent for purposes of IFP filing. In fact, one study of the Court’s certiorari
decision in bankruptcy cases justified the exclusion of IFP petitions from the study on the
grounds that “few, if any, bankruptcy cases can be found [on the IFP docket]. Moreover,
[since] enactment of the [Bankruptcy] Code, the Court has never granted certiorari in a
bankruptcy case off the IFP docket. Therefore, omission of the IFP docket should have
had little effect on our study” (Lawless and Murray, 1997: 116).
In response to the growing number of litigants seeking IFP status, both Congress
and the Supreme Court have undertaken measures over the past twenty years to curtail
the ability of certain litigants to obtain IFP status.
First, the 1996 Antiterrorism and Effective Death Penalty Act (Pub. L. 104-132,
110 Stat. 1214) placed limitations on the general grant of IFP status to indigent criminal
defendants. Specifically, that statute places some limits on the ability of incarcerated
criminal defendants to proceed in forma pauperis when seeking habeas relief. Moreover,
while the IFP statute still allows indigent litigants to seek review without prepayment of
fees, the Prison Litigation Reform Act of 1996, Pub. L. No. 104-134, 110 Stat. 1321
(1996), requires prison inmates who file IFP petitions to pay the fees eventually and
provides a means of garnishment to insure that the fees are eventually collected.
Second, the Court itself has begun to place limits on IFP eligibility. The federal
IFP statute provides that any petition or complaint filed by a litigant proceeding IFP may
be dismissed at any time if the court determines that the allegation of poverty is untrue or
if the action is frivolous or malicious. Because it appears to grant the Court greater
latitude in dismissing IFP petitions than paid petitions, this statutory provision is
11

controversial enough (Feldman, 1985). But the Court has taken this power further
through enactment of Supreme Court Rule 39(8).
Leave to proceed IFP is generally determined independent of the decision on the
request for substantive review, but Rule 39(8) now allows the Court to deny IFP status if
it determines that the request for substantive review or relief is frivolous or malicious.
The Court adopted this provision in 1991 in an effort to reduce the ability of “frequent
filers” to file IFP petitions (In re Amendment to Rule 39, 500 U.S. 13 (1991), per
curiam). Although the liberal minority on the Court objected to any limitation on
accessibility, the majority of the Court felt compelled to address the “plague[] by a dozen
or so seemingly unhinged souls who filed one frivolous application after another . . .
tak[ing] advantage of the Court’s open-door policy” (Lazarus, 1998: 280). Moreover, in
a dissenting opinion, Justice Stevens (joined by Justice Blackmun) questioned whether
the amendment to the rule would have a practical effect, as the new rule essentially
requires the Court to consider the merits of the petition at the same time the Court
determines whether to grant permission to proceed in forma pauperis.
One of the frequent filers who prompted the amendment of Rule 39 surfaced
twice during my own limited dip into the IFP pool: Vladimir A. Zatko. In Zatko v. The
United States District Court for the Northern District of California et. al (Doc. No. 765161), one of seven IFP petitions filed by Mr. Zatko during the 1975 and 1976 Terms of
the Supreme Court, the exact nature of Mr. Zatko’s claim is a bit unclear, though the
most lucid and succinct explanation states:
Much can be said about the expense and ignomity to which an individual
is necessarily exposed when he is required to submit to an clearly
unconstitutional indeterminate incarceration causing insanity instead of
12

rehabilitation and illegal arrest by the U.S. Immigration thereby seeking
vindication in a totally paralysed State Courtroom presided by an three
feet elevated judge without jurisdiction as to the sentence laws but with
apparently more than enough jurisdiction to categorically and arbitrarily
deprive such person of his right to proceed an appeal in propria persona (
this God given Right emanates from every verse of the Holy Bible )
shaking off the “Iron Heel ” of this cruel and unusual punishment.
(Petition at p. 20, errors in the original). However opaque Mr. Zatko’s legal argument,
his righteous indignation is clear: “The millions from their graves cry out for justice”
(Petition at p. 12).
By 1981, Mr. Zatko’s situation had not changed, but, if anything, his prayer for
relief had become even less coherent. In Zatko v. California (Doc. No. 81-5173), Mr.
Zatko asserts he is a descendent of the Slavic royal family and heir to the Russian throne,
and, in explaining why review should be granted, he explains: “Where Satan invades the
privacy of this Court, and threatens to spread lawlessness in the name of California, the
appropriate remedy is a humble prayer to the Holy Mother of our Lord Jesus Christ -- the
Supreme Justice.”
Between 1981 and 1991, Mr. Zatko filed seventy-three petitions with the U.S.
Supreme Court (Lane, 2003: 350). In 1991, the Supreme Court issued its first order
applying Rule 39.8 to deny Mr. Zatko and another frequent filer leave to proceed in
forma pauperis (Zatko v. California, 502 U.S. 16 (1991)). In its per curiam opinion, the
majority justified the procedure as a means “[t]o discourage abusive tactics that actually
hinder us from providing equal access to justice for all . . .” (Id. at 18)
In his dissent in Zatko, Justice Stevens questioned both the ability of the amended
rule to achieve the desired effect and the possible repercussions of the application of the
rule for equal access to the courts. Specifically, Stevens noted that in the months
13

between the effective date of the amendment and the Zatko decision nearly 1000 IFP
petitions were filed.
[W]ell over half of these petitions could have been characterized as
frivolous. Nevertheless, under procedures that have been in place for
many years, the petitions were denied in the usual manner. . . .
The Court has applied a different procedure to the petitioners in
these cases. . . . As a result, the order in their cases denies leave to
proceed in forma pauperis pursuant to Rule 39.8, rather than simply
denying certiorari. The practical effect of such an order is the same as a
simple denial. However, the symbolic effect of the Court’s effort to draw
distinctions among the multitude of frivolous petitions – none of which
will be granted in any event – is powerful. Although the Court may have
intended to send a message about the need for the orderly administration
of justice and respect for the judicial process, the message that it actually
conveys is that the Court does not have an overriding concern about equal
access to justice for both the rich and the poor.
By it action today, the Court places yet another barrier in the way
of indigent petitioners. By branding these petitioner under Rule 39.8, the
Court increases the chances that their future petitions, which may very
well contain a colorable claim, will not be evaluated with the attention
they deserve.
(Zatko v. California, 500 U.S. 16, 19-20 (1991) (Stevens, J., dissenting joined by
Blackmun, J.).
Justice Stevens’s concerns proved prophetic. The Court has used the rule to go
far beyond the merits of the particular petition before the Court to instead deny IFP status
to individuals with a history of frivolous or malicious actions; thus a petition that may not
be frivolous enough to justify denial of IFP status may nevertheless provoke the use of
Rule 39.8 if the petitioner has a history of such petitions (Lane, 2003).7 Moreover, in
1992, the Court began using the amended Rule 39 to prospectively deny IFP status to
7

“Prophetic” may be overstating the case. In point of fact, the Court had imposed a prospective ban
on IFP status for at least three “frequent filers” before amending Rule 39; in each instance, the majority
failed to cite any statute or rule to support its order, a fact which incited the ire of Justices Marshall,
Stevens, Brennan, and Blackmun (In re Demos, 500 U.S. 16 (1991); In re Sindram, 498 U.S. 177 (1991); In
re McDonald, 489 U.S. 180 (1989)).

14

petitioners in non-criminal matters: “As petitioner has repeatedly abused this Court’s
process, the Clerk is directed not to accept any further petitions in noncriminal matters
from petitioner unless the docketing fee required by Rule 38(a) is paid and the petition is
submitted in compliance with Rule 33.1” (Martin v. District of Columbia Court of
Appeals, 506 U.S. 1 (1992)). In other words, like the boy who falsely cried wolf so often
that no one came when the threat was real, petitioners who have filed numerous frivolous
petitions may be denied IFP status even if their most recent petition raises critical legal
issues (Lane, 2003).
Despite these limitations on the availability of IFP status, the number of litigants
filing IFP petitions shows no signs of decreasing. Indeed, as Figure 1.1, infra, shows, the
sharp upward trend in IFP petitions that began in the mid-1980s has continued unabated.
Still, for certain individual litigants, the changes in federal law and Supreme Court rules
have effectively foreclosed Supreme Court review of their claims.

BENEFITS OF IFP STATUS
IFP status has implications beyond the waiver of prepayment of fees. Pursuant to
the Supreme Court’s own Rule 39, relating to unpaid petitions, indigent petitioners not
only avoid prepayment of fees but they are also excused from many of the technical
requirements. Specifically, IFP petitioners may file briefs and petitions on regular 8 ½ by
11 paper rather than following the detailed “booklet format” typically required in
Supreme Court filings; rules relating to paper weight and colored covers are waived; and
IFP litigants need only file the original and ten copies of each document rather than the

15

usual 40 copies required of paying litigants8 (Stern et al., 2002: 492). While this
additional benefit to IFP status may seem trivial, for litigants who are proceeding pro se
(without the assistance of an attorney), this advantage may be even more important than
the fee waiver.
Of perhaps even greater importance, however, if the Court grants certiorari or
notes probable jurisdiction on the appeal, the Court may appoint counsel to represent the
indigent petitioner at oral argument, pay attorney fees, and pay the costs of reproducing
and serving briefs.

HISTORY OF THE IFP DOCKET
Although the IFP statute dates back to 1892, the number of IFP petitions rose
dramatically during the tenure of Chief Justice Charles Evans Hughes; Chief Justice
Harlan Fiske Stone, who served as an associate justice on the Hughes Court and followed
him as chief justice, attributed the increase in IFP petitions to the attention Chief Justice
Hughes gave those petitions and his willingness to grant them review (Mason, 1956:
639). Accordingly, in 1947, the Vinson Court began to docket IFP petitions separately,
on the so-called “miscellaneous docket”; in 1948, the number of IFP petitions placed on
the miscellaneous docket (690) nearly equaled the number of paid petitions filed on the
Court’s regular appellate docket (773). As illustrated in Figure 1.1, between 1948 and
roughly 1986, the number of IFP petitions filed closely tracked the number of paid
petitions filed. In the late 1980s, however, the number of IFP petitions began a dramatic
8

Originally, the Supreme Court required IFP petitioners to file only the original copy of briefs and
petitions; the advent of low-cost copying mechanisms has allowed the Court to impose the additional
burden of requiring 10 copies. Yet even today, incarcerated litigants proceeding pro se may file only the
original petition, in which case the burden of reproduction is on the Clerk of Court (Stern et al., 2002: 502).

16

PaidFilings
IFPFilings
6,000

Number of Petitions

5,000

4,000

3,000

2,000

1,000

0
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
1948

Year

Figure 1.1. Paid and Unpaid Filings with the U.S. Supreme Court by Year

trend upward while the number of paid petitions remained relatively stable. By the 2000
Term, 75% of the petitions filed with the U.S. Supreme Court were unpaid, and that
proportion shows no sign of declining (Epstein et. al, 1994: 65-67).
While the proportion of petitions filed with the Court without prepayment of fees
has increased, the proportion of the cases accepted for review that originate on the unpaid
docket has not followed a similar trend. Figure 1.2 illustrates the trend in filings in
comparison with the trend in acceptances.
As a practical matter, IFP petitions placed on the miscellaneous docket are
denoted by special docket numbers. The first two digits of the docket number assigned to

17

1.0

Proportion of the
Petitions Filed with
Court that are IFP

0.9

Proportion of the Cases
Granted Review that
are IFP

0.8

Proportion

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0

2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
1948

Year

Figure 1.2. Case Filings and the Court’s Acceptance of IFP Petitions

each Supreme Court petition represent the term in which the case was filed. These two
digits are followed by a hyphen and then by a number representing the order in which the
case was filed. Paid cases are numbered consecutively staring with “1,” while unpaid
petitions are numbered consecutively starting with “5000.”
In addition to the different numbering system, the Court’s IFP petitions are treated
administratively as quite distinct. Historically, the Court has maintained separate
procedures for reviewing IFP petitions. During the Hughes, Vinson, and Warren Courts,
if only one copy of an IFP petition was filed (and only one was required), that copy went
to the Office of the Chief Justice where it was summarized, and that summary was
disseminated to the other Justices (Tanenhaus et al., 1963).

18

The Chief Justice’s special control over the IFP petitions had practical
consequences. For example, Chief Justice Earl Warren reportedly told his law clerks to
“be their [the IFP petitioners’] counsel,” to search for, develop, and frame the legal issues
presented by the IFP petitions (Mauro, 1998; O’Brien, 1993: 172). During Earl Warren’s
tenure as Chief Justice, the Court accepted as many as 5 percent of the IFP petitions filed
each term, and cases originating on the unpaid docket comprised up to 40% of the total
cases accepted for review. In comparison, during the 2000 Term, 5897 IFP petitions
were filed, yet the Court granted full review to only 17 IFP petitions (only 0.3% of the
number filed); and, although just over 75% of the petitions filed were IFP petitions, only
12% of the cases granted full review originated on the miscellaneous docket.
During his first term, Chief Justice Warren Burger essentially eliminated the
Chief Justice’s monopoly over the IFP petitions when he created the “cert pool”
(O’Brien, 1993: 173). Under the cert pool system, all petitions for review--both paid and
unpaid--are distributed among the law clerks of the participating Justices. The law clerks
prepare memoranda for circulation to all of the Justices’ chambers. The cert pool thus
distributed the responsibility of taking a first cut at IFP petitions among the chambers of
all of the participating Justices. At present, all of the Justices participate in the cert pool
except for Justice John Paul Stevens, whose law clerks brief all of the petitions for the
Justice (Lazarus, 1998: 31).
Still, even though the IFP petitions are no longer funneled through the Office of
the Chief Justice, they are still segregated from the paid petitions, kept in physically
separate stacks. Paid petitions are delivered to chambers by the Clerk of Court’s office
on Wednesdays; IFP petitions are delivered on Thursdays (Lazarus, 1998: 30).
19

In short, the Court has drawn a clear line between the paid and unpaid petitions.
Yet does the Court’s administrative segregation of the IFP petitions reflect a genuine
qualitative difference between the petitions on the two dockets?
There is considerable anecdotal evidence to indicate that the Justices themselves
consider the two classes of cases to be quite distinct. As noted above, both Chief Justice
Harlan Fiske Stone and Justice William Brennan considered the overwhelming bulk of
the petitions on the unpaid docket to be frivolous. At least one Supreme Court law clerk
has confessed that the clerks “flip through [the IFP petitions] pretty fast” on the
assumption that few if any deserve close attention (Mauro, 1998). Even Chief Justice
Earl Warren’s direction to his law clerks--to mine the possibilities the IFP docket
presented--discriminated between the two types of cases; although this is a far more
benevolent form of discrimination between the two dockets, it nevertheless reflects an
opinion that the actual petitions filed on the IFP docket are somehow inferior, in need of
shaping and refining.
Still, despite the anecdotal evidence, the question remains: what sort of petitions
are on the IFP docket and do they differ systematically from the paid docket? The
remainder of this dissertation begins to try to answer that question as well as the more
specific question of whether, how, and to what extent the Court’s selection of IFP
petitions differs from its selection of paid petitions.
Chapter Two presents an overview of the literature on the Court’s agenda-setting
function. Specifically, the chapter describes case characteristics correlated with the
Court’s grant of review to paid petitions. While the existing literature has focused on the
simple fact of correlation, there is reason to believe that these case characteristics, or cue
20

characteristics, actually provide the Court with information about the certworthiness of
the petition and thus allow the Court to quickly and efficiently winnow out petitions
worth of review.
Chapter Three explores the theoretical implications of the existing agenda-setting
literature for selection of petitions from the IFP docket. In that chapter, I hypothesize
that, even controlling for these identified cue characteristics, IFP petitions will be less
likely to receive review. Moreover, I argue that considering the different effect of the
identified cue characteristics on IFP cases relative to paid cases will provide some insight
into whether the cue characteristics are mere correlates with the grant of certiorari or
whether the cue characteristics actually affect the Court’s decision to grant review;
specifically, the Court’s selection of cases in two different informational contexts creates
a sort of natural experiment that helps tease out the nature of the relationship between cue
characteristics and the Court’s agenda-setting process. Chapter Three also describes the
data collected as part of this research project and outlines the overall research design.
Chapter Four presents a descriptive analysis of the petitions on the unpaid docket,
focusing on the differences between the paid and unpaid dockets. This descriptive
analysis fills a gap in the literature as well as helping to refine the hypotheses generated
in Chapter Three. Specifically, the descriptive analysis empirically tests the assumption
made by scholars and jurists that IFP petitions are qualitatively different than paid
petitions, and, further, it defines the way in which the two dockets differ. Moreover, by
examining the patterns of cue characteristics on the two dockets, the descriptive analysis
lends specificity to the hypotheses about how the effect of the cue characteristics might
vary between the two dockets.
21

Chapter Five describes multivariate analyses of the Court’s selection of petitions
from the unpaid docket. Specifically, Chapter Five presents both a dynamic analysis of
the Court’s attention to IFP petitions and a case-level analysis of the Court’s relative
selection of paid and unpaid petitions. The dynamic analysis provides support for the
notion that the two dockets truly are distinct in terms of the Court’s agenda-setting
process, that the Court does differentiate between the two dockets during the case
selection process. The case-level analysis begins to explore the precise ways in which
the the Court’s selection of cases varies across the two dockets, considering both the
independent effect of IFP status on the probability of the Court selecting a case for
review and the differential effects of the cue characteristics across the two dockets.
Chapter Six summarizes the findings of this research and presents a plan for
further research on the Court’s unpaid docket.

22

CHAPTER 2
THE COURT’S CASE SELECTION PROCESS

INTRODUCTION
Procedurally, the U.S. Supreme Court’s case selection process is relatively
straight-forward. The petitions are delivered to the Justices’ chambers by the Clerk of
Court once each week (Wednesdays for paid petitions, Thursdays for unpaid petitions),
and the evaluation process begins:
Rather than each Justice considering every case independently, a clerk for
one Justice in the “cert. pool” circulates an advisory memo to all the
Justices in the pool.9 this “pool memo” summarizes a case and assesses
whether it is “certworthy”--that is, whether it raises a sufficiently
important and controversial issue to merit the Supreme Court’s attention.
Although the Justices do not follow the pool memo recommendations
slavishly, in practice they carry great weight (Lazarus, 1998: 31).
The law clerk memos are used to compile the “discuss list,” a list of cases that
deserve discussion at the Court’s biweekly conference.10 According to Justice Ginsburg

9

The “cert. pool” was created in 1972, under the aegis of Chief Justice Warren Burger, although
both Chief Justice Burger and Justice Powell have claimed credit for the idea (Palmer, 2001: 107). Before
Burger’s tenure, the clerks in each Justice’s chambers prepared memos for their Justice with respect to each
of the paid petitions for review; the Chief Justice’s chambers generated a single memo for the unpaid
petitions. After the Court instituted the cert. pool, Justices Marshall and Stevens continued to use their own
clerks’ memos; until his retirement, Justice Brennan did most of his certiorari work himself, not even
relying on his own clerks (Lazarus, 1998: 31).
10
The practice of circulating a conference list, or “special list,” began during the 1930s. From its
inception until sometime in the 1950s, the list was a “dead list,” a list of cases that did not require further
discussion and were denied certiorari. The Chief Justice would draft the initial dead list, but any Justice
could request that a case be removed from the dead list and given further consideration. The transition to a
“live list” or a “discuss list” procedure took place sometime around 1950. Interestingly, Provine (1980)
indicates that the switch from a dead list to a discuss list came early for the unpaid petitions.

23

(1994: 884), fewer than 15% of petitions make the discuss list; the rest are summarily
denied review.
Every other Friday, the Court meets in conference to decide which petitions it will
grant and which it will deny. “In conference, the chief justice or the justice who added a
case to the discuss list opens discussion of the case. In order of seniority, from senior to
junior, the justices then speak and usually announce their votes” (Baum, 1998: 109). It
takes an affirmative vote by four justices to grant certiorari; this is known as the “rule of
four.” Without those four affirmative votes, certiorari is denied. If any justice requests,
however, a particular petition may be “relisted,” to allow the justices to research the
issues further or to try to sway the votes of their fellow justices (Baum, 1998: 110).
The clear procedural rules that govern the agenda-setting process are deceptive.
The manner in which the Court determines which petitions to grant and which petitions to
deny is anything but clear.
The remainder of this chapter will outline what is known about the Court’s
agenda-setting process, what that knowledge tells us about how the selection of IFP
petitions might be different, and how the study of the selection of IFP petitions might
further illuminate the Court’s agenda-setting process more generally.

GOALS IN SUPREME COURT AGENDA-SETTING
The Court’s Rule 10 provides the only explicit guidance to would-be petitioners
about what types of cases the Court will and will not accept:

24

Rule 10. Considerations Governing Review on Writ of Certiorari
Review on a writ of certiorari is not a matter of right, but of judicial
discretion. A petition for a writ of certiorari will be granted only for
compelling reasons. The following, although neither controlling nor fully
measuring the Court’s discretion, indicate the character of the reasons the
Court considers:
(a) a United States court of appeals has entered a decision in conflict with
the decision of another United States court of appeals on the same
important matter; has decided an important federal question in a way that
conflicts with a decision by a state court of last resort; or has so far
departed from the accepted and usual course of judicial proceedings, or
sanctioned such a departure by a lower court, as to call for an exercise of
this Court’s supervisory power;
(b) a state court of last resort has decided an important federal question in
a way that conflicts with the decision of another state court of last resort or
of a United States court of appeals;
(c) a state court or a United States court of appeals has decided an
important question of federal law that has not been, but should be, settled
by this Court or has decided an important federal question in a way that
conflicts with relevant decisions of this Court.
A petition for a writ of certiorari is rarely granted when the asserted error
consists of erroneous factual findings or the misapplication of a properly
stated rule of law.
Litigants are certainly aware of the dictates of Rule 10, and they organize their
petitions for review accordingly. Even some pro se petitioners will mimic the language
of Rule 10 in their arguments, using the items listed in Rule 10 as headings in the body of
their petitions. For example, in Luna v. U.S. (Doc. No. 77-6896) petitioner pro se Luna
summarizes the reason for granting his writ in the index to his petition:
The decision below, which denies error in the admission of heroin at trial,
has decided an important question of federal law which has not been, but
should be, settled by this court, namely: What are the standards for
establishment of a complete chain of custody of a fungible substance
which must be met to warrant its admission as evidence?
25

Alas, despite his careful repetition of the language of Rule 10(c), the Court denied Luna’s
petition for review.
Similarly, respondents opposing review will argue that the Rule 10 factors are not
present. For example, the State of Arizona’s opposition brief in Mincey v. Arizona (Doc.
No. 77-5353) urges the Court to deny review, noting that the facts in the case are so
unusual11 that there is little likelihood for the case to have national impact. Just as the
invocation of Rule 10 does not guarantee the Court will grant review, denying the
applicability of Rule 10 does not guarantee the Court will deny review. Despite the
unusual factual scenario underlying the claim in Mincey, the Court granted certiorari.
Although Rule 10 does not provide a clear benchmark for determining which
cases will and will not receive review by the Supreme Court, Rule 10 does hint at the
Court’s objective in carrying out its agenda-setting function; “a careful reading of the rule
suggests that it is, if not a precise guide, at least an accurate guide to what the Court
does” (Teger and Kosinski, 1980: 845). First, the Court is concerned with clarifying the
law, resolving disputes among the lower courts and issuing decision to address new legal

11

The events leading up to Mincey’s arrest and conviction were, indeed, unusual. An undercover
police officer bought drugs from Mincey in Mincey’s apartment. The officer later returned to the
apartment with several other officers, forced an entry (without a warrant), and tried to enter Mincey’s
bedroom. A shootout ensued, and the undercover officer was mortally wounded. Mincey was also
grievously wounded and taken to the hospital. The police secured the apartment and, still without a
warrant, brought in a special forensics team. The police occupied the apartment for a remarkable four full
days, making a detailed analysis of the contents of the apartment. Throughout the course of the
investigation, the police never sought a warrant. Less than five hours after Mincey was admitted to the
hospital, the police attempted to question him. He was intubated and could not speak, but he was able to
write answers to the officers’ questions (which were not recorded). During the interrogation, Mincey, who
may or may not have been medicated, repeatedly requested an attorney, lapsed into unconsciousness
several times, complained of excruciating pain, and questioned his own lucidity. Both the hand-written
statements Mincey produced in the hospital and the fruits of the search of his apartment were admitted at
trial. Mincey’s petition asserted that the warrantless search of his apartment violated his 4th and 14th
Amendment rights and that the in-hospital interrogation violated his 5th and 6th Amendment rights.

26

issues. Second, the Court is concerned with the rule of law rather than the outcome of
individual cases; the Court explicitly states that it is not interested in correcting
misapplications of the law in particular cases, so long as the lower court articulates the
correct rule of law. Finally, the Court is concerned with issues of broad legal
significance rather than esoteric legal issues that have limited real-world application; in
the relatively short text of Rule 10, the Court uses the word “important” no less than five
times.
While Rule 10 offers some insight into the goals and motivations of the Justices in
the agenda-setting process, it is at best a partial explanation. Beyond the almost
administrative concerns articulated in Rule 10, scholars believe that the Court attempts to
clarify and solidify legal policy consistent with the Justices’ own policy preferences.12 In
short, the Justices view their docket as instrumental, the individual cases as a means to
establishing particular policies they favor and as mechanisms for asserting control,
assuring compliance, by the lower courts. As Caldeira and Wright (1988: 1111) so
succinctly put it, “justices of the U.S. Supreme Court are motivated by ideological
preferences for public policy and . . . they pursue their policy goals by deciding cases
with maximum potential impact on political, social, or economic policy.” Such a
conceptualization of the Court’s goal in agenda-setting is broadly accepted (McGuire and
12

There is no need to delve into the deeper mystery of whether Justices’ policy preferences are
motivated by purely political beliefs and attitudes or whether they are, instead, motivated by a genuine
belief that “the law” dictates those policy positions. Although the evidence that Justices are motivated by
more general political attitudes is quite persuasive, that proposition is not necessary to the argument
presented in this research. Following the advice of Baum (1994: 757), I assume only that Justices have
policy preferences, ideas about how they believe the legal landscape should look, that their overriding
concern is to bring the rest of the legal community into compliance with those beliefs, and that these beliefs
(as manifested in voting patterns) “generally approximate a unidimensional structure.” Although I will
refer regularly to “ideology,” a liberal Justice may be motivated as much by a liberal construction of
constitutional theory as by left-leaning political beliefs.

27

Caldeira, 1993; Segal and Spaeth, 1993: 165-207; Songer, 1979) and consistent with the
understanding of the Court’s decision-making more generally (Epstein and Knight, 1998;
Rohde and Spaeth, 1976: 70-95; Segal and Spaeth, 1993; Schubert, 1965).
Although this notion that the Court uses its docket as a means to policy ends
might raise the hackles of those who favor a constrained Court, one that leaves the
creation of policy to the elected branches, it should not. Such a role for the Court does
not necessarily undermine majoritarian democracy; rather, a policy-motivated court with
the power of judicial review may actually assist a majoritarian legislature in adopting
socially optimal policy. Specifically, such a court may play an informational role, being
able to “correct” policies with unintended bad consequences, thus enabling the legislature
to enact riskier policies (Rogers, 2001).
Regardless of normative concerns, however, there is little doubt that the Supreme
Court has policy objectives that it pursues through both the agenda-setting process and its
decisions on the merits. This is not to suggest that the goals of Justices are
straightforward and uncomplicated. Other factors, especially individual ideas and group
norms about the proper business and obligations of the Court (Provine, 1981), may
temper the Court’s pursuit of policy outcomes. However, policy objectives stand out as
the most discrete and prominent feature in the judicial decision-making landscape.

CONSTRAINTS IN SUPREME-COURT AGENDA-SETTING
Despite the broad latitude the Court enjoys in selecting petitions for review, the
Court is not entirely unconstrained in setting its agenda.

28

First, in deciding whether or not to grant plenary review in a particular case,
Justices are presumably constrained by certain legal considerations, such as issues of
justiciability and jurisdiction (Perry, 1991b). Specifically, the Court cannot hear cases in
which the parties are in complete agreement, such that there is no controversy between
them, or in which the controversy has been rendered moot. Similarly, the Court cannot
hear cases that are not yet ripe for consideration, where the controversy is foreseeable but
not yet developed, or cases in which one party lacks standing to assert his claim.
Moreover, the Court lacks jurisdiction to decide issues of pure state law. Finally, there
are procedural bars to jurisdiction, such as the party seeking review failing to do so in a
timely manner.
These legal considerations create a threshold constraint on the Court’s plenary
review. As Provine (1981: 325) notes, “[t]he memos [Justice] Burton’s clerk’s wrote for
him on each case and the case selection voting patterns suggest that cases with
jurisdictional defects, inadequate records, and no clearly presented issues were the most
likely to be [dead] listed.”
Still, recognizing that the Court is somewhat constrained by legal concerns raises
the question of exactly how constrained they are. The literature provides a range of
possible answers. For example, while Provine (1981) suggests that jurisdictional issues
form a critical threshold in the certiorari process, Segal and Spaeth (1993: 206) assert
that these legal considerations are open to broad interpretation and subject to judicial
manipulation, such that they form no real constraint at all. In other words, even the most
egregious jurisdictional and jurisprudential defects can be overlooked if the Justices are
sufficiently motivated by their policy objectives to accept a case.
29

The most pressing constraint on the Justices, however, is time. Deciding a case
on the merits involves many time-consuming tasks: reading the record and substantive
briefs, researching legal issues, preparing for and attending oral arguments, engaging in
debate with other Justices, preparing draft opinions, and engaging in the iterative process
of circulating opinions for comment and revising accordingly. Even with the assistance
of dedicated and well-trained law clerks, the Court simply cannot generate merits
decisions on more than a small fraction of the cases for which review is sought.
Moreover, the more petitions the Court must screen, the less time the Court has to
devote to creating policy through merits review. As discussed in Chapter One, the
number of litigants seeking review from the Supreme Court has risen dramatically over
the past 75 years. One study in the mid-1970s estimated that, at most, each Justice
spends an average of 9.5 minutes on each paid petition for certiorari and even less time
on each unpaid petition (Casper and Posner, 1976: 65-66). Indeed, as discussed in
Jucewicz and Baum (1990), the strain of the Court’s burgeoning workload during the
1970s and 1980s prompted calls for dramatic reforms--such as the creation of a National
Court of Appeals--and actually resulted in a number of legislative and administrative
measures to limit petitions (discussed in greater detail in Chapter One).

COGNITIVE TASK
The Court’s primary objective, then, is to maximize legal policy making within
the jurisprudential and logistical constraints imposed upon it.
“To the justices, a given case is not particularly important. As one clerk
explained, ‘Frankly, they don’t care enough about any individual case . . . . Another case
30

may come up with the [same]issue.’ To the justices, the process of selecting cases is, in
fact, the process of selecting issues” (Palmer, 1999: 58 (quoting Perry, 1991a: 153)). In
other words, the Court’s goals and motivations in the agenda-setting process are best
served by attention to issues rather than cases.
Still, the business of the Courts is conducted in the context of cases, and to
achieve its policy-making objective, the Court must quickly and efficiently identify a
handful of petitions, among the thousands of petitions received each year, that will allow
the Court to have the most significant policy impact.
[T]he justices maximize their positions subject to a complicated matrix of
legal rules, social norms, and institutional expectations. Nevertheless, the
efficient pursuit of policy objectives implies that justices will devote their
resources to those cases that have the most profound implications for
policy relative to the effort required to hear and to decide the cases
(Caldeira and Wright, 1988: 1111).
Efficiently solving that calculus, however, is no mean feat.
The fundamental problem is one of information. “[T]he members of the Court
must still rely upon lawyers [and pro se litigants] to provide them with reliable
intelligence regarding the legal issues raised in a case as well as their likely political
consequences. These informational needs are extensive. In the selection of cases . . . the
justices face uncertainty in locating genuinely meritorious cases from among the
thousands of petitions brought annually to the Court” (McGuire, 1998: 509). Thus, while
the Court may be interested in issues, it cannot create those issues out of whole cloth but
must, instead, find those issues in the cases brought before it.
One way in which the Court might manage the immense task of sifting through
petitions for review is to rely upon heuristics, or cognitive shortcuts, to identify those
31

cases worthy of review. “[H]euristics are rather parsimonious and effortless, but often
fallible and logically inadequate, ways of problem solving and information processing. A
heuristic provides a simplifying routine or ‘rule of thumb’ that leads to approximate
solutions to many everyday problems” (Manstead and Hewstone, 1995: 296; see also
Kahneman, Slovic and Tversky, 1982, for a detailed discussion of heuristics in cognitive
psychology). In the field of cognitive psychology, the literature on heuristics portrays
people as “cognitive misers”; faced with an immensely complex world and bombarded
with information, people must develop shortcuts, simplified decision rules, that allow
them to make reasonable decisions with limited cognitive effort (Manstead and
Hewstone, 1995: 296-300; Lau and Redlawsk, 2001: 952; Kahneman, Slovic and
Tversky, 1982). Thus, while heuristics may be fallible, they are nevertheless necessary:
“heuristics serve the important function of simplification and economy in an extremely
complex environment in which a systematic and exhaustive attempt at every routine task
would cause a permanent overload of information and a breakdown of adaptive behavior”
(Manstead and Hewstone, 1995: 299).
In the field of political science, interest in heuristics has focused primarily on their
use by the mass electorate in making political judgments, particularly with respect to the
evaluation of candidates (see Lau and Redlawsk, 2001, for an excellent overview of the
literature on the use of heuristics by voters; for similar overviews, see also Kuklinski and
Hurley, 1994, and Kuklinski et al., 2001). For example, in evaluating political
candidates, voters can and do rely on the candidate’s party affiliation (Rahn, 1993),
endorsements by interest groups (Brady and Sniderman, 1985), and even the candidate’s

32

appearance (Riggle et al., 1992), rather than expending the effort to obtain detailed
information about the candidate’s policy preferences.
However, judicial scholars have borrowed the concept of heuristics to help
explain the manner in which Supreme Court Justices slog through the mountain of
petitions they receive every year, searching for the best vehicles for legal policy-making.
Specifically, as first argued by Tanenhaus et al. (1963), the Court relies upon certain cues
in petitions for review, the presence or absence of particular factors that might correlate
with the amorphous quality of “certworthiness,” to select petitions for closer
consideration. “The cue theory of certiorari maintains that the justices of the Supreme
Court employ cues as a means of separating those petitions worthy of scrutiny from those
that may be discarded without further study” (Tanenhaus et al., 1963: 121). The presence
of one or more of these cues will draw the Court’s attention, increasing the probability
that the case will receive detailed consideration and, thus, increasing the probability that
the Court will ultimately grant review.
Tanenhaus et al. identified three cues that did, in fact, correlate strongly with the
probability that the Court would grant review: dissension, either “among the judges of
the court immediately below, or between two or more courts and agencies in a given
case”; the federal government as the petitioning party; and the presence of a civil liberties
issue. Subsequent analyses have shown these cues to be robust (Teger and Kosinski,
1980; Armstrong and Johnson, 1982).
Moreover, other research on the Court’s agenda-setting function, whether
explicitly invoking cue theory or not, has identified other case characteristics that may

33

signal Justices that a particular petition presents a prime opportunity to engage in legal
policy-making.13
The cues that have been identified in the literature are not arbitrary. Quite the
contrary, both the more general literature on heuristics and the more specific literature on
cue theory posit that the cues, signals, or heuristics relied upon do contain informational
content, although the accuracy of that informational content is not always high.14
Specifically, with respect to the cues identified in the agenda-setting literature, these case
characteristics send signals about the value of the case as a vehicle for legal
policymaking. In other words, these case characteristics are strongly associated, either
positively or negatively, with some intangible quality of “certworthiness,” and their
presence provides Justices with low-cost information about the value of the case.
It is important to note that, while the literature on cue theory has shown a strong
correlation between certain case characteristics and the probability that the Court will
grant plenary review, scholars have not been able to conclusively determine the existence
of a causal relationship between these case characteristics and the certiorari decision. As
noted above, the alleged informational content of the cues derives from their correlation
13

In this analysis, I will use the terms “cue” and “signal” interchangeably. However, Cameron,
Segal, and Songer (2000) note that “signal” is a term of art in game theoretic literature and refers to a
message that can be manipulated or chosen by the sender; some cues, such as participation by amici curiae,
are true signals in that they can be manipulated by the sender (the amici), but other cues, such as the pro se
representation of the petitioner, are exogenous and more properly considered “indices.” The extent to
which various cues are manipulable or not is beyond the scope of this discussion. But it is worth noting
that some cues that, on their face, appear exogenous, actually can be manipulated by participants in the
legal process. For example, the presence of a circuit conflict would seem to be exogenous. However, a
Circuit Court of Appeals may choose to acknowledge the conflict created by its decision and invite
Supreme Court resolution, or it may, instead, attempt to distinguish the case before it from the cases in
other circuits and thus mask the existence of the conflict.
14
For example, the party affiliation heuristic used by so many voters is not random. We can
generally make educated guesses about a candidate’s broad policy positions based on his party affiliation.
We might infer, for example, that a Democratic candidate is pro-choice while a Republican candidate is
pro-life. Thus, the heuristic does impart information to the voter, although that information is not always
accurate; there certainly are pro-life Democrats and pro-choice Republicans.

34

with some slippery underlying quality of “certworthiness.”15 Yet it is unclear whether the
Justices actually rely upon the presence of cues as an indication of certworthiness or
whether the Justices ascertain certworthiness quite independent of the cues, and the cues
simply correlate with the Justices’ choices.
For example, McGuire and Caldeira (1993) found that cases in which specialized
Supreme Court attorneys participated were more likely to be accepted for review. The
cue theory posits that Justices note the presence of these specialized attorneys and draw
some inference about the significance of the case from their involvement. To illustrate:

Certworthiness Ö Specialized Attorneys Ö Court Attention/Review

It is possible, however, that the specialized attorneys are simply better able to predict
which cases the Court will otherwise select for review and choose to align themselves
with those cases. To illustrate:

Specialized Attorneys
Ü
Certworthiness
Þ
Court Attention/Review

15

Indeed, Justice Harlan, commenting on the discretionary nature of the certiorari process at the
New York City Bar Association in 1958, stated that “[f]requently the question whether a case is
‘certworthy’ is more a matter of ‘feel’ than of precisely ascertainable rules” (Reimann, 1999: 169). Justice
Harlan’s comment suggests that the Justices do not make conscious reference to the cue characteristics.

35

Thus, specialized attorneys do, in fact, involve themselves with cases that are “better,” or
more “certworthy,” but it is unclear whether the Justices actually use the presence of
specialized attorneys in making their determination about certworthiness.
Indeed, Provine (1981: 324) criticizes the premise behind cue theory by noting
that Tanenhaus’s cues, while correlated with the grant of review, do not perfectly predict
deadlisting; specifically, Provine’s examination of the papers of Justice Burton revealed
that there were numerous cases presenting one or more of Tanenhaus’s cues that were
nevertheless deadlisted, denied with only the most cursory consideration. According to
Provine (1981: 324), “[i]f the cues actually served this short-circuiting purpose [described
by Tanenhaus et al.], cases containing cues should not appear on the special lists.”
Provine goes on to conclude that “neither the justices nor the clerks rely on a fixed set of
cues to separate cases into those worthy of scrutiny and those to be discarded summarily”
(1981: 325). However, the psychological literature on heuristics does not suggest that
they must be applied in a lock-step manner, without any deviation.
Provine also suggests that, because of the assistance provided by law clerks, the
Justices do not need to rely on such gross methods for screening cases. Palmer (2001)
examines the influence of the cert-pool memos on the Court’s certiorari decisions;
specifically, Palmer examined the cert-pool memos for cases granted review during the
1972-1974 and 1984-1985 Supreme Court Terms. Palmer notes that the law clerk memos
prepared by the cert pool did not consistently offer any recommendation on the ultimate
issue of review until 1981. Moreover, during the 1984-1985 Terms, the cert pool only
recommended the Court grant review in half of the cases in which the Court actually did
grant review, and the cert-pool memo recommended outright denial of review in almost a
36

quarter of the cases in which the Court granted review (Palmer, 2001: 111-114). The fact
that the Court granted review over the negative recommendation of the pool memo so
frequently indicates that the Court is looking beyond the cert-pool recommendation and
conducting some independent analysis of certworthiness. This suggests that, while the
pool memo may summarize the information set forth in the petitions, it does not supplant
the decision-making task of the Justices themselves.
Moreover, there is a strong theoretical basis for concluding that the identified case
characteristics are, themselves, signals to the Justices of the certworthiness of individual
petitions rather than merely manifestations of other more elusive qualities of
certworthiness. First, given the sheer magnitude of the task of selecting cases for review,
it appears eminently reasonable, if not absolutely necessary, that the Justices employ
some sort of cognitive shortcuts; even relying upon the summaries in the cert-pool
memos, the Justices have an enormous volume of information to consider. Second, the
logical connection between the identified cues and the objectives of the Justices, both
stated in Rule 10 and inferred from judicial behavior, lends the theory face validity.
The next section of this chapter discusses both the Tanenhaus cues and the
additional cues subsequently identified in the literature.

POTENTIAL CUES TO CERTWORTHINESS
Ideology of Decision Below
Starting from the premise that members of the Court are motivated, in part, by the
desire to move legal policy in a direction consistent with their own ideological
preferences, it follows that the majority of the Court has more to gain from accepting
37

cases decided in an ideologically inconsistent manner and reversing them than by
accepting cases decided in an ideologically consistent manner and affirming them.
Because it only takes four Justices to grant certiorari, it is possible that the
ideological minority could vote to accept cases at odds with its position. However, there
is evidence to indicate that Justices are strategic in this regard and generally unwilling to
vote to grant certiorari in cases where they have little likelihood of prevailing on the
merits (Caldeira, Wright, and Zorn, 1999; Segal and Spaeth, 1993: 192). For example,
the liberal minority during the Burger Court years was unlikely to grant certiorari to
conservative lower court rulings when the minority knew that the conservative majority
would likely prevail at the merits stage and, thus, the conservative position would
become further entrenched.
Thus, the Court can maximize the efficiency of its agenda-setting function--get
the most policy-making bang for the buck, so to speak--by focusing primarily on
correcting errors made by the Court below (Armstrong and Johnson, 1982: 149; Boucher
and Segal, 1995; Brenner and Krol, 1989: 832-33; Schubert, 1958; Schubert, 1962;
Songer, 1979; Ulmer, 1972). So long as the Courts of Appeals and the State Supreme
Courts are “getting it right,” where “right” is defined by the policy preferences of the
Court’s majority rather than by any objective legal benchmark, the Supreme Court is
generally content to allow the rulings of the lower courts stand, even on issues of
significant legal importance.16 In other words, the Court limits its workload by focusing
on correcting the errors of the courts below.17

16

Note, however, that the Court does have some incentive to issue opinions even on matters that the
lower courts are deciding in a manner consistent with the Court’s own policy preferences. Specifically, the

38

Accordingly, one cue the Supreme Court may rely upon in winnowing the
petitions brought before it is the ideological direction of the lower court decision (Songer,
1979). Specifically, when the majority of the Court is ideologically conservative, the
Court is more likely to grant certiorari when the decision below is liberal; conversely,
when the majority of the Court is ideologically liberal, the Court is more likely to grant
certiorari when the decision below is conservative. Although there are exceptions, the
ideological direction of most cases is relatively easy to ascertain, and thus it performs the
function of Tanenhaus et al.’s cues: providing a shortcut for Justices in identifying
petitions that require closer review.

Lower Court Behavior
Certain aspects of lower court behavior may signal the Court that a particular
issue is more appropriate for review. Most notably, disagreement among or within the
lower courts--as manifested by a conflict between U.S. Courts of Appeals and/or State
Supreme Courts,18 a dissent from the decision for which review is sought, or a reversal in
the case’s procedural history--may indicate that a case is more appropriate for Supreme
precedential value of a Supreme Court opinion is greater and presumably more enduring than decisions by
the lower courts, and the Supreme Court may be motivated to address and issue solely for the sake of
solidifying it’s policy position for the long term. There is some evidence that the Court does engage in
such so-called “aggressive grants” (Boucher and Segal, 1995).
17
As discussed in Smith (1999: 388), this ideological error correction differs from the sort of legal
error correction in which the Court is popularly believed to engage. Contrary to popular belief, the U.S.
Supreme Court is not able to, nor is it interested in, correcting every legal and factual error committed by
the lower courts. Indeed, Rule 10 specifically states that the Court is unlikely to grant review where “the
asserted error consists of erroneous factual findings or the misapplication of a properly stated rule of law.”
18
Note the distinction, here, between a conflict alleged by the petitioner and an actual conflict in the
law. While the former is obvious on the face of the petition, meeting the requirement of a cue or heuristic,
ascertaining the existence of an actual conflict requires some research (presumably on the part of law
clerks). Using actual conflicts coded by the New York University Supreme Court Project (1984), Caldeira
and Wright (1988) found that both alleged conflict and actual conflict independently increase the
probability that the Court will grant plenary review.

39

Court review; these case characteristics are explicitly mentioned in Rule 10, but they are
relevant to the task of identifying good vehicles for legal policy-making. First, these
procedural characteristics reflect uncertainty in the law, evidence that reasonable minds
can reach conflicting conclusions, and thus invoke the Court’s administrative mandate to
clarify legal rules and create consistency between jurisdictions. Second, from a policy
perspective, these factors may represent “close cases,” legal issues that, because they are
not yet settled or clear, provide fertile ground for the expression of legal policy
preferences. One consequence of the norm of stare decisis, which disfavors revisiting
legal issues that are already squarely decided, is that cases which are not clearly settled
provide the best opportunity to advance legal policy. Thus it is not surprising that these
indicia of conflict and uncertainty have been shown to correlate with the probability that
the Court will accept review (Caldeira and Wright, 1988; Tanenhaus et al., 1963; Teger
and Kosinski, 1980; Ulmer, 1984).
Another signal of legal conflict is an allegation that the lower court decision
conflicts with a prior decision of the U.S. Supreme Court. Such conflict not only signals
a need for the Supreme Court to reassert control over its agents, but it may also signal an
area of the law that is in flux, one that is ripe for reconsideration and fine-tuning.
Specifically, lower courts ignoring or openly defying Supreme Court precedent may be
an indication that existing legal principles no longer apply to changing social, economic,
and technological circumstances. Ulmer (1983, 1984) has shown that this other sort of
conflict, which is also mentioned in Rule 10, is associated with an increased probability
that the Court will grant review.

40

In addition to indicia of conflict or uncertainty, George and Solimine (2001)
demonstrate that the U.S. Supreme Court is more likely, ceteris paribus, to review a
decision by a Court of Appeals sitting en banc than a decision by a panel of a Court of
Appeals. They assert that the decision by a Court of Appeals to review a case en banc
sends a signal about the importance of or controversy surrounding the legal issues
presented by the case. Indeed, the criteria used by the U.S. Courts of Appeals to
determine whether or not to grant en banc review are theoretically similar to the criteria
used by the U.S. Supreme Court to determine whether or not to grant review; as a result,
the decision by the Courts of Appeals serves as a “first cut,” a trial run of the certiorari
process. When a Court of Appeals determines that a case is worthy of en banc review, it
sends a strong signal to the Supreme Court that the case is significant enough to warrant
review.
Accordingly, the Court is more likely to accept review when there is a dissent on
the court below, when there is a reversal in the case’s procedural history, when there is an
allegation of an interjurisdictional conflict, an allegation of a break from existing
Supreme Court precedent, and/or an en banc decision by a U.S. Court of Appeals.

The Parties: Attorney Representation and Solicitor General Involvement
The Court’s primary source of information about the certworthiness of a
particular case is the briefs filed by the litigants themselves, both the petitioner and the
respondent. Yet not all of that information is equally valuable. One source of variation
among cases on the Court’s docket is the credibility of the information presented to the
Court.
41

For example, experienced attorneys and “repeat players” (attorneys who appear
regularly before the Court) are better sources of information for Justices (McGuire, 1993:
197-198; McGuire and Caldeira, 1993; Galanter, 1974). These attorneys may possess
better legal training and advocacy skills. More importantly, however, these elite
attorneys are perhaps better able to size up the merits of their clients’ cases and are more
credible because they have strong disincentives to lie or exaggerate their clients’ claims.
As McGuire and Caldeira (1993: 719) explain:
Lawyers seek to gain access to the Supreme Court to win a lawsuit for
clients, and this no doubt provides a strong incentive to exaggerate the
importance of a case. To be sure, lawyers are “officers of the court,”19 but
the strong divergence of interest reduces their credibility as honest brokers
of the importance of a case. . . . Those who bring cases to the Court over
and over again [however] build up a reputation in the eyes of the Court, a
costly and valuable asset (McGuire 1993). Reputation not only helps to
attract and maintain clients but also translates into influence in the Court.
Thus, lawyers and other repeat players have a strong interest in
maintaining credibility in the Court.
Thus, the identity of the individual providing the information to the Court may itself
serve as a cue to the Court about the credibility of that information and, accordingly, the
degree to which the Court should pay closer attention to that case.
It is an unfortunate reality of the American legal system that some litigants, even
those with worthy claims, cannot afford to obtain representation, and some litigants who
can afford representation nevertheless choose to represent themselves.20

19

The effect of the ethical obligation attorneys have to the court should not be underestimated. In
Slater v. U.S. (Doc. No. 86-5650), Mr. Slater’s court-appointed attorney concludes the affidavit, submitted
in lieu of a petition for writ of certiorari, by noting that he “concluded that any issues in this case that could
be presented to this Court in a petition for writ of certiorari would be characterized as frivolous.”
20
As evidence of the difficulty in obtaining counsel, consider the case of John Young (Doc. No. 775620) who was on Georgia’s death row. Mr. Young’s court-appointed counsel withdrew after the
mandatory appeals, in part because the attorney had legal problems of his own (he was charged with
possession of marijuana with intent to distribute). Mr. Young sought habeas relief, challenging Georgia’s

42

“Unrepresented litigants may clutter up cases with rambling, illogical reams of
what purport to be pleadings, motions, and briefs. They may seek out courtrooms as
forums to vent strongly held but legally unfounded social and political theories or as
battlegrounds to satisfy private, legally unredressable vendettas” (Nicholas, 1988: 351).
Furthermore, pro se litigants often do not understand the procedural limitations on the
Court and have difficulty putting their issues into a proper legal framework. For
example, in Mondaine v. Wyrick (Doc. No. 84-5986), the petitioner pro se expressed his
frustration with the lower court’s assertion that his habeas claims were barred by his
failure to exhaust state court remedies: “The petitioner does not have the slightest idea of
what remedies he must exhaust to be able to maintain federal habeas corpus. . . . If this
court can find any available state remedies, if this court can direct the petitioner to any
remedy then petitioner would be happy to represent the perjury issue again; what remedy
is this?” (Brief of Petitioner at 6-7).
Even when they are able to present cogent legal arguments, pro se litigants are not
always able to explain why their case is worthy of review. Smith (1999) found that pro
se litigants were less likely to make reference to the Rule 10 factors than were attorneys.
Thus, pro se litigants may be less able to send the appropriate signals to the Court about
the merits of their petition; however certworthy the underlying claim may be, the
petitions filed by pro se litigants may not display the cues that will garner the Court’s
attention.

statute that allowed women to opt out of jury service. The record in the case indicates that, despite Mr.
Young’s grave situation and the relative merit of his claim, he was unable to obtain representation from the
NAACP, the ACLU, the Southern Poverty Law Center, the Christian Leadership Conference, or the
University of Georgia Law School clinic program. Eventually, Mr. Young was able to retain a pro bono
attorney, but his petition for review was nevertheless denied.

43

More importantly, for purposes of this discussion at least, pro se status may itself
be a “negative” cue of certworthiness, a factor that independently reduces the Court’s
interest in a petition. Specifically, pro se litigants may lack credibility with the Court.
Assertions of conflict or allegations of issue importance, coming from a lay person, may
lack the gravitas of similar assertions and allegations by members of the bar.
Furthermore, the fact that a pro se litigant was unable to obtain counsel, even through one
of the many interest groups that provide legal representation (such as the NAACP Legal
Defense Fund, the ACLU, and the American Family Association), may be a signal about
the case’s relative lack of importance.
Although there is a bright-line distinction between pro se petitioners and
petitioners who are represented by counsel, that distinction is not the end of the story.
Not all attorneys are created equal. McGuire (1993), McGuire and Caldeira (1993), and
Galanter (1974), have all argued that elite lawyers who appear regularly before the
Supreme Court enjoy special credibility and, accordingly, special success with the Court.
The most obvious example of a repeat player before the Supreme Court is the
U.S. Solicitor General. Judicial scholars have long recognized the peculiar advantage
that the Solicitor General has before the Supreme Court and have posited that the Court
pays special attention to, and is generally more favorable towards, petitions for review
authored by the Solicitor General (Tanenhaus et al., 1963; Teger and Kosinski, 1980;
McGuire, 1998).21 Similarly, amicus briefs filed by the Solicitor General seem to carry
more weight than other amici (Segal, 1988). The reason for the Solicitor General’s
21

This apparent advantage also applies to Supreme Court decisions on the merits, both when the
Solicitor General is a litigant (Canon and Giles, 1972) and when the Solicitor General submits an amicus
brief (Segal, 1988).

44

particular success is less certain. Some scholars attribute the Solicitor General’s success
at least in part to judicial deference to the executive branch and the collegial relationship
between the Solicitor General’s office and the Court (Tanenhaus et al., 1963; Caplan,
1987; Salokar, 1992; Segal, 1988). McGuire (1998), however, found that the “special
role” of the Solicitor General derives not from any special status but from litigation
experience; when McGuire controlled for litigation experience, the Solicitor General’s
elevated rate of success disappeared.
Whether the Solicitor General is simply a readily identifiable repeat-player or
whether the Solicitor General actually enjoys some advantage above and beyond that
born of litigation expertise, there is little question that the Solicitor General’s support for
a petition for review increases the probability that review will be granted. “Indeed, the
proportion of the Solicitor General’s petitions for certiorari that the Court grants is
consistently over fifty percent, whereas paid petitions filed by other parties are granted at
a rate of only about three percent”22 (Cordray and Cordray, 2001: 763). What’s more, the
Court will, on occasion, solicit the Solicitor General’s position on a petition for certiorari
in cases where the United States is not a direct party.
Taken together, these studies suggest a continuum of credibility in the information
presented to the Court, ranging from information presented by pro se litigants (who lack
the legal training to assess the certworthiness of their petitions and who lack disincentives
to exaggerate the importance of their claims) to information presented by the U.S.
Solicitor General (who brings remarkable experience to bear on the assessment of
22

It makes sense to compare petitions filed by the Solicitor General to paying private litigants, rather
than all private litigants, because the Solicitor General’s petitions are always on the paid docket. If, in fact,
there is some inherent bias against the IFP docket, comparing the Solicitor General petitions to both paid
and unpaid private litigants would be comparing apples and oranges.

45

certworthiness and whose professional relationship with the Court presents the ultimate
disincentive to exaggerate). If an argument presented by an ordinary attorney represents
a baseline of credibility, a petition filed by a pro se petitioner--which necessarily is an
argument in support of review--is less credible and thus less likely to be granted.
Conversely, an argument presented by the U.S. Solicitor General is more credible and is
thus more likely to sway the Court; if the Solicitor General supports review, the Court is
more likely to grant review, but if the Solicitor General opposes review, the Court is
more likely to deny review.23

Amicus Briefs
Cert-stage amicus briefs provide yet another source of information to Justices
about the issues implicated by and importance of particular petitions for review. As
Caldeira and Wright (1988: 1112) hypothesize,
the potential significance of a case is proportional to the demand for
adjudication among affected parties and that the amount of amicus curiae
participation reflects the demand for adjudication. . . . [A]micus curiae
participation by organized interests provides information, or signals-otherwise largely unavailable--about the political, social, and economic
significance of cases on the Supreme Court’s paid docket.
Because the preparation of an amicus brief is quite expensive,24 they are not just cheap
talk but real, reliable indicators of the significance of a case beyond its implications for

23

Obviously, when the Solicitor General is the petitioner, the Solicitor General supports review.
However, the Solicitor General sometimes supports review when the Solicitor General is the respondent,
i.e. when the Solicitor General prevailed in the Court below. Of the sample used for this dissertation, the
Solicitor General was the respondent in 257 of the cases. Of those 257 cases, the Solicitor General opposed
review in 136 of the cases, supported review in 21 cases, and expressed no position at all in 100 cases.
24
Caldeira and Wright’s research estimates that a single amicus brief costs an interest group between
$15,000 and $20,000. Their estimates are based on surveys of interest groups done in the late 1980s.
Presumably, the cost of filing a brief has risen considerably since then.

46

the real parties in interest. Moreover, like the Solicitor General, amici are often repeat
players and, consequently, have incentives to represent the merits and significance of
petitions accurately (Spriggs and Wahlbeck, 1997: 367-68).
It is important to note that the substance of these amicus briefs is generally
irrelevant. Both amicus briefs supporting a grant of review and amicus briefs opposing a
grant of review increase the probability that the Court will grant review. While this may,
at first blush, seem counter-intuitive, it fits squarely with the cue theory of agendasetting. The cue or signal involved is generated by the interest group paying attention to
the case and considering it significant enough to expend valuable resources on at the cert
stage, not by the arguments set forth in the briefs themselves. The latent characteristic
that the Court is attempting to ascertain by reference to the cues is the potential policy
impact of the case; if policy-oriented interest groups--especially interest groups with
experience surveying the legal landscape--identify a particular case as important enough
to justify expending their own resources, that fact serves as a cue to the Justices about the
potential the case possesses as a vehicle for legal policy-making. Accordingly, the Court
is more likely to grant review to petitions for which there is cert-stage amicus activity.

CONCLUSION
Starting with the relatively uncontroversial premise that Supreme Court Justices
view their docket as instrumental in advancing their policy preferences, and taking into
account the necessary constraints on the ability of the Justices to comb through the
mountain of petitions received to find those gems of policy-making potential, cue theory
posits that Justices employ cognitive shortcuts in their agenda-setting function.
47

Specifically, the Justices rely on heuristics, on readily ascertainable cues or signals about
the significance and potential for policy-making of a case. The more such cues present,
so the theory goes, the more attention the Justices will give to the case and, consequently,
the more likely the Justices are to accept review.
Although a few of the numerous empirical studies of agenda-setting have
included IFP petitions in their sample, there is no comprehensive study of cues contained
in the IFP petitions themselves, nor is there any research comparing the influence of cues
on agenda-setting on the two dockets, paid and unpaid. In the chapters that follow, I
begin to explore the relatively uncharted waters of the IFP docket with particular
attention to the signals of significance that have been identified in the agenda-setting
literature.

48

CHAPTER 3
HYPOTHESES AND DATA

INTRODUCTION
The ultimate question posed by this research is whether IFP petitions share equal
footing with paid petitions or whether, in the alternative, the Court treats IFP petitions
differently for purposes of granting or denying plenary review. More specifically, while
we know that IFP petitions are administratively segregated and that the Court and other
observers consider the IFP/paid distinction to have qualitative significance, we do not
know whether the distinction translates into disparate treatment during the agenda-setting
process.
As an initial step in seeking to answer this fundamental question, this dissertation
considers the relative effects of various cues or correlates to the Supreme Court’s grant of
plenary review on both paid and unpaid petitions. This chapter outlines the hypotheses,
data, and methods around which the analysis is built.

CUES IN THE CONTEXT OF THE IFP DOCKET: HYPOTHESES
Given the primary research question articulated above, the null hypothesis is that
paid and IFP cases do, in fact, enjoy equal opportunity before the Court. Put more
precisely:
49

Hypothesis I:
There is no difference between the IFP and paid cases with respect to the Court’s
agenda-setting function.
Moreover, given what we know about the cue theory of the Court’s agenda-setting
function, Hypothesis I can be expressed as two corollary hypotheses:
Hypothesis Ia:
The base probability of the Court accepting a petition in the absence of any
identified cue is the same for both paid and unpaid petitions.
Hypothesis Ib:
The marginal effect of each of the identified cues on the probability of the Court
accepting a petition is the same for both paid and unpaid petitions.
There is reason to expect, however, that the null hypothesis of equivalence
between the paid and unpaid dockets will prove false and that, instead, the alternative will
be supported:
Hypothesis II:
The calculus of the certiorari decision varies between the paid and unpaid
dockets.
There are two primary ways in which the agenda-setting process may be affected
by the IFP status of the petition.
First, the IFP status of the petitioner may, itself, be a cue utilized by the Supreme
Court. Specifically, as discussed in Chapter One, there is a common wisdom, shared by
scholars and Justices alike, that unpaid petitions are generally frivolous and
uncertworthy. As a result, the mere fact that a case is unpaid may create an a priori
assumption that the case is not certworthy. Conversely, the fact that petitioners who pay
50

the filing fee are, essentially, putting their money where their mouth is may serve as a
signal to the Justices that these petitions are more likely to raise legitimate legal issues.25
Indeed, the Court has explicitly noted that IFP petitioners, particularly those who file pro
se, “are not subject to the financial considerations . . . that deter other litigants from filing
frivolous petitions” (In re McDonald, 489 U.S. 180, 184 (1989) (per curiam)). Thus,
while the imposition of a filing fee creates an incentive for attorneys and petitioners to
independently screen the merits of their petitions, to file only when they have a legitimate
reason to believe that they might prevail, no such incentive exists for unpaid petitioners.
This leads us to Hypothesis IIa:
Hypothesis IIa:
The base probability of the Court accepting a petition in the absence of any
identified cue is lower for unpaid petitions than for paid petitions. In other words,
all else equal, the Court is less likely to accept an unpaid petition than a paid
petition.
In addition to the possibility of an independent effect of IFP status on the
probability of the Court accepting review, the marginal effect of other previously
identified cues may vary between the two dockets. Specifically, if the cue characteristics
do provide information to the Court about the certworthiness of petitions, we would
expect that the cues would have greater marginal effect in the context of the docket on
which they are less prevalent.

25

Note, however, that the filing fee required to submit a paid petition to the Supreme Court does not
reflect anything approaching the actual cost of litigation. Accordingly, Judge Richard Posner of the
Seventh Circuit Court of Appeals and former Solicitor General Rex Lee, among others, have advocated
raising the filing fees assessed by the Federal courts significantly in an effort to reduce the caseload of the
Federal judiciary and, more specifically, to winnow out cases that are either frivolous or trivial (See Beier,
1990, for an overview and critique of this argument). Given that the filing fee is so low, paying it may not
be a particularly strong signal about credibility.

51

By way of analogy, consider the effect of a job applicant having internship
experience. If there are only two or three individuals in the applicant pool with such
experience on their resumes, the internship really stands out, sets these individuals apart
from the crowd. If, however, three quarters of the applicants have internship experience
on their resume, then the value of internship experience as a signal of distinction or
exceptional quality is greatly diminished; a potential employer will focus more attention
on other aspects of the applicants’ resumes, specifically on those aspects that do a better
job of differentiating between applicants.
Thus, if certain cues are less prevalent on the unpaid docket than on the paid
docket and if the Justices are cognizant of the paid/unpaid status of the petitions as they
review them, we would expect the cues to have greater marginal effects in the context of
the unpaid docket.26 Conversely, if other cues--particularly negative cues, such as a pro
se petitioner, that reduce the likelihood of review--are less prevalent on the paid docket,
we would expect those cues to have a greater marginal effect (albeit a negative one) in
the context of the paid docket.
For example, before Warren Burger was appointed to the Court, the government
was required to respond to all petitions in which named the government as a respondent.
Chief Justice Burger, however, eliminated that requirement. The number of cases in
which the Solicitor General has waived response has risen over the years:

26

Teger and Kosinski (1980) and others assume that case factors that are positively associated with
the grant of plenary review are less prevalent on the unpaid docket than on the paid docket. However, to
date no one has empirically demonstrated the relative frequency with which most identified cues arise on
the paid and unpaid docket. The descriptive analysis presented in Chapter 4 makes such a comparison and,
in so doing, helps to identify which cues might have greater marginal effects in the context of the unpaid
docket.

52

In the 1983 Term, responses were waived in 54% of all cases in which the
Government was a respondent, in 79% of in forma pauperis criminal
cases, and in 60% of the paid criminal cases. But during the 1990 Term,
responses were waived in 70% of all Government cases, in 83% of in
forma pauperis cases and in 50% of paid criminal cases (Stern et. al, 1993:
434, n.47).
Now the government responds to only about 5% of the IFP petitions which name
the government as a respondent (Mauro, 1998). Generally, this trend has been viewed as
a sign that IFP petitions are unimportant; Alan Morrison of the Public Citizen Litigation
Group commented that the government’s failure to respond “is a not-so-subtle signal to
the court that it can ignore these cases” (Mauro, 1998). However, when the government
does respond to an IFP petition, the response may be a strong signal that the case merits
further consideration; on the paid docket, where government response is more routine, the
signal may lack much power.
Arguably, the effect of positive cues on the selection of IFP petitions may simply
be less significant across the board. To the extent that the Court views unpaid petitioners
as less credible than paid petitioners, they may disregard or discount the allegation of
positive cues provided by unpaid petitioners. However, any perceived “lack of
credibility” on the part of unpaid petitioners, as described in the explanation of
Hypothesis IIa, arises from a preconcieved notion that the petitions lack merit and from
the lack of any filing fee that would force petitioners and their attorneys to seriously
assess the overall merit of their petitions. In other words, IFP status may implicate the
decision to seek review in the first place, but there is no reason to believe that IFP status
implicates the substance of the petition, the invocation of particular cues or legal
arguments. With respect to allegations of cue factors, there is no reason to believe that
53

attorneys representing IFP petitioners are more likely to mislead the Court on the merits
of their clients’ petititions than are similarly situated attorneys representing paid
petitioners; similarly there is no reason to believe that indigent pro se petitioners are more
likely to misrepresent the merits of their case than are paying pro se petitioners. As a
result, there is no reason to believe that the Justices would put less faith in an allegation
of a positive cue by an indigent litigant than in a similar allegation by a paying litigant;
any difference in the manner in which the Justices weight the cues across the two
dockets, then, is more likely to be a result of the value of the cue in differentiating cases
than a result of the Justices simply not believing that the allegation of the cue is true.
Accordingly,
Hypothesis IIb:
The marginal effect of some or all of the identified cues on the probability of the
Court accepting a petition will vary between the unpaid docket and the paid
docket, with cues having a greater effect in the context of the docket on which
they are less prevalent.
Hypothesis IIa and IIb are not mutually exclusive. Indeed, a priori, I expect
support for both hypotheses. Specifically, I expect that the baseline probability that an
IFP petition will be granted review, in the absence of any other cues, will be lower than
the baseline probability that a paid petition will be granted review, and that the marginal
effects of various cues on those baseline probabilities will vary between the paid and
unpaid petitions.

54

DATA
Testing these hypotheses requires analysis of the relationship between case cues
and Supreme Court dispositions of both paid and unpaid cases. Because this research
requires case-level information on enough IFP petitions, both granted and denied, to
perform meaningful statistical analysis, it required creation of an original dataset. This
section describes the population, sampling, and coding involved in creating the dataset
used for the substantive analysis described in the following chapters.

Population
This study examines certiorari decisions of the Supreme Court from the 1976
through 1985 Terms. I selected these terms for several reasons.
First, they represent ten full Court terms with only a single change in the
composition of the Court (the retirement of Justice Potter Stewart near the end of the
1980-1981 Term and the subsequent appointment of Justice Sandra Day O’Connor just
prior to the commencement of the 1981-1982 Term). Controlling for the composition of
the Court is critical because the attitudes, values, and concepts of the Court’s role of
individual Justices affect the Court’s selection of cases (Provine, 1991; Segal and Spaeth,
206-207). On an individual level, a Justice may be more inclined to vote in favor of
certiorari for IFP petitions if she places particular emphasis on equality of access as an
end unto itself or if issues that arise more regularly on the IFP docket (especially the
rights of criminal defendants) are particularly salient to her.
Moreover, as the Court’s ideological complexion changes with change in the
Court’s composition, merits-based strategies regarding cases that raise ideologically
55

charged legal issues will shift. As the Court becomes more conservative, for example,
liberal Justices--who may generally favor access to the Court by the poor and by criminal
defendants--may be less inclined to grant review because of expectations regarding
merits outcomes.
A second reason for limiting analysis to these two consecutive natural courts is
that the entire timeframe is included in the papers of Justices Brennan and Powell, so
discuss lists and conference votes are available for future study. While this analysis
considers the behavior of the Court as a whole, individual-level analysis through
examination of individual certiorari votes may ultimately shed additional light on the
role of ideology and strategy in the certiorari process.
Finally, the use of these two natural courts reflects the need to control for both
law and process in the analysis. The terms included in the study all post-date the creation
of the cert pool, and the Court’s case management seems to have been fairly consistent
throughout the time period. Moreover, these terms all precede the changes in Federal law
and Court rules that were designed to narrow the availability of review for habeas
petitioners27 and to reduce the incidence of “frequent filers.”28 Thus, the use of these
particular years controls for legal and administrative factors that may affect the selection
of cases, particularly those factors that might have a different effect on the IFP docket
than on the paid docket.

27

Specifically, the 1996 Antiterrorism and Effective Death Penalty Act (Pub. L. 104-132, 110 Stat.
1214), which limited the availability of IFP status for habeas petitioners, and Teague v. Lane, 489 U.S. 288
(1989), which stated that the Court would no longer apply “new” constitutional doctrine retroactively and
thus effectively stripped many habeas petitioners of standing.
28
Specifically, a 1991 Amendment to Rule 39 allows the Court to deny IFP status to petitioners
whose petitions are particularly frivolous (In re Amendment to Rule 39, 500 U.S. 13 (1991), per curiam).

56

I collected data directly from the petitions for writs of certiorari, responsive
papers, and amicus briefs filed in support of or opposition to review. Petitions for IFP
cases granted review and all paid cases were available on microfiche at the Moritz Law
Library at The Ohio State University. IFP petitions denied review are not reproduced by
the U.S. Supreme Court; data on these cases was obtained through archival research at
the U.S. National Archives and Records Administration in Washington, D.C.

Sampling
The study employs a case-controlled, or choice-based, sampling scheme.
Specifically, the sample is stratified exogenously by docket (paid or IFP) and natural
court (the 1976 through 1980 Terms or the 1981 through 1985 Terms) as well as
endogenously by disposition (grant or deny certiorari), with disproportionate allocation
among the strata; in other words, the probability that a particular case will be included in
the sample will vary depending on the stratum in which the case is located (Tryfos,
1996). Because the number of IFP cases granted review is so small, that cell controlled
the sampling.29 Thus, for each natural court “j,” the sample consists of all IFP cases
granted review (nj); a sample of nj paid cases granted review; and samples of paid and
IFP cases denied review that ranged from 1.5(nj) to nearly 2nj.30 Table 3.1 provides
sample sizes for each stratum included in the study.

29

I identified IFP cases granted review from the U.S. Law Week Supreme Court Status Report,
which lists action taken on all paid cases and all unpaid cases granted review.
30
For the paid cases, both those granted review and those denied review, I identified the petitions
disposed of during each term (and thus each natural court) from the U.S. Law Week Supreme Court Status
Report. Unfortunately, there is no source available for identifying all unpaid petitions denied review in a
particular term. However, for each term covered in the study, I identified the number of unpaid petitions
docketed in that term and the number carried over from the previous term by using the workload summaries

57

Paid

Unpaid

Total

1976-1980 Terms
Granted
1981-1985 Terms

87

87

174

1976-1980 Terms
Denied
1981-1985 Terms

89
64

106

1976-1980 Terms
Total
1981-1985 Terms

176

259

153
66

153
66
172

153

261
278

306
130

306
132

431
170

431

737

172

Table 3.1. Sample Sizes for Strata in Study

For the six strata from which a sample was drawn (paid petitions granted, paid petitions
denied, and IFP petitions denied for each of the two natural courts in the study), the
sample was taken pursuant to the 1 in k linear systematic sampling method, also known
as the UNILESS method (Hedayat and Sinha, 1991:233). Specifically, I considered the
petitions in each stratum to be arrayed in sequence according to docket number. Given
that the population of a stratum is N = nk, where n equals the desired sample size for the
stratum, a starting point i was randomly selected, such that 1 ≤ i ≤ k. The sample then
consists of the petitions which hold the following positions in the sequence of docket

published in U.S. Law Week at the end of each Court term. I assumed that the Court disposes of unpaid
petitions roughly in order. Let dt be the number of unpaid cases filed in term t, and let ct be the number of
unpaid cases carried over from term t to term t+1. I thus assumed that the relevant population of unpaid
petitions for each term t is the highest ct-1 docket numbers from term t-1 and the first (dt-ct) docket numbers
from term t.

58

numbers: i, i + k, i + 2k, . . . , i + (nk - k).31 Tables 3.2 and 3.3 provide the population
size, sample size, k value, and i value for each stratum in the study.

Population Size
N
Unpaid Grants
87
Unpaid Denieds
9574
Paid Grants
706
Paid Denieds
8363

Sample Size
n
87
172
87
130

Interval
k
n/a
56
8
64

Starting Value
I
n/a
37
7
31

Table 3.2. Sampling Scheme for the 1976 through 1980 October Terms

Population Size
N
Unpaid Grants
66
Unpaid Denieds
10163
Paid Grants
761
Paid Denieds
9203

Sample Size
n
66
106
64
96

Interval
k
n/a
95
11
143

Starting Value
I
n/a
74
3
72

Table 3.3. Sampling Scheme for the 1981 through 1985 October Terms

Once the samples were identified, a simple replacement method was used to
replace cases identified in the sample that were inappropriate for the study. Specifically,
a few of the cases identified for inclusion in the sample were not petitions for certiorari

31

One troubling characteristic of systematic sampling designs is that, for any two observations, i and
j, the probability that both i and j will be selected may equal zero. Accordingly, the variance of the
Horvitz-Thompson Estimator cannot be unbiasedly estimated. Generally, then, more complex methods of
systematic sampling are necessary. However, for a population in random order--one in which there is no
functional relation between the labeling index and the variables of interest--a 1 in k UNILESS design is
equivalent to a simple random sample and will produce unbiased estimators (Hedayat and Sinha, 1991:
248-249). My use of a 1 in k UNILESS design, then, is based on the assumption that docket numbers are
essentially a random index, that the order in which petitions are filed (which determines docket number)
bears no relation to any substantive characteristic of the petitions or to the disposition of the petitions.

59

but, rather, appeals or petitions for extraordinary writs. Several other petitions were
withdrawn before any action by the Court, either because of the death of the petitioner,
settlement between the parties, or some undisclosed reason. When a case identified in the
sample could not be used, the docket number immediately following or immediately
preceding the unusable case was substituted.

Coding
Coding of individual petitions involved reading each petition and either
consulting the Clerk of Court’s docket sheet (where available) or paging through the
additional documents (responsive briefs, amicus briefs, and Court memoranda) contained
in the file. I coded all data on the petitions directly into a Microsoft Access database to
limit the opportunity for transcriptional errors. Moreover, I photographed the petitions
obtained from the National Archives using a digital camera to allow for spot-checks of
the coding and to allow for subsequent coding of additional variables, as needed.
Between direct coding and recoding, the following variables were constructed:

Identifying Information:
Docket Number: The docket number, assigned to the petition by the Clerk of
Court when the petition is filed, is the unit of analysis. The docket number consists of
two parts: a two digit number that represents the term in which the petition was filed,
separated by a hyphen from a number that represents the order in which the petition was
filed (numbered consecutively from “1” for paid petitions and from “5001” for unpaid
petitions.
60

Case Name: This string variable is a short version of the case name as it appears
on the original petition for certiorari. In many cases, the short version of the case name
omits some individual parties, both petitioners and respondents. Moreover, the case
name reflects the identity of the parties at the time the petition was filed; in some cases,
the case name changed before the Court disposed of the matter, especially when the
matter was ultimately decided on the merits.
Date Filed: This variable, coded as a date, represents the date (day, month, and
year) on which the petition for review was filed with Court and docketed by the Clerk of
Court. In some instances, the Clerk of Court back-dates a petition for administrative
purposes. As a result, the date filed may be considerably earlier than the date on which
the petition is first circulated to the Justices (and, accordingly, the docket number may be
inconsistent with the date filed).
Date of Disposition: This variable reflects the date (day, month, and year) on
which the Court issued an order disposing of the petition for review, either by denying
the petition; granting the petition; summarily reversing or affirming the lower court
decision; or granting the petition, vacating the lower court decision and remanding the
matter back to the lower court (commonly known as a “grant-vacate-remand” or “GVR”).
Term of Disposition: This variable, a four-digit number, represents the Supreme
Court term in which the Court disposed of the petition for review. I identified the
population from which the sample was drawn by reference to the term of disposition; for
purposes of weighting the sample for quantitative analysis, the term of disposition
variable dictated the sampling probability of the case and, accordingly, the weight
assigned.
61

Case Disposition (Dependent Variable):
Disposition: This string variable codes the manner in which the Court disposed of
the petition for review: “cert granted,” “cert denied,” “grant-vacate-remand,” “summary
affirmance,” or “summary reversal.”
Grant/Deny: This dichotomous variable, used as the dependent variable in the
multivariate analysis in Chapter Five, was constructed from the more specific
“disposition” variable. It is coded “1” if the Court granted certiorari or issued a
summary affirmance or reversal; it is coded “0” if the Court denied certiorari or issued a
GVR order.
The Court’s disposition of petitions for review is not a purely dichotomous
process, grant or deny; there are shades of gray. The Court sometimes issues an opinion
on the merits of a case without first granting certiorari and, hence, without benefit of
formal briefs or oral arguments. These summary affirmances and reversals, issued either
with or without a short per curiam opinion, are relatively rare32 (Stern et al., 1986: 279).
Hellman (1983) suggests that these summary decisions reflect the Court’s error
correction role rather than its policy-making role. Although the Court does not devote
much of its precious resources to these decisions, they are nevertheless merits decisions,
with precedential value,33 that imply consideration of the underlying legal issues and a
desire to intervene in a binding manner.

32

Of the 737 cases included in the dataset created for this analysis, three were summary affirmances
and 20 were summary reversals. Employing appropriate weights to account for the probability of selecting
cases from each stratum, the summary dispositions account for just 0.2% of the sample.
33
As Songer and Linquist (1996) note, the Court’s summary reversal in National Miners Corp. v.
West Virginia (1989) states that the decision is based on Ashland Oil v. West Virginia (1989), itself a
summary decision.

62

In addition to these summary decisions, the Court responds to a number of
petitions for review with an order to grant certiorari, vacate the decision below, and
remand the matter in light of some authority (typically another case recently decided by
the Court)34. The Court does not consider these “grant-vacate-remand” orders, or
“GVRs,” to be merits decisions: the Court records GVR votes with certiorari votes
rather than merits votes, the Court does not publicly announce dissents from the decision
to vacate in the U.S. Reports, and the Court has even specifically referred to GVRs as “an
appropriate exercise of our certiorari jurisdiction” (Lawrence v. Chater, 516 U.S. 163,
166 (1996) (per curiam); Segal and Spaeth, 1996b: 1078-79). More importantly, the
GVR orders do not direct the court below to reach any particular decision on the merits;
indeed, the lower court often reaffirms its original ruling on the merits, and the Supreme
Court rarely overturns the decisions on remand (Hellman, 1984).
Disposition of the petitions was coded as grant, deny, summarily affirm,
summarily reverse, or grant-vacate-remand. However, given the considerations above,
summary affirmances and reversals were treated as grants of review for purposes of
sampling, and GVRs were treated as denials of review.35 Moreover, most multivariate
analysis was conducted including summary decisions (coded as grants) and GVRs (coded

34

Grant-vacate-remand orders are somewhat more common than summary dispositions. Of the 737
cases collected for this study, eight were disposed of via GVR orders. Employing appropriate sampling
weights, the GVR orders account for 1.5% of the sample.
35
The decision to include these “middle ground” dispositions in the sample and to divide them as I
have done is not uncontroversial. The symposium on precedent in the 1996 American Journal of Political
Science highlights the debate among scholars about whether and how to examine these troublesome
dispositions (Segal and Spaeth, 1996a, 1996b; Songer and Lindquist, 1996; Brenner and Stier, 1996).
However, I am not alone in taking this approach (Teger and Kosinski, 1980: 838).

63

as denials), although some results--described in Chapter Five--were checked against
analysis that excluded the GVRs and summary decisions.
Dissent: Although dissents from dispositions are relatively rare, they do occur.
This variable, a number, codes the number of Justices, if any, who dissented from the
Court’s disposition. This variable was not used in the analyses in this dissertation, but
will be useful as a launching point for individual, Justice-level analysis of the certiorari
process.

Case Characteristics:
Case Type: This variable specifies whether the underlying dispute is “civil,”
“criminal” (including both direct appeals and habeas petitions), “prisoner civil rights”
(any issue relating to the terms and conditions of a sentence), or “bankruptcy.” This
variable was used in the descriptive analysis described in Chapter Four.
Factual Summary: This rather lengthy textual field was used to record a summary
of the factual predicate of the underlying dispute including any relevant procedural
history.
Summary of Legal Issues: This variable summarizes the legal questions
presented by the petition for review, including reference to any specific Constitutional
provisions or statutes. The factual summary and summary of legal issues, together,
formed the basis for a number of more specific issue-area variables involved in the
descriptive analysis in Chapter Four. Specifically, these textual descriptions of the cases
were used to identify cases that involved family law and custody issues, welfare benefits,

64

immigration, employment discrimination, application of the death penalty, and issues
involving minority rights.

Petitioner Characteristics:
IFP: This dichotomous variable, the primary independent variable of interest,
was coded “1” if the petition for review was filed without prepayment of fees (as
determined by the docket number assigned) and coded “0” for paid petitions. This
variable is used extensively in the descriptive and multivariate analyses that form the
bulk of the rest of the dissertation. Ultimately, while disposition is the dependent
variable of interest, the effect of IFP status--both independently and in conjunction with
other case characteristics--is the independent variable of greatest interest.
Pro Se: Although this variable was coded as a dichotomous variable--coded “1”
if the petition was filed pro se and coded “0” if the petition was prepared by an attorney
who was not also the petitioner--the latent variable, the actual characterization of the
petitioner, was more fluid. First, there were a very few cases in which the petitioner
signed the brief as “attorney for petitioner” but the individual signing the brief was
clearly the party seeking relief; those petitions were coded as pro se, although it was
possible that the party filing the brief on his own behalf was a member of the bar.
Moreover, there were a few petitions that expressly indicated that they were prepared by
“writ writers,” prison inmates who have developed a cottage industry preparing legal
documents for their fellow inmates; these petitions were also coded as pro se petitions.
Criminal: This dichotomous variable was coded “1” if the petition was under
some form of criminal sanction (incarceration, fine, or probation) at the time the petition
65

was filed, regardless of the nature of the petition, and coded “0” otherwise. There were
four civil cases filed by incarcerated petitioners: two cases involving interpretation of the
Freedom of Information Act, one case involving termination of parental rights, and one
employment discrimination case. Similarly, there were two prisoner civil rights cases
filed by individuals who were no longer under any form of criminal sanction; in both
cases, the petitioner filed suit seeking damages for injuries sustained during a past period
of incarceration. This variable was not used in subsequent analyses, but it was coded
with an eye toward future research.
Solicitor General Petition: This dichotomous variable was coded “1” if the
Solicitor General was the party petitioning for review and coded “0” otherwise.
Ultimately, this variable is not used in subsequent analyses. It is only relevant in the
context of the paid docket, because the Solicitor General never files without paying the
Court’s fees. Instead, variables describing the Solicitor General’s involvement in the
case, if any, and his position on the issue of review were used in both the descriptive and
multivariate analysis. However, in the interest of creating a dataset with applications
beyond this research, the Solicitor General’s status as petitioner was separately coded.

Responding Party:
Response Filed: This dichotomous variable was coded “1” if any respondent filed
a sua sponte responsive brief and coded “0” otherwise; in some cases, particularly where
the Solicitor General was the respondent, the responsive brief actually supported review.
I posit that presence of a sua sponte response is a potential cue of certworthiness;

66

accordingly, this variable is integral to both the descriptive analysis and multivariate
analysis presented in subsequent chapters.
Response Requested: In some cases, overwhelmingly criminal and prisoner civil
rights cases, the responding party declined to file a responsive brief but the Supreme
Court requested such a brief before disposing of the petition. This variable was coded
“1” when the Court actually requested a responsive brief and “0” otherwise. The Court’s
request for a response is more a result of the Court’s interest in a petition than a cause of
that interest. Accordingly, the variable is not important to this analysis; still, it was coded
with an eye toward future research.

Other Parties or Participants:
Solicitor General: This string variable codes the Solicitor General’s
involvement, if any, at the cert stage: “Petitioner,” “Respondent,” “Amicus,”
“Intervenor,” or “n/a.” This variable captures only the Solicitor General’s role in the
certiorari process, not the position taken regarding the propriety of review.
Solicitor General Pro/Con: This variable codes the position taken by the Solicitor
General, regardless of role, on the issue of granting review: “supports,” “opposes,” or
“silent.”
SG Support: This variable, recoded from “Solicitor General pro/con,” was coded
“1” if the Solicitor General supported review, regardless of his role in the litigation, and
coded “0” otherwise.

67

SG Oppose: This variable, recoded from “Solicitor General pro/con,” was coded
“1” if the Solicitor General opposed review, regardless of his role in the litigation, and
coded “0” otherwise.36
In Chapter Two, I discussed the special credibility the Solicitor General has with
the Court, and I would anticipate that the Solicitor General’s position--either support for
or opposition to a petition--would carry weight with the Court. Accordingly, these two
summary variables are important to the multivariate analysis. The other variables that
describe the Solicitor General’s involvement in the case are collected for the sake of
future analysis.
Amicus Briefs Pro: This variable reflects the number of amicus briefs, if any,
filed in support of a grant of certiorari. The variable is based on the number of briefs,
not the number of separate amici (at times, multiple amici will sign on to a single brief).
Amicus Briefs Con: Similarly, this variable reflects the number of amicus briefs,
if any, filed in opposition to a grant of certiorari.
Amici Pro/Amici Con: These two text variables list the identity of any amici
filing briefs in support of or in opposition to, respectively, a grant of certiorari.
Amici Pro/Con: This dichotomous variable, recoded from “amicus briefs pro”
and “amicus briefs con,” was coded “1” if there were any cert-stage amicus briefs, either
supporting or opposing review, and coded “0” otherwise. Because the presence of amici
is more significant than the positions stated in them (Caldeira and Wright, 1988), this

36

The base condition, in which both “SG support” and “SG oppose” is “0,” represents those cases in
which the Solicitor General does not take any position on the issue of review, either because the Solicitor
General is not involved or because the Solicitor General, as a party respondent, declines to file a responsive
brief.

68

summary variable is used in subsequent analyses rather than the more precise variables
above.

Decision Below:
Cite Below: This variable codes the official reporter citation of the decision from
which review was sought. This information was not always available, either because the
decision below was unpublished, the decision below was not published at the time the
briefs were filed with the Court, or the parties (petitioner and respondent) failed to
provide any citation to the decision below.
Court Below: This textual variable notes the identity of the court issuing the
decision from which review was sought; this variable was coded with as much specificity
as possible.
Tribunal Type: In contrast to the “court below” variable, this variable categorizes
the type of court issuing the decision from which review was sought: “State Court (trial
or appellate),” “Circuit Court (three judge panel),” “Circuit Court (en banc),” or “Special
Court of Appeals”; the “special” Courts of Appeals included U.S. Court of Claims, the
U.S. Court of Customs and Patent Appeals, and the U.S. Court of Appeals for the Federal
Circuit.
Federal: This dichotomous variable, recoded from “tribunal type,” was coded
“1” if the court issuing the decision from which review was sought was a federal court
and “0” otherwise. While “tribunal type” records more precise information for later
research, this study limits consideration solely to the state/federal distinction captured by
the “federal” variable. Although the distinction is not anticipated to be a cue of
69

certworthiness, and thus is not included in the multivariate analysis of disposition, the
variable is considered in the descriptive analysis in Chapter Four.
En Banc: This dichotomous variable, recoded from “tribunal type,” was coded
“1” if the court issuing the decision from which review was sought was a U.S. Court of
Appeals sitting en banc and coded “0” otherwise. This variable has been identified as a
potential cue of certworthiness by George and Solimine (2001) and is employed in the
multivariate analysis of the Court’s agenda-setting process.
Lower Court Direction: This variable, recoded from the procedural history
outlined in the “factual summary” and the legal issues described in “summary of legal
issues,” codes the ideological direction of the lower court decision. It is coded “1” for a
conservative lower court decision, “-1” for a liberal lower court decision, and “0” for a
lower court decision of indeterminate ideological direction. The decision rules
articulated by Spaeth (1999), in determining the lower court direction and direction of the
Supreme Court’s decision, governed coding of this variable, although Spaeth coded
decisions “0” for indeterminate ideology, “1” for liberal ideology, and “2” for
conservative ideology.
Although this variable is not particularly interesting in terms of descriptive
analysis (there is very little variation on this variable on the IFP docket), it is a critical
strategic factor in the Court’s agenda-setting process and is thus crucial in the
multivariate analysis that follows.

70

Rule 10 Factors:
The following Rule 10 factors are explored in both the descriptive analysis in
Chapter Four and the multivariate analysis in Chapter Five.
Dissent Below: This dichotomous variable was coded “1” if there was a dissent in
the opinion from which review was sought and “0” otherwise; the presence of a dissent
was determined by a reference to a dissent in the petition or responsive brief or, if the
decision below was included in an appendix to the petition, by a notation of dissent in the
opinion.
Reversal Below: This dichotomous variable reflects the presence of a reversal in
the case’s procedural history. It is coded “1” if such a reversal exists and “0” otherwise.
The presence of a reversal was determined by reference to a reversal in the petition or
responsive brief or, if the decision below was included in an appendix to the petition, by a
notation of a reversal.
Interjurisdictional Conflict: This dichotomous variable codes for an allegation of
a conflict in authority between two federal courts of appeals, two state supreme courts, or
between a federal court of appeals and a state supreme court. The variable is coded “1” if
such an allegation is made, in whatever language, and “0” otherwise.
Break from Precedent: This dichotomous variable was coded “1” if the petitioner
asserted that the decision below was in conflict with existing Supreme Court precedent
and coded “0” otherwise. Any allegation that the decision below ran contrary to earlier
decisions of the Supreme Court, even if those earlier decisions were not squarely on
point, was considered an allegation of a break from precedent. For example, in Riccardi
v. U.S. (Docket Number 77-5736), the petitioner pro se made reference to legal holdings
71

in a string of U.S. Supreme Court cases regarding the obligation of courts to conduct sua
sponte competency hearings; he then concluded that “[u]nder the circumstances of this
case, petitioner has been deprived of the protection this Court insists be accorded insane
and incompetent defendants . . .” (Petitioner’s Brief at 5). Although Riccardi never
explicitly stated that the lower court failed to follow on-point Supreme Court precedent,
his meaning was clear, and this case was coded as alleging a break from precedent.
Novelty/Importance: This dichotomous variable was coded “1” if the petitioner
asserted in the petition that the legal issue presented was novel or important and coded
“0” otherwise. This variable was coded “1” if the petitioner explicitly invoked the
language of Rule 10 or if the petitioner merely asserted either that the particular legal
issue was one the Court had never considered or that resolution of the legal issue would
have implications beyond the scope of the present conflict. Again, there were no “magic
words” involved, simply an allegation that the legal issue had never been addressed by
the Supreme Court or an allegation that the case would have repercussions beyond the
individual litigants.
Note that an individual petition might contain both an allegation of a break from
precedent and an allegation that the petition raised a novel legal issue; while those two
assertions might seem contradictory, they might be raised with respect to two separate
legal issues raised by a single petition.

CONCLUSION
With data in hand, this study can now turn to a deeper examination of the Court’s
unpaid docket. The remainder of this project attempts to address the hypotheses raised in
72

this chapter through analysis of the date described. The descriptive analysis in Chapter
Four serves two purposes.
First, it empirically tests the notion that the IFP petitions are qualitatively
different than the paid petitions. Specifically, the descriptive analysis will compare the
two dockets with respect to the frequency of case types (civil, criminal, prisoner civil
rights, and bankruptcy). For the civil cases, the analysis will compare the two dockets
with respect to both the “federal” variable (whether the court from which review is
sought is federal or state) and with respect to the frequency of particular issue areas
(family law and custody, welfare benefits, immigration, and employment discrimination,
all of which were coded based on the “summary of legal issues” variable). In addition,
the analysis will compare the two dockets with respect to the frequency of minority rights
issues being raised, that issue area also being coded from the “summary of legal issues”
variable.
Second, the analysis in Chapter Four determines the relative prevalence of the
different cue characteristics on the two dockets for purposes of anticipating the different
marginal effects we would expect to see from those cue characteristics in the context of
the two dockets. Specifically, the analysis compares the two dockets with respect to the
proportion of cases filed by pro se petitioners, the presence of a responsive brief, the
ideological direction of the lower court decision, the en banc status of the decision below,
the allegation of each of the Rule 10 factors, participation by amici, Solicitor General
support for review, and Solicitor General opposition to review.
Chapter Five continues the process by looking first at the Court’s attention to the
IFP docket over time, with an eye toward determining whether the Court treats the two
73

dockets as distinct for purposes of case selection. The chapter continues by looking at
case-level determinants of review on the two dockets to see whether and how the Court
weighs the cue characteristics differently across the two dockets. Specifically, the
analysis considers the effect on the probability of the Court granting review of IFP status
of the petitioner, pro se status of the petitioner, the presence of a responsive brief, the
ideological direction of the lower court decision, the en banc status of the decision below,
the allegation of each of the Rule 10 factors, participation by amici, Solicitor General
support for review, and Solicitor General opposition to review.

74

CHAPTER 4
A DESCRIPTIVE ANALYSIS OF THE COURT’S IFP DOCKET

INTRODUCTION
Empirical analysis of the United States Supreme Court’s docket has, in the past,
focused almost exclusively on the paid docket to the exclusion of the IFP petitions
(Caldeira and Wright, 1988; Caldeira and Wright, 1990; Teger and Kosinski, 1980). This
is true despite the fact that, between 1948 and 1986, roughly half of the petitions filed
with the Supreme Court were filed by indigent petitioners, and since 1986 the proportion
of filings by indigent petitioners has skyrocketed. Indeed, during the 2001-2002 Supreme
Court Term, 7972 petitions were filed with the Court, and 6043 (nearly 76%) of those
were IFP petitions.37
As noted in Chapter One, scholars justify the exclusion of IFP petitions from
analysis on the assumption that these petitions are generally frivolous and unimportant
(Tanenhaus et al., 1963; Segal and Spaeth, 1993: 192). To date, however, the
assumptions made by scholars about the composition of the IFP docket have been just
that: assumptions. Although there is anecdotal evidence about the petitions from
Supreme Court Justices and law clerks, no one has conducted a rigorous analysis of the
IFP docket to determine whether and in what ways it differs from the paid docket.
37

All data on the Supreme Court’s workload, unless otherwise specified, is from Epstein et al.
(2003: 62-71)

75

Such an analysis is interesting in its own right, but it also has implications for the
hypotheses articulated in Chapter Three. As discussed there, the relative prevalence or
scarcity of particular cues among the unpaid petitions may affect the weight the Court
attaches to those cues in the context of the unpaid docket and, accordingly, the marginal
effects of those cues on the probability of the Court granting plenary review to unpaid
petitions.
This chapter begins the process of exploring the petitions on the IFP docket in
greater depth. The data collected as part of this research project presents a unique
opportunity to test the validity of the assumptions made by scholars and legal
practitioners about the IFP docket as well as to investigate the ways in which the IFP
docket and the paid docket differ from one another.
On a methodological note, because of the stratified sampling used in collecting
the data, the data were weighted for purposes of this analysis. Specifically, the cases
were weighted based upon their proportion in the population as opposed to their
proportion in the sample, and the weights were constructed so that the weighted sample
summed to 737 (the actual number of observations collected). Thus, for each of the eight
strata, the weight attached to each case equals the k value for the stratum multiplied by
the sample size (737) and divided by the population size (36,125). For example, I
sampled one out of every 64 paid petitions denied certiorari during the 1976-1980 Terms.
Each of those petitions selected for review was weighted by 64 * (737/36,125), or by
1.31. All frequencies, percentages, and crosstabulations below are based on the weighted
data.

76

TYPES OF CASES ON THE IFP DOCKET
Each petition was coded for the type of case it presented: civil, criminal, prisoner
civil rights, or bankruptcy. Table 4.1 breaks down the number of cases from each docket
by case type,

Civil

Criminal

Prisoner
Civil Rights

Bankruptcy

Total

Paid Cases

61.75%

35.54%

1.81%

0.90%

100%

IFP Cases

12.07%

79.56%

7.88%

0.49%

100%

-.521*

.449*

.137*

n/a

Pearson’s R
* p<.001

Table 4.1. Paid and IFP Cases by Type of Case

and provides the correlation between each case category and IFP status. For example, the
correlation between IFP status and a civil issue area, as opposed to any other type of issue
area, is -.521. Percentages reflect the proportion of each docket that is comprised of each
type of case; for example, 61.75% of the paid cases were civil, while only 12.07% of the
IFP cases were civil. Figure 4.1 illustrates the difference in case types between the IFP
and paid dockets graphically.

77

Docket

100

IFP
Paid

Percent

80

60

40

20

0
Civil

Criminal

Prisoner Civil Rights

Case Type

Figure 4.1. Proportion of Paid and IFP Dockets Comprised by Different Case Types

As anticipated, the IFP docket is disproportionately composed of criminal and
prisoner civil rights cases, as opposed to civil cases. What’s more, of the remarkable 44
petitions in the sample brought by death-row inmates, only one was a paid petition.
Those civil cases which are on the IFP docket, moreover, differ from the civil
cases on the paid docket.
The civil cases on the IFP docket are less likely to be petitions for review of a
federal court decision than are the civil petitions on the paid docket. Specifically, over
78% of the paid civil cases arose from the lower federal courts, while only 53% of the
IFP civil cases sought review of federal court decisions. Table 4.3 illustrates the number
78

of IFP and paid civil cases based on the type of court, federal or state, from which review
is sought; Figure 4.2 provides a graphic illustration of the same information.

Review Sought of
State Court Decision

Review Sought of
Federal Court Decision

Total

Paid Cases

21.46%

78.54%

100%

IFP Cases

46.94%

53.06%

100%

χ2 = 13.216, df = 1 (p<.001); r = -.228 (p<.001)
Table 4.2. Paid and IFP Civil Cases by Court of Origin

IFP Cases

Paid Cases
Court
Below

Court
Below

Federal

Federal

State

State

Figure 4.2. Paid and IFP Civil Cases by Court of Origin
This finding may be relevant to the certiorari question because civil cases arising
in the lower federal courts, by definition, either raise questions of federal law or meet the
79

criteria for diversity jurisdiction, including the threshold amount-in-controversy. As a
result, it seems likely that civil cases arising in the lower federal courts are more likely to
pose the sort of high-stakes federal questions that the Supreme Court is interested in
reviewing than those civil cases arising in the state court system. Conversely, state court
civil matters are more likely to involve pure issues of state law and the administration of
the state courts that, as a matter of judicial federalism, the Supreme Court considers best
left to the state courts to resolve; at a minimum, the federal issues involved may be
overshadowed by state law considerations.
The civil petitions were more specifically coded for particular issue areas that we
might expect to find addressed with greater frequency on the IFP docket. Specifically,
the following issue areas were coded: (1) family law and custody38; (2) employment
discrimination39; (3) administration of and eligibility for welfare benefits40; and (4)
immigration and naturalization.
Figure 4.3 illustrates the breakdown of paid and IFP civil cases among the four
issue areas of interest; the “other” category represents all civil cases that did not fall
within one of the four areas of interest.

38

Although family law and custody are predominantly state law issues, these issue areas also raise
federal questions, both statutory (such interpretation of the Uniform Custody Jurisdiction Act) and
constitutional (particularly equal protection and due process concerns). Indeed, one need look no further
than the current controversy over gay marriage to see the potential federal dimension of family law.
39
For purposes of this analysis, employment discrimination was limited to constitutional and
statutory claims of race, gender, and religious discrimination in employment. This classification does not
include labor disputes or due process claims related to pre- or post-termination hearings.
40
Welfare benefits include food stamps, Medicare, Medicaid, HUD subsidies, and unemployment
benefits; workers’ compensation benefits are not included.

80

100

Paid
IFP

Percent

80

60

40

20

0
Employment

Family

Immigration

Other

Welfare

Issue Area

Figure 4.3. Civil Cases by Issue Area

Although there was no statistically significant difference between the paid and
IFP dockets with respect to the incidence of employment discrimination cases, and
immigration cases were actually more prevalent on the paid docket, both family law and
custody cases and welfare benefits cases arose with greater frequency on the unpaid
docket than on the paid docket. While fewer than 1.5% of the paid civil petitions raised
family law or custody issues, a full 10.42% of the IFP civil petitions were family law
cases. Similarly, only 1% of the paid civil cases involved administration of welfare
benefit programs, while 12.5% of the IFP civil petitions raised these issues. Table 4.3
81

illustrates the number of family law and non-family law civil cases on both the paid and
unpaid docket; Table 4.4 illustrates the same with respect to cases involving welfare
benefits.

Family Law and Custody

All Other Civil Cases

Total

Paid Cases

1.46%

98.54%

100%

IFP Cases

10.50%

89.58%

100%

χ2 = 10.182, df = 1 (p<.001); r = .201 (p<.001)
Table 4.3. Incidence of Family Law and Custody Cases on the Paid and IFP Dockets

Welfare Benefits Cases

All Other Civil Cases

Total

Paid Cases

1.00%

99.00%

100%

IFP Cases

12.50%

87.50%

100%

χ2 = 16.869, df = 1 (p<.001); r = .26 (p<.001)
Table 4.4. Welfare Benefits Cases on the Paid and IFP Dockets

The differences between the paid and unpaid docket are not limited to civil issue
areas. All cases in which the petitioner was a racial minority and where there was a

82

nexus between the legal issue and the petitioner’s minority status were identified,41
regardless of whether the cases were civil, criminal, or prisoner civil rights cases. These
“minority rights” cases included seven challenges to methods of witness identifications,
seven challenges to the racial composition of juries, three issues regarding latitude during
voir dire, and seven employment discrimination cases. Of the 34 cases identified as
involving “minority rights issues” in the sample, 26 were IFP cases, and only 8 were
paid. Based on the weighted data, the correlation between IFP status and minority rights
issues is .085, significant at the p<.05 level.
This analysis lends some credence to the concern that disparate treatment of the
unpaid docket may have implications for the substance of the Court’s docket. In other
words, if there is an a priori bias against IFP petitions, it may result in the Court’s failure
to give adequate consideration to legal issues surrounding family law and custody and the
administration of welfare benefits programs, issues which have tremendous practical
significance in the day-to-day lives of many people. Moreover, racial minorities seeking
to secure equal treatment may face an insurmountable hurdle in this, their last forum for
relief.

41

There are, however, other cases in the sample that involve issues of race, but in which the
petitioner is not a member of a racial minority. Specifically, there are three cases involving challenges to
affirmative action programs, one case in which an employer challenged the award of attorney fees in a Title
VII employment discrimination case, two cases involving discrimination in elementary and secondary
schools (one involving HEW funding eligibility and one a challenge to a desegregation plan), and one
rather unusual case in which the State of Iowa asserted that is was not a “white person” within the meaning
of a Congressional Act regulating relations between Native Americans and white persons.

83

REPRESENTATION
Although the majority of petitions filed with the Supreme Court, both paid and
unpaid, are prepared by attorneys, some petitioners file pro se, without assistance of
counsel.42 Pro se petitions presumably are less apt to frame legal issues in a compelling
way, are less apt to identify important legal considerations that would make the Court
more likely to accept review, and perhaps send a signal that no attorney considered the
case worthy of representation (Smith, 1999).
Although pro se petitions are far from rare on the paid docket, unpaid petitions
are significantly more likely to be pro se. Table 4.5 illustrates the number of petitions on
both the paid and unpaid dockets that were prepared by attorneys; Figure 4.4 provides a
graphic representation of the same data.

Pro Se

Attorney-Prepared

Total

Paid Petitions

7.83%

92.17%

100%

IFP Petitions

32.10%

67.90%

100%

χ2 = 64.388, df = 1 (p<.001); r = -.296 (p<.001)
Table 4.5. Representation of Petitioners on the Paid and Unpaid Dockets
42

For purposes of coding the cases in the sample, I considered a petition to be “filed by an attorney”
if an apparently licensed attorney who was not a party signed off on the petition for review. As a result, the
pro se category of petitions includes cases filed by litigants who happen to be attorneys (Smith, 1999). In
addition, this dichotomous coding scheme fails to account for those pro se petitions that were actually
prepared by “jailhouse attorneys,” or “writ-writers.” While some petitions prepared by these “lay
professionals” can be identified (such as docket number 85-6277, which was “Prepared by American
Inmate Paralegal Association Incorporated, by Melvin Leroy Tyler, Executive VP, For Petitioner Pro Se”),
most cannot. As a result, there was no way to code those petitions separately with any degree of
confidence.

84

IFP Cases

Paid Cases
Atty

Atty

Pro Se

Pro Se

Figure 4.4. Representation of Petitioners on the Paid and Unpaid Dockets
The correlation between IFP status and pro se status holds and is significant for
civil petitions (χ2 = 100.983, df = 1 (p<.001); r = -.631 (p<.001)), criminal petitions (χ2 =
7.892, df = 1 (p<.01); r = -.134 (p<.01)), and prisoner civil rights petitions (χ2 = 15.438,
df = 1 (p<.001); r = -.637 (p<.001)).

PARTICIPATION BY OTHER PARTIES
Caldeira and Wright (1990; 1988) demonstrate that briefs filed by amicus curiae
at the certiorari stage, either in support of or in opposition to the petition for certiorari,
increase the probability of the Court granting review. They argue that the participation of
amici, their expenditure of valuable resources to support or oppose a petition, sends a

85

signal about the importance of the particular case and the policy ramifications--beyond
the immediate controversy--of the case’s outcome.
Amicus participation at the certiorari stage is significantly less likely on the IFP
docket. Using weighted figures, cert stage amici briefs were filed in 4.52% of paid
petitions; more specifically, amici briefs in support of review were filed in 3.61% of paid
petitions, and amici briefs opposing review were filed in 1.00% of paid petitions. In
contrast, cert stage amici briefs were filed in only 0.04% of unpaid petitions; amici briefs
in support of review were filed in 0.04% of unpaid petitions and amici briefs opposing
review were filed in only 0.01% of unpaid petitions. The correlation between amicus
participation in support of a petition and IFP status is -.142 (p<.001); the correlation
between amicus participation in opposition to a petition and IFP status is -.071 (p<.05);
and the correlation between any amicus participation and IFP status is -.159 (p<.001).
Similarly, responding parties are significantly less likely to file a brief in response
to an IFP petition than to a paid petition. When the respondent is the United States, the
correlation between IFP status and a response is -.268 (p<.001); the Solicitor General
filed a response in 59.52% of the paid cases in which the United States was the
respondent, but filed a response in only 32.45% of the IFP cases in the United States was
the respondent. When the respondent is anyone other than the United States, the
correlation between IFP status and a responsive brief being filed is -.409 (p<.001); in
those cases, responses were filed in 81.12% of the paid cases and only 41.00% of the IFP
cases.43 Looking at all respondents, both the Solicitor General and others, the correlation

43

For purposes of these analyses, cases in which the Solicitor General was the petitioner were
excluded.

86

between IFP status and a responsive brief being filed is -.354 (p<.001); based on all
cases, respondents filed responsive briefs in 72.59% of the paid cases and only 37.04% of
the IFP cases. These results are set forth in Table 4.6. While other studies of the
Supreme Court’s certiorari process have generally not considered the effect of a
responsive brief on the Court’s selection of cases for review, it seems plausible that the
silence of a responding party may be a cue that a case is particularly uncertworthy,
unimportant, in the same way that the effort expended by amici is a cue of a case’s
significance (Smith, 1999).

SG Respondent

Non-SG Respondent All Cases

Paid Cases

59.52%

81.12%

72.59%

IFP Cases

32.45%

41.00%

37.04%

Table 4.6. Percent of Cases in Which Responsive Brief Was Filed

Looking specifically at cases in which the Solicitor General was involved but in
which the Solicitor General was not the petitioner (i.e., those cases in which the S.G. was
a respondent, intervenor, or amicus), the Solicitor General is less likely to express support
for IFP petitions than paid petitions (r=-.12, p<.05). Similarly, the Solicitor General is
less likely to express opposition to IFP petitions than paid petitions (r=-.169, p<.01).

87

LOWER COURT BEHAVIOR
Not surprisingly, there is a pronounced difference between the paid and unpaid
dockets with respect to the ideological direction of the decisions from which review is
sought. Specifically, Table 4.7 sets forth the percentage of petitions, based on weighted
data, on the paid and unpaid dockets that seek review from liberal, conservative, or
indeterminate lower court decisions.

Ideological Direction of
Lower Court Decision

Paid Petitions

IFP Petitions

Liberal

25.4%

0.8%

Conservative

70.6%

98.3%

Indeterminate

4.0%

1.0%

Table 4.7. Ideological Direction of Lower Court Decisions, Paid and Unpaid Dockets

There is no difference between the paid and unpaid dockets with respect to the
presence of an en banc decision below. Based on weighted data, there were six en banc
decisions being appealed on the paid docket and seven en banc decisions being appealed
on the unpaid docket, a difference that is not even remotely significant.

88

RULE 10 CONSIDERATIONS
While the scholarly interest in the Supreme Court’s certiorari process is
motivated in large part by the lack of clear-cut rules defining which cases will be
accepted and which will not, the Supreme Court has defined certain considerations that, if
manifest in a petition, increase the chances of review. Specifically, in its Rule 10, the
Supreme Court states that the Court is more inclined to grant certiorari when the lower
court’s decision (whether of a U.S. Court of Appeals or a state court of last resort)
conflicts with another court (either a U.S. Court of Appeals or a state court of last resort)
on an important question of federal law; when the lower court’s decision conflicts with
relevant decisions of the U.S. Supreme Court; or when the lower court “has decided an
important question of federal law that has not been, but should be, settled by” the
Supreme Court.
The substance of Rule 10 makes clear that the Court is interested in creating
stability and certainty in the law by ensuring that laws are being interpreted and applied
consistently across jurisdictions and by ensuring that “important federal questions”
promptly acquire some finality through resolution by the Supreme Court. Although Rule
10 does not explicitly state that a dissent on the court below or a reversal in the
procedural history of a case increase the likelihood of review, these case characteristics
may “signal ferment in the lower courts and suggest a problematic outcome, or perhaps
one worth of a closer look” (Caldeira and Wright, 1988: 1115).
Given, then, that dissension below, a reversal below, interjurisdictional conflict, a
break from Supreme Court precedent, and the novelty or importance of the federal
question presented may all increase the chances of the Court accepting review of a case,
89

it would behoove petitioners to draw attention to those case characteristics in their own
petitions. For each of the 737 petitions in this sample, I coded whether the petitioner
makes reference to each of these potentially important case characteristics.
IFP petitions were significantly less likely to reference dissension in the court
below (χ2 = 8.554, df = 1 (p<.01); r = -.108 (p<.01)), a reversal in the case’s procedural
history (χ2 = 66.965, df = 1 (p<.001); r = -.302 (p<.001)), an interjurisdictional conflict
(χ2 = 13.796, df = 1 (p<.001); r = -.137 (p<.001)), or the novelty or importance of the
issue presented (χ2 = 41.886, df = 1 (p<.001); r = -.238 (p<.001)). There was no
significant difference between the IFP petitions and the paid petitions with respect to a
reference to a break from Supreme Court precedent.
Controlling for the pro se status of the petitioner, however, the picture changes
considerably.
Looking only at those petitions prepared by attorneys, the petitions from the IFP
docket were significantly less likely to reference each of the Rule 10 factors, except for a
break from Supreme Court precedent, than were petitions from the paid docket. Table
4.8 shows the number of attorney-prepared petitions on both the paid and IFP dockets
that reference each of the Rule 10 factors and the correlation between IFP status and
reference to each factor for attorney-prepared petitions; Figure 4.5 provides a graphic
illustration of the incidence of the Rule 10 factors among attorney-prepared paid petitions
and attorney-prepared IFP petitions.

90

Dissension
Below

Reversal
Below

Conflict
Between
Jurisdictions

Break From
Precedent

Novelty /
Importance
of Case

Paid
Petitions

10.78%

30.39%

41.83%

44.77%

52.94%

IFP Petitions

5.09%

6.91%

30.55%

42.55%

33.45%

Correlation

-.104*

-.297***

-.117**

-.022

-.196***

* p<.05
** p<.01
***p<.001
Table 4.8. Reference to Rule 10 Factors in Attorney-Prepared Petitions

The pro se petitions present a very different picture. Looking exclusively at pro
se petitions, there is a statistically significant difference between the paid and IFP dockets
with respect to only two of the Rule 10 factors: break from precedent and the novelty or
importance of the issue. However, the IFP petitions are actually more likely to include a
reference to a break from Supreme Court precedent. Table 4.9 shows the number of pro
se petitions on both the paid and IFP dockets that reference each of the Rule 10 factors
and the correlation between IFP status and reference to each factor for pro se petitions;
Figure 4.6 provides a graphic representation of the same data.

91

Paid Cases

Proportion of Cases

1

0.8

0.6

0.4

0.2

0
Dissent Below

Interjurisdiction Split
Novel Issue
Split Below
Break from Precedent

IFP Cases

Proportion of Cases

1

0.8

0.6

0.4

0.2

0
Dissent Below

Interjurisdiction Split
Novel Issue
Split Below
Break from Precedent

Figure 4.5. Rule 10 Factors in Attorney-Prepared Petitions
92

Paid
Petitions
IFP Petitions
Correlation

Dissension
Below

Reversal
Below

Conflict
Between
Jurisdictions

Break From
Precedent

Novelty /
Importance
of Case

0%

0%

23.08%

11.54%

34.62%

3.08%

3.85%

20.77%

42.31%

15.38%

.073

.081

-.021

.237**

-.184*

* p<.05
** p<.01
Table 4.9. Reference to Rule 10 Factors in Pro Se Petitions

While there is no clear theoretical reason indigent pro se petitioners are better
able or more apt to identify a break from Supreme Court precedent than are more affluent
pro se petitioners, it is possibly driven by the availability of informal legal advice for
prison inmates, such as the writ writers discussed in footnote 42 above, and similar
networks of legal support for indigent civil litigants. Lay legal support services may
actually be more available to indigent communities, including prison populations, than
they are to citizens in more affluent communities. As a result, the collective knowledge
and advice of these groups may give indigent pro se petitioners an edge over paying pro
se petitioners. However, at this point such a contention is pure speculation; an
explanation may not be forthcoming from this data.

93

Paid Cases

Proportion of Cases

1

0.8

0.6

0.4

0.2

0
Dissent Below

Interjurisdiction Split
Novel Issue
Split Below
Break from Precedent

IFP Cases

Proportion of Cases

1

0.8

0.6

0.4

0.2

0
Dissent Below

Interjurisdiction Split
Novel Issue
Split Below
Break from Precedent

Figure 4.6. Rule 10 Factors in Pro Se Petitions
94

In sum, there does generally appear to be a difference between the IFP docket and
the paid docket with respect to references to the “Rule 10 factors.” Specifically, IFP
petitions are less likely to contain allegations of dissension below, reversal below,
interjurisdictional conflict, and the novelty or importance of the issue presented. Most
interesting, however, that disadvantage of the IFP petitions relative to the paid petitions
appears to be driven by the attorney-filed petitions rather than the pro se petitions.
Although attorneys representing IFP petitioners perform “better”--i.e., they reference
Rule 10 factors more frequently--than do pro se petitioners, either paying or IFP, they do
not fare well in comparison to the attorneys representing paying petitioners, and that gap
is significant.
Obviously, attorneys filing IFP petitions have no control over whether there was a
dissent or reversal in the case’s procedural history. On the other hand, the significant
differences in references to the novelty or importance of the issue, interjurisdictional
conflict, and break from controlling precedent present an interesting question. Clearly,
the attorneys filing IFP petitions are not making full use of the Rule 10 factors in the
same way that attorneys filing paid petitions are doing. However, it remains unclear
whether this disparity reflects a disparity in the advocacy skills of the attorneys at issue or
whether it reflects real differences between the issues raised by the two sets of petitions.

CONCLUSION
The results of this descriptive analysis generally support the common wisdom
about in forma pauperis petitions: the IFP docket is disproportionately comprised of
criminal and prisoner civil rights cases; the civil cases on the IFP docket are
95

disproportionately cases arising in the state courts (and thus perhaps do not reflect the
national scope and importance of cases the Supreme Court grants review); and the IFP
petitions are generally less likely to manifest those indicia of certworthiness that have
been associated with the Court’s grant of plenary review (reference to the Rule 10
factors, responsive briefs, amicus briefs, dissension on the court below, and reversals in
their procedural history).
However, several of the findings suggest strongly that the IFP docket deserves
closer attention. First, the disproportionate number of IFP petitions filed pro se, while
perhaps simply a further indication of the overall lack of merit of IFP cases, may in fact
reflect a growing legal underclass, a segment of society too poor to afford legal
representation. Indeed, the finding that the disparity between the IFP and paid dockets
with respect to citation to Rule 10 factors is driven by the attorney-filed petitions may
indicate that the legal representation the poor can find is nevertheless inferior. Certainly
neither of these observations is definitive, but the lack of representation on the IFP docket
and the differences between attorney-filed petitions on the two dockets raises a flag about
the possibility of a legal underclass.
The overrepresentation of certain legal issues--namely family law and custody
issues, issues relating to the administration of welfare benefits, and issues relating to
minority rights--on the IFP docket further suggests that the IFP docket requires closer
study. If there is an a priori bias against IFP petitions on the part of Supreme Court
Justices and their clerks, or even if the IFP petitions are approached and evaluated in a
substantially different manner than the paid petitions, the implications for the substantive

96

development of these legal issues are thought-provoking, and the implications for equal
access to the courts is grave.
In sum, this analysis is only a first tentative step in attempting to study
systematically the Court’s IFP docket. Yet even this initial foray has revealed disparities
between the IFP and paid dockets that underscore the concern that the poor may not be
receiving equal justice.
With respect to the hypotheses set forth in Chapter Three, this descriptive analysis
reveals significant disparities between the paid and unpaid dockets in terms of the
prevalence of the cue characteristics. Eight of the anticipated cue characteristics--the
presence of a sua sponte responsive brief, a dissent in the decision below, a reversal in the
case’s procedural history, an allegation of an interjurisdictional conflict, an allegation that
the issue presented is novel or important, cert-stage amicus participation, Solicitor
General support, and Solicitor General opposition--are significantly less common on the
IFP docket than on the paid docket. Accordingly, based on the theory presented in
Chapter Three, if the cue characteristics do provide the Court with information about the
certworthiness of petitions or act as heuristics in the case-selection process, I would
expect these cues to have a greater effect on the Court’s selection of IFP petitions than on
the Court’s selection of paid petitions. In contrast, one negative cue (pro se status) is less
prevalent on the paid docket; accordingly, I would expect pro se status to be a more
significant determinant of case selection in the context of the paid docket.
With that further refinement of the hypotheses set forth in Chapter Three, the
study can now turn to empirical analysis of the Court’s selection of IFP and paid

97

petitions. Chapter Five provides several analyses of the Court’s selection of cases in an
effort to test the hypotheses described in Chapter Three.

98

CHAPTER 5
THE SUPREME COURT’S SELECTION OF IFP PETITIONS

INTRODUCTION
This chapter turns to the central question posed by this project: how does the
Court’s selection of IFP petitions differ, if at all, from its selection of paid petitions?
Using both the dataset compiled for the dissertation research as well as data compiled
from secondary sources, this chapter considers determinants of case selection, in the
aggregate and individually, with a particular eye toward understanding the Court’s
selection of cases from its unpaid docket.

SELECTION OF IFP CASES IN THE AGGREGATE
Before turning to the Court’s selection of individual cases, both paid and unpaid,
it makes sense to undertake dynamic analysis of the aggregate selection of IFP cases.
Specifically, looking at the total number of cases granted review each year, what
determines what proportion of those cases originate on the unpaid docket? Using time
series regression analysis, we can get a sense of the gross determinants of the Court’s
acceptance of IFP petitions. Moreover, this analysis may indicate whether the
administrative and qualitative differences between the paid and unpaid dockets, described

99

in Chapters One and Four, translate into a distinction between the two dockets during the
Court’s case selection process.
The dependent variable in these analyses represents the percentage of all petitions
accepted by the U.S. Supreme Court during a given term—from the 1948 through the
2001 terms—that originated on the IFP docket. For example, during the 2001 term, the
Court granted 88 petitions of which 6 originated on the IFP docket, so the value of the
dependent variable for that term equals 100 * (6/88), or 6.82.
The analyses employ two primary independent variables. First, the analyses
include a measure of the ideology of the median member of the Supreme Court, on a
scale from -1 (most conservative) to 1 (most liberal) (Segal et al., 1995).44 Cases filed
IFP are, by definition, cases in which indigent parties lost in the lower courts. If we
assume that liberal justices are more sympathetic to the positions of all manner of
indigent parties (including criminal defendants), and that the Court is more inclined to
grant certiorari to petitions that represent errors by the courts below, then we would
expect a more liberal Court to be more interested in providing IFP petitioners with a
remedy. Moreover, the analysis in Chapter Four revealed that, at least for the 1976
through 1985 Terms, roughly 98% of IFP petitions sought review of a conservative lower
court decision.
44

The measure of ideology used here is derived from content analysis of newspaper editorials
written about the Justices during the time between their nomination by the President and their confirmation
by the Senate (Segal et al., 1995). The benefit of this measure of ideology is that it is independent of,
albeith highly correlated with, the Justices’ votes on the Court, so it provides a measure of ideology that is
not endogenous to the voting behavior in which scholars are most interested (certiorari and merits votes).
The drawback to using this measure is that it is static, a measure taken at one relatively brief point in time,
and does not account for changes in a Justice’s world view and attitudes while the Justice sits on the Court.
Given the remarkably long tenure of many Supreme Court Justices, this drawback cannot be minimized.
However, given the alternatives, the Segal et al. measure of ideology presents the best solution to a thorny
problem.

100

Since support of only four of the nine justices is required to grant a petition for
certiorari, it may seem that the average ideology of the four most liberal justices would
be a more appropriate measure. However, there is some support for the notion that
justices are strategic in their certiorari decision, taking the likelihood of prevailing on the
merits into consideration when deciding whether to grant review in a particular case
(Ulmer, 1972; Boucher and Segal, 1995). As a result, an approximation of the
ideological climate of the Court as a whole would best capture this “merits component”
of the certiorari decision.
Second, the analyses employ two different measures of the availability of IFP
petitions relative to paid petitions. In other words, to the extent that the Court is
constrained by the petitions brought before it, these variables take into consideration the
relative number of petitions in that pool that are IFP petitions. The first analysis uses the
proportion of new petitions filed each term that were filed IFP; this variable, then,
captures the immediate demand for adjudication of IFP petitions. The second analysis, in
contrast, uses the proportion of petitions on the docket, both newly filed and carried over
from previous terms, that are IFP ; this variable accounts for the availability of paid and
IFP petitions for review. These two measures of availability are highly correlated
(r=.981, p<.001) so they cannot be included in the same analysis. Instead, separate
analyses were performed using the two availability variables.

Dynamic Analysis with Proportion of Cases Filed
Regressing the precentage of cases granted review originating on the IFP docket
on the proportion of cases filed that are IFP and the ideology of the median Justice results
101

in a Durbin-Watson statistic of 1.15, which is below the lower bound for the DurbinWatson statistic for n=54 and k´=2 (dL = 1.49). Moreover, Figure 5.1 is the partial
autocorrelation graph of the residuals from the basic regression, and the PACF suggests
positive serial correlation. Accordingly, the null hypothesis of no positive
autocorrelation must be rejected.
There are a number of methods for correcting for positive serial correlation. One
such method involves including a lagged dependent variable. From a theoretical
perspective, a lagged dependent variable might account for the development of norms on
the Court. For example, if criminal procedure issues—which are frequently raised by
indigent petitioners—are given significant attention during one term of the Court, it is

Coefficient

1.0

Upper
Confidence
Limit
Lower
Confidence
Limit

Partial ACF

0.5

0.0

-0.5

-1.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Lag Number

Figure 5.1. Partial Autocorrelation Function of Residuals of Basic Regression
102

possible that that fact alone will affect how the Court behaves in the next term. The
presence of such norms may account for the serial correlation noted in the initial
regression. Including a lagged dependent variable as an independent variable may
eliminate the serial correlation problem detected in the original model through a
substantively meaningful way, pinpointing the source of the AR process rather than
simply compensating for it.
Including a lagged endogenous variable in the analysis yields the results in Table
5.1. The Durbin-Watson statistic is inappropriate in models employing a lagged
endogenous variable, but Durbin’s h statistic can be calculated from available
information and used to test for serial autocorrelation.
Durbin’s h is calculated with the following formula:
h = (1-DW/2) * [T/(1 – T[var B])]0.5
where T is the number of observations, DW is the Durbin-Watson statistic, and var B is
the square of the standard error of the estimated coefficient of the lagged endogenous
variable. Durbin’s h is normally distributed. Here, the calculated value for Durbin’s h is
1.9474; because this value is not statistically significant, the null hypothesis (no serial
correlation) cannot be rejected. The PACF graph of the residuals of this model, below in
Figure 5.2, lends support to this conclusion.

103

Variable

Coefficient
(B)

Standard
Error

-1.888

5.088

Proportion of All Cases Filed That are IFP

21.363*

10.314

Ideology of the Median Justice on the Supreme
Court (Coded -1 for Most Conservative, +1 for
Most Liberal)

8.327**

2.505

Lagged Endogenous Variable (Percentage of
Cases Granted Review Originating on the IFP
Docket in the Previous Term)

0.397**

0.132

Dependent Variable:
Percentage of Cases Granted Review
Originating on the IFP Docket
Constant

R2
Adjusted R2
N
* p<.05
** p<.01

0.601
0.577
53

Table 5.1. Predictors of Aggregate IFP Petition Selection (Proportion of Cases Filed IFP)

The explanatory coefficients are both significant and appropriately signed. Specifically,
as the proportion of cases filed with the Supreme Court as IFP petitions increases, the
proportion of cases accepted by the Supreme Court which are IFP cases also increases.
This finding may seem surprising given that the Court’s attention to IFP petitions has
remained relatively stable since 1990 while the proportion of the filings with the Court
that are IFP has continued to climb.
104

Coefficient

1.0

Upper
Confidence
Limit
Lower
Confidence
Limit

Partial ACF

0.5

0.0

-0.5

-1.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Lag Number

Figure 5.2. PACF of Residuals from Model Employing Lagged Endogenous Variable

Figure 5.3 plots both the proportion of cases filed with the Court that are IFP (converted
to a percentage by multiplying the raw number by 100) and the dependent variable--the
percentage of cases granted review that originate on the IFP docket--over time, using a
centered moving average of five years to smooth the curve. That figure illustrates how
closely the two variables tracked one another until roughly 1990, when the grant measure
levels off and the availability measure continues to rise. This apparent disparity,
however, can be reconciled by the fact that we are controlling for the ideology of the
median Justice on the Supreme Court, a value that remained remarkably low during this
timeframe.
105

Grantperc

80.00

Fileprop * 100

60.00

40.00

20.00

0.00

2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
1948

Year

Figure 5.3. IFP Filings and IFP Acceptances, as Percents of Totals, Smoothed

Moreover, as overall ideology of the Supreme Court, as determined by the
ideology of the Court’s median member, becomes more liberal, the Court devotes more
of its attention to IFP petitions.
In addition, the lagged endogenous variable is also significant and positively
signed, indicating the existence of some form of “norm” building on the Court. In other
words, the Court’s attention to IFP petitions in a given term is positively correlated with
its behavior in the term immediately precedent.

106

Overall, the model explains nearly 60% of the variance in the dependent variable,
a rather impressive amount given the simplicity of the model.

Dynamic Analysis with Proportion of Cases on Docket
Not surprisingly, regressing the dependent variable on the ideology measure and
the second availability measure (the proportion of cases on the Court’s docket that are
IFP) results in a Durbin-Watson statistic of 1.267, which is below the lower bound for the
Durbin-Watson statistic for n=54 and k´=2 (dL = 1.49). Accordingly, the null hypothesis
of no positive autocorrelation must be rejected in this model as well.
Including a lagged endogenous variable in the analysis yields the results in Table
5.2. Durbin’s h cannot be calculated in this instance because T[var B] is greater than 1,
so the term under the square-root is less than zero. Instead, I used the Breusch-Godfrey
test for higher-order autocorrelation.
To perform the Breusch-Godfrey test, I saved the residuals from the primary analysis and
regressed those residuals on the original regressors and four lags of the residuals.45 The
Breusch-Godfrey test statistic is then calculated:
(n-p)* R2 ~ χ2p
where “n” is the number of observations in the underlying regression (here, 53) and “p” is

45

The use of four lags is somewhat arbitrary. Ideally, if there is some a priori reason to expect
autocorrelation of some particular order, the regression would include sufficient lags to satisfy those
autocorrelation concerns. Here, however there is little reason to expect any particular order of
autocorrelation; I chose, then, to use four lags to be certain that any latent autocorrelation.

107

Coefficient
(B)

Standard
Error

-2.597

4.650

Proportion of the Court’s Docket that is IFP

24.597*

9.949

Ideology of the Median Justice on the Supreme
Court (Coded -1 for Most Conservative, +1 for
Most Liberal)

9.258**

2.566

Lagged Endogenous Variable (Percentage of
Cases Granted Review Originating on the IFP
Docket in the Previous Term)

0.338*

0.138

Variable

Dependent Variable:
Percentage of Cases Granted Review Originating
on the IFP Docket
Constant

R2
Adjusted R2
N
* p<.05
** p<.01

0.615
0.591
53

Table 5.2. Predictors of Aggregate IFP Petition Selection
(Proportion of Cases on Docket that are IFP)
the number of lags of the residual used in the test regression (here, 4). The R2 of the test
regression was 0.132, so the test statistic equals 6.468. The critical chi-square value at
the p=.05 level of significance for 4 degrees of freedom is 9.48773. Because the
Breusch-Godfrey test statistic (6.468) is less than this critical value, we cannot reject the
null hypothesis of no autocorrelation.
108

The PACF graph of the residuals, Figure 5.4, supports the result of the BreuschGodfrey test that there is no autocorrelation.

Coefficient

1.0

Upper
Confidence
Limit
Lower
Confidence
Limit

Partial ACF

0.5

0.0

-0.5

-1.0
1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16

Lag Number

Figure 5.4. PACF of Residuals from Model Employing Lagged Endogenous Variable

The results of this model are consistent with the earlier model. The explanatory
coefficients are both significant and appropriately signed. Specifically, the Court’s
attention to IFP cases increases as the Court becomes more liberal and as the proportion
of available cases arising on the IFP docket increases. Once again, the lagged
endogenous variable is also significant and positively signed, indicating the existence of
some form of “norm” building on the Court or continuity in the Court’s attention to IFP
petitions. The predictive power of this analysis is comparable to the analysis using the
109

other measure of availability; again, the model accounts for roughly 60% of the variance
in the dependent variable.

Implications
This analysis has several interesting implications. First, it seems to confirm that
there is an ideological aspect to the Court’s selection of IFP petitions. Specifically, as the
Court becomes more liberal, IFP petitions are more likely to receive review. Although
this conclusion is quite intuitive, what we would expect given our understanding of the
Court’s error-correction approach to the certiorari process and the types of petitions on
the IFP docket, the strong empirical support for this conclusion is noteworthy.
Second, this analysis indicates that there is some continuity between terms of the
Court with respect to attention to IFP petitions, that each term’s case selection is not an
independent event but, rather, is shaped in part by the Court’s focus from the term before.
This observation has implications for our understanding of the Court’s agenda-setting
process that extend beyond consideration of the IFP docket. Specifically, analysis of the
Court’s case-selection process tends to assume that each decision, to grant or deny
certiorari with respect to a particular petition is an independent event (this despite the
understanding voiced in the literature of the workload constraints under which the Court
operates). If there is a temporal interdependence of case selection from term to term, it

110

follows that there may be within-term interdependence of the decision to select certain
petitions for review.46
Such interdependence is interesting in its own right, but it also has important
implications for the analysis that has been done of the Court’s agenda-setting in the past,
because it would suggest that one of the primary assumptions that underlie both
multivariate regression analysis and maximum likelihood estimation, independence of
irrelevant alternatives, may be violated. In other words, there may be some form of
“spatial autocorrelation” at work that has not, as yet, been taken into account in statistical
models of the Court’s selection process.
For purposes of addressing the primary question posed by this study--whether,
how, and to what extent the Court’s selection of IFP petitions differs from its selection of
paid petitions--this dynamic analysis lends support to the underlying assumption that the
Court’s selection of IFP petitions is different in kind from its selection of paid petitions.
In Chapter One, I described the administrative separation of the paid and IFP petitions,
and in Chapter Four, we learned that there are real differences between the cases on the
IFP docket and the paid docket. This analysis, however, suggests that the administrative
and qualitative differences between the two dockets translate into disparate treatment by
the Court during the agenda-setting process itself. Specifically, the importance of
ideology to the Court’s attention to IFP petitions and the temporal trends in that attention
point to a distinction between the IFP petitions and the paid petitions in the Court’s
agenda-setting process that transcends the qualitative differences between the two
46

To draw an analogy, the Court’s agenda-setting function may be akin to the composition of a
symphony, where each note is not chosen in isolation but in reference to the notes that went before and the
notes to come, in an effort to create a unified whole.

111

dockets that were outlined in Chapter Four. The significance of the lagged endogenous
variable, in particular, suggests that the Court operates under evolving norms of behavior
with respect to the IFP petitions.

CASE-LEVEL ANALYSIS OF THE SELECTION OF IFP PETITIONS
Having determined that there are substantive differences between the IFP
petitions and the paid petitions and that there is some real distinction between the Court’s
selection of IFP petitions and its selection of paid petitions, at this point the analysis turns
to whether and how IFP status affects the probability that the Court will grant certiorari
in an individual case. In Chapter Three, I identified two specific questions designed to
ascertain whether and how the decision calculus used by the Supreme Court in the
agenda-setting process varies between the paid and the unpaid docket. First, is the base
probability of the Court accepting a petition in the absence of any identified cue the same
for both paid and unpaid petitions? Second, is the marginal effect of each of the
identified cues on the probability of the Court accepting a petition the same for both paid
and unpaid petitions?

Independent Effect of IFP Status on Probability of Review
To address the first of these questions, I used logit analysis of the complete
dataset. The dependent variable, grant/deny, is coded “1” if the Court granted review or
summarily disposed of the case on the merits and coded “0” if the Court denied review or
issued a “GVR” order. In addition to IFP status, the independent variables employed
reflect the “cues” identified in other empirical analyses of the Court’s agenda-setting
112

function: the ideology of the lower court decision (“-1” for a liberal decision, “1” for a
conservative decision), the presence of a sua sponte responsive brief, whether the
petitioner proceeded pro se, whether the U.S. Solicitor General supported the petition for
review, whether the U.S. Solicitor General opposed review, whether any cert-stage amici
were filed, whether the decision from which review is sought is an en banc decision of a
U.S. Court of Appeals, and whether the petition references each of the Rule 10 factors
(dissent below, reversal below, interjurisdictional conflict, break from precedent, and the
novelty/importance of the case). Cases were weighted based on the probability of a case
being drawn from the population; in other words, they were weighted based on the
proportion of cases sampled from each stratum relative to the number of cases in each
stratum). Because theory dictated the direction of the effect of each independent
variable, significance was determined using a one-tailed test. The results of this analysis
are set forth in Table 5.3.
The pseudo-R2 of the model is 0.4346, indicating that the model overall does a
reasonable job of predicting case acceptance given the large stochastic and fact-specific
component of the agenda-setting process. Moreover, the sign associated with each
coefficient was consistent with a priori expectations.
The model indicates that the majority of the identified cues were significantly
correlated with the Court’s decision to grant or deny review.47 Specifically, the Court
was significantly more likely to accept a case if: (1) the decision below was ideologically
liberal; (2) the respondent filed a responsive brief; (3) the petitioner noted an
47

The en banc cue approached, but did not achieve, significance; the one-tailed p-value for the en
banc cue was 0.051, just barely above the threshold for statistical significance. Dissent below, a reversal in
the case’s procedural history, and an allegation of the novelty or importance of the issue did not even
approach statistical significance.

113

Variable

Coefficient

Standard Error

-5.054

1.165

-0.909***

0.265

2.767**

0.975

-1.706***

0.466

Dissent on Court Below?

0.419

0.499

Reversal in Case’s Procedural History?

0.084

0.343

Interjurisdictional Conflict Alleged?

0.858**

0.301

Break from Supreme Court Precedent Alleged?

0.813**

0.314

0.214

0.316

4.400***

0.497

Solititor General Opposition to Petition?

-0.681*

0.341

Cert-Stage Amicus Participation?

1.480*

0.649

En Banc Decision Below?

0.995

0.656

-0.465*

0.273

Constant
Ideological Direction of Lower Court Decision
(Coded -1 for Liberal, +1 for Conservative)
Responsive Brief Filed Sua Sponte?
Pro Se Petitioner?

Allegation of Novelty/Importance of Issue?
Solicitor General Support for Petition?

Petitioner IFP?
*
**
***

p<.05
p<.01
p<.001

Table 5.3. Logit Analysis of All Cases, GVRs and Summary Dispositions Included
114

interjurisdictional conflict; (4) the petitioner noted a break from existing Supreme Court
precedent; (5) cert-stage amicus briefs were filed; or (6) the Solicitor General supported
review. The Court was significantly less likely to accept a case if: (1) the petitioner filed
pro se or (2) the Solicitor General opposed review.
These findings are generally consistent with existing scholarship on the Court’s
agenda-setting literature. It is interesting to note, however, that dissent on the court
below and reversal in the case’s procedural history were not significantly associated with
the Court’s decision to grant review. Of the Rule 10 factors, these two are arguably the
most “objective”; these factors reflect the behavior of other judges rather than allegations
or arguments made by attorneys or pro se litigants.
For purposes of this dissertation, the most important result is that, holding all
identified cues constant, IFP status does significantly reduce the probability that the
Court will grant review in an individual case. In the absence of any other cues, either
positive or negative,48 the probability of the Court granting review to an indigent
petitioner seeking review of a conservative lower court decision is 0.16%, while the
probability of the Court granting review to a similarly situated paying petitioner is 0.26%
(a difference of 0.1%). In the absence of any other cues, either positive or negative, the
probability of the Court granting review to an indigent petitioner seeking review of a
liberal lower court decision is 0.99%, while the probability of the Court granting review
to a similarly situated paying petitioner is 1.56% (a difference of 0.57%). While these
differences may not appear large in any absolute sense, given the improbability of the
48

In other words, the variables coding for a pro se petitioner, cert-stage amici, S.G. support, S.G.
opposition, an en banc decision below, a responsive brief, and each of the five Rule 10 factors were all set
to “0”.

115

Court granting any petition for review, the relative disadvantage of IFP petitions is
striking.
As noted in Chapter Three, the decision to include GVR and summary disposition
cases is somewhat controversial. Accordingly, I re-ran the analysis of all cases, this time
excluding the GVR and summary disposition cases. In this more constrained analysis,
Solicitor General support for the petition was dropped from the model because there was
no variation on the dependent variable when Solicitor General support was present; in
other words, the Court granted review in every case in which the Solicitor General
supported review.49 Results of this second analysis on the smaller dataset, including only
outright grants or denials of review, are presented in Table 5.4.
This additional analysis differs from the first in several interesting ways. First,
the presence of a dissenting opinion below becomes significant, while the Solicitor
General’s opposition to review loses significance (p=.087, one-tailed). More importantly,
the effect of IFP status on the probability of the Court granting review becomes
statistically insignificant, though just barely (p=.051, one-tailed).
In the case of IFP status and the Solicitor General’s opposition to review, both of
which are negatively signed and both of which lose significance when the GVR and
summary disposition cases are dropped from the analysis, it appears that these variables
are affected by the loss of GVR cases. Table 5.5 presents a crosstabulation (using
weighted data) of IFP status and disposition, and Table 5.6 presents a crosstabulation of
Solicitor General position and disposition.
49

S.G. support did not predict success perfectly in the first analysis because there were three cases in
which the Solicitor General supported review and the Court issued a GVR order (coded as a denial of
review). Once I excluded these “gray area” cases, however, there were no cases in which the Solicitor
General supported review yet the Court denied review.

116

Variable

Coefficient

Standard Error

-3.316

0.729

-1.282***

0.275

1.219**

0.474

-2.128***

0.513

1.153**

0.439

0.295

0.370

Interjurisdictional Conflict Alleged?

0.687**

0.323

Break from Supreme Court Precedent Alleged?

0.781**

0.348

Allegation of Novelty/Importance of Issue?

0.402

0.329

Solicitor General Opposition to Position?

-0.511

0.377

Cert-Stage Amicus Participation?

1.290*

0.652

En Banc Decision Below?

0.827

0.644

Petitioner IFP?

-0.471

0.306

Constant
Ideological Direction of Lower Court Decision
(Coded -1 for Liberal, +1 for Conservative)
Responsive Brief Filed Sua Sponte?
Pro Se Petitioner?
Dissent on Court Below?
Reversal in Case’s Procedural History?

*
**
***

p<.05
p<.01
p<.001

Table 5.4. Logit Analysis of All Cases, GVRs and Summary Dispositions Excluded

117

Cert Denied

Cert Granted

GVR

Summary
Affirm

Summary
Reverse

Paid Cases

297

27

6

0

1

IFP Cases

396

3

6

0

0

Table 5.5. Crosstab of IFP Status and Case Disposition, Weighted Data

Cert Denied

Cert Granted

GVR

Summary
Affirm

Summary
Reverse

SG Supports

0

10

3

0

1

SG Opposes

154

4

3

0

0

SG Silent

540

15

6

0

1

Table 5.6. Crosstab of Solicitor General Position and Case Disposition, Weighted Data

Presumably, the loss of significance for both variables results from the reduction
in “denials” or “failures” in the dependent variable, specifically those denials or failures
in which the independent variable of interest (Solicitor General opposition or IFP status)
is present. Although it is not entirely clear from the tables presented, it seems that the
effect of IFP status and Solicitor General opposition on the probability of the Court
granting review is driven in part by the Court’s predilection to afford those cases cursory
118

treatment by issuing a GVR order; GVR cases are not solely responsible for the effect, of
course, but it is the GVR cases that push the variables over the threshold of significance.
This is a particularly interesting finding. It suggests--although it certainly does
not prove--that even when the Court believes that the lower court decision deserves
closer inspection, the Court is more likely to take the relatively easy way out by issuing a
GVR order when the Solicitor General argues against review or when the petitioner is
IFP. In other words, the Court is more likely to “pass the buck” in these cases, send the
matter back to the lower court rather than expend the effort to grant full plenary review.
Thus, even though IFP status loses significance in this second analysis, the analysis
nevertheless lends support to the notion that the Court treats IFP petitions differently.
As noted above, one cue characteristic--dissent on the court below--which was not
significant in the first analysis is significant when the summary dispositions and GVRs
are excluded from the analysis. There is a positive correlation between the existence of a
dissent below and a “gray area” disposition (summary disposition or GVR): r=.167,
p<.001. Moreover, that correlation seems to be driven primarily by the GVR orders.
This seems to suggest that a dissent below increases the probability that the Court will
either grant review or send the case back to the lower court for reconsideration. Although
this finding could lead to a number of possible explanations, one story that seems quite
plausible is that dissent below is correlated with an incorrect decision below rather than
the importance of the issue presented. In other words, the concept of “certworthiness”
incorporates both legal factors (creating a consistent legal doctrine) and policy factors
(moving policy in the preferred direction and having a large effect on legal policy); a
dissent on the court below implicates the legal aspect of certworthiness but not
119

necessarily the policy aspect. Accordingly, a dissent on the court below triggers either
plenary review or a GVR order, the latter counting as a denial of review in the first
analysis and thus obscuring the effect of the cue.
Although this second analysis complicates the findings of the initial model
somewhat, the overall story remains the same. The majority of the cues identified in
prior studies are significant determinants of the Court’s grant of plenary review, and,
more importantly, IFP status does affect the Court’s disposition of cases even when those
other cues are held constant. All told, this first set of analyses supports Hypothesis IIa,
that the base probability of the Court accepting a petition for review (particularly for
granting full plenary review) in the absence of any identified cue is lower for unpaid
petitions than for paid petitions.

Interaction Between IFP Status and Other Cue Characteristics
Turning next to the question of whether the effect of the cues (other than IFP
status) varies between the IFP and paid dockets, I approached the issue two different
ways. First, I performed separate logit analyses on the IFP petitions and paid petitions.
For purposes of analyzing the IFP docket, I had to exclude two independent
variables--cert-stage amici and Solicitor General support--because they perfectly
predicted “success” (a grant of review). In other words, among the IFP petitions included
in the sample, the Court accepted every case in which a cert-stage amicus brief was filed
or for which the Solicitor General expressed support. Results for the logit analysis of the
remaining cues are set forth in Table 5.7.

120

Variable

Coefficient

Standard Error

Constant

-7.923

0.723

Ideological Direction of Lower Court Decision
(Coded -1 for Liberal, +1 for Conservative)

-0.572

0.501

3.077***

0.423

-1.279*

0.573

2.280***

0.533

-0.140

0.640

Interjurisdictional Conflict Alleged?

2.308***

0.418

Break from Support Precedent Alleged?

0.976**

0.407

Allegation of Novelty/Importance of Issue?

0.801*

0.399

-2.538***

0.515

Responsive Brief Filed Sua Sponte?
Pro Se Petitioner?
Dissent on Court Below?
Reversal in Case’s Procedural History?

Solicitor General Opposition to Petition?
En Banc Decision Below?
*
**
***

0.639

0.910

p<.05
p<.01
p<.001
Table 5.7. Logit Analysis of IFP Petitions Only
For purposes of analyzing the paid docket, I had to exclude another independent

variable--pro se status--because it predicted failure perfectly. In other words, among the
paid petitions included in the sample, the Court denied every pro se petition. Results for
the logit analysis of the remaining cues are set forth in Table 5.8.
121

Variable

Coefficient

Standard Error

-5.727

1.123

-1.104***

0.210

Responsive Brief Filed Sua Sponte?

2.274*

1.074

Dissent on Court Below?

0.539

0.502

Reversal in Case’s Procedural History?

-0.149

0.406

Interjurisdictional Conflict Alleged?

0.696*

0.372

Break from Supreme Court Precedent Alleged?

0.731*

0.403

Allegation of Novelty/Importance of Issue?

0.115

0.390

4.804***

0.526

Solicitor General Opposition to Petition?

-0.754*

0.469

Cert-Stage Amicus Participation?

1.421*

0.680

En Banc Decision Below?

1.683*

0.804

Constant
Ideological Direction of Lower Court Decision
(Coded -1 for Liberal, +1 for Conservative)

Solicitor General Support for Petition?

*
***

p<.05
p<.001
Table 5.8. Logit Analysis of Paid Petitions Only
Differences between the two analyses are immediately apparent. At the outset, it

is worth noting that, while the direction of the lower court decision is a significant
predictor of a grant of certiorari for paid petitions, it has no significant effect on the
122

disposition of the unpaid petitions. This is not altogether surprising given the relatively
small number of IFP petitions that seek review of liberal lower court decisions. In other
words, there is relatively little variation in the lower court direction variable with respect
to the IFP petitions.
To further illustrate the different effects of cues on the paid and the unpaid
dockets, Table 5.9 shows the change in probability of the Court granting review
attributable to each cue, assuming a base model in which the lower court decision is
conservative and no other cue, either positive or negative, is present; Table 5.10 shows
the same first-differences assuming a base model in which the lower court decision is
liberal and no other cue, either positive or negative, is present. Because the
overwhelming majority of IFP petitions are requests for review of conservative lower
court decisions, the figures presented in Table 5.9 have the most practical significance.
The probabilities and changes in probability reported in Tables 5.9 and 5.10 are
somewhat misleading. For example, it appears at first blush that the presence of a
responsive brief affects the probability of the Court granting review to a paid petition
more than the probability of the Court granting review to an IFP petition. Specifically,
when the decision below is conservative, the presence of a responsive brief increases the
probability of the Court granting review to a paid petition by 0.93 percentage points,
while it increases the probability of the Court granting review to an unpaid petition by
only 0.42 percentage points. However, this raw change fails to take into consideration
the difference in base probability of the Court granting review to a paid or unpaid
petition. A more meaningful comparison of the cue characteristics requires consideration

123

IFP Cases Only
(base probability = 0.02%)
Probability
with Cue

Paid Cases Only
(base probability = 0.11%)

Change

Probability
with Cue

Change
+0.93%

124

Responsive Brief

0.44%

+0.42%

1.04%

Pro Se Petitioner

0.01%

-0.01%

n/a

Dissent Below

0.20%

+0.18%

0.18%

+0.07%

Reversal Below

0.02%

0.00%

0.09%

-0.02%

Interjur. Conflict

0.20%

+0.18%

0.22%

+0.11%

Break from Precedent

0.05%

+0.03%

0.22%

+0.11%

Novel/Imp. Issue

0.05%

+0.03%

0.12%

+0.01%

---

11.64%

+11.53%

-0.02%

0.05%

-0.06%

---

0.45%

+0.34%

0.58%

+0.47%

SG Support

n/a

SG Opposition

0.00%

Cert-Stage Amici

n/a

En Banc

0.04%

+0.02%

---

Table 5.9. Changes in Probability Associated with Cues, Conservative Lower Court Decision

IFP Cases Only
(base probability = 0.06%)

Paid Cases Only
(base probability = 0.97%)

125

Probability
with Cue

Change

Probability
with Cue

Change

Responsive Brief

1.37%

+1.31%

8.71%

+7.74%

Pro Se Petitioner

0.02%

-0.04%

n/a

Dissent Below

0.62%

+0.56%

1.66%

+0.69%

Reversal Below

0.06%

0.00%

0.84%

-0.13%

Interjur. Conflict

0.64%

+0.58%

1.93%

+0.96%

Break from Precedent

0.17%

+0.11%

2.00%

+1.03%

Novel/Imp. Issue

0.14%

+0.08%

1.09%

+0.12%

SG Support

n/a

---

54.51%

+53.54%

SG Opposition

0.01%

-0.05%

0.46%

-0.51%

Cert-Stage Amici

n/a

---

3.91%

+2.94%

En Banc

0.12%

5.02%

+4.05%

+0.06%

---

Table 5.10. Changes in Probability Associated with Cues, Liberal Lower Court Decision

of the magnitude of the effect they have on the probability of granting review relative to
the probability of granting review in their absence.
We can get a sense of the relative effect of the cues on the unpaid and paid
dockets by looking at the factor change in the odds of the Court granting review
associated with each cue. With logit analysis, a unit change in an independent variable,
xk, (here, moving from “0” to “1,” or absent to present) changes the odds of “success”
(the dependent variable equaling “1”) by a factor of exp(βk) (Long, 1997: 79-82). Table
5.11 illustrates the factor changes associated with each independent variable across the
two groups, IFP and paid cases.
Recall from the conclusion of Chapter Four that, based on the relative prevalence
or scarcity of various cues on the two dockets, I hypothesized that the presence of a
responsive brief, a dissent on the court below, a reversal in the case’s procedural history,
an allegation of an interjurisdictional conflict, an allegation of a novel or important issue,
Solicitor General support, Solicitor General opposition, and cert-stage amicus
participation would all have greater effects on the Court’s selection of unpaid cases than
paid cases. A pro se petitioner was hypothesized to have a greater effect on the Court’s
selection of paid cases than unpaid cases. This analysis generally supports those
expectations.
Looking at the magnitude of the change in odds, the effects of the presence of a
responsive brief, a dissent on the court below, an allegation of interjurisdictional conflict,
an allegation of a novel or important issue, and Solicitor General opposition on the
selection of IFP petitions are all remarkably greater than the effects of those variables on
the selection of paid cases. Moreover, the descriptive analysis in Chapter Four showed
126

Factor Change
IFP Cases

Factor Change
Paid Cases

Responsive Brief

21.69

9.72

Pro Se Petitioner

0.28*
(3.57)

--

Dissent Below

9.78

1.71

Reversal Below

0.87*
(1.15)

0.86*
(1.16)

Interjurisdictional Conflict

10.05

2.01

Break from Precedent

2.65

2.08

Novel/Important Issue

2.23

1.12

--

122.0

0.08*
(12.5)

0.47*
(2.13)

--

4.14

1.89

5.38

SG Support
SG Opposition
Cert-Stage Amici
En Banc Decision Below

* Negative effects result in factor changes between 0 and 1. For the sake of comparison,
the magnitude of the factor change from a negative effect is calculated by taking the
inverse of the factor change. That inverse value is reported in parentheses.
Table 5.11. Factor Change Associated with Cue Characterstics, Paid and IFP Cases

127

that there was no significant difference between the two dockets with respect to an
allegation of a break from precedent; accordingly, we would not expect the magnitude of
the effect of that variable to differ significantly between the two dockets and, while it is
higher on the IFP docket than on the paid docket, the difference is not as stark as with
other variables.
Head-to-head comparisons of the effects of Solicitor General support, cert-stage
amicus participation, and the pro se status of the petitioner are not possible because the
first two variables perfectly predict success on the IFP docket (where they were expected
to have a greater effect) and the last variable perfectly predicts failure on the paid docket
(where it was expected to have a greater, albeit negative, effect). Thus, while an outright
comparison of these variables is not possible, it is not possible precisely because of the
overwhelming magnitude of the effects of these variables on the dockets where their
effects were expected to be largest.
For example, nearly 8% of the paid petitions were filed pro se, and all of them
included in this sample were denied review; essentially, the probability of review
dropped to “0.” Pro se status does significantly reduce the probability of the Court
granting review to an IFP case, but the magnitude of the effect is not so stark: the
presence of a pro se petitioner cuts the probability of success in half.
The one variable that did not perform as expected was the presence of an en banc
decision on the court below. The descriptive analysis in Chapter Four indicated that there
was no difference in the prevalence of this cue between the two dockets. Accordingly,
we would not expect the effect of this cue to vary significantly between the two dockets.
However, the factor change in odds attributable to this cue is remarkably larger in the
128

context of the paid docket (5.38) than the IFP docket (1.89). It is an interesting finding,
one certainly worth closer consideration; however, at this point, no theoretical
explanation presents itself.
An alternative way to assess the difference in the effects of the cue characteristics
across the two dockets is to construct a series of interaction terms and perform logit
analysis on the full dataset. Such analysis allows for a more fully specified model;
specifically, while certain variables had to simply be dropped from either the analysis of
IFP cases or the analysis of paid cases because they perfectly predicted success or failure
on one docket, they can be included in the overall model (although they cannot be used to
create interaction terms).
Thus, I performed logit analysis on the full dataset including all of the cue
characteristics as well as terms to capture the interaction between IFP status and a
responsive brief, a dissent below, a reversal in the case’s procedural history, an
interjurisdictional conflict, an allegation of a break from precedent, an allegation of a
novel or important issue, and the Solicitor General’s opposition to the petition. The
results of that analysis are in Table 5.12.
Using the coefficients from this model, Table 5.13 shows the probability that the
Court will grant review to a conservative lower court decision when the case is filed IFP
or paid, based on the presence of a single cue characteristic and the change in probability
associated with the cues relative to the base probability of a case without cues.

129

Variable

Coefficient

Standard Error

-5.007

1.454

-0.909***

0.261

-1.653

1.480

-1.755***

0.550

Responsive Brief Filed Sua Sponte?

2.793*

1.327

IFP/Response Interaction

0.124

1.396

Dissent in Court Below?

0.223

0.545

IFP/Dissent Interaction

2.036**

0.743

Reversal in Case’s Procedural History?

0.102

0.359

IFP/Reversal Interaction

-0.217

0.696

Interjurisdictional Conflict Alleged?

0.741*

0.342

IFP/Conflict Interaction

1.505**

0.546

Break from Supreme Court Precedent Alleged?

0.820*

0.365

IFP/Break from Precedent Interaction

0.255

0.533

Allegation of Novelty/Importance of Issue?

0.152

0.360

IFP/Novelty/Importance Interaction

0.585

0.520

4.515***

0.494

-0.466

0.388

IFP/SG Opposition Interaction

-1.889**

0.630

Cert-Stage Amicus Participation?

1.564**

0.629

0.923

0.701

Constant
Ideological Direction of Lower Court Decision
(Coded -1 for Liberal, +1 for Conservative)
IFP Petitioner?
Pro Se Petitioner?

Solicitor General Support of Petition?
Solicitor General Opposition to Petition?

En Banc Decision Below?
*

p<.05, ** p<.01, ***p<.001
Table 5.12. Logit Analysis of All Cases, Including Interaction Terms
130

IFP Cases Only
(base probability = 0.05%)
Probability
with Cue

Paid Cases Only
(base probability = 0.27%)

Change

Probability
with Cue

Change

131

Responsive Brief

0.94%

+0.89%

4.22%

+3.95%

Dissent Below

0.49%

+0.44%

0.34%

+0.07%

Reversal Below

0.07%

+0.02%

0.30%

+0.03%

Interjur. Conflict

0.49%

+0.44%

0.56%

+0.29%

Break from Precedent

0.15%

+0.10%

0.61%

+0.34%

Novel/Imp. Issue

0.11%

+0.06%

0.31%

+0.04%

SG Opposition

0.00%

-0.05%

0.17%

-0.10%

Table 5.13. Changes in Probability Associated with Cues, Conservative Lower Court Decision

Moreover, Table 5.14 presents the factor change associated with the cue
characteristics, for both paid and unpaid petitions, based on the analysis of all petitions
and the use of interaction terms.
Again, we see that Solicitor General opposition, an allegation of
interjurisdictional conflict, and a dissent on the court below have a significantly greater
effect on the selection of IFP petitions than on the selection of paid petitions (the
coefficients on the terms capturing the interaction of these cues with IFP status are
statistically significant). Although the effect of a responsive brief, an allegation of a
break from precedent, and an allegation of a novel or important issue is greater on the IFP
docket than on the paid docket, those interactions are not statistically significant.
Most interestingly, controlling for the interaction effects between IFP status and
cue characteristics actually causes IFP status to seem to lose its significance as a
determinant of case selection. In other words, at first glance, it would appear that the
independent effect of IFP status on the probability of the Court granting review was
actually a result of the difference in weights attached to cue characteristics across the two
dockets. Once the model was corrected to allow for the more significant influence of
three of the cue characteristics, the apparent bias against IFP petitions became
statistically insignificant.
However, the loss of significance seems to be the result of an increase in the
standard error associated with the coefficient. If we look at the factor change associated
with IFP status in the model without interaction terms (Table 5.3) and the factor change
associated with IFP status in the model including the interaction terms (Table 5.12), we
see that, in fact, the magnitude of the effect of IFP status actually increases. Specifically,
132

Factor Change
IFP Cases

Factor Change
Paid Cases

Responsive Brief

18.49

16.33

Pro Se Petitioner

0.17*
(5.88)

0.17*
(5.88)

Dissent Below

9.57

1.25

Reversal Below

0.89*
(1.12)

1.11

Interjurisdictional Conflict

9.45

2.10

Break from Precedent

2.93

2.27

Novel/Important Issue

2.09

1.16

SG Support

91.28

91.28

0.09*
(11.11)

0.63*
(1.59)

Cert-Stage Amici

4.78

4.78

En Banc Decision Below

2.52

2.52

SG Opposition

* Negative effects result in factor changes between 0 and 1. For the sake of comparison,
the magnitude of the factor change from a negative effect is calculated by taking the
inverse of the factor change. That inverse value is reported in parentheses.
Table 5.14. Factor Change Associated with Cue Characteristics, Paid and IFP Cases

133

the factor change associated with IFP status in the more parsimonious model is 1.59
(1/exp(-.465)) while the factor change associated with IFP status in the more complete
model is 5.26 (1/exp(-1.653)).
These results, while certainly not conclusive, suggest two things. First, these
results strongly suggest that the Court’s agenda-setting decision calculus varies between
the paid and unpaid dockets. Not only does IFP status have an independent effect on the
probability of the Court granting review, but also IFP status affects the weight attached to
other cue characteristics in the Court’s decision-making process.
The second, and perhaps more theoretically interesting, observation to make about
these results is that the manner in which the Court’s decision calculus varies between the
two dockets suggests that the “cues” that scholars have identified in the agenda-setting
literature really are cues, or heuristics, and not merely correlates of some latent quality of
“certworthiness.” If the correlations between identified cues and the grant of certiorari
were nothing more than a correlation, we would not expect to see any systematic
difference between the relationship between the presence of cues and the grant of
certiorari on the two dockets. In fact, there is a pattern to the differences between the
two dockets: the cues that are less common on the IFP docket generally have a greater
effect on the probability of success for IFP cases (and in several cases that greater effect
is statistically significant), while the one cue that is less common on the paid docket than
on the IFP docket (pro se status) appears to have a much greater effect on the probability
of success for paid cases. In other words, the data seem to support the hypothesis that
less common cues provide more information, a better means of distinguishing among
cases.
134

Again, this evidence is far from conclusive. Yet it does begin to build a case for
“cue theory” being something more than a neat descriptive tool. The pattern of
differences in the effect of identified cues between the unpaid and paid docket lends
credence to the notion that the cues themselves are a source of information to the Justices
during the agenda-setting process, and the value of that information depends in part on
the extent to which it serves to differentiate between cases and winnow down the pool of
potential cases to something more manageable. Clearly, these Justices do not use cues to
execute hard-and-fast decision rules, but it seems that the cues may play a causal role in
the agenda-setting process.

CONCLUSION
While each individual statistical test in this chapter provides only a brief glimpse
of the Court’s case selection process, taken as a whole they paint an interesting portrait.
The overarching theme is that, for purposes of agenda-setting, the Court’s IFP docket is
truly distinct from its paid docket. In the aggregate, the Court’s acceptance of IFP
petitions is influenced by both the Court’s ideological composition and the Court’s past
behavior towards IFP petitions. Looking at specific cases, the cue characteristics that
have been identified or implied by previous studies generally are correlated with the
probability of the Court granting review, and the relative effect of those factors varies
between the paid and unpaid docket in a manner that generally corresponds to the relative
prevalence of those factors on the two dockets. In short, the Court’s selection of unpaid
cases seems to be a process similar to yet distinct from its selection of paid cases.

135

From a theoretical standpoint, this conclusion is significant. One specific
conclusion of note is that the evidence suggests that the cues identified in other agendasetting analyses really do serve an informational role in the Court’s decision process.
More generally, however, the consistent divergence between the paid and unpaid dockets
suggests that the two dockets provide a sort of natural experiment in which the Court is
conducting the same general task—selecting cases—under different conditions. Sussing
out the nature and extent of the differences in the way the Court behaves under these
conditions will hopefully provide further insight into “black box” of the Court’s agendasetting function.
From a normative standpoint, the implications of this analysis are still more
significant. First, it appears that IFP cases truly are at a disadvantage relative to paid
cases; the IFP petitions are, ceteris paribus, less likely to be accepted by the Court.
Second, the Court seems to employ a different decision calculus between the two
dockets, weighting the cue characteristics differently in the two separate contexts, in a
manner that makes. Overall, this analysis lends credence to the fact that the IFP docket is
treated differently than the paid docket, and that difference in treatment calls into
question whether there really is equal access to the Court.

136

CHAPTER 6
CONCLUSIONS AND FUTURE RESEARCH

SUMMARY OF FINDINGS
At the outset, the concern of this research was the effect of the U.S. Supreme
Court’s agenda-setting function on a particular class of litigants: those who are so
impoverished that they file their petitions in forma pauperis. Such inquiry moves the
agenda-setting literature into relatively uncharted waters; the overwhelming majority of
empirical studies of the Court’s case selection have focused exclusively on the Court’s
paid docket, and those few studies that have incorporated the unpaid docket have not
made any comparison between the two groups of cases.
The purpose of pursuing this line of inquiry is twofold. First, there are important
normative considerations at stake; any discrimination against or even disparate treatment
of IFP petitions calls into question whether citizens, particularly poor citizens, truly enjoy
equal access to justice. Second, the Court’s selection of petitions from two separate
dockets creates a sort of natural experiment that allows us to better understand the microlevel cognitive process that occurs during agenda-setting.
In Chapter Two, I summarized the existing scholarship on the Court’s case
selection process. Although constrained by certain threshold legal considerations, the
Justices generally use their docket as a means to move broad legal policy in the direction
137

of their personal preferences (regardless of whether those preferences are born of genuine
legal ideals or more pragmatic political attitudes). Accordingly, the task the Justices face
is to winnow through the enormous number of petitions for review they receive each year
to select those cases that provide them with the best opportunities to affect and solidify
legal policy. Scholars have identified certain case characteristics--cue characteristics-that correlate with the likelihood of the Court granting review, although the jury is out on
whether the Justices actually use the presence of those cues as a heuristic device or a
source of information in their decision process or whether, in the alternative, the
correlation between the cue characteristics and case selection is incidental.
In Chapter Three, I set forth the central hypotheses of the dissertation. The null
hypothesis was that IFP petitions and paid petitions enjoy equal opportunity before the
Court; any difference in the rates of acceptance between the paid and unpaid dockets is a
result of the (presumed) inferior quality of the IFP petitions rather than any differential
treatment of those petitions by the Court. The alternative hypothesis, of course, was that
there is a difference between the Court’s selection of paid petitions and its selection of
IFP petitions. That difference might take two forms: (1) all else equal, the Court is less
likely to accept unpaid petitions than paid petitions and/or (2) the weight attached to other
determinants of the Court’s acceptance of cases--the cue characteristics--varies between
the two dockets. I further hypothesized that, if the Justices do in fact derive some
information about certworthiness from the presence of the cue characteristics, the relative
weight attached to those characteristics in the context of the two separate dockets would
be inversely proportional to the prevalence of the characteristic on that docket. Thus, if
one cue characteristic is less common on the IFP docket than on the paid docket, it will
138

have more informational value, be a more useful heuristic, in the context of the IFP
docket than the paid docket, and, accordingly, we would expect to see a stronger
correlation between that cue characteristic and a grant of review in the context of the IFP
docket.
Using the data collected specifically for this research, Chapter Four undertook a
detailed descriptive analysis of the Court’s unpaid docket, testing the assumptions
scholars and jurists have made about the overall quality of the IFP petitions and
determining which cue characteristics are significantly more or less prevalent on the IFP
docket than on the paid docket. That descriptive analysis yielded relatively few surprises.
Indeed, the common wisdom that the majority of IFP petitions are criminal matters or
prisoner civil rights matters was confirmed.
The civil cases that were on the IFP docket disproportionately presented issues of
family law and welfare benefits availability. What’s more, issues of minority rights (in
both the civil and criminal context) were disproportionately raised on the IFP docket.
These findings highlight the significance of the IFP docket; the Court’s treatment of in
forma pauperis petitions has important implications for the Court’s attention to issues
that affect minorities and indigent communities, both groups traditionally disenfranchised
from the electoral political process.
Moreover, with respect to potential cues of certworthiness, the analysis indicated
that a disproportionate number of the IFP petitions were filed pro se. Not only does this
suggest that pro bono and government-subsidized legal services are failing to meet the
demand for attorney representation among indigent litigants, but it also means that this
negative cue is less prevalent on the paid docket and, thus, should have a more
139

pronounced effect on the probability of paid petitions being accepted. Sua sponte
responsive briefs, a positive cue, were less prevalent on the IFP docket. With the
exception of an allegation of a break from precedent, the Rule 10 factors (all positive
cues) were all less prevalent on the IFP docket. Amicus participation, yet another positive
cue, was less prevalent on the IFP docket. Accordingly, if the cues do, in fact, provide
information to the Justices or otherwise directly affect the agenda-setting process, these
positive cues should all have more significant effects on the selection of IFP cases than
paid cases.
Chapter Five turned to analysis of the Court’s selection of IFP petitions. The first
section in the chapter considered determinants of the Court’s aggregate attention to IFP
petitions, specifically the proportion of cases ultimately accepted for review that
originated on the IFP docket. That analysis indicated that the proportion of the Court’s
filings that are IFP as well as the ideology of the median Justice on the Court are
correlated with the Court’s attention to IFP petitions. In other words, as the proportion of
the total petitions filed with the Court that are IFP rises, so does the proportion of
petitions accepted by the Court that originate on the IFP docket. Moreover, when the
Court is more liberal it devotes more attention to IFP cases, and when the Court is more
conservative it devotes less attention to IFP cases.
In addition, the aggregate analysis showed that the Court’s attention to IFP
petitions in one term is affected by its attention to IFP petitions in the term immediately
previous. This seemingly innocuous finding actually has important implications for the
remainder of the analysis. Specifically, the comparison of selection of IFP petitions with
the selection of paid petitions is premised at least in part on the notion that the two
140

dockets are, in fact, distinct in the eyes of the Justices. The fact that the Court’s attention
to IFP petitions shows some degree of continuity from term to term suggests that the
cases on the IFP docket form a cohesive group, that the Court treats them as a class apart
from the paid petitions.
At the outset, the case-level analysis of the Court’s selection of unpaid petitions
revealed that, holding cues of certworthiness constant, the Court is less likely to select an
unpaid petition than a paid petition.
More detailed analysis demonstrated that the Court attaches different weights to
the identified cue characteristics in the context of the IFP docket than in the context of the
paid docket. This analysis provides some modest support for the proposition that cues-both positive and negative--have stronger effects when they are less prevalent which, in
turn, suggests that cues do, in fact, provide information to the Justices about
certworthiness. It is impossible to determine whether the Justices use the presence or
absence of cues as quick decision rules or whether the cues each provide a fragment of
information that is aggregated across all cues and included in a more subtle calculus.
Nevertheless, the analysis indicates that the correlation between the cues and case
selection is not simply spurious.
Moreover, when the model controlled for the interaction between IFP status and
cue characteristics, the independent effect of IFP status on the probability of the Court
granting review, identified in the more parsimonious model, dropped below the threshold
for statistical significance. Still, the magnitude of the effect of IFP status on the
probability of the Court granting review was actually greater in the more complex model.
In other words, although IFP status dropped below the threshold of statistical
141

significance, it nevertheless seems to exert a strong effect on the probability of the Court
granting review.
As with all empirical research, the results of the analyses and the conclusions
drawn from those results are only as good as the data on which they are based. Although
considerable care was taken in the collection of data, undoubtedly mistakes were made.
Moreover, coding some of the variables required an exercise in judgment, and there was
no second set of eyes to ensure reliability. Accordingly, it is entirely possible that the
cut-points for some of those coding judgments shifted over the course of the data
collection process. Hopefully any errors in coding are white noise, but the possibility of
some degree of systematic error cannot be dismissed.
In addition, the sample size employed here requires some caution in accepting the
inferences drawn from analysis. Collecting data on the 737 cases in this sample took
hundreds of hours, and any further data collection would have been impractical.
However, 737 cases is a tiny fraction of the population about which these inferences are
being drawn, and, given the large number of independent variables used in the
multivariate analysis, a larger sample size would have been preferable.

NORMATIVE AND THEORETICAL IMPLICATIONS
Taken together, the analyses presented in the preceding chapters have significant
normative and theoretical implications.
From a normative standpoint, the analyses strongly suggest that the Court views
the IFP docket as distinct from the paid docket, that the Court’s selection process varies
between the two dockets, and that the IFP petitions are at a relative disadvantage to the
142

paid petitions. This raises serious concerns about whether access to the Court is
conditional on economic status and whether financial resources give some litigants an
added advantage. While it is generally accepted that money buys influence in the elected
branches, the courts have traditionally been seen as the political refuge for the little guy,
the playing field that is even for all participants. Any systematic disadvantage to a class
of litigants threatens that ideal of equality, but a disadvantage to a class of individuals
that is already disadvantaged in the broader political arena is particularly troubling.
The magnitude of that normative concern is increased by the fact that there appear
to be certain issues that are raised with greater frequency on the IFP docket than on the
paid docket. Family law issues, issues involving access to welfare benefits, and issues
involving the rights of minorities are more prevalent on the IFP docket. Accordingly, if
indigent petitioners are at a disadvantage in the Court’s agenda-setting process, so, too,
are these particular legal issues. Any bias against unpaid petitions thus has systemic
ramifications beyond discrimination against a particular class of litigants; such bias has
implications for the overall development of the law. Specifically, if the number of
petitions brought to the Court on a given issue represents the overall demand for Supreme
Court adjudication of that issue, and if the Court’s attention to that demand is skewed by
the Court’s bias against unpaid petitions, then the Court’s decisions will not keep pace
with demand for adjudication in these issue areas.
Not surprisingly, these are issue areas that disproportionatly affect groups that are
otherwise politically disadvantaged, who do not necessarily have any other political
forum for their grievances. Thus, not only do indigent individuals have less opportunity

143

to have their particular grievances addressed by the Court, but also they are less likely to
have the issues important to them addressed by the Court in the context of other disputes.
The descriptive analysis of the Court’s IFP docket further raises concerns about
the availability and quality of legal representation for indigent litigants. Given the
prevailing disdain for public defenders, the differences between paid and unpaid attorneyprepared petitions described in Chapter Four do not come as any great surprise.
However, seeing empirical evidence of such disparity, even at the level of Supreme Court
advocacy, is sobering.
Moreover, the number of indigent petitioners who file pro se suggests that
existing programs to provide legal services to indigent communities--programs ranging
from government-subsidized legal aid programs and public defender offices, to law
school clinical programs, to pro bono legal services provided by interest groups and law
firms--are insufficient to meet the demand for such services among indigent
communities. The deficit of attorney services is particularly troubling given that pro se
litigants are less likely to have their petitions accepted by the Court.
Of course, given the limitations of the data involved in this study, it is impossible
to unravel the extent to which the pro se litigants are pro se because their claims lack
sufficient merit to lure an attorney to their defense. Similarly, it is impossible to
determine whether the apparent lack of quality in the attorney-prepared unpaid petitions
is the result of poor advocacy as opposed to lack of material to work with. Nevertheless,
the number of pro se petitions and the disparity between attorney-prepared paid petitions
and attorney-prepared unpaid petitions raises red flags about the access of indigent
communities to that most important of legal resources, competent legal counsel.
144

From a theoretical standpoint, this research both reaffirms and advances the
existing scholarship on the Court’s agenda-setting process. Past research on the Court’s
case selection has identified a number of case characteristics that are correlated with the
Court’s decision to grant review, and the findings of this research generally support the
findings of those other studies. Even with respect to the bare-bones identification of
correlates of plenary review, this analysis adds to the existing research by employing
slightly different operationalizations of previously identified cues and by combining a
larger number of those cues into a single analysis. Specifically, where past studies have
focused on the success of the Solicitor General as a petitioner before the Court, this study
considers the more specific variables of Solicitor General support for review (regardless
of the government’s role in the case) and the Solicitor General’s opposition to review.
Both variables were found to be significant correlates of plenary review, and I believe
they better capture the nature of the hypothesized cause of the Solicitor General’s
remarkable success with the Court: the Court’s trust in the merit of the Solicitor
General’s arguments. In addition, this analysis included several of the lesser-studied cue
characteristics, such pro se status of the petitioner and the presence of an en banc
decision below; again, both of these cues were significant determinants of plenary
review.
Moreover, past research has done little to explore the meaning behind the
correlation between cue characteristics and plenary review: do those case characteristics
actually play a role in the Court’s decision-making process or are they merely correlated
with some underlying quality that really drives the Court’s decision?

145

Obviously, this study cannot conclusively answer that question. However, the
marginal effects of the cues in the context of the two separate dockets create a pattern that
suggests that the cue characteristics do provide some level of information to the Justices
during the agenda-setting process. Specifically, the general pattern indicates that cue
characteristics have greater effects when they are less prevalent, suggesting in turn that
they have more effect when they do a better job of distinguishing between cases. The
exact mechanism remains unclear, but these results suggest that the cue characteristics do
play an important role in the cognitive task undertaken by the Justices.

DIRECTIONS FOR FUTURE RESEARCH
This research suggests a number of avenues for future research, some closely
linked to the fundamental issue of the Court’s agenda-setting function and others more
loosely related to the normative concerns that underlie this dissertation.
First, ideally the dataset should be extended both backward and forward in time.
Extending the collection of case-level data backward into the Warren Court years would
create a dataset with more variation on the ideology of the Court and allow further
investigation of the role of ideology in the selection of unpaid petitions. Extending the
dataset forward, to cover more recent terms of the Court, would allow empirical
exploration of the effect, if any, of legislative and administrative measures designed to
curtail the explosion of frivolous litigation. Moreover, collecting data from a broader
range of Supreme Court terms would allow evaluation of how the substance of petitions
has changed over time; in other words, it would allow exploration of trends in both civil

146

and criminal issues raised by IFP petitioners. As a result, such research might begin to
explain the remarkable upswing in IFP petitions that began in the 1980s.
Before expanding the time period of the study, however, I believe it is necessary
to create a more complete picture of the Court’s agenda-setting during the ten terms
already examined. The next step I plan to take along this research path is to mine the
papers of Justices William Brennan and Lewis Powell for more detailed data on the
Court’s agenda-setting process. Indeed, the particular terms involved in this study were
chosen precisely because records of the Court’s agenda-setting decisions are available for
these terms in the Justices’ papers. Specifically, I would like to collect data about which
cases were placed on the Court’s discuss lists (a larger number than were actually granted
review), and I would like to collect data on the certiorari votes of the individual Justices.
Tanenhaus’s original theory of the role of cue characteristics in the certiorari
process dictated that the presence of cues drew the Court’s attention and ensured that the
Court would give those cases manifesting the cues further consideration; in other words,
the presence or absence of cues acted as a first cut in the agenda-setting process
(Tanenhaus et al., 1963). As Provine (1981) notes, if Tanenhaus is correct, the effect of
cues should be strongest during the Court’s initial screening process, when the Justices
create the discuss list. While the analysis in this dissertation lends some credence to the
notion that cues do, indeed, serve an informational role in the Court’s case selection
process, empirical examination of the effect of cues on paid and unpaid cases making the
discuss list would further illuminate the connection between the cue case characteristics
and the cognitive task of selecting petitions for review.

147

Similarly, collecting data on the certiorari votes of individual Justices on both
paid and unpaid petitions would allow further exploration of the role of strategy in the
Court’s agenda-setting process. In particular, do the Justices engage in the same type and
degree of strategic behavior in selecting unpaid petitions as paid petitions or, in the
alternative, does the distinction in the selection process revealed by this research extend
to the exercise of strategy?
Beyond expanding the dataset for the purposes of delving even deeper into the
puzzle of the Court’s agenda-setting process, this research also raises further questions
worthy of consideration. First, given the apparent disparity between the success of paid
and unpaid petitions, do unpaid petitioners or attorneys who represent indigent litigants
have lower opinions about the accessibility, procedural justice, and substantive fairness
of the Court than do paying petitioners and their attorneys? Direct survey research of
litigants and counsel could provide invaluable evidence of the effect the treatment of
unpaid petitions has on the Court’s institutional support and legitimacy.
Second, the relatively high quality of IFP pro se petitions relative to paid pro se
petitions, detailed in Chapter Four, raises intriguing questions about the availability and
skill of lay legal support networks available to indigent communities. In particular, how
skilled are the inmates who serve as “writ writers,” or jailhouse lawyers, and how
institutionalized have their services become? The fact that some writ writers expressly
claim authorship of the briefs they write and make reference to “corporations” that
provide inmate paralegal services suggests that there is an untapped underground of legal
activity that may be having untold influence on the development of the law. Survey of
these lay legal networks in the prisons and similar support groups that focus on family
148

law, probate, and other areas of civil law would help to guage the extent to which
individuals who are too poor to afford representation are nevertheless able to navigate the
legal system.
In short, this results of this initial dip into the IFP pool indicate that further study
of the unpaid docket will help develop a richer understanding of Court’s agenda-setting
process. This often-overlooked group of cases may hold the key to the black box of the
Court’s case selection process.
Moreover, this research suggests the need for further study about the gap between
rich and poor within the legal system. Is there, in fact, a growing legal underclass whose
issues and concerns are being given short shrift by the judicial branch? Have indigent
communities, such as the prison population, developed informal systems to compensate
for their lack of resources when dealing with the justice system? Do indigent litigants
perceive that they are treated differently within the court system and, if so, how does that
perception affect their belief in the fundamental fairness of the courts?

149

APPENDIX A
RULE 39, RULES OF THE SUPREME COURT OF THE UNITED STATES (1999)

Rule 39.

Proceedings In Forma Pauperis

1.
A party seeking to proceed in forma pauperis shall file a motion for leave to do
so, together with the party’s notarized affidavit or declaration (in compliance with 28
U.S.C. § 1746) in the form prescribed by the Federal Rules of Appellate Procedure, Form
4. The motion shall state whether leave to proceed in forma pauperis was sought in any
other court and, if so, whether leave was granted. If the United States district court of the
United States court of appeals has appointed counsel under the Criminal Justice Act of
1964, 18 U.S.C. § 3006A, or under any other applicable federal statute, no affidavit or
declaration is required, but the motion shall cite the statute under which counsel was
appointed.
2.
If leave to proceed in forma pauperis is sought for the purpose of filing a
document, the motion, and an affidavit or declaration if required, shall be filed together
with that document and shall comply in every respect with Rule 21. As provided in that
Rule, it suffices to file an original and 10 copies, unless the party is an inmate confined in
an institution and is not represented by counsel, in which case the original, alone,
suffices. A copy of the motion shall precede and be attached to each copy of the
accompanying document.
3.
Except when these Rules expressly provide that a document shall be prepared as
required by Rule 33.1, every document presented by a party proceeding under this Rule
whall be prepared as required by Rule 33.2 (unless such preparation is impossible).
Every document shall be legible. While making due allowance for any case presented
under this rule by a person appearing pro se, the Clerk will not file any document if it
does not comply with the substance of these Rules or is jurisdictionally out of time.
4.
When the documents required by paragraphs 1 and 2 of this Rule are presented to
the Clerk, accompanied by proof of service as required by Rule 29, they will be placed on
the docket without the payment of a docket fee or any other fee.
5.
The respondent or appellee in a case filed in forma pauperis shall respond in the
same manner and within the same time as in any other case of the same nature, except
that the filing of an original and 10 copies of a response prepared as required by Rule
33.2, with proof of service as required by Rule 29, suffices. The respondent or appellee
may challenge the grounds for the motion for leave to proceed in forma pauperis in a
separate document or in the response itself.
150

6.
Whenever the Court appoints counsel for an indigent party in a case set for oral
argument, the briefs on the merits submitted by that counsel, unless otherwise requested,
shall be prepared under the Clerk’s supervision. The Clerk also will reimburse appointed
counsl for any necessary travel expenses to Washington, D.C., and return in connection
with the argument.
7.
In a case in which certiorari has been granted, probable jurisdiction noted, or
consideration of jurisdiction postponed, this Court may appoint counsel to represent a
party financially unable to afford an attorney to the extent authorized by the Criminal
Justice Act of 1964, 18 U.S.C. § 3006A, or by any other applicable federal statute.
8.
If satisfied that a petition for a writ of certiorari, jurisdictional statement, or
petition for an extraordinary writ is frivolous or malicious, the Court may deny leave to
proceed in forma pauperis.

151

APPENDIX B
EXAMPLE OF PETITION TO PROCEED IN FORMA PAUPERIS
AND AFFIDAVIT OF INDIGENCY

152

153

154

LIST OF REFERENCES

Armstrong, Virginia and Charles A. Johnson. 1982. “Certiorari Decision Making by the
Warren and Burger Courts: Is Cue Theory Time Bound?” Polity 15: 141.
Baum, Lawrence. 1994. “What Judges Want: Judges’ Goals and Judicial Behavior,”
Political Research Quarterly 47: 749-768.
Baum, Lawrence. 1998. The Supreme Court, Sixth Edition. Washington, D.C.: CQ
Press.
Beier, Martin D. 1990. “Economics Awry: Using Access Fees for Caseload Diversion,”
University of Pennsylvania Law Review 138: 1175-1207.
Boucher, Robert L., and Jeffrey A. Segal. 1995. “Supreme Court Justices as Strategic
Decision-Makers: Aggressive Grants and Defensive Denials on the Vinson
Court,” Journal of Politics 57: 824.
Brady, Henry E., and Pual M. Sniderman. 1985. “Attitude Attribution: A Group Basis
for Political Reasoning,” American Political Science Review 79: 1061-1078.
Brenner, Saul and John F. Krol. 1989. “Strategies in Certiorari Voting on the United
States Supreme Court,” Journal of Politics 51: 828.
Brenner, Saul and Mark Stier. 1996. “Retesting Segal and Spaeth’s Stare Decisis
Model,” American Journal of Political Science 40: 1036-1048.
Caldeira, Gregory A. and James L. Gibson. 1992. “The Etiology of Public Support for
the Supreme Court,” American Journal of Political Science 36: 635-664.
Caldeira, Gregory A., and John R. Wright. 1988. “Organized Interests and Agenda
Setting in the U.S. Supreme Court,” American Political Science Review 82: 1109.
Caldeira, Gregory A., and John R. Wright. 1990. “The Discuss List: Agenda Building
in the Supreme Court,” Law and Society Review 24: 807.

155

Caldeira, Gregory A., John R. Wright, and Christopher J.W. Zorn. 1999. “Sophisticated
Voting and Gate-Keeping in the Supreme Court,” Journal of Law, Economics,
and Organization 15: 549.
Cameron, Charles M., Jeffrey A. Segal, and Donald Songer. 2000. “Strategic Auditing
in a Political Hierarchy: An Informational Model of the Supreme Court’s
Certiorari Decisions,” American Political Science Review 94: 101-116.
Canon, Bradley C. and Michael Giles. 1972. “Recurring Litigants: Federal Agencies
Before the Supreme Court,” Western Political Quarterly 25: 183-91.
Caplan, Lincoln. 1987. The Tenth Justice. New York, NY: Knopf.
Casper, Gerhard and Richard A. Posner. 1976. The Workload of the Supreme Court.
Chicago, IL: American Bar Foundation.
Cordray, Margaret Meriwether and Richard Cordray. 2001. “The Supreme Court’s
Plenary Docket,” Washington and Lee Law Review 58: 737-794.
Dahl, Robert. 1957. “Decision-Making in a Democracy: The Supreme Court as a
National Policy Maker,” Journal of Public Law, 6: 279.
Ely, John Hart. 1980. Democracy and Distrust: A Theory of Judicial Review. Harvard
University Press.
Epstein, Lee and Jack Knight. 1998. The Choices Justices Make. Washington, D.C.:
Congressional Quarterly, Inc.
Epstein, Lee, Jack Knight, and Andrew Martin. 2001. “The Supreme Court as a
Strategic National Policymaker,” Emory Law Journal 50: 593.
Epstein, Lee, Jeffrey A. Segal, Harold J. Spaeth, and Thomas G. Walker. 2003. The
Supreme Court Compendium: Data, Decisions, and Developments, 3rd Edition.
Washington, D.C.: Congressional Quarterly, Inc.
Feldman, Stephen M. 1985. “Indigents in the Federal Courts: The Informa Pauperis
Statute--Equality and Frivolity,” Fordham Law Review 54: 413.
Fodacaro, Mark R. 1995. “Toward a Synthesis of Law and Social Science: Due Process
and Procedural Justice in the Context of National Health Care Reform,” Denver
University Law Review 72: 303.
Frankfurter, Felix and James M. Landis. 1928. “The Supreme Court Under the Judiciary
Act of 1925,” Harvard Law Review 42: 1.
156

Friedman, Barry. 1993. “Dialogue and Judicial Review,” Michigan Law Review, 91:
577.
Galanter, Marc. 1974. “Why the ‘Haves’ Come Out Ahead: Speculations on the Limits
of Social Change,” Law & Society Review 9: 95-160.
George, Tracey E. and Michael E. Solimine. 2001. “Supreme Court Monitoring of the
United States Court of Appeals En Banc,” Supreme Court Economic Review 9:
171-204.
Gibson, James L. 1989. “Understandings of Justice: Institutional Legitimacy, Procedural
Justice, and Political Tolerance,” Law & Society Review 23: 469-496.
Ginsburg, Ruth Bader. 1994. “Remarks for American Law Institute Annual Dinner May
19, 1994,” Saint Louis University Law Journal 38: 881.
Hedayat, A.S. and Bikas K. Sinha. 1991. Design and Inference in Finite Population
Sampling. New York, NY: John Wiley & Sons, Inc.
Hellman, Arthur. 1983. “Error Correction, Lawmaking, and the Supreme Court’s
Exercise of Discretionary Review,” University of Pittsburgh Law Review 44: 795873.
Hellman, Arthur. 1984. “Granted, Vacated, and Remanded: Shedding Light on a Dark
Corner of Supreme Court Practice,” Judicature 67: 392-95.
Hellman, Arthur. 1985. “Case Selection in the Burger Court: A Preliminary Inquiry,”
Notre Dame Law Review 60: 947.
Hellman, Arthur D. 2001. “Never the Same River Twice: The Empirics and
Epistemology of Intercircuit Conflicts,” University of Pittsburgh Law Review 63:
81-157.
Issacharoff, Samuel. 2001. “Political Judgments,” University of Chicago Law Review
(Symposium: Bush v. Gore), 68: 637.
Jucewicz, Joseph and Lawrence Baum. 1990. “Workload Influences on Supreme Court
Case Acceptance Rates, 1975-1984,” The Western Political Quarterly 43: 123135.
Kahneman, Daniel, P. Slovic and Amos Tversky, eds. 1982. Judgment Under
Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press.

157

King, Gary and Langche Zeng. 2001. “Logistic Regression in Rare Events Data,”
Political Analysis 9: 137.
Kuklinski, James H. and Norman L. Hurley. 1994. “On Hearing and Interpreting
Political Messages: A Cuationary Tale of Citizen Cue-Taking,” The Journal of
Politics 56: 729-751.
Kuklinski, James H., Paul J. Quirk, Jennifer Jerit, and Robert F. Rich. 2001. “The
Political Environment and Citizen Competence,” American Journal of Political
Science 45: 410-424.
Lane, Cristina. 2003. “Pay Up or Shut Up: The Supreme Court’s Prospective Denial of
In Forma Pauperis Petitions,” Northwestern University Law Review 98: 335-366.
Lau, Richard R. and David P. Redlawsk. 2001. “Advantages and Disadvantages of
Cognitive Heuristics in Political Decision Making,” American Journal of Political
Science 45: 951-971.
Lawless, Robert M. and Dylan Lager Murray. 1997. “An Empirical Analysis of
Bankruptcy Certiorari,” Missouri Law Review 62: 101.
Lazarus, Edward. 1998. Closed Chambers: The Rise, Fall, and Future of the Modern
Supreme Court. New York, N.Y.: Penguin Putnam, Inc.
Long, J. Scott. 1997. Regression Models for Categorical and Limited Dependent
Variables. Thousand Oaks, CA: SAGE Publications, Inc.
Maguire, John MacArthur. 1923. “Poverty and Civil Litigation,” Harvard Law Review
36: 361.
Manstead, Antony S.R. and Miles Hewstone, eds. 1995. The Blackwell Encyclopedia of
Social Psychology. Cambridge, Mass.: Basil Blackwell Inc.
Mason, Alpheus Thomas. 1956. Harlan Fiske Stone: Pillar of the Law. New York,
N.Y.: The Viking Press.
Mauro, Tony. 1998. “Pauper Petitions a Long Shot,” USA Today, Wednesday,
December 23, 1998, p. 10A.
McGuire, Kevin T. 1993. The Supreme Court Bar: Legal Elites in the Washington
Community. Charlottesville, VA: University Press of Virginia.
McGuire, Kevin T. 1998. “Explaining Executive Success in the U.S. Supreme Court,”
Political Research Quarterly 51: 501-526.
158

McGuire, Kevin T. and Gregory A. Caldeira. 1993. “Lawyers, Organized Interests, and
the Law of Obscenity: Agenda Setting in the Supreme Court,” American Political
Science Review 87: 717-726.
Mishler, William and Reginald Sheehan. 1996. “Public Opinion, the Attitudinal Model,
and Supreme Court Decision Making: A Micro-Analytic Perspective,” Journal of
Politics, 58: 169.
New York University Supreme Court Project. 1984. New York University Law Review
59: 823-1003, 1403-1929.
Nicholas, Eric J.R. 1988. “Note, Preserving Pro se Representation in an Age of Rule 11
Sanctions,” Texas Law Review 67: 351.
O’Brien, David. 1993. Storm Center, 3rd Edition. New York, N.Y.: W.W. Norton.
Palmer, Barbara. 1999. “Issue Fluidity and Agenda Setting on the Warren Court,”
Political Research Quarterly 52: 39-65.
Palmer, Barbara. 2001. “The ‘Bermuda Triangle?’: The Cert Pool and Its Influence
Over the Supreme Court’s Agenda,” Constitutional Commentary 18: 105-120.
Perry, H.W. Jr. 1991a. Deciding to Decide: Agenda Setting in the United States
Supreme Court. Cambridge, MA: Harvard University Press.
Perry, H.W. Jr. 1991b. “Agenda Setting and Case Selection,” in John B. Gates and
Charles A. Johnson, eds., The American Courts: A Critical Assessment,
Washington, D.C.: Congressional Quarterly, Inc., pp. 235-253.
Provine, D. Marie. 1980. Case Selection in the United States Supreme Court. Chicago,
IL: University of Chicago Press.
Provine, D. Marie. 1981. “Deciding What to Decide: How the Supreme Court Sets Its
Agenda,” Judicature 64: 320.
Rahn, Wendy M. 1993. “The Role of Partisan Sterotypes in Information Processing
about Political Candidates,” American Journal of Political Science 37: 472-496.
Reimann, Glenn W. 1999. “Sour Grapes or Sound Criticism: Is the Supreme Court
Really Not Taking Enough Non-Tax Business Cases?,” University of Miami
Business Law Review 8: 161.
Riggle, Ellen D., Victor Ottati, Robert S. Wyer, James Kuklinski, and Norbert Schwartz.
1992. “Bases of Political Judgments: The Role of Sterotypic and Nonsterotypic
Information,” Political Behavior 14: 67-87.
159

Rogers, James R. 2001. “Information and Judiical Review: A Signaling Game of
Legislative-Judicial Interaction,” American Journal of Political Science 45: 8499.
Rohde, David, and Harold Spaeth. 1976. Supreme Court Decisionmaking. San
Francisco, CA: Freeman.
Salokar, Rebecca Mae. 1992. The Solicitor General: The Politics of Law.
Philadelphia, PA: Temple University Press.
Schubert, Glendon. 1958. “The Study of Judicial Decision-Making as an Aspect of
Political Behavior,” American Political Science Review 52: 1007-1025.
Schubert, Glendon. 1962. “Policy Without Law: An Extension of the Certiorari Game,”
Stanford Law Review 14: 284.
Schubert, Glendon. 1965. Judicial Policy-Making. Glenview, IL: Scott, Foresman.
Segal, Jeffrey A. 1988. “Amicus Curiae Briefs by the Solicitor General During the
Warren and Burger Courts,” Western Political Quarterly 41: 135-44.
Segal, Jeffrey A., Lee Epstein, Charles M. Cameron, and Harold J. Spaeth. 1995.
“Ideological Values and the Votes of U.S. Supreme Court Justices Revisited,”
Journal of Politics 57: 812.
Segal, Jeffry and Harold Spaeth. 1993. The Supreme Court and the Attitudinal Model.
New York, N.Y.: Cambridge University Press.
Segal, Jeffrey and Harold Spaeth. 1996a. “The Influence of Stare Decisis on the Votes
of the United States Supreme Court Justices,” American Journal of Political
Science 40: 971-1003.
Segal, Jeffrey and Harold Spaeth. 1996b. “Norms, Dragons and Stare Decisis: A
Response,” American Journal of Political Science 40: 1063-1082.
Smith, Kevin. 1999. “Justice for All?: The Supreme Court’s Denial of Pro Se Petitions
for Certiorari,” Albany Law Review 63: 381-425.
Songer, Donald R. 1979. “Concern for Policy Outputs as a Cue for Supreme Court
Decisions on Certiorari,” The Journal of Politics 41: 1185-1194.
Songer, Donald R. and Stefanie A. Lindquist. 1996. “Not the Whole Story: The Impact
of Justices’ Values on Supreme Court Decision Making” American Journal of
Political Science 40: 1055-1058.
160

Spaeth, Harold J. 1999. United States Supreme Court Judicial Database, 1953-1997
Terms [Computer File], 9th ICPSR Version. East Lansing, MI: Michigan State
University, Department of Political Science [producer]. Ann Arbor, MI: Interuniversity Consortium for Political and Social Research [distributor].
Spriggs, James F. and Paul J. Wahlbeck. 1997. “Amicus Curiae and the Role of
Information on the Supreme Court,” Political Research Quarterly 50:365-386.
Stern, Robert L., Eugene Gressman, Stephen M. Shapiro, and Kenneth S. Geller. 2002.
Supreme Court Practice, 8th Edition. Washington, D.C.: The Bureau of National
Affairs, Inc.
Taft, William Howard. 1925. “The Jurisdiction of the Supreme Court Under the Act of
February 13, 1925,” Yale Law Journal 35: 1.
Tanenhaus, Joseph, Marvin Schick, Matthew Muraskin, and Daniel Rosen. 1963. “The
Supreme Court’s Certiorari Jurisdiction: Cue Theory,” in Glendon A. Schubert,
ed., Judicial Decision-Making, New York, NY: Free Press, pp. 111-132.
Teger, Stuart H. and Douglas Kosinski. 1980. “The Cue Theory of Supreme Court
Certiorari Jurisdiction: A Reconsideration,” The Journal of Politics 42: 834-846.
Tryfos, Peter. 1996. Sampling Methods for Applied Research: Text and Cases. New
York, NY: John Wiley & Sons, Inc.
Ulmer, S. Sidney. 1972. “The Decision to Grant Certiorari as an Indicator to the
Decision on the Merits,” Polity 4: 429.
Ulmer, S. Sidney. 1983. “Conflict with Supreme Court Precedent and the Granting of
Plenary Review,” The Journal of Politics 45: 474-478.
Ulmer, S. Sidney. 1984. “The Supreme Court’s Certiorari Decisions: Conflict as a
Predictive Variable,” The American Political Science Review 78: 901-911.
Yingling, Kevin L. 1999. “Note: Justifying the Judiciary,” Journal of Law and Politics,
15: 81.
Yoo, John C. 2001. “In Defense of the Court’s Legitimacy,” University of Chicago Law
Review (Symposium: Bush v. Gore), 68: 775.

161