University of Utah

University of Utah Judicial Review by the Burger and Rehnquist Courts: Explaining Justices' Responses to Constitutional Challenges Author(s): Stefani...
Author: Annabelle Hall
3 downloads 3 Views 4MB Size
University of Utah

Judicial Review by the Burger and Rehnquist Courts: Explaining Justices' Responses to Constitutional Challenges Author(s): Stefanie A. Lindquist and Rorie Spill Solberg Source: Political Research Quarterly, Vol. 60, No. 1 (Mar., 2007), pp. 71-90 Published by: Sage Publications, Inc. on behalf of the University of Utah Stable URL: http://www.jstor.org/stable/4623808 Accessed: 16/07/2009 20:07 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=sage. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

University of Utah and Sage Publications, Inc. are collaborating with JSTOR to digitize, preserve and extend access to Political Research Quarterly.

http://www.jstor.org

Judicial Review by the Burger and Rehnquist Courts

Political Research Quarterly Volume60 Number 1 March2007 71-90 ? 2007 University of Utah 10.1177/1065912906298579 http://prq.sagepub.com hosted at http://online,sagepub.com

Explaining Justices' Responses to Constitutional Challenges Stefanie A. Lindquist

VanderbiltUniversity,Nashville, Tennessee

Rorie Spill Solberg OregonState University,Corvallis In thisarticle,theauthorsassessvariousinfluenceson U.S. SupremeCourtjustices'behaviorin casesinvolvingjudicial reviewof federal,state,andlocal statutes.Focusingon challengesto the constitutionality of statutesconsidered by the BurgerandRehnquistCourtsduringthe 1969to 2000 terms,the authorsevaluatethe impactof attitudinal, andcontextualvariableson individual institutional, justices'votesto strikeor upholdstatuteschallengedbeforethe Court.The authorsfindthatthejustices'ideologicalresponsesto the challengedstatutes,the extentof amicussupportfor the statute,the supportof the solicitorgeneral,congressional preferences,andthe existenceof a civil liberties challengeto the statuteareall significantly relatedto thejustices'votesto invalidateorupholdstatutes.Theyalso findthatin the RehnquistCourt,conservative justicesareless likelyto strikestatestatutesbutmorelikelyto strike dimensionemergesin the BurgerCourt. federallaws thantheirliberalcounterparts, while no similar"federalism" thanliberaljustices Indeed,in theBurgerCourt,a distinctpatternemergeswithconservative justicesmorerestraintist in bothstateandfederalcases. Keywords:judicial review;federalism; constitutionalchallenges; U.S. SupremeCourt

1. Introduction Scholars of judicial politics and law have demonstrateda continuingfascinationwith the "countermajoritarian difficulty" posed by the institutional position and characteristics of the U.S. Supreme Court within American democracy (see Friedman 2002). Since the Court's decision in Marbury v. Madison (1803),' many academics and politicians have lamented the tension between democraticgovernanceand an unelectedjudiciary'sexercise of judicial review. As Friedman (2002) pointed out, criticism of the Court's exercise of judicial review has emerged both from the Right and the Left, dependingon the historicalcontext. Recent decisions by the Rehnquist Court striking down progressive legislation like the Violence againstWomenAct,2for example, have drawn criticism from liberals; the Warren Court's decisions in the area of criminal rights and civil liberties drew the wrathof many conservatives. Without question, the Court's authorityto review the constitutionality of legislative enactments is

among its most important powers. A decision to invalidatea federal statuteoften raises issues related to separationof powers and democratictheory (see, e.g., Dahl 1957; Bickel 1986; Ely 1980; Choper 1980). On the otherhand, when the Courtinvalidates state or local legislation underthe U.S. Constitution, its decision implicates issues of federalism and the scope of state sovereigntyand autonomy.Thus, decisions involving constitutionalchallenges to federal, state, or local laws are often among the most politically sensitive and consequentialacts the Court can undertakeand continueto provide fuel for the debate regardingthe properrole of the SupremeCourt and the appropriateapplicationof its power. While much has been theorized regarding the implications and consequences of the use of judicial review, less is known about the likelihood of the applicationof this considerablepowerby the individual justices. What factors lead the justices to eschew deference and strike a law promulgatedby a duly elected body? What factors motivate the justices to restraintheirpersonalpreferencesandbow to the will of a majority?In this article, we seek to evaluatethe 71

72 PoliticalResearchQuarterly

influences that affect the individualjustices' choices to strikeor uphold federal, state, or local statutesand ordinances.We explorethe individualjustices'responses to such constitutionalchallengesbroughtin 238 cases during fifteen terms of the Rehnquist Court (19862000) and 458 such constitutionalchallenges brought during the sixteen terms of the BurgerCourt (19691985). We also compare the manner in which the Burgerand RehnquistCourtshave exercised this significant power. In particular,we constructa model of the justices' voting behaviorthat draws upon theoreticalperspectives stemming from attitudinal,strategic,and neoinstitutional models of judicial behavior.In doing so, we address several researchquestions. First, how do the justices' attitudesinfluence theirvotes to strikeor uphold federal and state legislation? Do their attitudes toward states' rights and federalism modify or condition their ideological reactions to these laws, and does any such dynamic vary between the Rehnquistand BurgerCourts?Second, do thejustices demonstrate any strategic or institutional concerns with respect to the position of the solicitor generalon the statute in question or to congressional preferences? Similarly,are the justices responsiveto interest group pressure in these cases? Finally, do legal standardssystematically influence the justices' decisions to strike legislation? In the following sections, we explainthe theoretical underpinningsof our study, present our hypotheses, and specify and estimate a model to test the hypothesized relationships empirically. The results of our study demonstrate,first, that decision makingin judicial review cases in the Burgerand RehnquistCourts is largely shapedby the same forces, indicatingsome enduringcharacteristicsassociated with these unique disputes. We also find that the membersof both the Burger and Rehnquist Courts are responsive to a numberof differentfactorswhen assessing the constitutionality of legislative enactments,including their own ideological predispositionstoward the substantive policy embedded in the statute, congressional preferences, the solicitor general's position on the statute, interest group pressurein the form of briefs amicus curiae, and the natureof the legal challenge broughtagainst the statuteat issue. The results from both Courts suggest that liberals are more likely to strikestate statutesthanare conservatives,but only on the Rehnquist Court are conservatives more likely than liberals to strike federal statutes.In contrast,in the BurgerCourt,conservativeswere more restrained with respect to both federaland state statutes.

2. Theoryand Hypotheses Political scientists who have studied the Supreme Court's exercise of judicial review have typically focused on the question whether the Court does, fashionto protect indeed, act in a countermajoritarian the rightsof minorities(e.g., Dahl 1957;Casper1976). In the seminalstudyin this tradition,Dahl (1957) concluded that the SupremeCourt, as a member of the "national ruling coalition," generally exercises its power of judicial review in ways that conformto the dominantcoalition'spolicy agenda.Thus, ratherthan protectingminorityrights againstinfringementby the majority,Dahl arguedthat the Courtis typically supportiveof the policies of otherpolitical institutionsat the time the decision is rendered-largely because the Court'smembershiphas been at least partiallyformed by the dominantcoalition.As evidence for this thesis, Dahl presentedfindingsthatthe bulkof cases in which federal legislation was declared unconstitutional occurredmorethanfour years afterthe legislationwas enactedand suggestedthat the Courtwas reluctantto invalidatelegislationenacted by the "live"or current nationalmajority.Dahl also presentedevidence thatof those federalstatutesdeclaredunconstitutionalby the Court within four years of passage, Congress often respondedby reversingthe Court'sdecisions through legislation or constitutionalamendment.As a result, Dahl concludedthatthe Court,throughits exercise of judicial review, does not play a particularlycountermajoritarianrole. Other scholars have disputed Dahl's (1957) findings, especially in light of post-1957 decisions by the Court. According to Casper (1976), for example, Dahl's characterizationof the Court fails to account fully for decisions of the WarrenCourtstrikingdown state or local legislation,recognizing that such decisions often had far-reachingnational consequences. Others have used Dahl's work as a springboardto evaluatethe influence of the Court'sexercise of judicial review on the process of partisan polarization priorto criticalelectoralrealignments(Funston1975; Gates 1984, 1987). In general, these scholars have arguedthat,farfromservinga "legitimacy-conferring" function for majoritypolicies, the Court often contributes to political realignment through decisions that destabilize the majority coalition (Adamany 1973; Gates 1987; but see Canon and Ulmer 1976). On the other hand, researchers have also recently demonstratedthat while most Supreme Court decisions invalidatingfederal statutesare not reversedby Congress, determinedand united majoritiesare often

Lindquist,Solberg/ JudicialReview by the Burgerand RehnquistCourts 73

quiteeffective at reversingor circumventingindividual decisionsfromtimeto time(MeernikandIgnagni1997). Basedon theirfindingsregardingcongressionalattempts and successes at reversing the Court, Meernik and Ignagni(1997, 464) concluded,for example, that"the constitutionalsystem of checks and balances is keeping the Supreme Court from dominating the other branches through the use of judicial review." This conclusion is also consistent with the notion that the SupremeCourt often rendersdecisions that are consistent with public opinion (Marshall 1989; Barnum 1985; Flemming andWood 1997; Stimson,Mackuen, and Erikson 1995). This debate among political scientists reflects the importanceof the Court'srole in the nationalpolicymaking process through its authority to invalidate federal, state, and local statutes.3Yet existing studies' focus on the broaderpolitical or theoreticalimplications of the Court's judicial review decisions have often ignored the micro-level decision-making processes involved in those decisions. Recently,however, several studies have modeled the individualjustices' votes in judicial review cases. Segal and Spaeth (2002) explored the natureof the individualjustices' choices to invalidate actions of federal, state, and local authorities,4finding that, for the Rehnquist Court,the justices' votes to invalidateor upholdwere dominatedby ideological considerations,with liberal justices voting to strike conservative statutes, and vice versa for conservative justices (pp. 415-16). Howardand Segal (2004) followed this analysis with an evaluation of the justices' responses to requests from litigants to declare laws unconstitutional.To identify litigant requestsfor the Courtto exercise its power of judicial review, Howardand Segal contentanalyzed briefs filed in all cases decided between 1985 and 1994. They found that the Court does not declarelaws unconstitutionalsua sponte, nor does the Courtutilize this importantpower often. Finally,Sala and Spriggs (2004) considered the influence of congressionalpreferenceson the justices' attitudinalvoting behavior in challenges to federal legislation by separatingthe period between 1946 and 1999 into differentregimes depending on the likelihood of the best legislative response to the Court's decisions. They found that, in those regimes where the justices are more likely to vote strategically,attitudinalconsiderationsstill dominatetheir voting behavior. In this study,we build on this existing researchby constructinga model that incorporatesa number of potentialinfluences on the justices' behavior,including variables drawn from the attitudinalmodel and

from theoreticalperspectiveshighlightinginstitutional andcontextualfactorsthataffectthejustices'decisions. We do so in the context of challenges to both federal and state laws, thus distinguishing our study from existing research.We test our model using decisions involving the potential exercise of judicial review in the Burger Court (1969 to 1985 terms) and the RehnquistCourt(1986 to 2000 terms).Below, we set forth these theoreticalexpectationsregardingthe justices' decision making in cases involving challenges to federal,state, and local statutes.

A. The Justices'Ideology We begin with the dominant model of Supreme Courtdecision makingin the literature:the attitudinal model. Withoutquestion,the justices' votes are often substantiallyinfluenced by their ideological predispositions (Segal and Spaeth 2002). Especially given Segal and Spaeth's (2002) recent investigation of judicial review cases described above, we therefore expect that the justices' votes to uphold or invalidate federal,state, or local legislation will be substantially influencedby theirattitudinalresponses to the policy positions furtheredby those statutes (Howard and Segal 2004). For example, we expect that, when considering a constitutional challenge to a statute that restricts abortion rights for women, liberal justices will be more likely to vote to invalidatethe law, and conservativejustices will be more likely to uphold the law. This attitudinalexpectationthus involves the justices' reactionsto the substantivepolicy embedded in the statuteat issue; where the justice's ideology is consistent with the policy outcome furtheredby the statute,he or she will be less likely to vote in invalidate the law. Hence, our first hypothesis is, Hypothesis 1: When considering the constitutionality of federal, state, or local laws, the justices' votes will be influenced by the consistency between the directionof the statuteand the justices' ideological preferences. While the literaturegenerallycharacterizesthe justices' attitudinalpredispositionsalong a single liberalconservative ideological dimension, other case characteristicsmay stimulatethe justices to respond to countervailingconsiderationsbased on their attitudes towardfederalismand states' rights (for a thorough analysis of the multidimensionalnatureof the justices' attitudes, see Rohde and Spaeth 1976). Althoughthe BurgerCourtis known for its more balanced treatmentof congressionalpower vis-i-vis the

74

Political ResearchQuarterly

states (see, e.g., E. M. Martin 1985), one of the RehnquistCourt'smost distinguishingcharacteristics is its preoccupationwith federalismand states'rights (Fallon 2002; Massey 2002; Whittington2001). Most of the Court's pro-state-rights decisions may be attributedto the influence of the conservativejustices on the Court. Indeed, many of the most recent decisions involving the Eleventh Amendment and state sovereignimmunityhave been renderedby minimumwinning coalitions of conservativejustices including Justices Rehnquist, Scalia, Thomas, Kennedy, and O'Connor.These cases illustratethe critical ideological divide within the Rehnquist Court concerning states' rights and the proper scope of federal power. Although we have already hypothesized that the statute's ideological direction (i.e., its substantive policy impact)will influencethejustices'votes injudicial review cases, we also expect that, at least in the RehnquistCourt, ideological differences among the justices may also manifest themselves in relation to the source of the statute. That is, the justices' attitudes toward federalism may accentuateor mitigate the influence of their policy-specific preferences. For example, while we expect that liberal justices will be predisposed to strike conservative statutes, we also expect that liberals will be less inclined to strike federal as opposed to state statutes. Similarly, we expect that conservative justices will be more likely to strike federal as opposed to state statutes, especially in the RehnquistCourt.Thus, we offer our second hypothesis: Hypothesis 2: The justices' ideological responses to the individual statutes will also be conditioned on the source of the legislation in question, with conservativejustices more likely to strike federal statutes and liberaljustices more likely to strike state statutes.

B. Strategicand InstitutionalInfluences In addition to attitudinalfactors, recent research has focused scholars'attentionon the extent to which the justices act strategicallyin response to their institutionalenvironmentand to the anticipatedactions of other political actors (Epstein and Knight 1998; Maltzman, Spriggs, and Wahlbeck 2000; Heberlig and Spill 2000). As Epstein and Knight (1997, 4) observed, the justices "realize that their ability to achieve their goals depends on a considerationof the preferences of others, of the choices they expect others to make, and of the institutional context in

which they act."In the context of judicial review, the justices may be particularlysensitive to the preferences of those actors in the coordinate branches in Congress and the executive branch, as well as to public opinion. While one might assume that the Court would be most likely to engage in strategic interactionswith the coordinatebranches(especially Congress) in statutorycases-where the legislature may most easily alter the Court's outcome-the Courtis not unconstrainedin constitutionalcases. As Friedmanand Harvey (2003, 127) suggested, "The sharpdistinctionbetween constitutionaland statutory cases is flawed . . . [as] [t]herearenumerousweapons a sitting Congresscan apply againsta SupremeCourt deemed to be recalcitrant, including jurisdiction stripping,budgetcutting,Courtpacking,and even the impeachmentof Supreme Court Justices."Although these threats are not often realistic, in combination with action by other political actors, they may shape Courtbehaviorat some level. For example, the executive branch may also participate or initiate such retaliatory action should the president find the Court'sactions unacceptable.Moreover,public opinion and interestgrouppressuremay similarlyprovide importantconstraintson the Court'schoices to invalidate legislation to the extent such public sentiment may circumscribe or undermine implementationof the Court's decisions (see Rosenberg 1991), as well as provide the impetus for a political response by Congress and the president.Thus, our model of the justices' votes in judicial review cases should account for the influence of these institutional or external constraintson the Court'sdecision making. In constructingourhypotheses,then,we begin with the executive branch.As its representative,the solicitor generalhas a remarkablerecordof success before the Courtat both the certiorariand merits stages (see Segal 1990; CaldeiraandWright 1988; Pacelle 2003). This record reflects the good working relationship between the office and the justices, the significant deference the Court tends to afford the opinion of the executivebranch,as well as the lawyers'expertise in the solicitor's office (McGuire 1995, 1998). Moreover,the solicitor'srecordof success extends to all types of participation,whetheras counsel or amicus curiae.Therefore,we anticipatethat the presence of the solicitor general, either as a counsel or as an amicus participant,will greatlyinfluence the votes of the individualjustices. If the solicitorgeneralappeals or intervenes and supportsa challenged statute, we anticipatethe law will survive the challenge. On the otherhand,if the solicitoropposes a statute,we expect

Lindquist,Solberg / JudicialReview by the Burgerand RehnquistCourts 75

that the justices will more likely vote to invalidate the statute. Hypothesis 3: Support for a law by the solicitor general, as partyor amicus curiae,increasesthe likelihood of that the justices will vote to uphold the law; opposition by the solicitor increases the likelihood that the justices will vote to strikethe law. The extent to which the Supreme Court defers to the legislative branch is less clear; the evidence is mixed (see, e.g., Segal 1999; Segal and Spaeth2002; Epstein and Knight 1998; Sala and Spriggs 2004). As noted above, most such studies focus on statutory interpretationby the Court, on the assumption that the Court is more constrained strategicallyby congressional action because Congress has greaterretaliatory power in such cases (it can amend legislation with a simple majority and the president's acquiescence). With respect to the invalidationof a statute, however,cross-institutionalinfluencesare also potentially important. Unlike straightforwardstatutory interpretation,judicial review results in a more direct rebuke to lawmakers. Moreover, Meernik and Ignagni (1997) have demonstratedthat Congress can and often does respondto the SupremeCourtby trying to reversepolitically unpopulardecisions involving the exercise of the Court's power of judicial review (see also Hausegger and Baum 1999). Meernik and Ignagni furtherfound that this "retaliatory" response is particularlypronounced in situations in which the Court invalidatesa state statute,a Furthermore, findingthatwould seem counterintuitive. as Epstein and Knight (1998, 141) observed,even if the legislative constraintis theoreticallymore salient in statutorycases, we should not "ignorecompletely the external constraintimposed by the separationof powers in constitutionalcases." Their review of the justices' papers and correspondence provided evidence that the "the externalconstraintof the separation of powers system is in fact operative in some constitutional cases" (p. 157). Finally, in a recent study, Friedman and Harvey (2003) found that the Courtwas sensitive to the ideological composition of the sitting Congress when choosing whetherto invalidate federal laws. Thus, the justices may be sensitive to such crossinstitutional pressure as they evaluate whether to strikea law challenged on its face. At the same time, suchresponsivebehaviormay not be completelystrategic in nature.As noted above, Dahl (1957) postulated

that the Court does not always act in a countermajoritarianfashion but more typically conforms to the preferences of the dominant political coalition. Moreover,researchon the direct influence of public opinion on the justices' behavior suggests that Court members are not immune from trends in the public mood-trends that are highly correlatedwith the policy preferencesof those in Congress. We note that some studies contradict the notion thatthe Courtrespondsto the ideological preferences of members of Congress. For example, Howard and Segal (2004, 135-36) found that the justices do not appearconstrainedin years 1993 and 1994, at which time Democrats controlled both the presidency and Congress. Sala and Spriggs (2004) drew similarconclusions in their analysis of particularregimes (configurationsof congressionalpreferences)most likely to generatestrategicjudicial voting behavior,finding no such influence. Segal and Spaeth(2002) and Segal (1999) also arguedthat separationof powers models do not adequately explain decision making on the Court in statutory cases. These studies generally characterizethe influence of Congress in strategic terms.While we recognize thatexisting researchcalls into questionstrategicaccountsof thejustices' behavior, we conceptualizethe influence of Congressmore broadlyto reflect both the preferencesof membersof Congress as well as general trendsin public opinion. This perspective comports with other research,such as thatof Neal Devins (2004), arguingthat,at least in the Rehnquist Court, the justices' decisions striking congressional statutes may be often viewed as ratherthan "countermajoritarian." We "majoritarian" propose that the justices may be more reticent to invalidate statutes that conform to prevailing congressional preferences,5 either because they make strategic calculations regarding those preferences, because the justices constitutemembersof the dominant political coalition (see Dahl 1957), or because they respondto public opinion: Hypothesis4: The justices are less likely to invalidate a statute that is ideologically congruent with prevailingcongressionalpreferences. Dahl's (1957) studydebatedthe extentto which the justices are willing to strikelegislationenactedby the dominant majority coalition, especially when that coalition remains in power at the time the statute is challenged. As a result, Dahl focused on the age of federal statutes challenged before the Court, finding thatthe Courtmost often invalidatedstatutesthatwere

76

Political ResearchQuarterly

more than four years old at the time of the Court's decision. Dahl concluded that the Court's deference to newer statutesis a functionof the judicial selection process.As new justices join the Court,they are more likely to sharethe values andattitudesof the dominant coalition in the elected branchesat the time of their appointment.Thus, as Dahl noted,membershipchange will drive the Court to favor more modem policies simply because the Court is often itself a part of the current political coalition. If Dahl's hypothesis is valid, then, we might expect the Courtto be similarly reluctantto invalidatestatutesenactedby more recent legislative majorities. In addition to reflecting Dahl's (1957) ideas, the age of the statutemay be relevantfor anotherreason related to legal considerations.To the extent that the Supreme Court's interpretationof the Constitution changes over time, which it surelydoes, older statutes may be morelikely to fall subjectto invalidationunder a new constitutional formulation created at a later date. Thus, statutoryage may be positively related to judicial review in at least two ways. On the other hand,one might also arguethatolder statutesaremore immune to constitutionalreview because they have "stood the test of time."This is consistent with findings regardingthe Court's interpretationsof its own precedents:initially,as precedentages, the probability of being positively interpreteddecreases until the precedentbecomes very old (seventy-nineyears), at which point positive interpretationsincrease (see Hansfordand Spriggs 2006, 73, n. 24). On balance, however,we conclude thatthe following hypothesisis most supportedby the existing literature: Hypothesis 5: As a statute or ordinanceages, the justices will be more likely to strikethe law. In addition to external influence from Congress and the solicitor,interestgroupinvolvementmay also serve as a source of informationto the justices concerning the public's reaction to Courtdecisions. The presence of amicus curiae, like that of the solicitor general, is a well-documented and common occurrence before the SupremeCourt.Amici briefs filed by interestgroups, researchsuggests, reveal the breadth and scope of the issue for the justices as well as provide a gauge of currentpublic opinion. At least one study has documentedthe influence of amicus curiae on the decision on the merits (McGuire 1995), althoughgenerally,interestgroupsparticipatingamicus curiae are most influential at the certioraristage (Caldeira and Wright 1988, 1990). At the merits

stage, the influence of amicus curiae could arise in several ways. First, the justices could be directly influenced by the arguments in the amicus briefs (Epstein and Kobylka 1992). Second, the justices may be influencedby the display of supportfor a particularposition in termsof the numberof briefs filed, or in terms of the prestige and reputation of the groups involved. Third, a selection effect may exist where more groupsjoin the party with the most persuasive argument.For our purposes, we rely on differential supportof amici for the partieschallenging or defending the statute. We therefore hypothesize that the law of numberswill influencethe perceptions and votes of the justices. If more briefs are filed on behalf of the statuteor ordinancethan opposing, the justices will be less prone to invalidatethe law. Hypothesis 6: Greater support by amicus curiae for the law decreases the likelihood of that the justices will vote to invalidatethe statute.

C. Legal Influencesand SelectionEffects We also expect that the natureof the legal issues before the Court will affect its decision making. Legal challenges to statutesor ordinancesbased upon a civil libertiesclaim are simply less likely to survive judicial scrutiny. Such cases are often examined under strict standardsof review, as opposed to the rationalbasis standardreservedfor non-civil liberties challenges based on the Commerce Clause. Under the strictestlevel of scrutiny,in particular,the burden of the governmentto defend its statuteis quitehighas some have claimed, it is "strictin theory, fatal in fact."As a result, more statuteschallenged as violative of citizens' civil liberties will fail to pass constitutional muster, even before a more conservative SupremeCourt. Hypothesis 7: When reviewing statutes or ordinances challenged on the bases of civil liberties violations, the justices will be more likely to strikethe law. Finally, we note that the judicial review cases heard by the Supreme Court do not arise on the Court'sdocket at random.Instead,the Court'scertiorari process is complex and often involves strategic calculations on the part of the justices (Boucher and Segal 1995). As Brace, Hall, and Langer (2001) noted, if court interventionis nonrandom,ignoring this selection process raises the likelihood that conclusions about the forces affecting subsequentvotes

Lindquist,Solberg/ JudicialReview by the Burgerand RehnquistCourts 77

to invalidatelegislation will be inaccurate(see, e.g., Heckman 1979). Although specifying a two-stage model would be optimal to account for this selection bias statistically,the data are not availableon the certiorariprocess to enable us to do so. Like Brace et al. (2001), therefore, we incorporateda covariate into our model to capture the dynamics of the selection process. In the Supreme Court, the nature of the lower courts' decisions in a case's proceduralhistory affects the likelihood that the Courtwill hearthe case (Caldeira,Wright,and Zorn 1999). In this context, in particular,we expect that the Court would be more likely to docket cases where the lower court had struck down a statute as unconstitutional.We have therefore incorporatedthis variable in an effort to control for the dynamics associated with the docketing process.

3. Data and Methods A. DependentVariable To evaluate these hypotheses, we identified all cases heardby the Burgerand RehnquistCourtsduring the 1969 to 2000 terms in which the constitutionality of a federal, state, or local statute was challenged. To select our cases, we first used the "uncon" variable in the U.S. Supreme Court Database (Spaeth 2003) to find those cases in which a statute or ordinance had been found unconstitutional. We then used the "authdec"variables in that databaseto identify all cases where there may have been a constitutional challenge to the statute and examined the opinion in each such case to determine whetherit should be included.We cross-checkedthis list against those cases in the Benesh and Spaeth (2003) Burgerand RehnquistCourtJustice-Centered Databasesin which a justice was recordedas having voted to strike a statute or ordinance. This process resulted in the identification of 458 cases in the BurgerCourtand 238 cases in the RehnquistCourtin which the Courtconsidereda constitutionalchallenge to a federal, state, or local law. Once we had identifiedthe cases involving constitutional challenges, we used the Justice-Centered Databases to construct our dependent variable. We relied on the Benesh and Spaeth databases and our own review of each case to identify whetherthe individual justices voted on the constitutionalityof the statuteat issue. In particular,we evaluatedconcurring and dissentingopinions for each justice to ensurethat he or she actuallyvoted on the constitutionalityof the

statute. This coding process resulted in a dichotomous dependent variable, coded as 1 if the justice voted to strikethe statuteand 0 if the justice voted to uphold the statute. Our data included 3,938 votes from the Burger Court and 1,992 votes from the RehnquistCourtfor analysis.

B. IndependentVariables Conformitybetweenstatuteandjustice ideology.We have hypothesizedthatthejustices'choices to strikeor upholda statutewill be influencedby their attitudinal predispositions.Justices with conservative attitudes should be more likely to strike"liberal"statutes,and liberaljustices shouldbe more likely to strike"conservative"statutes.Thus, we needed to devise a measure thatreflectsconsistencybetweenthe ideologicalorientationof thejustice andof the statuteitself. We began by assigning each justice an ideological score using the JudicialCommon Space Scores (JCS scores) developed by Epstein et al. (forthcoming), a measurementstrategy that has been used in existing research (Sala and Spriggs 2004). The values of the JCS scores range from -.674 to +.801. We switched the signs on these scores such that positive values reflected liberalism, negative values conservatism. Although the JCS scores are calculatedbased on the justices' voting behavior, and thus may raise endogeneity problemsfor some analyses, we are less concerned aboutthatproblemwith respectto our model. In particular,we are not predictingthe directionality of the justices' votes, but ratherwhetherthey voted to strike a challenged statute.Use of the JCS scores to constructan independentvariablethereforedoes not raise the same degree of concern as would be present if we were predictingthe ideological directionof the justices' votes. In addition, estimation of the same model using the Segal-Cover scores to measurejustice ideology-which are based on preappointment commentaryabout the justices' attitudesprimarilyin civil liberties cases-returned essentially the same substantive results (see Segal et al. 1995). (We include those results in the appendix.) Next, we assigned an ideological score to the individualstatutesat issue using the directionalitycodes in the U.S. SupremeCourtDatabase.Thus, if the majority invalidatedthe particularstatute,and the direction of the outcome was liberal, we coded the statute as "conservative,"and vice versa for liberalstatutes(see Segal and Spaeth 2002, chap. 10, using the same methodology6).Liberal statutes received a value on this statute ideology variable of I and conservative

78

Political ResearchQuarterly

statutes-1. To constructa variablemeasuringthe consistency between the statuteand the justices' ideologies, then, we simply multipliedthe statutoryideology variableby thejustices' JSC scores. Since those scores were coded with positive values reflectingliberalorientationsand negative values reflecting conservative orientations,the resultingfigurewill be positive when thereis ideological congruityand negativewhen there is not. Moreover,the value of the term will vary with the intensity of the justices' preferences. Indeed, extremeliberalsmay be expectedto reactmore vociferously to a conservativestatutethan a moderateliberal or moderate conservative; this independent variablewill reflect such variationamongthe intensity of the justices' attitudesalong the liberal-conservative dimension.The more ideological consistencybetween the justice and the statute,the more likely the justice will vote to uphold the statute.We thereforeexpect a negative coefficient on this variable,which we have labeled "Consistencybetween StatutoryDirectionand JusticeIdeology." Orientationtowardfederalism. In addition to the straightforwardattitudinalresponse we expect on the basis of the statutorypolicy, we also hypothesized a relationship between the justices' ideology and the source of the statute.According to the conventional wisdom, conservativejustices on the RehnquistCourt are attitudinallypredisposedto reduce federal power vis-a-vis the states and to enhance or protect state authority.We have no similar a priori expectations regardingthe Burger Court but neverthelesstest the relationshipbetween ideological and statutorysource for justices on that Court as well. Thus, we expect that the justices' reactions to the challenged statutes will vary depending on the source of the statute, as well as on the justices' ideological reaction to the underlyingstatutorypolicy. To evaluatethis hypothesis, we createda multiplicativetermcomposed of two variables:(1) whetherthe statutewas a state statuteor local ordinance (coded as 1) or a federal statute (coded as 0) and (2) the justices' JCS score. Interactingthese two variableswill enable us to evaluate whetherliberaljustices are more likely to strike state statutes and conservativejustices less likely to strike state statutes, while use of the inteff program developed by Norton, Wang, and Ai (2004) will enable us to determinewhetherthat interactioneffect is significance across all observations. Because we expect that liberal justices will be more likely to invalidatestate statutesand conservativejustices less so, we expect a positive coefficient on this variable.

Solicitorsupport.The solicitormay participatein a case either as counsel or as amicus curiae. In these judicial review cases, the solicitor'ssupportfor either petitioneror respondent,whether as counsel or amicus, reflected his supportor opposition to the statute at issue. For that reason, we created a variable to reflect whetherthe solicitor supportedor opposed the statute in his capacity either as amicus or counsel. Ultimately, there were very few cases in which the solicitoropposedthe statutewhile acting as partyrepresentative(where, for example, a state statute was challenged by the United States as violative of the Commerce Clause). As a result, we created three dichotomous variables that reflect whether (1) the solicitor supportedthe particularstatuteor ordinance when appearingas a directrepresentativeto a partyin the case (the U.S. government),(2) the solicitor supported the statute or ordinance when appearing as amicus in the case, or (3) the solicitor opposed the statuteor ordinancewhen appearingas amicus. The variableswere coded 1 if the solicitor's participation fell into one of the describedcategories,and 0 otherwise. Ultimately,we chose not to include the solicitor as Party variable in the Burger Court models (althoughit is included in the RehnquistCourt models) because it consistentlyfailed to achieve statistical significance and because it was collinear with our variablereflecting the source of the statute (and did not achieve significance even in the absence of the statutory source variable). Thus, we evaluated the impactof the solicitoron BurgerCourtdecision making only when he entereda case as amicus curiae. Congressionalpreferences.In our fourthhypothesis, we postulatedthat the justices would respond to congressional preferences. We conceptualized this variable as a measure of the ideology of Congress, underthe suppositionthat the SupremeCourtwould be less likely to strikea statutethat was ideologically consistent with the prevailingmajorityin the legislature. To measure the ideology of Congress, we used the median legislator's Judicial Common Space Score from each chamberfor the 91st through 106th Congresses(Epsteinet al., forthcoming).These scores are calculatedsuch thatnegative scores reflect liberal legislators and positive scores reflect conservative legislators.We calculatedthe mean of these ideological medians to create a measure of congressional preferences across both chambers.To construct our measureof the congruencebetween the statute'sideological direction and congressional preferences, as noted previously we first switched the signs on the

Lindquist,Solberg / Judicial Review by the Burgerand RehnquistCourts 79

Judicial Common Space scores such that positive scores were associated with liberal preferences.We then multiplied the JCS Congressional score corresponding to the Court'sterm by the directionof the statute(coded as 1 for liberal statutes,-1 for conservative statutes). Positive values on this measure thus reflect contemporaneous ideological congruence between the statuteand the sitting Congress,and negative values reflect a lack of congruence.We expect a negative coefficient on this variable. Age of statute.As we note above, if the justices are indeed part of the governing coalition as suggested by Dahl (1957), we would expect that statutoryage would be related to the justices' votes to invalidate legislation. To measure the age of each statute, we simply subtractedthe date of each statute'senactment from the Courtterm in which the case was decided.7 Because we expect the justices to be more reluctant to invalidatelaws from more recently elected legislative bodies, we expect a positive coefficient on this variable. Because of the extreme skewness in this variable'sdistribution,we employ the naturallog of this variablein our models.8 Other specifications of this variable,including quadraticformulations,were not significantin any of our models. Amicus support.We hypothesizedthat the justices would be responsiveto public opinion and potentially to interest group supportor opposition to statutesas expressedthroughamicuscuriaeparticipation.To measurethe influence of amici, we countedthe numberof amicus briefs (not signatories) filed in support or opposition to the statuteat issue. We did so by crossreferencingamicus data recordedin both Lexus and Westlawto ensurethatour countingwas accurate.We then subtractedthe numberof opposing briefs from the number of supportingbriefs in each case.9 We expect that the resulting variable will be negatively related to the likelihood that a justice will vote to invalidatethe statute. Civil liberties cases. Where the constitutional challenge raises civil liberties issues, we hypothesized that the justices would be more likely to strike the challenged statute.To identify those cases involving civil libertiesissues, we relied on the "issue"variable in the U.S. SupremeCourtDatabase,identifying all cases thatraisedconstitutionalchallengesrelatingto the FirstAmendment,Due Process, Equal Protection, Civil Rights, and Privacy (categories 210 to 537 in the database).All cases that fell in these categories

received a value of 1 on the civil liberties variable; otherwise the variablewas coded 0. Selection bias control.As we discussed above, we are not able to control for the actual dynamics of the certiorariselection process, as relevant data on the certiorari process are not readily available.'o However,we did constructa variablethat we believe captured some of those selection dynamics, in an effort to control partially for selection bias (see Brace, Hall, and Langer 2001). The variablelabeled "LowerCourtDisposition"is measuredto reflect the disposition of the court below"-if a lower court invalidatedthe statute (whether at the district court, circuit court, or state court), we coded the variableas 1, and 0 otherwise. We coded the lower court's disposition of the constitutional challenge from the SupremeCourt'sopinion. Statistical method. Because our dependent variable reflects the dichotomouschoice to invalidateor uphold a statuteon its face, we utilized logit to estimate our model; given the potentialthat the justices' individual votes would not be independent across cases, we clustered on the justice (see Long and Freese 2001).12 We presentthe resultsof our datacollection process and model estimationbelow.

4. Results Table 1 provides a breakdownof the characteristics of the cases in the database.Of the 458 constitutional challenges brought before the Burger Court, 199 resulted in the law's invalidation, at a rate of about 12 statutesstruckper termon average.In comparison,in the RehnquistCourt,of the 238 challenges brought, 109 resulted in the invalidationof the challenged statute,at a rate of about 7 statutesstruckper term.Moreover,the BurgerCourtwas more reticentto strikefederal statutes,invalidatingonly 21 percentof federal laws challengedbefore the Court.In contrast, the RehnquistCourtinvalidated41 percentof federal laws challenged from 1986 to 2000. When it came to state or local laws, however, the two courts were aboutequally likely to strikea statuteor ordinanceas violative of the federal constitution(50 percentin the Burger Court, 47 percent in the Rehnquist Court). Thus, the BurgerCourtwas more lenient towardfederal statutesbut, in general, evaluated the constitutionality of more laws than did the RehnquistCourt. On the other hand, the Rehnquist Court considered the constitutionality of fewer laws but was more

80 Political ResearchQuarterly

Table1 DescriptiveStatistics:CasesRaisingConstitutionalChallengesto Federal,State, and Local Statutes--Burgerand RehnquistCourtsCompared Variable Federallaw declaredunconstitutional?

State/local law declaredunconstitutional?

Origin of statutechallenged Age of federal laws challenged

Age of federal laws struck

Age of state/local laws challenged

Age of state/local laws struck

Position of solicitor general

Amicus brief differential(supporting amicus-opposingamicus) N

BurgerCourt, 1969-1985 Terms: Frequencies Yes = 23 No = 85 (Mean = .21) Yes = 176 No = 174 (Mean = .50) Federal= 108 State/local = 350 Mean = 18.61 years (Standarddeviation= 18.44) Median = 11.5 years Mean = 17.2 years (Standarddeviation= 16.7) Median = 9 years Mean = 14.20 years (Standarddeviation= 19.52) Median = 7 years Mean = 14.15 years (Standarddeviation= 19.17) Median = 6 Supportas amicus: 7.4 percent Oppose as amicus: 2.8 percent Maximum= 17 Minimum= -7 Mean = 0 458

activist in striking federal statutes. At least in part, these differences may be due less to the justices' propensityto strike statutes and more to changes in the docket between the two Courts. During the BurgerCourt,the SupremeCourt'sdocket was determined by Public Law 86-3 (28 U.S.C. ? 1252), which provided for direct appeals to the Supreme Court from decisions invalidatingActs of Congress. This law was repealed in 1988. As a result, more appeals from decisions involving federal laws may have been heard by the Supreme Court, although it seems unlikely that the Court would have denied certiorari in such cases in any event. Moreover,the Rehnquist Courthas consideredfewer cases each term than the Burger Court and has devoted a lesser portion of its docket to constitutionalcases as opposed to matters involving statutoryconstruction. The two courts also differ with respect to the age of the laws challenged and struck. Although the BurgerCourtwas subjectto a more expeditedprocess

RehnquistCourt, 1986-2000 Terms: Frequencies Yes = 33 No = 46 (Mean = .41) Yes = 76 No = 83 (Mean = .47) Federal= 79 State/local = 159 Mean = 13.92 years (Standarddeviation= 14.11) Median = 7 years Mean = 14.30 years (Standarddeviation= 15.82) Median = 7 years Mean = 17.46 years (Standarddeviation = 22.07) Median = 9 Mean = 16.25 (Standarddeviation= 21.48) Median = 7 Supportas party:27.31 percent Supportas amicus: 12.61 percent Oppose as amicus: 5.04 percent Maximum= 17 Minimum= -10 Mean = .029 238

regardingchallenges to federal laws (i.e., the process of directappealdescribedabove), it neverthelessconsidered somewhat older federal statutes than the Rehnquist Court. The median age of federal laws challenged in the RehnquistCourt was only 7 years, comparedto a 11.5-yearmedianage for statuteschallenged in the Burger Court, suggesting that litigants in recentyears have been fairly swift in bringingconstitutionalchallenges to new legislation, and that the Supreme Court has also been quick to take these appeals. The Burger Court had a slight tendency to strike newer federal laws, however. On the other hand, the RehnquistCourtevaluatedolder state laws than the Burger Court, on average, although both courts had a tendency to strike newer state laws. Thus, it appearsthatage is negativelyratherthanpositively associatedwith the likelihood of invalidation. Table2 and 3 presentsthe resultsof the logit models of the justices' voting behaviorin judicial review cases in the BurgerandRehnquistCourts.These tables

Lindquist,Solberg / JudicialReview by the Burger and RehnquistCourts 81

Table2 Logit Estimationof Justices'Votesto StrikeChallengedStatute/Ordinance, BurgerCourt, 1969-1985Terms(with HypothesizedDirectionof Coefficient) IndependentVariable Consistencybetween statutory directionandjustice ideology (-) Source of Statute(State) x Justice Ideology (+/-) State statute(+) Justice ideology (+/-) Solicitor supportfor statute(amicus) (-) Solicitor opposition to statute(amicus) (+) Consistency between statutedirection and congressionalpreferences(-) Age of statutein years (logged) (+) Amicus supportfor statute(differential)(-) Civil liberties challenge (+) Lower court disposition Constant Log-likelihood Pseudo R2 N

Coefficient/StandardError (Full Model) -1.74 (0.215)***

Coefficient/StandardError (FederalModel) -2.30 (0.231)***

Coefficient/StandardError (State Model) -2.21 (0.144)***

0.464 (0.305) 0.870 0.550 -0.362 -0.123 -1.65

(0.103)*** (0.260)** (0.120)** (0.153) (0.227)***

-1.55 (0.622)**

-0.339 (0.109)** - 0.113 (0.144) -1.79 (0.285)***

-0.047 (0.032) -0.050 (0.011)*** 0.224 (.079)** -0.448 (0.051)*** -0.693 (0.104)*** -2,272.99 .166 3,938

0.034 (0.041) 0.049 (0.034) 0.275 (0.175)t -0.877 (0.166)** -0.526 (0.195)** -504.81 .145 935

-0.060 (0.031) -0.069 (0.014)** 0.162 (0.095)* -0.347 (0.066)*** 0.174 (0.128) -1,795.52 .135 3,003

-

Note: All coefficients generated in Stata 8.2, with robust standarderrors, clustering on the justice. Model estimated using Judicial CommonSpace Scores. The federal model could not be estimatedwith the solicitor amicus variables,as such participationwas too small in number. p = .06. *p < .05. **p < .01. ***p < .001.

Table3 Logit Estimationof Justices'Votesto Strike ChallengedStatute/Ordinance, RehnquistCourt, 1986-1999Terms2(with HypothesizedDirectionof Coefficient) IndependentVariable Consistency between statutorydirection andjustice ideology (-) Source of Statute(State) x Justice Ideology (+/-) State statute(+) Justice ideology (+/-) Solicitor supportfor statute(party)(-) Solicitor supportfor statute(amicus) (-) Solicitor opposition to statute(amicus) (+) Consistency between statutedirectionand congressionalpreferences(-) Age of statutein years (logged) (+) Amicus supportfor statute(differential)(-) Civil liberties challenge (+) Lower court disposition Constant Log-likelihood Pseudo R2 N

Coefficient/StandardError (Full Model) -1.742 (0.141)***

Coefficient/StandardError (FederalModel)

Coefficient/StandardError (State Model)

-1.99 (0.263)***

-1.90 (0.144)***

0.829 (0.326)** -0.054 (0.190) -0.197 (0.330) -0.643 (0.163)*** -0.770 (0.156)*** 1.37 (0.213)*** -0.966 (0.419)** -.0039 (0.050) -0.032 (0.016)* 0.411 (0.134)*** -0.359 (0.076)*** -0.130 (0.203) -1,159.64 .134 1,992

-0.761 (0.183)*** -0.844 (0.505)*

-0.808 (0.146)*** 1.86 (0.192)*** -1.001 (0.447)**

-0.141 (0.074) 0.006 (0.032) 0.372 (0.132)** 0.176 (0.130) -0.096 (0.275) -382.03 .117 644

0.043 (0.061) -0.039 (0.011)*** 0.469 (0.222)* -0.547 (0.098)*** -0.096 (0.326) -803.36 .140 1,348

Note: All coefficients generated in Stata 8.2, with robust standarderrors, clustering on the justice. Model estimated using Judicial Common Space Scores. One-tailed significance tests used with directionalhypotheses. a. JudicialCommon Space Scores are availableonly for the years 1953 to 1999; hence one year of the RehnquistCourtdata was dropped for purposesof estimatingthis model. Alternativeestimationsusing differentideology variablesin the appendixinclude the final year of our data (2000). *p < .05. **p < .01. ***p < .001.

82 Political Research Quarterly

Table4 in Predicted of Vote to Strike Resultingfrom ChangingValuesof Probability Change IndependentVariables,HoldingOtherVariablesat Their Median,BurgerCourt Variable Ideological congruence (statuteandjustice) Ideological congruence(statuteandjustice) Solicitor support/amicus Congressionalcongruence Congressionalcongruence Amicus support Amicus support Civil liberties challenge

Change in VariableValue

Change in PredictedProbability

10th percentile - 50th percentile 50th percentile - 90th percentile 0- 1 10th percentile 50th percentile 50th percentile--> 90th percentile -5 0 0 --> 5 0- 1

-26.4% -22.5% -8.7% -6.9% -6.6% -6.2% -6.1% +5.5%

Table5 in of Vote to StrikeResultingfrom ChangingValuesof Predicted Probability Change IndependentVariables,HoldingOtherVariablesat Their Median,RehnquistCourt Variable Ideological congruence (statuteandjustice) Ideological congruence (statuteandjustice) Solicitor support/party Solicitor support/amicus Solicitor opposition/amicus Congressionalcongruence Congressionalcongruence Amicus support Amicus support Civil liberties challenge

Change in VariableValue 10th percentile 50th percentile -0 0 ---

50th percentile 90th percentile 1 1 0 1 10th percentile -50th percentile 50th percentile- 90th percentile -5 0 0 -- 5 0 -- 1 --

also provide results of models estimating the likelihood of a vote to invalidatefederalandstatelegislation separately,allowing us to comparewhetherthe impact of our variablesdepends on whetherthe law emerges from the federalor a state government.Tables4 and 5 provide additionalinformationregardingthe interpretation of the logit results for the full models by presenting changes in the predictedprobabilityof a vote to strikegiven a particularchange in the value of the independentvariables(with the othervariablesheld at theirmedians).These values enablethe readerto comparethe impactof the variableson thejustices'propensity to invalidatestatuteschallengedbefore them.

A. The Burger Court Full model. Almost all of our initial hypotheses received support from the data. First, as expected, consistency between the statute's direction and the justices' ideologies is statistically significantand the

Change in PredictedProbability -23.2% -21.9% -15.9% -18.9% +26.4% -1.2% -7.6% -3.8% -3.9% +10.2%

coefficient is negative.Thus, as the justices' attitudes become increasinglyconsistentwith the statute'sideological content,the justices demonstratea decreased willingness to strike the statute. As the statistics in Table4 reveal, the relationshipbetween ideology and individual votes to strike statutes has a significant substantiveimpact,changingthe probabilityof a vote to strike by almost 49 percent over the range of the variable from its 10th to 90th percentiles. However, the multiplicative term reflecting the interaction between the justices' ideologies and the source of the statutedid not producea statisticallysignificantcoefficient. Thatis, on the BurgerCourt,the justices' ideological reactionsto legislative enactmentswere not conditioned on whether the statute emerged at the state or federallevel. Figure 1 illustratesthis point by providing a graphicalrepresentationof the relationship between liberal and conservativejustices' willingness to strikefederal and state laws in the Burger Court.Note thatfor both federal and state legislation,

Lindquist,Solberg / JudicialReview by the Burgerand RehnquistCourts 83

Figure 1 between Relationship Ideological Congruity and Propensity to Strike Federal and State Laws-Burger Court

A

State Statutes

d

C-

o. N

LO (II -1

-.5 -----

B

0 IdeologicalCongruity

LiberalJustices

-A

.5

1

ConservativeJustices

FederalStatutes

v. r-

o

IdeologicalCongruity A LiberalJustices ConservativeJustices =-

the liberal justices' voting patternsreveal the more pronouncedimpactof ideology. In the case of federal legislation in particular,conservativejustices demonstrate a far more restraintistorientation toward the evaluationof its constitutionality.The same patternis less pronounced but still discernible in the case of state legislationas well. Thus, in the case of both state

and federallaws, liberalswere more activist. In addition, the coefficient for the State Statute variable is significant and positive. Although interpretationof this variable is somewhat complicated by its inclusion in an interactiveterm, it neverthelessreflects the Court's greater willingness overall to strike state as opposed to federal statutes.

84 PoliticalResearchQuarterly

We also hypothesized that the executive branch would influence the justices' voting behaviorthrough the solicitor general. Although the impact of the solicitor's participationis well documented,in these judicialreviewcases, the solicitor'simpactwas limited to those cases in which he participatedas an amicus in supportof the challenged statute's constitutionality. As noted above, we initially created a variable reflectingthe solicitor's involvementrepresentingthe United States as a partyto the litigation;this variable never achieved statistical significance in any of our model specifications. On the other hand, where the solicitor supportedthe statute as an amicus curiae, the justices were less likely to vote to strike the statute,reducingthe probabilityof a vote to strikeby 8.7 percent. Justices on the BurgerCourtwere also apparently influenced by cross-institutional considerations involving the legislature (whether consciously or unconsciously). Statutes that were consistent with congressional preferenceswere less likely to survive constitutional scrutiny by the individual justices. This variable has a 13.5 percent substantiveimpact on the likelihood of a vote to strike over its 10th to 90th percentile range. In contrast, the age of the statute was not significantly related to a vote to strike.Finally, where more amici briefs were filed in support of the statute than in opposition to it, the statute was more likely to generate a positive reaction from the justices. Civil liberties challenges produced more votes to strike laws as unconstitutional, perhaps due to the enhanced level of scrutiny employed in those cases. Moreover, the control we added for selection bias was also statistically significant, althoughwe did not specify an expected direction for this variable a priori. Nevertheless, the negative coefficient indicates that, where the lower court struck the statute, the Supreme Courtjustices were more likely to vote to uphold the law. This result is not surprising.First, where the lower court has struck the statute, it suggests that reasonable people could differ regardingthe statute's constitutional infirmities. These circumstances should drive both the certiorariprocess (one would expect that a declarationof unconstitutionalitywould raise a case high on the Court's radar screen). And since the Supreme Court often takes cases to reverse, it is not surprisingthat the justices are more likely to uphold statutes that were struck below. The variable's significance also gives us some comfort that we have controlled for the selection bias resulting from the Court's discretionarydocket.

Separate models: Federal and state. The separate models-estimated by separating the data set into cases involving state and federal legislation-reveal thatin the BurgerCourt,thepatternsof influenceremain fairly constantregardlessof the sourceof the legislation,withtwo exceptions.The first exception involves the influence of amicus curiae. When the Burger Court was faced with a constitutional challenge to state legislation, a preponderanceof amicus briefs in favor of the legislation mitigated against its being struck down. Not so with respect to federal legislation, where the differentialbetween numbersof amicus briefs supportingand opposing the statutehad no significantinfluence on the justices' voting behavior. Moreover,the presence of a civil liberties challenge to a federal enactmentdid not mitigate in favor of a vote to strikefederallaws, perhapsbecausetherewere fewer such challenges to federal statutes, many of which instead addressed constitutional questions underthe commerce and supremacyclauses.

B. The RehnquistCourt Full model. Like the BurgerCourtmodel, most of the variablesin the RehnquistCourtmodel were also significant and in the hypothesized direction. First, the variable reflecting consistency between the statute's direction and the justices' ideologies was statisticallysignificantand the coefficient is negative. The greaterthe consistency between the statute'sideological directionand the justices' policy preference, the less likely it is that the justice will vote to strike the statute. Figure 2 graphicallyrepresentsthe relationship between this ideological congruity and the justices' propensity. Note that there appears to be some (albeit mild) relationshipbetween the justices' ideology and their propensity to strike federal as opposed to state laws, with conservativesmore influenced by their ideological reactionto federal legislation, and vice versa for liberal justices. This relationshipwas statisticallysignificant, as the interaction term in the logit model reveals (and as confirmed using the inteff program;Norton, Wang, and Ai 2004). Thus, the conservative justices' rhetoric suggesting an increaseddeference to state legislation is manifested in their voting behavior, but they are less restrainedthan their liberal counterpartsin the case of federallegislation, as Figure 2 reveals.This is true even after controllingfor ideological effects. As for cross-institutionalinfluences, the solicitor's influenceis even moreimportantin the Rehnquistthan in the BurgerCourt.When the solicitor supportedthe

Lindquist,Solberg/ Judicial Review by the Burgerand RehnquistCourts 85

Figure 2 between Relationship Ideological Congruity and Propensity to Strike Federal and State Laws-Rehnquist Court

State Statutes

A

0

V-

-1

-.5 -

----

0 IdeologicalCongruity

LiberalJustices

B

-A--

.5

1

ConservativeJustices

FederalStatutes

0•

-.5

-1 0

0 IdeologicalCongruity

LiberalJustices

statute as amicus, the probabilityof a justice voting to strike the statute increasedby 18.9 percent;when he supportsthe statute as party,the probabilityof a vote to strike is reduced by 15.9 percent. He is also effective when opposing a statuteas amicus, with an impressive26.4 percentincrease in the probabilityof a vote to strikewhen he does so (which is a fairly rare event). In addition, our measure of the congruence between the statute and congressional preferencesis significant and in the negative direction. Thus, it

A

.5

1

ConservativeJustices

appearsthat as the consistency between the statute's ideological direction and congressional preferences increases,thejusticesareless likely to strikethe statute. And as in the Burger Court, no statistically significant relationshipexists between a statute's age and the likelihood that a justice will vote to invalidateit. The Rehnquist Court's members also respond to pressure from interests filing amicus curiae briefs. Since the measure reflects the number of briefs, ratherthanthe numberof interestsresponsiblefor the

86 Political ResearchQuarterly

briefs, it is not a direct measureof individualinterest group support.But indirectly,the coefficient's significance indicates that, in this important subset of cases, amicus involvementmay have a criticalimpact on the justices' votes. Where the differentialbetween supportingand opposing briefs shifts from -5 to 5, it decreasesthe probabilityof a vote to strikeby 7.7 percent. Also, where a civil liberties- based challenge is brought against the statute, the justices are more likely to vote to invalidate it--by 10 percentage points. And where the lower court struck down the statute,the SupremeCourtwas more likely to uphold the statute,a finding thatprobablyreflects the Court's propensityto reverse. Separate models: Federal and state. Once the data set is divided into two parts,the dynamics associated with the independent variables in the Rehnquist Court are somewhat clarified. While the models are fairly consistent, the federalmodel has less predictive power and two of the variablesno longer achieve statistical significance. In the case of federal legislation, the justices appearto be less influenced by interest group pressure (this was true in the Burger Court as well). Nor does lower court disposition affect the RehnquistCourt's voting patternsin cases involving federal statutes.

5. Discussion This article presents a comprehensive model of SupremeCourtjustices' votes to invalidateor uphold federal, state, and local statutes in cases involving constitutional challenges during the Burger and RehnquistCourt's 1969 to 2000 terms.The portraitof the justices' voting behaviorpresentedhere reflects a variety of influences on the choice to invalidate statutes. Moreover, many of these findings are consistent with existing research on cross-institutional influences, interestgroup pressure,and the influence of the solicitor general. First, the justices' ideologies clearly affect their vote choice in these cases, thus once again confirming the preeminenceof the attitudinalmodel as a predictor of the justices' behavior. Where the statute's ideological direction (liberal or conservative)is consistent with a justice's ideological orientation, the justice is less inclined to strike the statute. In the RehnquistCourt,this ideological influence is moderated by or conditioned on the source of the statute, with liberals more likely to strike statutestatutesand conservatives more likely to strike federal statutes.

This finding indicatesthatrhetoricalsupportfor judicial restraintis not manifestedin the justices' voting behaviorin cases involvingjudicial review of federal statutesin the RehnquistCourt. As far as activism in generalis concerned,conservative politicians and commentatorsoften note the importanceof "judicialrestraint"and "strictconstruction of the Constitution."The implicationis that conservative justices will be more restrained in their choices to invalidate legislation emerging from the democraticprocess. While this rhetoric has become particularlyheated in more recent years, recently appointedconservativejustices to the SupremeCourt do not appearto act in accordancewith this policy. The patternsrevealedin Figure2 indicatedthat,on the RehnquistCourt,conservativesandliberalsare almost equally likely to vote ideologically in choosing to strikeboth state and federal legislation. On the other hand, the relationship between conservatism and restraintis more pronouncedin the Burger Courtespecially in the case of federallegislation.As Figure 1 reveals, conservativejustices are less inclined to invalidate both federal and state legislation, even when they disagreewith thatlegislationideologically. In the BurgerCourt,restraintistconservativesactually behavedaccordingly.The same cannotbe said for the RehnquistCourt,regardlessof any rhetoricalcommitment to restraintby those on the Right. Second, the justices are sensitive to their institutional context when they render their decisions in cases involvingjudicial review.As one would expect from the existing literature,the solicitor general is influentialin these cases, particularlywhen he enters the case as amicus in supportof a statute'sconstitutionality (Segal 1990). This suggests that when the solicitor is able to select cases in which to participate (with the exception of those cases in which the Supreme Court requests him to serve as amicus, in which cases the amicus participationis essentially mandatory;see Pacelle 2003), he is quite adept at choosing cases in which his chosen position is likely to prevail. As in other cases, the justices appear to respect the position of the solicitor even when exercising the coveted power of judicial review. In general, however, it is also interesting to note that the solicitor's influence in these cases appears to have increasedover time; the impact of the solicitor is far more pronouncedin the RehnquistCourtthan in the BurgerCourt,a finding that warrantsfurtherinquiry. The executive branchis not the only political institution that affects the Court's decision-making process. In our model, we find a contemporaneous

Lindquist,Solberg/ JudicialReview by the Burgerand RehnquistCourts 87

influence of congressional preferences on the justices' voting behavior,a finding thatis consistentwith Epstein and Knight's (1998) proposalthat separation of powers constraintsmay affect the Court even in constitutional cases, as well as the findings of Friedmanand Harvey (2003) in the context of challenges to federal statutes (but see Sala and Spriggs 2004). In that sense, this result may reflect strategic behavior by the justices, but that is not clear. The influence of this variable may reflect the indirect impact of the public mood on the justices' voting behavior,or the justices' own evolving perceptions concerning appropriatepublic policy as part of the dominantpolitical coalition. At the very least, it presents a potent challenge to the position thatthe Court will be sensitive to congressional preferences in statutorycases only-a position thathas been repudiated by Segal (1999). Moreover,in both Courts, the impact of Congressionalpreferences appearedto be more slightly pronounced with respect to state statutes.While this result may seem counterintuitive, Meernikand Ignagni(1997) found thatCongresswas more likely to respondto the Court'sdecisions invalidating state as opposed to federal statutes.Our finding is thus consistent with existing research suggesting thatretaliationby Congressin the form of bills to overridejudicial decisions is more likely to occur when state legislation is invalidated. To the extent our findings indicatethe influence of Congresson the Court'sconstitutionaldecisions, they are inconsistent with conclusions drawn by Howard and Segal (2004) in their study of the Rehnquist Court,as well as by Sala and Spriggs (2004). Howard and Segal concludedthatthe conservativejustices did not appearconstrainedby their political environment in the years 1993 and 1994, when Democraticdominance of the elected branches could have affected their responsiveness to constitutionalchallenges to liberal statutes.The reason for our dissimilarfindings may be methodological.Howardand Segal evaluated conservativejustices' supportfor liberalandconservativerequeststo strikestatutesin constrained years (19931994) and unconstrainedyears (1985-1992, 1995) by comparingpercentagesupportfor liberaland conservativestatutesacrossthosetimeperiods.In ourmodel,on the otherhand,we evaluatedthe influenceof congressional preferenceson all justices while controllingfor other factors. It is possible, therefore, that political constraintsoperatedmore profoundlyon liberaljustices than on conservativejustices. Similarly,Sala and Spriggs (2004) concluded that the justices' ideological reactions to congressional

enactmentswas not eliminatedduringparticularperiods, andthe ideologicalconfigurationsof the two institutions created potentially greaterconstraintson the justices' willingness to vote in accordancewith their ideological predispositions.While Sala and Spriggs presenteda rigorousanalysisof the separationof powers thesis, their approachmay not have allowed for a more subtle cross-institutionalinfluence even after controlling for the justices' ideologies. We reiterate that the impactof our congressionalvariablemay not reflectstrategicbehavioron the partof thejustices, but ratherthe processby which thejustices' assimilationin the social fabric,as well as membershipchange on the Court,affects the justices' perceptionsof the constitutionalityof legislativeenactments.Moreover,our findings, like those of Sala and Spriggs, indicatethat the attitudinalvariablesexercise a far greaterimpact on the likelihoodof a vote to strikethandoes ourmeasure of congressionalinfluence. Curiously,while the justices do appearsomewhat responsive to congressional preferences (or public opinion), our variablereflectingthe age of the statute does not performin the mannerhypothesizedby Dahl (1957). Dahl suggestedthatthejustices-as partof the currentgoverningcoalition-would be more inclined to strike statutes that had been enacted by previous majorities in Congress. Yet we find that the Court strikesyoungerratherthan older statutesand that the relationshipbetween age and votes to strikeis not significantin ourmultivariatemodels. Since the periodof Dahl's researchended in 1956, it is possible that his findingsare time-bound. The justices also respondto othercontextualinfluences and to legal factors. In particular,the differential in the numberof amicus briefs filed in supportor in opposition to the particularstatute at issue influences the justices' voting behavior. This finding is relatively unusual, since the existing researchgenerally supportsthe conclusion that amicus briefs influence the certiorari process more than they do the Court's decisions on the merits (but see McGuire 1995). In contrast, our findings indicate that, as the numberof supportiveamicus briefs increasesrelative to the numberof briefs in opposition to the statuteor ordinance, the less likely the justices will vote to invalidatethe law. However,this influence is limited to cases in which a state statute was challenged. In the case of federal legislation, the justices appear more impervious to interest group pressures. And while our amicus variablepresentsa relativelycoarse measure of amicus influence, the raw numbersmay indeedreflectthe intensityof interestgroups'focus on

88 Political ResearchQuarterly

statutes.Finally,giventhehigherdoctrinal particular hurdlesto a statute'sconstitutionality whencivil liberties challengesare broughtagainstthe statute,it comesas no surprisethatcivil libertieschallengesare most effectiveas the bases for constitutionalchallengesbeforethe Court.

6. Conclusion The model presentedhere suggeststhat the justices' voting behaviorin cases involvingjudicial reviewis the productof a nuancedprocessthat is stronglygovernedby thejustices'policypreferences. that At the sametime, the modelalso demonstrates instituare also affected the justices' decisions by tionalandcontextualinfluencesin additionto those preferences.In short,thejusticesdo notrenderthese importantdecisionsin a vacuum,but ratherdo so while consideringor at least while subjectto the As influenceof theircomplexpoliticalenvironment. Segal andSpaeth(2002, 174) noted,the Court"has

not marchedto the beatof alienor enigmaticdrums, eventhoughthosedrumshavetypicallybeatena stridentlypartisancadence"(see also RohdeandSpaeth 1976, 145). Moreover, these cross-institutionaland

contextualinfluenceshave an importantsubstantive impacton thejustices'votesin thesecases.Forthose criticswhocomplainthattheCourtactsin a strongly fashion without consideration countermajoritarian for the preferencesof otherpoliticalactorsor the public, this portraitof the justices' votes in cases of stateandfederal challengingthe constitutionality laws shouldprovidesome, albeit limited,comfort. Clearly,thejusticesrenderthesedecisionsin a highly deferenceor at contextualized fashion,demonstrating leastsensitivityto Congress,thepresident,andinterest groupsalike.AlthoughDahl's(1957)proposition thatthe Courtconformsto the dominantmajorityby invalidatingolderstatuteswas not borneout by our analysis,his overarchingconclusionthat the Court does not buckcurrentpoliticaltides is nevertheless by ourfindings,atleastfortheCourtunder supported the two mostrecentchiefjustices.

Appendix Votes to Strike ChallengedStatute/Ordinance, of Estimation Justices' Logit Burger and RehnquistCourts(with HypothesizedDirectionof Coefficient) Using Segal and CoverScores IndependentVariable Consistencybetween statutorydirection andjustice ideology (-) Source of Statute(State) x Justice Ideology (+/-) State statute(+) Justice ideology (+/-) Solicitor support(party)(-) Solicitor supportfor statute(amicus) (-) Solicitor opposition to statute(amicus) (+) Consistency between statutedirectionand congressionalpreferences(_)a Age of statutein years (logged) (+) Amicus supportfor statute(differential)(-) Civil liberties challenge (+) Selection bias control Lower court disposition Constant Log-likelihood Pseudo R2 N

Coefficient/StandardError(Burger)

Coefficient/StandardError(Rehnquist)

-0.743 (0.226)***

-0.575 (0.250)**

-0.038 (0.172) 0.863 (0.116)** 0.415 (0.145)** -0.287 (0.114)** -0.124 (0.178) -2.51 (0.952)**

0.198 (0.188) 0.159 (0.224) 0.008 (0.146) -0.531 (0.151)*** -0.772 (0.137)*** 1.35 (0.208)*** -0.185 (1.13)*

-0.044 (0.030) -0.054 (0.012)*** 0.254 (0.090)**

-0.151 (0.041) -0.025 (0.018)* 0.473 (0.118)***

-0.390 (0.052)*** -0.674 (0.124)*** -2,399.30 .12 3,939

-0.275 (0.072)*** -0.001 (0.188) -1,347.01 .06 2,089

Note: All coefficients generatedin Stata 8.2, with robuststandarderrors,clusteringon the justice. One-tailedsignificance tests used with directionalhypotheses. a. Congressionalpreferences measuredusing Poole-RosenthalScores ratherthan the Judicial Common Space scores; values per term reflect mean of the median in the two chambers. *p < .05. **p < .01. ***p < .001.

Lindquist,Solberg/ JudicialReview by the Burgerand RehnquistCourts 89

Notes 1. 1 CR. (5 U.S.) 137 (1803). 2. UnitedStates v. Morrison, 529 U.S. 598 (2000). 3. Clearly,statesupremecourts'exerciseof judicialreviewmay be equallyconsequentialat the statelevel (see Langer2002; Emmert

1992). 4. Segal and Spaeth (2002) examined "actions" declared unconstitutional,which could include challenges to administrative action as well as direct challenges to particularlegislation (p. 415). For this reason, their figures for the numberof actions declared unconstitutionaldiffer significantly from ours, which involve directchallenges to the constitutionalityof specific statutory provisions. 5. Heberlig and Spill (2000) examined whether the justices respondedto amicus curiae briefs filed by membersof Congress and found little effect of such individualized participation. However,we are testing whetherthe Courtrespondsto the overall disposition of Congress ratherthan the voices of one or several members. 6. We confirmedthat Segal and Spaeth (2002) had used this methodthroughan e-mail exchangewith HaroldSpaeth(February 20, 2003). 7. Identifying the age of the statute often requiredhistorical researchregardingthe enactmentdate, since the West Code citation contains dates that do not necessarily reflect the actual date of enactment(but ratherthe most recent compilation and publication date of the particularcode volume). 8. Quadraticspecifications of this variable did not produce significantresults. 9. We recognizedthat a simply differencescore may not fully reflect the influenceof amicusbriefs, since it treatsa 6-3 differential the same as a 3-0 differential,which presenttwo quitedifferent scenarios.We thereforeconstructedan alternativemeasurerelying on a methodologydevelopedby David Klein (2002) in his analysis of circuitpanel'sadoptionsof new legal rules.His concernwas with a variablereflectingthe numberof circuitsthathad adopteda particularlegal rule where a conflict existed. Because a 3-0 split amongthe circuitswas differentthana 6-3 split,he adopteda strategy of subtractingthe numberof circuitsthathad adoptedthe particular rule from the number that had rejected the rule. Where adoptionsoutnumberedrejections,he multipliedthe differenceby the proportionadopting.Whererejectionsoutnumberedadoptions, he multipliedthe differenceby the proportionrejecting.We followed the same principle with amicus supportingand amicus rejectingthe rule.This new variablewas similarlyinsignificantand had no significanteffect on the remainingvariables. 10. Data on the cert votes for individual cases is available prior to the Rehnquist Court, but to fully account for selection effects, we would need cert dataon all cases broughtto the Court. 11. In most cases, the lower court was the circuit court. In cases involving a direct appeal from the districtcourt, we coded the Lower CourtDisposition variablebased on the disposition of the constitutionalissue in the districtcourt. In appeals from state supremeor appellatecourts, we followed the same procedure. 12. We considered clustering on the case as an alternative specificationbutrejectedit. Accordingto Zorn (forthcoming),the decision to cluster on the justice versus the case is dependent upon the model specification. Here, we anticipatethat our model does a reasonablejob of capturingwithin-casevariationgiven the numberof case-relatedvariablesincluded,and we expect thatthe

latent dependencein the model is likely to lie between the individualjustices ratherthan within included cases. In other words, thereis consistency in thejustices' views in judicial review cases; therefore, it is more appropriateto cluster on the justice rather than the case.

References Adamany,David. 1973. Legitimacy,realigningelections, and the SupremeCourt. WisconsinLaw Review 3:790-846. Barnum,David G. 1985. The SupremeCourtand public opinion: Judicial decision making in the post-New Deal period. Journal of Politics 47:652-65. Benesh, SaraC., and HaroldJ. Spaeth.2003. The SupremeCourt Justice-Centered Database-The Rehnquist Court. http://www .polisci.msu.edu/pljp/. Bickel, Alexander. 1986. The least dangerous branch. New Haven, CT:Yale UniversityPress. Boucher,Robert,and JeffreyA. Segal. 1995. SupremeCourtjustices as strategicdecision makers:Offensive grantsand defensive denialson the VincentCourt.Journalof Politics 57:54-50. Brace, Paul, Melinda Gann Hall, and Laura Langer. 2001. Placing state supreme courts in state politics. State Politics and Policy Quarterly1 (1): 81-108. Caldeira,GregoryA., andJohnR. Wright. 1988. Organizedinterests and agenda setting in the U.S. SupremeCourt.American Political Science Review 82 (4): 1109-27. . 1990. Amicus curiae before the Supreme Court: Who participates, when and how much? Journal of Politics 52 (3): 786-806. Caldeira,GregoryA., John Wright,and ChristopherZorn. 1999. Sophisticatedvoting and gate-keepingin the SupremeCourt. Journal of Law, Economics and Organization15 (3): 549-72. Canon, BradleyC., and Sidney Ulmer. 1976. The SupremeCourt and critical elections: A dissent. American Political Science Review 70:1215-18. Casper,Jonathan.1976. The SupremeCourtand nationalpolicymaking.AmericanPolitical Science Review 70:50-63. Choper,Jesse H. 1980. Judicial reviewand the nationalpolitical process. Chicago: Universityof Chicago Press. Devins, Neal. 2004. The majoritarianRehnquistCourt?Law and ContemporaryProblems 67:63-81. Dahl, Robert. 1957. Decision making in a democracy: The Supreme Court as a national policymaker.Journal of Public Law 6:267-90. Ely, John Hart. 1980. Democracy and distrust:A theory ofjudicial review.Cambridge,MA: HarvardUniversityPress. Emmert,CraigF. 1992. An integratedcase-relatedmodel of judicial decision-making: Explaining state supreme court decisions in judicial review cases. Journal of Politics 54:543-52. Epstein, Lee, and Jack Knight. 1997. The new institutionalism, partII. Law and Courts7:4-9. I. 1998. The choices justices make. Washington, DC: CongressionalQuarterlyPress. Epstein, Lee, and Joseph Kobylka. 1992. TheSupremeCourtand legal change: Abortion and the death penalty. Chapel Hill: Universityof North CarolinaPress. Epstein, Lee, Andrew D. Martin, Jeffrey A. Segal, and Chad Westerland. Forthcoming. The judicial common space. Journal of Law, Economics and Organization.

90

Political ResearchQuarterly

Fallon, Richard H., Jr. 2002. The "conservative"paths of the RehnquistCourt'sfederalismdecisions. Universityof Chicago Law Review 69 (Spring):49. Flemming, Roy B., and B. Dan Wood. 1997. The public and the Supreme Court: Individual justice responsiveness to American policy moods. American Journal of Political Science 41:468-98. Friedman,Barry.2002. The birthof an Americanobsession: The history of the countermajoritarian difficulty,partV. YaleLaw Journal 112:153-259. Friedman, Barry, and Anna L. Harvey. 2003. Electing the SupremeCourt.Indiana Law Journal 78:123-51. Funston, Richard. 1975. The Supreme Court and critical elections. AmericanPolitical Science Review 69:795-811. Gates, John B. 1987. Partisanrealignment,unconstitutionalstate politics, and the U.S. Supreme Court, 1837-1964. American Journal of Political Science 31 (2): 259-80. . 1984. The AmericanSupremeCourtand criticalrealignment:A critical review. Social Science History 8:267-90. Hansford,Thomas, and James F Spriggs III. 2006. The politics of precedent on the United States SupremeCourt. Princeton, NJ: PrincetonUniversity Press. Hausegger,Lori,andLawrenceBaum. 1999. Invitingcongressional action:A study of SupremeCourtmotivationin statutoryinterpretation.AmericanJournalof PoliticalScience43 (1): 162-85. Heberlig, Eric S., and Rorie L. Spill. 2000. Congress at court: Membersof congress as amicus curiae.SoutheasternPolitical Review 28 (2): 1-24. Heckman, James J. 1979. Sample selection bias as specification error.Econometrica47:153-62. Howard,Robert,andJeffreyA. Segal. 2004. A preferencefor deference? The Supreme Court and judicial review. Political Research Quarterly57 (1): 131-43. Klein, David. 2002. Making law in the United States courts of appeals. Cambridge,UK: CambridgeUniversityPress. Langer,Laura.2002. Judicial review in state supremecourts: A comparativestudy.Albany:StateUniversityof New YorkPress. Long, J. Scott, and Jeremy Freese. 2001. Regression modelsfor categorical dependentvariables using Stata. College Station, TX: Stata Press. Maltzman,Forest,James F Spriggs II, and PaulWahlbeck.2000. Crafting law on the Supreme Court: The collegial game. Cambridge,UK: CambridgeUniversityPress. Marshall, Thomas R. 1989. Public opinion and the Supreme Court. New York:Longman. Martin, Elain M. 1985. The Burger Court and preemptiondoctrine: Federalism in the balance. Notre Dame Law Review 60:1233-54.

Massey, Calvin. 2002. Federalism and the Rehnquist Court. Hastings Law Journal 53 (January):431. McGuire, Kevin T. 1995. Repeat players in the Supreme Court: The role of experiencedlawyers in litigation success. Journal of Politics 57:187-96. . 1998. Explainingexecutive success in the U.S. Supreme -Court. 1998. Political Research Quarterly51:505-26. Meernik, James, and Joseph Ignagni. 1997. Judicial review and coordinateconstructionof the Constitution.AmericanJournal of Political Science 41:447-67. Norton, Edward C., Hua Wang, and Chunrong Ai. 2004. Computinginteractioneffects and standarderrorsin logit and probitmodels. The Stata Journal 4 (2): 154-67. Pacelle, Richard 2003. Between law and politics: The solicitor general and the structuringof race, gender and reproductive rights litigation. College Station: Texas A&M University Press. Rohde, David, and HaroldSpaeth. 1976. SupremeCourtdecision making.San Francisco:W.H. Freeman. Rosenberg, Gerald. 1991. The hollow hope: Can courts bring about social change? Chicago:University of Chicago Press. Sala, Brian R., and James F. Spriggs II. 2004. Designing tests of the Supreme Court and the separation of powers. Political Research Quarterly57 (2): 197-208. Segal, JeffreyA. 1990. SupremeCourtsupportfor the solicitorgeneral:The effect of presidentialappointments.WesternPolitical Quarterly43 (1): 137-52. . 1999. SupremeCourtdeferenceto Congress:An examination of the Marksist Model. In Supreme Court decision making:New institutionalistapproaches,ed. Cornell Clayton and HowardGillman.Chicago: Universityof Chicago Press. Segal, Jeffrey A., Lee Epstein, Charles Cameron, and Harold Spaeth. 1995. Ideological values and the votes of U.S. Supreme Courtjustices revisited. Journal of Politics 57 (3): 812-23. Segal, JeffreyA., and HaroldJ. Spaeth.2002. TheSupremeCourt and the attitudinal model revisited. Cambridge, UK: CambridgeUniversityPress. Spaeth,Harold.2003. The original UnitedStates SupremeCourt database. http://www.polisci.msu.edu/pljp/. Stimson, JamesA., Michael B. Mackuen,and RobertS. Erikson. 1995. Dynamic representation.American Political Science Review 89:543-65. Whittington,Keith.2001. Takingwhatthey give us: Explainingthe Court'sfederalismoffensive.Duke Law Journal51:477-520. Zorn, Christopher.Forthcoming. Comparing GEE and robust standard errors for conditionally dependent data. Political Research Quarterly.