[N.B. Footnotes cannot be accommodated in HTML format. Accordingly, the 223 in this review have been removed. Readers should consult the original text for them.]

Georgetown Law Journal
Volume 92, Page 941, June, 2004
Book Review

FEDERAL COURTS, STATE COURTS AND CIVIL RIGHTS:
JUDICIAL POWER AND POLITICS

REVIEW OF DANIEL R. PINELLO, GAY RIGHTS AND AMERICAN LAW,
CAMBRIDGE UNIVERSITY PRESS, 2003, PP. 349. $23.00

Nan D. Hunter
Professor of Law
Brooklyn Law School

INTRODUCTION

With Lawrence v. Texas, in which the Supreme Court struck down a Texas law prohibiting same-sex sexual acts, and Goodridge v. Department of Public Health, in which the Massachusetts Supreme Court ruled that same-sex couples could not be excluded from marriage under that state's constitution, fresh on the books, Daniel Pinello's book, Gay Rights and American Law, arrives like a cold shower. Pinello, a lawyer and political scientist, studied every appellate decision in gay rights cases in both federal and state courts – approximately 400 in all – between 1981 and 2000. His analysis – that judicial hostility was pervasive – takes us back to the not so long ago era when Bowers v. Hardwick dominated the law of homosexuality and AIDS dominated the culture. Although Pinello's conclusions are not uniformly grim, the data he amassed point to rampant anti-gay prejudice among judges.

The simplistic notion that judges inexorably follow precedent in a noble quest to ascertain and apply the best and truest interpretation of common law principles and statutory text to the case at hand has long since been superseded: by legal realism, by the legal process school, by critical legal theories, by law and economics, by rational choice models, and by cognitive legal studies scholars. Yet it is still true that no challenge strikes closer to the heart of the legal profession than the proposition that the rule of law is nothing more than an artifact of power. Those in the legal academy may wonder what the "jurisprudes" will come up with next. But we implicitly trust that they will come up with something, some new complex theory of judicial decisionmaking more intellectually satisfying than the conclusion that, with a few quirky exceptions, judges simply side with the party whose political or policy views they find the most congenial.

What legal scholars seldom consider is whether there is a method to prove or disprove, to verify or falsify, any of these theories. By contrast, the study of law among political scientists such as Pinello has become centered on (or, depending on one's view, mired in) empirical analysis of judicial decisionmaking. Pinello's book is based on the application of methods of statistical analysis to the several hundred cases in his database, and from that he makes two claims of institutional dimension about the politics of the American judiciary, as reflected in gay civil rights cases.

Pinello's first major conclusion is that federal courts not only were less receptive than state courts to gay rights claims, but that they were systemically hostile. What Pinello describes as his most surprising finding is that state courts do a better job of protecting gay civil rights than federal courts do, a conclusion that he extrapolates from his data to cover "other beleaguered minorities" as well. In tackling this issue, Pinello takes himself and the reader directly into the longstanding debate on institutional parity: whether state court systems are on par with the federal courts in providing the appropriate level of protection for constitutional rights.

Pinello's second institutional claim is that the personal policy preferences of judges control the outcome of cases almost as much as law does. Gay Rights and American Law seeks to build upon what is now a substantial body of empirical literature on the question of whether judges' personal beliefs have a greater impact on judicial decisions than the law does. Pinello's goal is to enrich the political science discussion with consideration of a bigger pool of possible variables that could predict judicial bias. He has chosen to test his model in a field where major principles of law are still a work in progress, where at least one issue being litigated is viewed as dangerous electoral dynamite for both political parties, and where a major point of debate in popular discourse is whether secular principles should even be the primary touchstone in setting society's rules. Surely one can expect gay-rights cases to provide a rich source of decisions upon which to base an assessment of whether law truly governs law.

The book speaks to scholars in three fields: federal courts, empirical legal studies, and civil rights. This review addresses the issues raised by the book in that order.

In Part I, I provide the historical context for the parity debate, which is unfortunately missing from the book, and find that Pinello's work makes an important contribution. He demonstrates that federal courts consistently blocked the door for gay and lesbian litigants pressing claims for equal treatment, more so than state courts, even when one controls for subject matter and statutory versus constitutional claims. This pattern is directly traceable to the political party affiliation of the President who appointed the judges in each given case.

Pinello's book adds to the evidence of systematic hostility by Republican appointees to civil rights claims, which has been building for some time. His focus on one subject matter area allows him to consider a host of other possible factors and demonstrate the overwhelming impact of party affiliation. He fails, however, to fully analyze the logical conclusion of his findings: that the politicization of the judicial appointment process has punctured the most important wall shielding the federal courts from partisan infection. The fundamental question raised by all of this work is whether the longstanding belief that, on the whole, the federal judiciary, with its guarantee of lifetime tenure, is less susceptible than state courts to ideological manipulation has been overtaken by political events. If so, what in the past has been an accepted starting point for debates over parity now amounts to a myth of insulation.

Part II of the review tackles the methodology questions, and finds a number of problems, some stemming from the nature of empirical work and others that flow from Pinello's approach. Many of the same factors that make gay rights cases so fertile for investigation of the relationship between law and politics also make them exceedingly difficult to fit into an empirical model. The litigation context may involve abstract principles of constitutional law or the particulars of a custody case or the technicalities of immigration law. It is not even always clear which decisions should count as "gay rights" determinations. Nor is it necessarily obvious whether to code certain cases as wins or losses. Pinello did not create these shortcomings in the method, but unfortunately he also did not entirely succeed in surmounting them. Most significantly, such a study cannot measure the structures of consciousness that shape not only what courts decide, but what lawyers argue, which cases they elect to bring, whether they appeal, and the terms upon which cases are litigated and decided.

In Part III, I explore civil rights issues more broadly, and conclude that Gay Rights and American Law consists of questions left both unasked and unanswered, and these are in many ways the most interesting. Questions about the relationship between claims of liberty and equality based on sexual orientation and a series of structural constitutional questions lie just beneath the surface of the book's text. The recent history of gay rights law provides a context not only for a comparison between federal and state courts, but also for an understanding of the evolving relationship between judicial and legislative branches in the development of civil rights laws. How judges interact with legislatures and the extent to which statutes, rather than case law, dominate the field are both important. Likewise, the tension between the executive and legislative branches over appointments to the federal bench is beginning to intersect with the politics of gay rights, and is likely to do so increasingly in the future, which will lead to heightened debates over the extent to which views on homosexuality affect a candidate's fitness to serve in the judicial branch. Pinello touches on these topics, but because of its empirical approach, the book leaves one wishing for more and deeper analysis.

Sexual orientation issues can no longer be cabined to questions of individual rights. We are poised to enter a period in which the ramifications of sexual orientation law for debates about the capacity and legitimacy of core legal institutions will become manifest. Gay rights has emerged as a litmus test issue for both political parties, and as a focal point for large-scale political mobilization. Sexual orientation discrimination is no longer a trivial or peripheral issue in American politics and law. Whatever its shortcomings, Gay Rights and American Law raises important questions about the purpose, function, and vitality of our structures of governance.

I. REVERSE DISPARITY?

A. PARITY

In the 1970s, the Supreme Court began to cut back the jurisdiction of the lower federal courts in those cases, such as habeas corpus, in which lower federal courts have the power to review and reverse a state court's interpretation of federal constitutional law. In so doing, the Court signaled that it was returning to a more passive approach after a period in which federal judges were encouraged to view the outcome of state court proceedings involving assertions of constitutional rights with significant skepticism. The seeming return to a pre-civil-rights-movement philosophy was an early sign of what is by now a full-fledged states' rights jurisprudence.

Disagreement about the proper relationship between the two systems dates from the constitutional convention, when Madison won the power to allow a federal body, Congress, to make the final decision about whether any inferior federal courts should exist, against arguments that state courts provided all the capacity necessary for a judiciary, short of one national Supreme Court. Even after Madison's victory, there was no original federal court jurisdiction over claims arising under federal law until 1875. In practice, the assumption of state court parity continued until the shift to skepticism of state courts that occurred in the Warren Court era.

Civil rights lawyers howled in protest over the change in direction in the 1970s. This marked the beginning of the third wave of parity debates. The most influential reaction came from Burt Neuborne, who wrote what became the classic argument that the Supreme Court's new faith in the capacity of state courts was misguided. In “The Myth of Parity,” Neuborne asserted that federal courts were institutionally superior to state courts in protecting individual liberty from infringement by state actors, so much so that deference to state court judgments and fact-finding by federal courts amounted to a retrenchment of constitutional protections.

Neuborne based his argument on three assertions. First, he argued that "a competence gap exists between the state and federal courts," both because federal judges typically display a "higher level of legal talent" than state court judges and because of the greater resources, such as a higher caliber of judicial clerks, available to federal judges. Second, he argued that federal judges had "a series of psychological and attitudinal characteristics" that produced a greater willingness to enforce constitutional rights vigorously, ranging from elite "socioeducational" backgrounds and a corresponding sense of obligation to uphold the elite traditions of the federal bench, to greater attentiveness to Supreme Court jurisprudence and less contact with the unsympathetic facts presented by some civil liberties plaintiffs. Lastly, Neuborne asserted that the lifetime tenure of federal judges, as compared to the selection processes for many state court judges, rendered them "as insulated from majoritarian pressures as is functionally possible." Neuborne's institutional claim was sweeping: "the only judicial forums in our system capable of enforcing countermajoritarian checks in a sustained, effective manner are the federal courts." The three grounds upon which he based this conclusion became the starting point for the later debate.

Much has changed since then. Most notably, the philosophical complexion of the federal courts has shifted dramatically from the prevailing norm in the 1960s and 1970s, when Neuborne engaged in federal court litigation, the experience that shaped his argument. After Republican control of the Executive Branch for all but four years of the 1970s and 1980s, the protectiveness toward individual liberty that Neuborne associated with federal judges diminished dramatically. Neuborne himself has acknowledged that he would no longer argue that the two appellate court systems lack parity. Neuborne now finds the federal and state appellate courts to be equivalent, although he also has continued to assert that federal trial courts are measurably better than state trial courts. Other scholars have joined in declaring that a presumption favoring federal courts for adjudication of constitutional issues has become questionable.

With the realization that federal courts were becoming increasingly conservative, scholars began to argue over whether there really was an intrinsic difference in quality, or whether perceived quality fluctuated as an artifact of political direction. Central to this question was the threshold question of how quality could be defined: Should quality refer solely to rights-protective outcomes? Would erudite opinions count as quality, even if they repeatedly rejected arguments for expanded individual rights? Was it correct to assume that higher quality would inevitably result from greater competition for the much smaller number of federal judgeships?

These issues raised in the 1980s-to-1990s phase of the parity debate led in turn to questions of measurement. Political scientists began to develop empirical models by which they claimed to measure quality objectively and to be able to compare state and federal court systems. The most significant of these studies, conducted by Solimine and Walker, found little support for the belief that federal courts were consistently more rights-protective in adjudicating constitutional claims. In a study of more than 1,000 cases raising claims under the First, Fourth, and Fourteenth Amendments between 1974 and 1980, Solimine and Walker compared litigants' success rates in state and federal courts. They found a success rate of 31.4% in state courts and 41.7% in federal courts, a difference that, while statistically significant (that is, would not happen by chance), was too small to have significant predictive power. Moreover, when broken down by type of claim, several subsets of comparison lacked even statistical significance. They concluded that there was "weak parity" between the two court systems: that state and federal courts were not fungible, but that federal constitutional claims got at least a fair hearing in state courts. Accordingly, "the case for systematic bias simply can't be made."

Erwin Chemerinsky sought to move past these debates by asserting that it was impossible to prove parity or its absence by empirical methods because of the confounding factors that haunt litigation. Moreover, he asserted that parity was the wrong question. Chemerinsky argued that the proper way to resolve the debate, regardless of which political philosophy dominated the federal judiciary at any given moment, was to expand litigant choice. Using the principle of litigant autonomy and prioritizing the right to have federal courts determine the proper application of federal law, he proposed a variety of changes in federal law so that either plaintiff or defendant could, if that party asserted a potentially dispositive issue of federal law, ensure that a federal district court would accept jurisdiction over the case. Such changes would enable all litigants who invoked federal law to, in effect, determine parity for themselves, in light of when and where their case was brought and which particular issues it entailed.

What now seems most striking about the history of the modern parity debate is what is missing: From the beginning, the assumption has been that the only question was whether the federal courts were significantly better, or whether the state courts were sufficiently capable so that both systems gave equivalent levels of consideration to questions of constitutional interpretation. What if the differential cut in the other direction? What if state courts were systemically more sympathetic to individual claims of liberty and equality? Responding to arguments that the parity debate was pointless either because it was unresolvable or because it was really about something else (one's preferences for substantive law outcomes) anyway, Susan Herman suggested that "there might be more interest in a disparity debate." Herman argued that if the tables were turned, so that the Supreme Court or Congress began to expand the jurisdiction of the predominantly conservative federal courts, many constitutional law scholars might suddenly decide that it was important to revisit the question of whether state courts were entitled to deference.

The possible reverse of disparity is made all the more salient by another parallel development in rights jurisprudence: the rise of state constitutions as the source of substantive guarantees of individual freedom. The spiraling growth in the importance of state constitutions also dates from the 1970s and 1980s, when state courts began to interpret state constitutional provisions more expansively than the correlative provisions of the U.S. Constitution. Not surprisingly, plaintiffs' lawyers then began relying more on state law, which in turn provided more opportunities for protective rulings, and on it went. In 1977, Justice Brennan heralded the potential of state constitutional law to protect individual rights. In 1986, he identified 250 instances of state courts interpreting the scope of individual rights in their own charters in a more protective fashion than the Supreme Court had interpreted the federal constitution. Thus, on substantive law grounds, state systems had begun by the end of the century to rival federal courts as the preferred arenas for constitutional adjudication, at least in certain fields.

With the political tables thus turned, the possibility that Herman suggested of attempts to expand rather than retract federal court jurisdiction in order to achieve more conservative policy ends has in fact begun. After decades of conventional wisdom that diversity jurisdiction was a poor use of precious federal judicial resources, Congress has begun to extend the reach of federal district court jurisdiction. In 2002, for example, Congress enacted the Multiparty, Multiforum Trial Jurisdiction Act (MMTJA). The MMTJA grants original federal jurisdiction in cases arising from accidents in which at least seventy-five persons died, subject to a finding of "minimal diversity." Part of the impetus for the expansion of federal jurisdiction to large-scale accidents was Congressional concern over forum-shopping by plaintiffs to select more sympathetic state courts and a perceived need to provide for "an effective one-time determination of punitive damages." Still pending in Congress is the Class Action Fairness Act, which passed the House in June 2003, which would create both original and removal federal jurisdiction over all class actions in which minimal diversity and an enhanced amount in controversy is met. Like other recent bills intended to enlarge federal control over state courts, both pieces of legislation reflect the interests of large corporate defendants.

Has a new reverse disparity supplanted both the myth of parity and the claim of federal superiority? If Pinello's data and conclusions are correct, Herman's half tongue-in-cheek comment was stunningly prescient.

B. A CASE STUDY IN FEDERAL AND STATE COURTS

Although Pinello may believe that his conclusion regarding state court superiority will surprise gay rights lawyers, in fact they long ago joined the migration to state courts, pushed by the hostility of the Supreme Court in Bowers v. Hardwick and lured by the same possibility of more expansive interpretations of state constitutions that had drawn other public interest lawyers. The trend to state fora in gay rights litigation was apparent by the early 1980s, and only accelerated after the Hardwick decision in 1986. The results were impressive: nine state court decisions striking down sodomy statutes as unconstitutional on state constitutional grounds, including the Georgia Supreme Court's invalidation of the statute upheld by the U.S. Supreme Court. Of course advocates continued to bring cases in federal court as well; but the recognition of potential advantages in state courts is by now old news.

Despite his misperceptions about the conventional wisdom among gay rights lawyers, Pinello can rightly claim to have made the first attempt to scientifically compare how the two systems have adjudicated this category of claims. To create a database upon which to conduct his analysis, Pinello identified published decisions of state and federal appellate courts between 1981 and 2000 in which the courts ruled on "legal issues having direct impacts on the rights of homosexuals." The mean outcome in what Pinello deemed to be "essential" cases was .503, meaning that the gay rights claim prevailed half the time. His data showed state courts to be substantially more supportive: The mean outcome in "essential cases" in state courts was .572, compared with a mean of .256 in federal courts. Overall, the trend over time was toward stronger rights protection: Pro-gay rulings occurred in 42.4% of the 1980s cases and 53.6% of the 1990s cases. However, no such trend developed in federal court decisions, where the success rate remained static.

Much of Pinello's analysis concerns the relative impact of legal precedent as compared to various personal characteristics of the judges (race, religion and sex, for example), as well as geographic location. However, there was essentially no difference overall in demographic characteristics between federal and state court judges. Therefore, the variances he found in those factors do not explain the discrepancy between results in the federal and state court systems.

Instead, one factor – the party affiliation of the official who appointed the judges – dominated all others. Among federal judges, this factor produced an impact that dwarfed all other personal characteristics. Appointment by a Democratic president produced a probability impact of 40.5%. The closest comparison was the impact of racial minority status, at approximately half that level. As Pinello described the meaning of this finding, a judge's appointment by a Democratic president determined approximately 40% of "the probability space" between success and failure for a gay rights claim. Pinello found that in cases without binding precedent on point, federal judges appointed by Democrats were 125% more likely to support a gay rights claim than Republican appointees. Tellingly, the story was very different in state courts. Among state court appellate judges with an identifiable party affiliation, Pinello found an incidence of pro-gay votes of 58.9% among judges appointed by Democrats, compared with 47.8% among judges appointed by Republicans. Thus, Democrats were only about 25% more likely to support the gay rights claim in comparable cases.

Pinello tested the federal-state court discrepancy by two further comparisons. First, he identified nineteen federal and twenty-five state court decisions involving federal constitutional questions. He found that state court interpretations of the U.S. Constitution were one and half times more likely than federal court interpretations to produce a pro-gay result. Second, Pinello separated constitutional and statutory claims, since the former often encompass broader, more generic standards than do cases involving interpretation of more precise statutory texts. He found that compared with all nonconstitutional claims in either court system, "the presence of causes of action based on the federal constitution increased the probability of gay people's losing by .19."

From these data, Pinello draws extremely broad conclusions as to the superiority of state courts: "Any federal court decision was likely to be predisposed against gay rights." "Litigating in federal courts, compared with state fora, increased the probability of dispositions against lesbian and gay rights claims by the staggering factor of .31." He admonished in dire terms: "[G]ay people who vote for, or otherwise support, Republican candidates engage in acts of self-immolation."

Subtract the hyperbole, and one is left with two questions: Are these conclusions about state court superiority justified by the data? And if they are substantially supported by the data, what are the broader meanings for the parity debate?

C. PROBLEMS OF MEASUREMENT

Recoding the cases and recalculating the numbers is beyond the scope of this review essay. Moreover, because I come to this review as an insider to law but an outsider to political science, I will not attempt recomputations of statistical significance and probabilities. Without engaging in that exercise, however, one can utilize the complete listing of cases that comprise the database, which is contained in Appendix 1.2 of Pinello's book, to assess whether the cases have been fairly counted and whether the comparisons make sense. One significant weakness is the questionable coding of cases as wins or losses; at least two federal cases seem to have been improperly counted as losses.

The second major problem in measurement is that differences in subject matter which are intrinsic to the divide between federal and state court jurisdiction were given little weight. Because federal courts are courts of limited subject matter jurisdiction and state courts are not, the kinds of cases that are adjudicated in the two systems are often different. All but three of the twenty-nine military cases, for example, were heard in federal court. Because of the strong deference courts give to military judgments when personnel policies are challenged, it is no surprise that courts upheld military policies that discriminated on the basis of sexual orientation in virtually every case: twenty-two of the twenty-six federal cases were recorded as losses. This was the only subject matter category of cases that skewed the overall outcome in the federal courts downward. Removing the military cases from the database would have resulted in a total of sixty-one federal court cases rather than eighty-seven, with eighteen rather than twenty-two wins for gay-rights plaintiffs, and forty-three rather than sixty-five losses. The change would have lifted the overall federal court mean from .256 to just under .3, still substantially below the mean for state court cases.

Two subject matter categories were, as a practical matter, as limited to state courts as military cases were to federal courts. Family law disputes accounted for roughly half of all the cases in Pinello's database, mostly consisting of custody, visitation, adoption and foster care ("CVAF") cases. All but four of the 234 family law cases were heard in state court. The second state court dominion was a smaller cluster (twenty-two cases) of challenges to criminal laws, mostly prohibitions on sodomy. After 1986, when the Supreme Court ruled in Bowers v. Hardwick that Georgia's sodomy law did not violate substantive due process protections under the U.S. Constitution, litigation of challenges to those laws in federal courts stopped. Thus, for most of the period under study, sodomy law and similar challenges were brought only in state courts and only on state constitutional law grounds.

Bill Rubenstein has suggested that the jurisdictional difference between state and federal courts, which in turn leads to the differential subject matter patterns in the two systems, may explain why state courts are more receptive than federal courts to gay rights claims. Specifically, he has asserted that the predominance of family law issues among lesbian and gay rights cases makes the assessment of which system is the better protector of individual rights different for gay rights litigators than for other civil rights lawyers. Proposing variations on each of Neuborne's arguments, Rubenstein argued that state court judges have greater technical competence in the subject matter areas most frequently involved in gay rights cases, especially family law; that there is a greater likelihood that state court judges will have dealt with lesbian and gay litigants in all types of cases, providing them with more positive attitudes toward this class of claims; and that because state judges are less insulated from local politics, there is a greater chance that they will respect the need for responsiveness to the political demands of minority groups.

Of Rubenstein's three reasons, only the second – that state court judges come into contact with lesbian and gay litigants more often than federal judges do – would help explain a relatively more open reaction to gay rights claims in state rather than in federal court. His first reason – that state judges have greater competence in family law matters – would tend to explain only the extent to which lesbian and gay parties obtained positive outcomes in the great mass of family cases. It is unclear whether Pinello's study, which found that the mean outcome in state courts in family law cases was approximately .535, only slightly better than 50%, can be said to support this argument. Of course, without the collective experience cited by Rubenstein, the success rate could have been lower. In any event, federal judges' presumed lack of expertise in family law would not seem consequential on the question of technical competence, since federal judges almost never hear family law cases. According to Pinello, Rubenstein's third proposed factor – that state court judges tend to have held other public office – cuts the other way. Pinello found that prior service in an elected position was a significant factor in predicting anti-gay decisions on the bench.

In searching for an appropriate body of cases upon which to base a comparison of state and federal court outcomes, what seems most fair is to find a subject matter area in which there are a reasonable number of cases litigated in each system, one which is not definitionally limited to one system or the other. Applying this criterion leaves the cases that Pinello classifies as raising (non-military) sexual orientation discrimination claims, of which there were thirty-two in federal court and thirty-eight in state court. Examining these cases might reveal patterns of pro- or anti-equality sentiments in federal and state court judges. At least, one might suppose, we wouldn't be comparing apples and oranges.

Or would we? Approximately two-thirds of the state court cases involved interpretations of state statutes prohibiting sexual orientation discrimination. Because there is no such federal statute, none of the federal court decisions did. Although a judge's proclivities could lead her to interpret a civil rights statute narrowly or generously, a state court judge would be starting from a position of deciding how to apply a legislatively determined policy, rather than working under the much more amorphous framework upon which a federal court would draw for constitutional interpretation.

If one pares the cases down further just to constitutional claims, however, the discrepancy between federal and state court systems holds. One methodological question that arises at this point is whether the sample size is sufficiently large to carry the full weight of Pinello's conclusions. Nonetheless, one must acknowledge that the same pattern of federal-state court difference is present both for all cases and for the body of the most directly comparable cases. Moreover, even if the state and federal courts were hearing different kinds of cases, if one looks only at the federal court database, the link between party affiliation and case outcome is still extreme.

A second way to approach the question of whether subject matters is to ask how Pinello's study compares with other studies of federal and state court outcomes in particular kinds of civil rights cases. Solimine and Walker found virtually no difference in outcome between state and federal courts adjudicating constitutional issues in criminal cases, or equal protection claims in civil cases. The only category of cases where federal courts produced pro-rights rulings with significantly greater frequency were those raising first amendment claims in civil cases.

Solimine and Walker's findings suggest support for a distinction that has emerged in several anecdotal accounts as well. In his own study, Neuborne described the cases with which he was concerned as those "advancing federal constitutional claims against local officials." In finding that "[o]n balance, the Establishment Clause plaintiff continues to be better off in federal court, most of the time," Doug Rendleman joined Rubenstein in speculating that the reason was that such parties tended to be iconoclasts rather than seekers of equality and fairness, and that federal judges might be more open to claims made by the former, whereas state court judges could be equivalent to federal judges for the latter types of claims. Rubenstein suggested that federal courts might be more receptive to liberty claims and state courts to equality claims, the latter often being more obviously relevant to the needs of daily life. In one real life example, these factors coalesced into a decision by plaintiffs' attorneys in Romer v. Evans to challenge an anti-gay amendment to the Colorado constitution in state, rather than federal court.

Despite intuitions like Rubenstein's and Rendleman's, no one has articulated persuasive reasons why state courts would do a better job than federal courts in anti-discrimination or equal protection cases. It is not clear whether federal courts were slower than state courts in recognizing gay equality claims, or whether the two systems were moving in opposite directions. Again, Pinello is not shy: "[S]tate courts did not necessarily become more progressive. Rather, the point of reference [federal courts] migrated far right."

D. QUESTIONS OF MEANING

One must consider the possibility that federal court superiority may have been simply a temporarily necessary aberration, which the Supreme Court was correct to jettison at the earliest appropriate moment. A number of scholars have argued that returning to parity as the operating norm was an appropriate course correction, restoring federal jurisdictional principles to the ex ante norm of reliance on state courts for the bulk of constitutional adjudication. Richard Fallon argued that the paradigm for jurisdiction has oscillated between two ideal types – the federalist model privileging state courts and the nationalist model preferring federal courts – producing a fluctuation which has served to foster philosophical debates, but which cannot account for the reality of how the federal and state court systems actually interrelate.

Although these concerns may instill some caution as to Pinello's conclusions about the greater receptiveness of state courts to gay rights claims, they do not refute them. At a minimum, his findings are strong enough that they should cause us to return to the question of whether a new reality of reverse disparity between federal and state courts has developed.

Despite his use of florid rhetoric at other places in the text, Pinello, ironically, understates his most important conclusion relevant to the parity debate: "[N]o one can make a credible argument that appointed, life-tenured federal judges necessarily are institutionally preferable [to state court judges] to resist majoritarian forces opposing minority rights." Later, he adds that the effects of single-party dominance in national elections demonstrate that "federal courts are far more susceptible to ideological manipulation by a single source than is the state bench."

If true, these assertions break down what have been the accepted parameters for parity analysis. All the voices in this debate have accepted that the political independence of the federal judiciary is the starting point for analysis and comparison. That was the Framers' intention: "[I]t may safely be affirmed that, together with the permanent tenure of their [federal judges'] offices, [Article III] affords a better prospect of their independence than is discoverable in the constitutions of any of the states, in regards to their own judges." Contemporary scholars who have defended state courts also have conceded that "[t]he superiority of a lifetime appointment in countering majoritarian pressures is rightly praised." There has been a near-universal mantra that lifetime appointment functions as the gold standard and benchmark against which other systems should be judged.

Pinello is making a fundamentally different, and more radical, claim: that the federal courts are systemically vulnerable to ideological manipulation; that as an institution, federal courts have become weaker guardians of freedom than state courts and may well remain so. This is a serious charge, and the degree of support for it in Gay Rights and American Law entitles it to serious consideration.

One possible way to understand the data is to see them as merely reflecting the inevitable direction of constitutional interpretation during a period when single-party dominance of the Executive Branch led to the appointment of generally conservative judges who were usually unreceptive to novel constitutional claims. One may believe that the trend Pinello identifies will be remedied by cycles in American politics, which will produce future eras of more liberal, pro-rights appointees who will push the pendulum back toward greater protection of individual rights. Or one may assume that lesbian and gay rights claims have been so extraordinarily controversial that they are not likely to constitute more than an exception to the rule, an outlier status that is rapidly changing.

Unfortunately, however, other recent studies confirm Pinello's analysis. Based on quantitative analysis of more than 45,000 federal district court opinions between 1933 and 1987, C.K. Rowland and Robert Carp found that not only were district judges appointed by Democrats more likely than those appointed by Republicans to issue a pro-rights ruling in civil rights and liberties cases, but that what had been a minor gap between the two groups in the earlier years of that period widened into a "chasm" starting in 1969. Between 1981 and 1986, the partisan difference in civil rights and liberties cases reached 18%, with Democratic appointees more than twice as likely to take pro-rights positions. A particularly compelling comparison is between Carter and Reagan appointees in race discrimination cases: Carter appointees upheld the position of minority litigants in 78% of the cases, Reagan appointees in 18%.

Although partisan differences do not emerge in every area of law, "an impressive array of empirical studies has tended to confirm that the political party of the appointing president or the judge makes a difference in a number of discrete subject matter areas." Rowland and Carp's findings of partisan predilection, for example, were consistent with Tracey George's analysis of en banc voting by judges on the U.S. Court of Appeals for the Fourth Circuit. One subject matter area where some of the sharpest differences have emerged is civil rights, especially cases in which plaintiffs seek to extend the equality principle to new situations. A study of federal appellate decisions beginning in the mid-1990s found that ideology (as defined by political party) consistently predicted a judge's vote in five categories of civil rights cases; the study did not test for voting in gay rights cases.

It is easy to see how the timing of this trend towards anti-civil rights decision making would have had a particularly acute impact on gay rights claims. Unlike race and sex discrimination plaintiffs, those challenging discriminatory actions against lesbians and gay men could not invoke the precedents set during the pro-civil rights era of the Warren Court. Instead, gay rights advocates were in the position of making arguments to establish or extend rights protections in the 1980s, just as a wave of ultraconservative federal judges was beginning to crest.

Although the extent to which President Clinton's appointees moderated the "conservatization" of the federal bench during the 1990s is unclear, one should hesitate before reaching the conclusion that reverse disparity is simply a phase of only momentary import. One must address the possibility that there are deeper ramifications. Consider this less sanguine explanation: Gay rights claims are the canary in the cage, the sentinel marker of a breakdown in one central function of the federal courts. If this is true, it is a signal that the extreme focus on partisanship and ideology in federal judicial appointments for the last thirty-five years has inflicted serious, long-term damage to the integrity of the federal judiciary.

What is truly important here is not whether federal courts are systematically hostile to any particular kind of claim, or even the number of disfavored claims. The fundamental question is whether the wall of judicial independence has been breached in any reliable pattern, sufficient to serve as an invitation for whatever political impulses might produce future assaults.

Whatever happens to the myth of parity, the time has come to bury the myth of insulation. There is no longer any reliable buffer between the federal courts and electoral politics. This loss is a change which has been gradually developing and has now emerged in full force. It upends the basic terms of the debate about the relative value of the federal and state courts because it overthrows virtually the only component of the comparison about which all the parties agreed. Moreover, there is no sign on the horizon of a mitigation in the level of partisanship that produced this result. If nothing else, perhaps Gay Rights and American Law will help force a confrontation with this reality.

II. THE EMPIRICISM QUESTION

A. THE LAW, POLITICS AND SCIENCE DEBATE

Two schools of thought collide in the political science literature on law and the courts. The first, generally more familiar to lawyers, is known as the legalist school or legal model. Legalists seek to determine the forces driving judicial decisions through close textual readings that attempt to position judicial texts in various intellectual universes: legal doctrine, jurisprudential philosophy, and institutional role. This interpretive approach produces the vast majority of the law review literature, and saturates the study of law, not only by students, but also by the professoriate. Behavioralists, by contrast, contend that judges act based on their policy preferences, rather than in any attempt to adhere to doctrine or professional norms. To behavioralists, judges are simply political actors in judicial robes, advocates trapped in a neutral's body. They conceive of judicial voting as expressions of preference, and direct their energy to the coding and statistical analysis of those votes rather than to the interpretation of text.

I share the view of legalists that the structure and process of analogical reasoning do not easily disaggregate into measurable markers, which can then be recorded as present or absent in a given case. Precedent and doctrine shape decisions and constrain judges, even while allowing space for nonlegal factors to influence decisions, sometimes strongly. Ronald Dworkin, for example, argues that law exerts a "gravitational pull" on judges deciding cases. He describes judicial discretion in tiers: lower when the judge is operating under preset authoritative standards; higher in cases with little applicable precedent.

Further, from the perspective of critical theory, for example, one should focus on the structures of consciousness that shape all legal actors: the deeper cultural and psychic sources of attitudes which operate in the unconscious, as well as what Gillman calls "the structure of legal argumentation associated with particular conflicts at particular times." To critical theorists, only the details of law are determined by the adjudication of lawsuits or the actions of legislatures. The demands of power elites, filtered through deeply embedded structures of legal consciousness, set the parameters of what is or is not possible in the legal system.

Critical theorists combine skepticism toward the rule of law, which is at least as strong as that of the behavioralists, with the legalists' rejection of empirical methods. This approach precludes application of both the false consciousness intrinsic to traditional textual interpretation and the false neutrality of scientific method. Pinello's study, like all quantitative analyses, fails to account for this deeper critique.

Interwoven with the debate over theory are questions of method. Empiricism dominates the field of political science, and it is especially congenial to behaviorialist studies. Behavioralist analysis accepts almost as a given that doctrine and precedent have minimal impact on judicial decisionmaking. Behaviorialist studies typically identify independent variables, such as whether a judge is appointed or elected, and determine whether they are correlated with particular outcomes. Such studies rely on regression analyses to isolate relationships among certain variables and outcomes, and therefore require that those variables and outcomes be defined, coded, and entered into a database. In order to be coded, cases must be sorted into discrete categories, such as whether a judge was appointed by a Democrat or a Republican, or whether a judgment favors a plaintiff or a defendant.

Confidently armed with regression analysis, behavioralists seek to reveal exactly which kinds of "external," that is non-legal, factors have the most power to influence court decisions. Among the primary contenders are various rational choice models, which often focus on theories as to whether judges act strategically in their rulings, both to further favored policy goals and to maximize institutional power vis-a-vis legislatures and other government entities. Other frequently studied approaches center on judges' personal characteristics, such as political leanings, demographic characteristics, and indicia of personality or other psychological traits.

Arguments over methodology reverberate in debates over theory, and vice versa. Behavioralists automatically reject arguments that cannot be "proved" in the sense of mathematical analysis and replication. For behavioralists, the sort of textual analysis that permeates law reviews is largely speculative and anecdotal. To legalists, the empirical methodology so appealing to behaviorialists results in a dumbing down of understandings of law. The contrast between empiricism's dominance of political science and the still only budding openness to empirical studies among those in the legal academy explains much of the silence between scholars in the two fields.

A good example of the internal conflict over theory and method among political scientists lies in the long effort to prove whether judges are governed by precedent. Two of the scholars most frequently associated with behaviorialism – Harold Spaeth and Jeffrey Segal – published a study in 1999 in which they reviewed 2,425 Supreme Court cases with dissenting opinions to see if the dissenting Justices accepted and applied the ruling of the initial case when similar issues arose in later cases. They found that such shifts occurred only about 12% of the time. Their conclusion presented a dramatic challenge to legalists and the profession generally:

Stare decisis is the lifeblood of the legal model, and the legal model is still the lifeblood of most legal scholars' thinking about law. Yet there has been virtually no real testing of the model ... [W]e must conclude this book with what the data so strongly suggest: in the realm of stare decisis, minority will does not defer to majority rule.

Spaeth and Segal's challenge to legalists – that stare decisis was an empty shibboleth with virtually no importance – was then itself attacked on both methodological and conceptual grounds. Attempts to test and replicate its findings produced significantly different results. One study recoded what counted as preferential behavior (behavior not bound by precedent), and reduced the Spaeth-Segal result by almost 20%. Another study singled out four moderate Justices for analysis and found an even greater reduction in the preferential voting of those Justices. Howard Gillman argued that the Spaeth-Segal study had a fatal flaw at its core: the presumption that Justices of the Supreme Court have the same obligation to adhere to precedent that judges in lower courts do. On Gillman's understanding, a Justice who persists in his or her dissenting view can no more be characterized as engaged in preferential voting than the Justices who voted in the initial case to declare what became the majority view.

It is in this debate that Gay Rights and American Law seeks to intervene. Pinello's strategy is to attempt a sophisticated synthesis of legalist and behavioralist theories. His goal is to ascertain the extent to which personal, institutional and cultural characteristics drive outcome, without excluding the role of precedent and doctrine. In this world, his conclusion that stare decisis and precedent do matter is a challenge to the conventional wisdom.

B. ATTITUDES AND LAW IN GAY RIGHTS DECISIONS

Pinello does not state any particular hypotheses as beginning points for his study, but he does articulate two goals: to analyze the impact of various factors on judicial decision-making in the adjudication of gay rights claims, and to formulate an improved methodology for "the meaningful quantitative empirical study of law." As to the latter, which serves as a path to the former, he aspires to meld the competing legal and attitudinal schools of scholarship in this field, and, by combining the two, produce a result which is greater, that is, more sophisticated, than the sum of its parts. His attempt to ratchet up the complexity and quality of empirical scholarship rests on both the incorporation of a greater number of attitudinal variables and a deeper reading (and presumably understanding) of judicial reasoning.

Pinello found that both legal and attitudinal factors are statistically significant, and concluded that the two models are complementary rather than mutually exclusive. Contrary to the claims of Spaeth and Segal, Pinello found that stare decisis matters, but that it matters differently at different levels of the judiciary. Pinello ran separate comparisons for courts of last resort and for intermediate appellate courts of four clusters of variables: legal, attitudinal, environmental and institutional. Not surprisingly, legal variables were the most important determinants of outcome in intermediate appellate courts, and the least important in courts of last resort. At the highest level of appeals, environmental and attitudinal factors yielded the greatest predictive capacity. The impact of attitudinal factors was greatest, and the impact of law lowest, in the family law cases.

The category of attitudinal variables included a broad range of characteristics of the judges themselves, and the particular attitudinal variables that Pinello found to be most salient will surprise no one. More than any factor except party affiliation, racial diversity on the bench helped those litigants pressing civil rights claims for lesbians and gay men. The second most significant factor was the presence of Jewish judges. To a lesser extent, the presence of women judges was also significant, especially in the CVAF cases. The statistical power of gender as a factor in producing pro-gay-rights results was roughly equaled by the statistical significance of a Roman Catholic religious affiliation in producing negative results. Prior service in elective office (other than as a judge) correlated strongly with negative outcomes. Age almost, but not quite, met the criteria for statistical significance.

Of the remaining factors studied, Pinello found that environmental ideology had the strongest predictive value. He scored states on the degree of political liberalism or conservatism that dominated each state's political culture based on an earlier political science study that measured "citizen and government ideology," primarily by election results. Similarly, Pinello found that statutory law – either the (positive) presence of an anti-discrimination law or the (negative) presence of a sodomy law – correlated with outcome in the whole range of gay rights cases. Examining institutional factors, i.e. the attributes of the court systems, produced mixed and confusing results. Overall, appointed judges were not more liberal than judges subject to election, although that outcome shifted with term length and environmental ideology. For the most part, a longer term length correlated with more pro-gay outcomes, although that was not true for merit-selected judges.

The patterns become somewhat more interesting when one examines which factors are significant at different levels of appeal or in different categories of cases. For example, the presence of judges of color and Jewish judges was significant at both levels of appeal, while the presence of female judges registered such an impact only in courts of last resort, and age and Catholicism mattered at the level of statistical significance only in intermediate courts. All five of those characteristics carried statistical significance in the CVAF category of cases, including at the appellate level. In the broader category of all family cases, however, Catholicism was not a significantly predictive characteristic, although it reappeared as such in the non-family law category.

What does all this mean, and does it mean anything more than what most readers would consider to be statements of the obvious? Lawyers will be surprised neither that precedent carries the most weight in predicting the outcome of litigation, nor that a judge's personal characteristics will often incline her to rule for or against certain types of claims. In family law, where greater discretion attaches to judicial decisionmaking in part because of broad standards such as the best interests of the child, a judge's life experiences and sense of proper norms are likely to carry more weight than in most other matters.

As Pinello states, his findings translate into statements of probabilities; the many exceptions that spring to mind do not disprove the degree of probability. In the end, Pinello's claim is rather modest: that his models "predict case outcomes with reasonable degrees of accuracy." He measured his own system by calculating the probability for the outcome of each case, based on the statistical models, and found that for 48.3% of the cases, the outcome was essentially too close to call. The models accurately predicted the outcome for 41.4% of the cases, and got it wrong 10.3% of the time.

Perhaps the most significant conclusion in the book is the unpredictability of gay rights cases in the time period that Pinello studied. For almost half of the cases, Pinello found that there was, in essence, a 50-50 chance of success or failure. Not even his complex array of variables could produce a prediction for half of the gay rights litigation efforts in the United States at the end of the twentieth century.

C. A CRITIQUE OF DEFINITIONS

One of the recurring confusions at the heart of Pinello's study is the relationship between party and claim. Pinello describes his study as one of claims – how the courts adjudicated claims of lesbian and gay rights. It does not purport to be a study of how the courts treated parties – litigants known for some reason to be lesbian or gay who were involved in any and all manner of disputes that made their way into the court system. Yet the confusion between the two is evident in Pinello's own description of what was being studied:

I coded court decisions as 1 if decided in favor of the lesbian or gay claim asserted or defended and 0 if against. For most cases, the coding process was forthright, in that a homosexual litigant clearly won or lost. However, when there was no such litigant, but a decision nonetheless affected the rights of gay people as a class, the coding rule became whether the court treated homosexuals as the legal equal of heterosexuals."

Obviously types of claims and types of litigants can merge in many instances. A lesbian mother fighting for custody, for example, may assert that the best interests of the child require placement with her, yet consideration of the impact of her sexuality becomes inextricably bound up with the assessment of best interests. Such a case does not involve assertion of a gay rights "claim" in the narrow sense of how a cause of action is framed, in the way that a case challenging a firing based on sexual orientation would. Yet the custody case is as relevant to a study of anti-gay judicial bias as the employment case. Thus one can understand how Pinello had to consider facts, law, and party identity in determining which cases met the definition for inclusion.

Although the answers to this conundrum are not easy, unfortunately Pinello seems to have resolved it in ways that are sometimes both simplistic and inconsistent. Beyond the basic subject matter categories, Pinello injected his own assessment of the importance of various kinds of cases. He considered certain cases to be essential and others nonessential. All of his conclusions are based upon a sample of only the "essential" cases. The issue of which legal disputes are essential to the treatment of gay rights claims cuts to the core of a deeper definition: what it is that Pinello's study is designed to investigate. Pinello's decision to treat same-sex harassment cases as nonessential illustrates how easily cases with important ramifications for changing the law's regulation of homosexuality can be missed if neither the cause of action nor the plaintiff is identified as gay.

The largest set of cases excluded as nonessential was a set of thirty-one same-sex sexual harassment cases. In this kind of lawsuit, a plaintiff sues complaining of harassment "because of sex" by a person or persons of the same sex as the plaintiff. These cases can be a fascinating window into legal and social understandings of gender and discrimination, and they are winnable. Same-sex sexual harassment cases had the highest (the most pro-gay) mean outcome (.742) of any other specific subject matter category. Moreover, they are the fastest-growing category of cases related to sexuality; a comprehensive listing of 107 such cases in federal courts reveals that there were virtually none prior to 1990.

Why were the same-sex sexual harassment cases excluded? Pinello asserts that "this area of the law evolved in large measure to protect heterosexual males from being solicited by gay men." He then notes that "virtually all targets of the alleged harassment are either explicitly described as heterosexual or presumptively so." He concludes that the high mean outcome provides evidence that "clearly" judges viewed this body of law as a penalty for gay men soliciting heterosexual men, and not as an extension of civil rights protection for sexual minorities. Pinello points to one case in which the Fourth Circuit interpreted "because of sex" to mean that male-male harassment claims would be cognizable only when the harasser was gay. No other federal court adopted that interpretation, however, and the Supreme Court rejected it, in another case included in Pinello's database.

Pinello's characterization seriously misreads the sexual harassment cases. Many of the cases that Pinello excludes involve efforts to protect men from harassment, often including assault, by other men who use the accusation of homosexuality or effeminacy as a means to persecute their targets. Some of the plaintiffs assert their heterosexuality; and for others, the record on this point is fuzzy. What is politically important about these cases is that they provide an opening for pressing the connection between degrading women because they are female and degrading men who are perceived as in some way un-masculine.

It is true that in quid pro quo harassment cases, in which the harasser is seeking to pressure the target into having sexual relations, the typical same-sex case would fall into the scenario that Pinello describes: A male seeking to pressure another male into a sexual encounter would likely be gay, and the target might be straight. But in hostile environment harassment, harassers often operate in groups and are not seeking sex, but are attempting to use some form of sexualized abusiveness to drive the target out of the workplace. The dynamic is the opposite of sexual attraction; it is intimidation that can border on terrorism. Many of the cases in Pinello's sample fall into that category of hostile environment cases in which verbal or physical abuse drew on anti-gay hatred.

The men in those cases, whether straight or gay or of ambiguous orientation, are suing to broaden the scope of sex discrimination so that it includes discrimination based on gender stereotypes. Using the principles of sex discrimination law to challenge anti-gay policies has become a major theme in sexual orientation law, a strategy that began before the time period of the study. Thus excluding this entire body of cases signifies a mechanistic and fairly narrow conception of what the study sought to capture.

Defamation cases provide another example of oversimplification of the range of cases that might be classed as discrimination cases. Defamation cases were also excluded by Pinello as nonessential because most involved plaintiffs asserting harm because of a false attribution to them of homosexuality. Admittedly, these cases are complex; most of the plaintiffs are seeking to renounce any association with homosexuality. Pinello codes them as victories for gay rights if the court rejects the principle that false "accusations" of homosexuality are slander per se, that is, considered defamatory without proof of special damages. Yet he excludes them from the database upon which his findings are based, a coding which overlooks the deeper meaning of the cases. Rulings that rejected the common law defamation per se approach required judges to accept that they "should not classify homosexuals with those miscreants who have engaged in actions that deserve the reprobation and scorn which is implicitly a part of the slander/libel per se classifications." Pinello, however, concludes that "the bulk of decisions ... involve heterosexuals concerned about false accusations of homosexuality. Again, this isn't a lesbian and gay rights category."

In other instances, however, plaintiffs asserting no connection with homosexuality were counted as presenting essential gay rights claims. Powell v. State, in which the Georgia Supreme Court struck down the state's sodomy law, is such a case. Powell arose from the criminal prosecution of a man accused of rape and sodomy; the victim was a woman. There is no suggestion of homosexuality, nor does the theory of the case suggest an objection to the social codes of gender. The constitutionality of the sodomy law was before the court because the defendant was acquitted of rape but convicted of sodomy, a crime for which consent was no defense. Because the statute covered acts by homosexuals as well as by heterosexuals, its invalidation eliminated prosecutions of gay people as well.

Powell was one of a series of sodomy and related criminal law challenges, almost all of which did involve gay litigants who asserted recognizably gay rights claims. I do not fault Pinello for including Powell; an analysis of sodomy law challenges during this period would be incomplete without it. It is equally true, however, that if one is seeking to measure the degree of hostility or openness to assertions of gay equality, such a study is also incomplete if it excludes cases ruling on whether men tormented as "fags" could bring anti-discrimination claims.

Another borderline set of cases involved challenges to municipal laws granting partner benefits to city employees. Often after such a law was passed, a group of conservatives sued the city to challenge its validity on the ground that state insurance law preempted it. Such lawsuits were brought in seven states during the time period of this study, and all are included in the database as essential lesbian and gay rights claims, under the category of family law. Here, the underlying subject matter concerned issues of equality for gay employees, but neither the identity of the litigants nor the legal claims and defenses directly implicated sexual orientation per se. One can justify the inclusion on the ground that assessing a court's adjudication of these lawsuits unavoidably involved the context of gay political action. Yet there is something strange in coding a court's decision as essential to gay rights, at least under Pinello's definition, when its text discusses only issues of home rule and preemption.

The complexities of how the state has regulated homosexuality make definitional questions quite difficult. They also may make this area of law an especially problematic one for empirical studies in which each case must be coded into either-or categories – either a "real" gay rights claim or not, either a victory or not. Pinello's efforts at categorization do not overcome problems inherent in the methodology. Moreover, with no explanation, Pinello excluded all AIDS cases, and does not mention transgender cases. After reading Gay Rights and American Law, one is left with the intriguing question of whether one can even confidently identify which judicial texts comprise America's law on sexual orientation.

D. A CRITIQUE OF APPLICATIONS

Statistical proof of the saliency of nonlegal variables such as race, gender, religion, age or environmental ideology demonstrates patterns in the relationship between case outcomes and the particular variables. Sufficiently strong correlations yield probabilities, and form the basis for a model that can predict the outcomes of cases with "reasonable accuracy." To most lawyers, this amounts to a lot of proof for conventional wisdom about who would constitute a tough bench for a gay rights lawyer. But most lawyers would expect that, except in rare cases, even a tough bench would not ignore the law.

Pinello attempts to account for the power of precedent by coding some cases as controlled by precedent; he concludes, as noted earlier, that stare decisis is the single most important factor influencing decisions, especially in intermediate appellate courts. But as any lawyer knows, precedent is not a binary game – there are degrees of precedent. Unfortunately, the study makes no attempt to distinguish cases on that basis. As a result, statements about causation in many cases leap from a threshold finding that the court was not limited by binding precedent to conclusions that characteristics of the judges must surely explain the outcome.

These logical gaps occur because Pinello's definitions of legal variables are seriously impoverished. Unlike the Spaeth-Segal study, this book does not pinpoint particular decisions and then examine whether judges in later cases applied the precedent established by those particular holdings. Rather, Pinello identifies a particular outcome – the vindication or rejection of gay rights claims – and seeks to assess why those outcomes occurred. Even assuming a correct assessment of what the outcome in each case was in those terms, this method will not capture the role of particular cases in the creation of legal doctrines which may be applied elsewhere.

In Gay Rights and American Law, the variables which mark the impact of the law, as opposed to characteristics of the judges or of their environment or of the particular judicial system, take account only of whether "the same or higher court decided comparable facts authoritatively." Pinello supplies a list of "cases decided under the influence of controlling precedent"; he apparently simply accepted each court's declaration as to what was "controlling precedent." He then used six codes for precedent: negative or positive from a higher court, negative or positive from the same court, and negative or positive from another district.

As a result of this methodology, and as is evident from his list of "controlling precedents," the only precedents that counted were other gay rights cases. Apparently, only cases involving homosexuality were thought to involve "comparable facts." It would have been considerably more interesting had Pinello looked beneath the surface of what each court declared was controlling precedent to offer an analysis of how legitimately controlling it was. Here, even more than in the assignment of cases to the essential or nonessential categories, a narrow view of analogy drains the study of much of its potential value. Under such an approach, the very structure of the study points to an over-registering of attitudes toward homosexuality. Pinello concludes that "[j]udicial awareness of the political, religious, and social imperatives surrounding rights claims by lesbians and gay men in late-twentieth-century America transcended the discrete subject matters" of the cases. In other words, the reader is instructed to assume that anti-gay or pro-gay judges brought often decisive attitudes to the adjudication of the entire variety of disputes in the study.

No doubt there are instances where such attitudes did determine the outcome, especially in cases where the controlling doctrine was as malleable as the "best interests of the child" principle. But without consideration of the full range of precedent with which courts were dealing, in addition to prior gay rights cases, some of Pinello's conclusions seem absurd. For example, Pinello compares two cases involving gay partners. In Posik v. Layton, a lesbian couple split up and one partner sought to enforce a written contract that promised her $2,500 a month for life – in essence, alimony – if the other partner breached the terms of the contract, which she did. The Florida appellate court enforced the contract. In Matter of Cooper, a New York court refused to treat the surviving partner in a gay couple as entitled to the statutory share of the decedent's estate to which a surviving spouse would have been entitled.

Pinello's analysis begins with what most lawyers would consider to be more than a rhetorical question: "Was the fact that Emma Posik documented her relationship with Nancy Layton by means of a written contract – and Ernest Chin did not do the same regarding his with William Cooper – the dispositive issue causing Posik to win and Chin to lose?" Quite possibly yes, you might think, or perhaps the outcome turned on the fact that the New York court would have had to interpret the statutory phrase "surviving spouse" to include the gay partner, while the Florida court had only to enforce a contract.

Pinello's analysis of this case, however, does not even bother to answer the question he posed. He supplements the question by noting that environmental ideology factors – the dominant political gestalt in New York versus that in Florida, plus the fact that Florida had a sodomy law and New York did not – would lead one to expect the reverse pair of outcomes: a gay rights win in New York and a loss in Florida. He then explains the anomaly as follows:

What happened?

Attitudinal forces most likely overcame the environmental context. The four New York appellate judges denying Chin's claim were Roman Catholic, while the Florida jurists acknowledging the worth of Posik's relationship with Layton were Episcopalian and Presbyterian. Catholic Church dogma is less tolerant of gays than the policies of liberal Protestant denominations such as the Episcopal and Presbyterian churches. In short, the attitudinal force of religion apparently played a decisive role in determining these domestic partners' legal rights.

What is more, three of the four New York judges served in prior non-judicial elective office, while none of the Florida jurists did.

The book is peppered with other examples of the author's apparent rush to reach certain conclusions. His analysis of Commonwealth v. Wasson, for example, is presented as self-evident when it is far from persuasive. In Wasson, the Supreme Court of Kentucky ruled that the state's sodomy law was unconstitutional under both the due process and the equal protection provisions of the state constitution. Like the five-four split in the United States Supreme Court in Hardwick, there was a one-vote margin in the Kentucky Supreme Court, but in the opposite direction: four to invalidate, three to uphold. Three of the court's Democratic appointees together with one Republican were in the majority; two Democrats and one Republican dissented. Pinello points out that the majority consisted of one Jewish judge and three Protestants (labeled as "moderate" and "liberal," based on their denominations); the dissenters included two Catholics and one "fundamentalist" Protestant. To Pinello, the reason for the outcome is clear: "Religion offers the only interpretive prism to understand Wasson's split result."

One obvious way to test for the plausibility of these explanations would be to inquire whether the same judges ruled differently in other cases involving gay rights claims. Inconsistency in outcomes reached by an individual judge would suggest that some variable associated with the particular case had a decisive impact as well, as calling into question what variables should be counted as decisive. There is no attempt in Gay Rights and American Law to test for this. Although re-analyzing the results of more than 400 cases is beyond the scope of this review, some examples stand out. Presumably without changing their biases or personal characteristics, judges do vote inconsistently – both pro- and antigay – a reality that a model such as Pinello's cannot explain. Thus, Judge Flaum in the United States Court of Appeals for the Seventh Circuit is seemingly engaged in schizoid jurisprudence, as are Judges Foley and Forsberg, of the Minnesota Court of Appeals; Judge Crawford of the Tennessee Court of Appeals; and Judge Kennedy of the Washington Court of Appeals.

Selection bias and coding errors come with the territory of this kind of empirical study. Large numbers of unremarkable cases that could shed light on judicial attitudes, such as prosecutions for solicitation, are missing from the database because they were defined out of the category of gay rights claims. Per curiam rulings or judicial decisions to deny review escaped consideration because votes by individual Justices were not identifiable. Coding – even when done correctly – eliminates the capacity to consider highly suggestive changes in the tone of a court's analysis, such as the contrast between the contemptuous language of the Supreme Court in Bowers v. Hardwick and the respectful tone of Hurley v. Irish-American Gay, Lesbian, and Bisexual Group of Boston, because both were losses. This difference, however, sent a signal that made Romer v. Evans much more predictable than Pinello's system would suggest.

The book's strength is its insistence that a rich mix of variables lies behind the construction of social rules – the law – on important political matters. Ironically, however, despite its elaborate statistical analyses, it fails to capture the complexity of law.

III. JUDICIAL POWER AND POLITICS

As with most studies, the contribution of Gay Rights and American Law will not be to answer with certitude the questions which it addresses. Nor will it end the debates, such as that over parity between federal and state courts, which others have sought unsuccessfully to terminate. Rather, its primary value will lie in whether it moves the study of judicial power and politics to new or different or more refined questions.

A. THE POLITICS AND NORMS OF JUDICIAL SELECTION

In Part I of the Review, I examined Pinello's assertion that partisan politics has distorted the process of selecting federal judges to the point of greatly weakening, if not extinguishing, what has long been a cardinal principle of our judiciary: that the structural safeguards afforded federal judges effectively insulate the federal bench from majoritarian pressures. In Part II, I analyzed the claim that, in both the state and federal systems, judges' policy preferences and their immediate (geographic) political environments drove their decisionmaking as much as legal doctrine and, in courts of last resort, more than legal doctrine. If the claim examined in Part II is correct, is the allegation described in Part I inevitable and unfixable?

If the non-law factors that Pinello found had the most statistical explanatory power to determine the outcomes of cases, then it is hardly any wonder that political leaders have intuited the same conclusion and acted accordingly. Perhaps what most legal scholars would regard as vulgar partisanship is in fact only pragmatic policymaking. If so, then we might need to consider developing a new process for federal judicial appointments, and begin arguing for a constitutional amendment to enact it. Or, less grandly, perhaps we should rethink what factors we should accept as the significant attributes for Presidents to consider in nominating and for the Senate to consider in confirming judicial nominees.

For example, judicial nominees could be asked to disclose if they belong to any clubs or other entities that exclude gay people, as they are now asked to disclose whether they belong to clubs that exclude women, persons of color, or members of religious minorities. Simply asking the question would be an enormous step forward, and would signal to lawyers hoping for such nominations that sexual orientation would be treated in at least some ways as on a par with other bases for bias, regardless of the source, religious or otherwise, of nominees' anti-gay beliefs. More powerful would be the addition of sexual orientation to anti-discrimination clauses in all judicial codes of ethics. Although direct enforcement of such codes seldom occurs, they serve an important educative function.

What is problematic about Gay Rights and American Law is not that it would help to puncture the hypocrisy of current political discourse about judicial nominations, in which Republicans and Democrats trade mutual accusations that the other allows ideology to dictate their positions on federal judgeships, while using coded terms such as "judicial activists" and "mainstream" to promise key constituency groups that only those with politically agreeable philosophies will be considered for nomination or accepted through confirmation. Rather, the risk is that the book could invite even more stereotyping by conflating too completely ideology and demographics. Although Pinello fully acknowledges that there are many individual exceptions to his statistical averages, the zeal with which he attacks certain backgrounds as predictive of bias is somewhat disturbing.

A case in point is the manner in which he describes Roman Catholic judges. Although religious background is only slightly more powerful than gender in predicting a judge's votes, Pinello himself lapses into what seems like significant hostility in how he describes this factor: "antihomosexual church dogma filtered into Catholic judges' official action"; "these data provide striking evidence of Catholic judges' hostility to lesbian and gay rights"; "Roman Catholic judges were some of the most hostile to the civil rights of homosexuals"; "Catholic judges deciding civil rights cases ... may have acted on their church's antipathy to lesbians and gays."

His repetition and choice of language is unfortunate, because it obscures an important and legitimate question about the impact of religion on judging. In one of the most interesting chapters in the book, Pinello describes the controversy that erupted when he sought to obtain information about the religious affiliation of some judges for whom there was no public documentation. To his credit, Pinello reprints extensive comments by colleagues both attacking and supporting him. What is most dramatic is how opposite the reactions are of political scientists and legal scholars. Pinello sent the names of twenty-five judges for whom he lacked a religious identification to a listserve of political science professors, requesting assistance in obtaining personal information about any of the judges. Not one of the political scientists who responded criticized Pinello for seeking the information. He then sent the same request to a listserve of law professors. The second request triggered a lengthy back-and-forth as to whether Pinello was attempting to invade the privacy of the judges or was violating professional or cultural norms in even suggesting that religious affiliation could shape a judge's decisions on matters of law.

The exchange is fascinating for what it tells us about the sharpness of disciplinary differences in norms and beliefs about what information constitutes legitimate knowledge. This gap probably reflects the contrast between assumptions common to empiricists oriented to pursuing whatever variables might logically be expected to produce greater than random variations in results, as opposed to the assumptions of interpretivists who find empirical approaches inherently reductionist. But the difference in reactions to Pinello's inquiry may also suggest a hesitancy to ascribe public and professional actions to private religious philosophy.

Of course Pinello was not suggesting that Roman Catholic judicial nominees should be disqualified on account of their religion. Nor, given the intensive individualized scrutiny that such nominees receive, does there seem to be a genuine risk that a blanket presumption of bias will attach to Catholic nominees. If there is any detailed and intensive "whole person" examination of candidates for any position, it is surely the process which has evolved for federal judicial nominations.

But the question of whether a nominee's religion should be a question raises a host of discomforting secondary issues. If it is not an invasion of privacy to attempt to learn the religious affiliation of a candidate for a lifetime appointment to the federal bench (who has chosen not to list one publicly), on the theory that one's religious philosophy should be considered a legitimate factor to include in the public knowledge about that candidate, why would it be wrong to inquire as to sexual orientation? Shouldn't the same principle control both? David Richards has drawn an extended analogy between sexual orientation and religion, arguing that gay rights claims are best understood as expressions of a right to conscience.

In Lawrence v. Texas, the Supreme Court removed the last vestige of criminality from homosexuality. Currently anti-gay stigma is still strong, and other legal disabilities short of criminal penalties attach, such as the exclusion from marriage. As those diminish further, however, sexual orientation will become just another characteristic like religion, as suitable for coding, counting, and statistical analysis as race or gender.

B. OF JUDGES AND LEGISLATURES

Gay Rights and American Law is a study of courts and judicial decision-making, seen through the prism of gay rights cases. As I suggested in Part I, however, it may well make its biggest mark as a contribution to the literature offering comparative assessments of federal and state courts. However, its coverage of "American law" is incomplete; to be correct, the title should read "American case law." What is missing is coverage of the role of legislatures and of statutory law. The impassioned advocate battling over constitutional rights in major test-case litigation may be the traditional image of the civil rights lawyer, but any analysis of contemporary law-reform campaigns is seriously lacking unless it accounts not only for the shift of a significant amount of litigation from federal to state courts, but also for the change in venue, for many breakthroughs, from courtrooms to legislatures.

The relationship between the judicial and legislative branches forms a critical part of the story, as much as the relationship between federal and state systems. This is true both in the process of change and in the role of different products of change. The process by which laws change, especially in controversial areas of liberty and equality, often involves a back-and-forth between judicial and legislative action, a dance which shapes the strategies of all participants – judges, legislators, litigators and legislative advocates. Statutes, as products of parts of that dance, have a profound impact on litigation. This interbranch dynamic has probably been more important in the campaign for gay civil rights than it was in any prior civil rights movement.

Studies such as Pinello's miss the process point by depicting judges as both powerful enough to impose their policy preferences and as strangely passive, acting in the thrall of attitudes and biases that overwhelm rational arguments. Yet judges are also strategic agents, not just in securing particular policy outcomes, but also in the more sophisticated ways that actors in complex institutions shape their behavior in order to provide incentives for other actors to behave in certain ways. Judges can, and perhaps should, be alert to the actions and reactions of legislatures.

Judges may need to engage in statesmanship, as well as interpretive dexterity, to ensure that the ultimate outcome in public law will be what they determine to be the legally proper one. Ironically, it is difficult to think of a more salient example of judicial statesmanship vis-a-vis the legislature than Baker v. Vermont, the challenge to that state's exclusion of same-sex couples from marriage, and one of the cases which fell within the boundaries of Pinello's database. In Baker, the court in effect "remanded" the case to the legislature, by suspending its judgment to give the legislative branch time to enact a civil union law. Yet the extraordinary story of the debate between the justices on Vermont's supreme court and their consciousness of the legislature cannot be reduced to empirical variables, and thus its richness is lost in what purports to be a definitive study of gay rights law during this period.

Although Pinello documents that whether a state has an anti-discrimination statute covering sexual orientation or a criminal law prohibiting sodomy has an impact on judicial decisions, his methodology does not capture the full import of this factor. To his credit, deciding to measure it at all signals a much deeper understanding of the extent to which contemporary law is dominated by statutes than shown by other analyses of judicial decisions. Even the comprehensive Spaeth-Segal study acknowledged this factor only by dividing its sample of cases into those interpreting the Constitution and those interpreting statutes.

The assumption that test-case litigation which produces invalidation of discriminatory government policies operates as the leading edge of a law-reform effort applies much more accurately to race and gender civil rights movements than to the gay rights movement. In part because anti-discrimination statutes were enacted throughout the United States during the race and gender movements, it was often easier for gay rights advocates to secure amendments of those already-existing statutes than to secure positive interpretations of constitutional guarantees in the courts. Thus, much more so than for race or gender issues, statutory protection against discrimination based on sexual orientation often preceded constitutional protection.

Lastly, study of the institutional dimensions of civil rights law should not overlook the role of local government. Advocacy of gay equality incubated in legislatures, especially at the municipal level. This history demonstrates that a majoritarian municipal body can function as a counter-majoritarian institution vis-a-vis the state, similar in some respects to the dynamic between legislatures and courts.

Gay rights cases provide a fertile context for scholars to probe even further into how actions of the separate branches, as well as of the different levels in the federal system, affect the development of new individual rights that later come to be seen as fundamental.

C. THE LAWRENCE QUESTION: A PARADIGM SHIFT?

Lastly, one might ask whether this study of gay rights law, and especially its gloomy conclusions about federal court supineness, continues to hold water after the Supreme Court's decision in Lawrence v. Texas. It is too soon to know what the full ramifications of Lawrence will be. To date, the only courts to interpret it have done so narrowly. No one can yet confidently assert whether Lawrence will make the "dis-parity" analysis of Part I obsolete, but there are reasons to doubt that it will.

Historically, the Supreme Court has not functioned as the leading edge of individual freedom. Even during the Warren Court era, the Court was never far out of sync with popular opinion. Most of the Court's most radical opinions – on school desegregation and interracial marriage, for example – were truly radical for their social consequences only in the South, where they inflicted body blows to a legal system organized around racial apartheid. The Court functioned in these cases to eliminate anachronisms, laws that had lost their claim to legitimacy after a long period of social and legal change.

Lawrence can be understood as having performed the same function. In 1986, when the Court decided Bowers v. Hardwick, twenty-four states and the District of Columbia had sodomy laws. By the time Lawrence was decided, that number was cut in half, to thirteen. When Brown was decided in 1954, the number of states that had de jure segregated public school systems was only slightly higher, and they were concentrated in one region. Indeed, in 2003 almost as many states had laws that prohibited sexual orientation discrimination as had sodomy laws. In 1986, Georgia's sodomy law was not an outlier; in 2003, Texas's sodomy law was.

My sense is that the Supreme Court engages in transformative instances of judicial review only after there has been a political tipping point, a signal from some other center of state power that a particular counter-majoritarian claim has achieved a critical mass of political support. I think that there is critical significance in the fact that the signal emanates from a separate institution of government, and not merely from the cumulative impact of developments within the same institution, such as a trend in lower federal court opinions. Following this instinct, I would identify President Truman's Executive Order in 1948 desegregating the armed forces as the institutional tipping point that enabled the Court's decision in Brown; and Congressional passage of the Equal Rights Amendment as the tipping point that presaged Roe v. Wade. Assuming at least for the moment that Lawrence is entitled to inclusion in the list of major judicial paradigm shifts, what served as the tipping point? There is no federal statute that prohibits sexual orientation discrimination; President Clinton's attempt to end the exclusion of openly gay and lesbian service members failed dramatically.

As I look back over the period that led to Lawrence, the same period under study in Gay Rights and American Law, I would locate the tipping point in the system that Pinello rightly, if perhaps overheatedly, praises: the state courts. The remarkable series of state supreme court rulings overturning sodomy laws – including decisions by the courts in such conservative bastions as Arkansas, Georgia, Kentucky, Montana, and Tennessee – put the U.S. Supreme Court in the embarrassing position of itself seeming like the outlier, the holdout against an ever louder claim for individual freedom.

Having rectified this aberration in Lawrence, however, the Supreme Court, together with the lower federal courts, may or may not continue to expand the principle of equality for lesbian and gay Americans. It is depressingly obvious that the federal courts reflect a markedly different perspective on civil rights today than they did in the years after Brown or even in the first years after Roe. Many federal court judges in the South during the 1950s and 1960s placed their allegiance to the Constitution above more parochial concerns. My own sad guess is that many federal judges of our era, in ruling on this equality claim, will not rise to the occasion, and may well continue instead to trail behind members of the state judiciary, even in their home jurisdictions.

IV. CONCLUSION

Gay Rights and American Law is both useful and disappointing. It entailed an enormous documentation project, and it makes a definite contribution to our understanding of judicial reactions to lesbian and gay rights claims during the 1980s and 1990s. In the end, however, the book offers an interpretation lacking the analytic bite that such a dynamic area of law demands.