[N.B. Footnotes cannot be accommodated in HTML format. Accordingly, the 171 in this review have been removed. Readers should consult the original text for them.]

Michigan Law Review

2005 Survey of Books Related to the Law

Volume 103, Page 1186, May, 2005


Frank B. Cross

Herbert D. Kelleher Centennial Professor of Business Law

The University of Texas at Austin

Review of

Gay Rights and American Law

By Daniel R. Pinello

New York: Cambridge University Press. 2003. Pp. xv, 349. Cloth, $70; Paper, $23.

One can find many analyses of the development of gay rights law in America but none are so illuminating as Daniel Pinello's in his book, Gay Rights and American Law. More significantly, while it offers a superb understanding of the recent record of gay rights litigation, the book provides a fine-grained and sophisticated understanding of judicial decisionmaking in this important and developing area of the law. Indeed, the value of the book for students of judicial decisionmaking even transcends its value for students of gay rights jurisprudence.

Quantitative empirical studies of judicial decisionmaking, well established in political science, have steadily burgeoned in the legal literature. Such research is enormously important to understanding the law. Unfortunately, much of this research has been incomplete and too discipline-centered. Thus, economists empirically study the law from their own perspective without sufficient appreciation of the claims of law and political science, while political scientists too often fail to appreciate the importance of legal and economic considerations. The disciplines have much to learn from one another. While legal researchers may have done less empirical work than those in other disciplines, and may have erred in the conduct of this research, they have been more integrative in this research and done a much better job of illuminating the importance of law itself. Pinello is a political scientist but conducts his research with an insightful appreciation of the law that is often lacking in political science research.

The context in which Pinello studies judicial decisionmaking is that of gay rights. His focus on this relatively narrow slice of the law helps us understand judicial decisionmaking more generally. First, as the focus narrows, one can better capture detailed variables that may be unavailable in a much broader context. Second, the narrower focus can show how the determinants of judicial decisionmaking are not uniform across the entire body of cases but may vary by case type. In the end, his book tells us a great deal about the development of gay rights law in America, much more than we can learn from a traditional analysis of cases and purely legal arguments.

The book begins with "narratives" of some of the interesting gay rights decisions, but I will focus on the empirical analyses. The crux of the book is in this extensive empirical research, which I summarize and place in the context of prior analyses in Section One of this Review. Section Two addresses the practical implications of the empirical findings of Pinello and others on legal and public-policy matters of contemporary concern. In Section Three, I examine the book's findings in the context of the recent Supreme Court decision in Lawrence, striking down the Texas anti-sodomy statute. The fourth and final Section departs from gay rights to consider the broader implications of Pinello's research methods and findings on legal research. Legal academics have been unduly and unfortunately averse to empirical analyses, which may be critical to our research enterprise.

I. The Empirical Analysis of Gay Rights

In his attempt to isolate the determinants of decisions advancing gay rights, Pinello conducts an empirical study of decisions on various gay rights topics. He analyzed 391 decisions, including 1439 separate votes by 849 appellate judges, over a thirty year period (p. 74). For each of these votes, Pinello coded numerous variables. He coded each case by type of issue and other factors such as amicus participation, coded each judge according to numerous individual characteristics, coded various institutional circumstances (such as court characteristics) and coded whether there was clear legal precedent governing the case.

The gay rights issues considered by Pinello included as "essential" are family matters (such as adoption rights and rights of domestic partners), sexual orientation discrimination in employment, gays in the military, the constitutionality and enforcement of laws criminalizing consensual sex practices, and free speech and free association rights of gays (p. 8). He also includes some cases he considered nonessential, such as defamation cases and same-sex sexual harassment. The family matters category contains the most cases in the database (p. 10). The book breaks down results for cumulative gay rights actions and for categories of case types. Pinello presents considerable detailed data on all his variables and correlations in extensive appendices (pp. 163-215). He also lists all the cases used in his study, giving considerable transparency to his analysis, and permitting the evaluation and replication of his research (pp. 167-213).

The central aim of the book was to ascertain which factors (judge, casetype, legal, etc.) best determined whether a case would yield an outcome favorable to gay rights. The book studies the effect of a substantial number of variables, which may be broken down into attitudinal factors (those involving the individual judge and her background), institutional factors (those involving judicial selection and surrounding institutions), and legal precedents that are supposed to be driving these decisions. Pinello found all these factors to be of some importance, in varying degrees, as discussed below.

A. Attitudinal Factors

Attitudinal factors are the individualized characteristics of the particular judge. These have sometimes been referred to as the "background" of the judge. The background factors that may affect judicial decisions are the manifold influences that enter into every person's makeup. They include all the genetic and environmental factors that may influence a person. Empirical research has consistently confirmed that individual judicial attitudes are a key determinant of judicial decisions. While judges profess fealty to the law, they have at times conceded that their decisions are influenced by attitudinal background factors. The most significant attitudinal factors addressed in Gay Rights and American Law are religion, ethnicity and gender, political party affiliation, and cultural environment.

1. Religion

In the earlier days of empirical research on judicial decisionmaking, it was believed that such a fundamental background factor as religion would play some role in decisions. Religion appears to play a role in voting for political offices. Establishing the strength of this connection for judicial decisions proved elusive for some time, however. There is some evidence that a judge's religion influences his decisions in some types of cases, though, and Pinello builds on this with his study of gay rights decisions. His findings on religion are possibly the most controversial in the book.

The pattern of voting by judges of different religions is fairly stark. Jewish judges were by far the most likely to cast pro-gay rights votes (68.5% in all cases and over 71% in family cases) (p. 88). This is a greater success rate than for judges who subscribe to no religion (66% and 55% respectively). Catholics were the least likely (under 45% in all cases and only 43% in family cases). In prior research that found some religious difference in judicial voting, Catholic judges have been more liberal, so the gay rights context may be a unique context (pp. 88-89).

When Pinello sought to break out the Protestant dominations that might be considered fundamentalist, he found some difference. 52% of the fundamentalist votes were against gay rights claims, while 45% of the votes of other Protestants were against gay rights claims (p. 89). This difference is not a dramatic one and perhaps a little surprising, as it seems that fundamentalist protestants were reasonably open to gay rights claims. In short, religious background clearly seems to matter in judicial decisionmaking, though not in a clear or simplistic manner. Religion does not appear to drive judicial decisions on gay rights but exerts some influence at the margins, depending upon the type of case and particular religious philosophy.

Pinello himself cautions against oversubscribing to his results, noting that "some Catholic judges such as William Brennan, Jr., supported gay rights, and some African Americans such as Clarence Thomas did not." (p. 100). His results can support only a probabilistic judgment about the effects of a particular background on judicial decisions. The data clearly demonstrate that a judge's religious background does matter somewhat, though, in how he or she will vote on gay rights claims, and in a fairly consistent predictable way.

2. Ethnicity/Gender

Another background factor that might be expected to influence judicial decisions is the judge's ethnicity or gender. Ethnicity and gender have some correlation with political attitudes and voting, but empirical studies of their effects on judicial decisionmaking have been mixed. One review and analysis of the research concluded that "empirical studies show only slight, if any, differences between the overall voting behavior of male and female judges along the dimension of gender." Indeed, most female judges dispute that their gender makes any difference in their decisions.

Pinello tested the hypothesis that ethnicity and gender could make a difference in judicial voting patterns on gay rights cases. He found that minority judges (operationalized as African Americans, Latinos, and Latinas) were about twenty percent more likely to vote in favor of gay rights claims than their majority counterparts (p. 78). Gender also had a significant association with the judge votes on gay rights. Women were twelve percent more likely to vote for such rights on all the essential claims and twenty-seven percent more likely on cases involving child custody, visitation, adoption, and foster care (p. 88). This aspect of the judicial background seemingly had an effect on how judges ruled.

Ethnicity and gender did not have an across-the-board association with decisions. For example, neither was significant in decisions involving a federal constitutional issue (p. 279). Nor were these variables statistically significant in federal court decisions (p. 290). Nevertheless, there was sufficient difference based on ethnicity and gender to demonstrate an effect for diversity in the composition of a judiciary. The course of the law will be different, depending on the degree to which the judiciary is diversified by gender and ethnicity. While cultural environment, discussed below, probably has an impact on both gay rights claims and on judicial diversity, the variables for ethnicity and gender had significance even over and above the effects of cultural environment.

3. Political Party

By far, political party is the clearest attitudinal determinant of judicial decisionmaking identified in prior studies, considered a proxy of ideology on a liberal to conservative spectrum. This attitudinal variable has consistently appeared as significant in empirical studies that analyze judicial decisionmaking. One survey of the research concluded that "only political party affiliation seems to have any significant and consistent capacity to explain and predict the outcome of judicial decisions." While judicial voting is not seen as partisan, in the sense of party loyalty, the consistence of party affiliation as a determinant of judicial outcomes is ascribed to the ideological preferences associated with the Democrat or Republican parties.

In contrast to the great bulk of preexisting research, Pinello did not find a statistically significant effect for party affiliation by itself in his results (p. 276). The party effects did show up, though, when party was combined with other variables. This is captured through what is known as an interaction variable. Such a variable is created by multiplying the values for two variables and then entering that sum as its own separate variable. When this was done, Pinello found a statistically significant effect for the combination of party and his measure of the state's cultural environment (p. 276). This suggests that, at the state court level of judges, many Democrats come from conservative states and that they vote conservatively on gay rights issues. In states where the cultural environment is more amenable to gay rights claims, Democrats tend to be more responsive to gay rights claims than Republicans. Similarly, the association of party affiliation and the length of the judicial term was significant, suggesting that some Democratic state judges may vote against gay rights for electoral concerns.

In contrast to the overall results, the effect of political party is very pronounced at the federal level. Pinello finds that "federal judges selected by Democratic presidents, compared with Republican appointees, positively determined an astonishing 40.5% of the probability 'space' between complete success and utter failure of lesbian and gay rights claims in federal appellate courts." (p. 151). Thus, the party effect was much greater in the federal courts, as has been found in much of the prior research that has involved federal courts. The politicization of federal court rulings may be due to the greater independence of those courts. Political scientists have suggested that "federal judges, because they have life tenure, may even be more ideological than are legislators or executives." Pinello's findings lend some credence to that claim.

4. Cultural Environment

Among other background factors is the cultural environment in which a judge lives. Cultural environment is a vague term, but it may be captured geographically and amplify the effects of party affiliation. It is well known that a Northeastern Republican will generally be more liberal than her Southern counterpart and possibly more liberal than even a Southern Democrat. The "red states" that voted for George W. Bush are often regarded as culturally different from the "blue states" that voted for John Kerry. Just as party affiliation may attitudinally determine some judicial decisions, so might the different geographical cultural environments of judges. Some prior research has found a significant geographic effect on judicial votes.

Pinello found a significant association between cultural environment and outcomes. Indeed, there are some remarkable disparities among states in the success rates of gay rights claims. In Missouri, gay rights claims had only a 9% success rate, while in Massachusetts they succeeded 78% of the time. Not every state was as predictable, though: gay rights claims had a significant 56% success rate in Texas (higher than New York's 52%) (p. 12). The number of decisions per state was not consistently high, though, and Pinello broke the data into regions so that it could be more confidently analyzed for the effects o cultural environment.

The overall results for cultural environment showed it to be a "highly predictive" variable for outcomes on gay rights claims (p. 92). Pinello concluded that location was a central factor in the outcome of cases, with particular states in the South and Midwest being about fifty percent less favorable to gay parents (p. 144). The presence of a state statute banning consensual sodomy had significance even over and above the regional variables. This suggests that the regional variables might not have fully captured the effects of cultural environment (being imperfect proxies), so that even the substantial effect for regional cultural environment could understate the effect of this variable on outcomes of gay rights cases.

The evidence that individual attitudinal factors influence the decisions of judges is clear. Pinello's data confirms this fact and adds new information about the role of attitudinal variables in gay rights cases at the federal and state levels. Pinello, however, did not test for the influence of the attitudes of other judges on the panel. This "peer effect" has a significant effect on outcomes. Appellate judges have emphasized that they are strongly influenced by other members of their panels. Some significant empirical results confirm that the decisions of federal circuit judges are affected by the ideological attitudes of other members of their appellate panel, in addition to their own attitudes. For example, Republican judges will moderate their conservatism when sharing a panel with a Democrat. Another study finds that racial diversity on a circuit court panel changed the voting behavior of other judges on the panel. Pinello did not attempt to identify this effect, which could have a substantial effect on voting. This effect could mean that he understated the importance of judicial diversity for judicial decisions.

B. Institutional Factors

Beyond individual attitudinal predispositions, judicial decisions may be influenced by a judge's institutional setting. Some of Pinello's most interesting findings came in his research on the effect of institutional factors on judicial gay rights decisions. While judicial background was clearly a significant factor in such decisions, the judicial background predispositions were substantially affected by the institutional environment in which they operated. The institutional factors on which he focused were: the different state methods for judicial selection, the choice between the federal and state courts as fora for gay rights claims, and the external role of gay rights interest groups on outcomes.

1. State Selection Methods

Pinello tested whether the different state methods for judicial selection have a significant effect on the outcome of gay rights claims. The states use a wide range of methods for the selection and retention of judges. These include selection by partisan election, nonpartisan election, selection by gubernatorial appointment, selection by legislative appointment, or selection with a system called "merit selection," in which a lawyers' group screens nominees. Within each system of selection, states provide for different judicial terms and many states in which judges are initially appointed require them to run in retention elections. The state's system of judicial selection could logically influence judicial decisions, and the diversity of these systems offer a valuable context for research.

Much of the existing research on state judicial selection methods has not shown a substantial effect from different methods. A few studies have found at least some apparent effect from selection method, and they have generally concluded that elected judges are more activist than appointed judges. One recent test discovered that merit-plan judges were less likely to hold statutes unconstitutional, while elected judges were somewhat more likely to do so. Elected judiciaries also produce more dissents, which might be a proxy for activism. However, there is also evidence that appointed judges were more receptive to sex discrimination claims than those who were elected. A study of public-utility regulation found that appointed courts were more likely to side with consumers in disputes, while elected judges were more likely to side with regulated firms and large commercial customers. The extant empirical research does not confirm any clear theories about the effects of judicial selection systems.

One might hypothesize that elected judges would be relatively less sympathetic to the rights claims of a fairly small minority of the population, such as gays. Chief Justice Abrahamson of the Wisconsin Supreme Court asks: "How can elected judges remain impartial and rule on minority interests when they depend on the majority of the electorate to stay in office?" Under the circumstances, majoritarian elected judges might be expected to disfavor gay rights. The data in this study did not bear out this hypothesis, however. Elected judges were somewhat more likely to vote for gay rights claims than were appointed judges (p. 91).

Pinello's results on state selection methods do not present a clear picture of their institutional effect. While appointed judges in general were more likely to oppose gay rights claims, the interaction variable of appointed judges and cultural environment and term in office were positive. Thus, appointed judges in a cultural environment favorable to gay rights claims were more activist in support of those claims. Moreover, while elected judges tended to be more positive to gay rights, judges with longer terms of office also were more positive (pp. 91-92). The positive results for elected judges suggest that electoral accountability may favor gay rights claims, but the positive results for term length suggest the contrary. Consequently, it is difficult to understand the implications of this finding, which does not resolve the division in the preexisting research. The results of this study do at least suggest that electoral accountability is not a strong negative factor in gay rights decisions.

2. State vs. Federal

There is a generalized preference for the federal courts in individual rights or civil rights litigation, including gay rights litigation. A few years back, an article suggested that in gay rights claims, the state forum was the more favorable one. This article was largely impressionistic, though. Pinello's empirical results provide strong confirmation of its conclusions and demonstrate that during the time period studied, gay rights claims were more successful when brought in state court.

Gay Rights and American Law devotes much of its analysis to addressing the differences between the federal and state fora. The book's comparison of federal and state courts controlled for a variety of variables, including the nature of the claim (under federal or state constitutions) and the other variables discussed above. Pinello found a clear difference in the receptivity of state and federal courts to gay rights claims. The coefficient for claims heard in federal court was negative and statistically significant. Its overall impact was a negative thirty percent probability of success (p. 275). State courts decided in favor of gay rights more than twice as often as did federal courts, at both the intermediate and supreme court levels (p. 111). Moreover, claims brought under state constitutions were resolved more favorably than federal constitutional claims.

Significantly, the federal context also muted the effect of many of the background variables. Gender and ethnicity were not significant, age lost significance, cultural environment had no significance, and the Jewish religion variable lost significance (though Catholic religion remained statistically significant and negative) (p. 290-91). Most dramatically, the variables for negative and positive precedent also lost statistical significance in the federal court context (p. 290). The author found that "any federal court decision was likely to be predisposed against gay rights" (p. 97; emphasis omitted).

3. Interest Groups

Pinello also considered the effect of organized interest groups promoting gay rights. One would expect to see some effect from their participation, since interest groups presumably wouldn't spend the time and money to engage in litigation if they did not believe their participation offered some benefit. Interest groups such as the NAACP and ACLU have participated vigorously in litigation and are involved in a majority of important constitutional cases. David Truman's political classic, The Governmental Process, emphasized that "[t]he activities of the judicial officers of the United States are not exempt from the processes of group politics," so "few organized groups can afford to be indifferent to its activities." Groups obviously perceive this importance and participate in litigation accordingly. The power of interest groups is somewhat disputed, though, with some research showing little effect. An examination of the role of the NAACP Legal Defense Fund concluded that there was no evidence demonstrating that it significantly influenced judicial decisions.

Although the empirical evidence on the role of interest groups in court is not conclusive, there is reason to think that strategic participation by interest groups could have an effect on the path of the law. Such groups have a broader perspective and interest in the state of the overall law, as opposed to the direct parties to litigation, who may be centrally interested in the outcome of a particular case. Interest groups are able to use selective strategic litigation to produce precedents favorable to their constituents.

In addition to the strategic coordination of litigation, one might also expect interest groups to contribute to the success of gay rights claims, merely through their greater resources committed to strategic litigation. Some gay rights plaintiffs may lack the funds or legal representation required to present their case to a court in the best light. Backing from an interest group can provide those funds and that representation. It seems likely that a case's success could hinge at least in part on the strength of the legal claims presented to the court, in which case interest group support could make a difference.

Pinello considered whether the role of gay rights interest groups made a difference in the outcome of litigation and found some evidence that such groups had a positive association with outcomes. The participation of a gay rights interest group in the case increased the probability of a pro-gay rights ruling by nearly 8% overall and by over 15% in decisions by courts of last resort, after controlling for other variables (p. 78, 80). Pinello concluded that interest groups played a vital role such that gays would be "wise to support the worthy efforts of these diligent organizations" (p. 150).

The role of interest groups requires further exploration, and it would be premature to conclude from Pinello's data that such participation materially enhances the probability of success for gay rights claims. When interest groups participated, such claims were more likely to succeed, but that may not be due to the interest group participation. For example, it is possible that interest groups simply were wise enough to involve themselves in gay rights claims that were relatively strong on the merits (a selection effect), and this explanation could also explain Pinello's results. Such groups have a self-interested reason to associate themselves with successful litigation in order to please their membership. Given the theoretical reasons to believe that interest group participation should matter, though, and Pinello's empirical support for the value of such groups, it seems fair to conclude tentatively that their role in litigation matters to some degree.

C. Legal Precedent

In some respects, the most significant aspect of Pinello's book is his analysis of the importance of legal precedents. Prior empirical analyses have given relatively short shrift to the role that legal variables--such as precedent--may have in determining judicial decisions. Some have called the belief in the controlling power of precedent "a low form of rational behavior" analogous to "necromancy." One view held by some political scientists is that "rules based on precedent were little more than smokescreens behind which judges hide their values." An empirical test at the Supreme Court suggested that individual justices give relatively little heed to the power of their prior decisions. Yet precedents are supposed to influence decisions and judges claim that precedents determine decisions.

Judges typically testify as to the importance of legal precedent, within their opinions and without. Surveys of judges find that the law is central to decisions. Some empirical research suggests that reliance upon precedents may in fact play a significant role in decisions. Most of the research involves the vertical application of precedent and whether higher court principals can effectively control the outcomes in lower courts (pp. 119- 22). Studies of lower courts have found that they appear to respond to the precedential directives of the Supreme Court.

Recent empirical evidence also suggests a role for horizontal precedent. Some intriguing recent research identifies certain "jurisprudential regimes" adopted by the Supreme Court that appear to have structured subsequent decisions even by the Court itself, indicating that the rationale of an opinion influences future votes. Another study found evidence that circuit courts heeded new legal rules established by other circuits. Intra-circuit district court precedents have been found influential in subsequent decisions. This research suggests that precedents may matter, at least in major Supreme Court cases, but the studies have yet to capture the routine use of precedent by courts.

The studies mentioned above have only scratched the surface of necessary quantitative empirical analysis of law, however. The legal precedent variable is a notoriously difficult one to capture in quantitative empirical research. Yet efforts to capture the variable are important, as "[t]he lack of empirical support is the greatest shortcoming for those who believe in the legal model" of judicial decisionmaking. Pinello, by focusing on the context of gay rights decisions, has been able to capture an effect for legal precedent and enhance the understanding offered by the prior research.

Pinello first identified clear precedents on gay rights (either positive and negative) that come from either the same court making the decision studied or from a higher court within the same judicial hierarchy. His coding for precedent was a demanding one and limited to precedents that might be considered "on point." Prior to 1970, there were very few precedents addressing gay rights, and those that existed were almost uniformly negative (pp. 128-29). Consequently his study period captures virtually the entire evolution of gay rights law. Given the paucity of historic doctrine, many of the decisions he studied were largely precedent-free, especially the earlier decisions. As the law developed, though, some courts had available clear precedent on the gay rights issue of concern. Pinello conducted several separate tests on the role of these precedents on judicial voting.

First, the book reports a standard regression of precedent along with his other attitudinal and institutional independent variables. Precedent played a relatively consistent role in predicting decisions and appeared to determine around a quarter to a third of the judicial votes. Of all the types of variables considered in all types of cases and court levels, positive precedent was the "most consistently consequential variable" (p. 86). Negative precedent also mattered and was statistically significant, though its overall impact was not so great.

In a second test of precedent, Pinello used his other variables to construct a probabilistic measure for the success of gay rights claims for a given judge, based on predictons from those variables. He then divided these probabilistic predictions into three sectors--those predicting less than a one-third probability of success, those predicting a one-third to two-thirds probability of success, and those predicting a greater than two-thirds probability. Having done so, Pinello then introduced precedents to see how they affected the results in a large table of frequency distributions. These results were most revealing. For example, take the set of judges with a predicted probability of success of less than thirty-three percent, labeled conservatives. When there was no precedent, these judges voted against gay rights claims nearly eighty percent of the time, and favored them around twenty percent, as predicted. When there was a negative precedent, confirming their predispositions, the conservative judges again voted against the gay rights claims about eighty percent of the time. However, in the presence of a positive precedent, the conservative judges voted against the gay rights claim only one-third of the time and voted for the claim two thirds of the time. Judges who attitudinally were negative toward gay rights claims, such as Catholics, became likely to vote in favor of such claims in the presence of a strong precedent. Roughly comparable results were found for liberal judges, though they showed somewhat less effect for negative precedent.

This test directly models precedents in operation. In the typical case, you have judges who are predisposed, to some degree, to favor one party or the other, absent any legal constraints. The key issue is the degree to which legal constraints such as precedent cause judges to modify their decisions. Pinello found a clear effect of precedent frequently overriding attitudinal and institutional preferences. The effect was not a universal one, as some judges apparently ignored precedent, but precedent nevertheless had a very substantial effect on outcomes.

Pinello conducted yet a third test in which he examined the votes of judges that appeared to ignore precedents (whom he labeled the "precedential atheists") (pp. 138-40). In this test, he examined whether the precedents in question were vertical or horizontal. This study produced the interesting result that conservative judges were more deferential to vertical precedent than horizontal precedent, but liberal judges actually deferred more to horizontal anti-gay rights precedents. These results varied by case area, though, and the number of votes in the database was small, so it is not clear that they are salient.

As the legal model would predict, precedent mattered more to intermediate appellate courts, and courts of last resort were more likely to make new precedents (p. 79). The role of legal variables was the most significant determinant for intermediate courts, but was not as great in supreme courts (p. 82). Overall, the results confirmed the theory that stare decisis was a powerful force in explaining the judicial decisions and in roughly the way predicted by the legal model (e.g., less powerful for courts of last resort). Pinello reported that his results reflected "the high courts' greater policy-making role and their resultant propensity to follow attitudinal, environmental and institutional forces more than legal ones" (p. 141).

While both negative and positive precedent had a statistically significant impact on outcomes, they were not equal in power. With controls for other attitudinal and institutional variables in his full sample of cases, Pinello found that the presence of a negative precedent reduced the probability of a subsequent pro-gay rights by around ten percent, but the presence of a positive precedent increased the probability of a subsequent pro-gay rights decision by thirty-six percent (p. 275). For decisions by courts of last resort (typically state supreme courts), the result was more remarkable--a positive precedent increased the likelihood of a pro-gay rights ruling by 34%, while a negative precedent also increased the likelihood of a pro-gay rights ruling, by around 18% (p. 80). In cases involving federal constitutional claims, precedent had a somewhat greater effect, for both negative and positive precedents.

At first glance, the findings about the greater power of positive precedents might seem to confirm some conservatives' worse fears about the judicial system--that activist liberal precedents breed activism, while contrary conservative precedents have only a weak countervailing effect, because conservative judges feel bound by stare decisis. Most of the general empirical data on judicial decisionmaking does not support this hypothesis; this research shows that conservative judges are no more restrained by precedent than are liberals. In some areas of the law, though, the same phenomenon is found. Some aspects of First Amendment law, such as the obscenity decisions, have witnessed a one-way ratchet in the direction of expanding rights, without countervailing conservative retrenchment. One can hypothesize several different reasons for these results in discrete legal areas, including (a) changing social mores that are followed by the courts, (b) the dynamics of a judiciary that is hesitant to pursue its preferences by creating a new precedent on a significant social policy but eager to pursue those preferences by following such a precedent, once set, (c) a libertarian bias in favor of extending individual rights, or even (d) the expansion of judicial power. Understanding which of these hypotheses (or which combination of them) explains the results would provide valuable information and represents an important topic for future research.

Pinello's data on precedents contain one other very important finding, in that precedent had a much weaker impact on cases involving child custody, visitation, adoption, and foster care. In these cases, consideration of legal variables had less effect and improved the predictive ability of his model by less than four percent, while the effect of attitudinal variables was much greater than in other gay rights cases. He noted that these cases involved an amorphous "best interests of the child" standard, and inferred that such a standard meant that the law placed less constraint on decisionmaking, giving free play to the judges' attitudinal predispositions (p. 84). This finding supports the logical conclusion that the effect of precedent will depend upon its clarity and precision and gives some insight into the importance of precedential language.

D. Other

Pinello coded for the effect of a few other variables that did not readily fit into one of his three main categories. The case type clearly made a difference, as family issues were much less likely to be decided in favor of gay rights, as were claims regarding gays in the military (p. 275). When the plaintiff was a gay male litigant, the claims were more likely to succeed than when the plaintiff was a lesbian. Older judges were more likely to rule against gay rights claims than were younger judges. The probability of a favorable gay rights decision also increased over time, during the decades studied.

Pinello's results tell us much about judicial decisionmaking but leave us wanting still more. While he considered many variables, including some interaction variables, some of the most intriguing interactions were not tested. For example, we know that precedent could override substantial attitudinal predispositions to the contrary, but Pinello does not report the effects on most of the precise groups defined by other variables. For example, did precedent have a different power in appointed judiciaries than in elected ones? The use of multiple regression analysis serves to address the interactions among variables to some degree. Thus, because precedent and religion were variables in the same equation, we know that different religions reach different results, with the presence of precedents held constant. The methodology does not tell us, however, if the role of precedent had a differential impact, in the presence of particular attitudinal or institutional variables.

* * * *

In the end, the data analysis of Gay Rights and American Law clearly demonstrates that attitudinal factors, institutional factors, the law, and additional factors all matter in judicial outcomes. Rather than pitting these theories of decisionmaking against one another, research should focus upon how they interact. The law was clearly the most important factor in intermediate appellate courts. By contrast, the law was the least important factor for courts of last resort. Other intersections of the law with other variables were not clearly tested. For example, it would be valuable to know if the judicial selection method of a state appeared to influence the significance of legal precedent as a variable. The wealth of findings contained in the study still did not address many crucial questions.

Another addition to these results would be a consideration of case outcomes. Pinello tested individual judge votes, but not overall case outcomes, in order to avoid a risk known as the ecological fallacy. Yet it is outcomes, and not votes, that ultimately matter. A judge's vote cast in dissent is not of the same significance as a swing vote, so outcomes should also be considered. Consider a panel of nine judges, five of whom are conservative and four liberal. Suppose that all vote as expected, except for one conservative who casts a liberal vote. While ideology would have predicted eight of the nine votes (88% accuracy), the ultimate outcome would be contrary to that predicted by the ideological variable. The possible presence of peer effects also demonstrates the need for additional analysis focused on case outcomes rather than individual judicial votes.

II. The Implications

The results found in Gay Rights and American Law are certainly intriguing as factual matters, but their ultimate importance lies in how they can inform political and judicial decisions. Pinello, a straightforward empiricist, presents relatively little analysis of such implications, leaving it to the reader to draw his or her own conclusions. This Section addresses some of the most salient implications.

A. Judicial Selection

The role of the Senate in the selection of federal judges continues to be a controversial one, and we are currently embroiled in a vigorous political dispute over the ideological screening of judges during the confirmation process. Senator Hatch has declared himself

heartened to read the scores of editorials all across this country that have addressed the notion of injecting ideology into the judicial confirmation process, because this notion has been near universally rejected--except, of course, for a handful of well paid, special interest liberal lobbyists in Washington, and a few other diehards.

The issue is not so clear cut, though. When Clinton was appointing judges, Republican Senators declared that it was "perfectly legitimate to vote against someone for a lifetime appointment based on ideology."

Before looking to the Senate's role, it is crucial to realize that presidents, to varying degrees, have historically nominated judges based on their "political and ideological compatibility." Presidents overwhelmingly select nominees from their own party and there is considerable anecdotal evidence of presidential screening of nominees for ideology. This presidential screening occurs outside the public view, though, and the controversy has swirled around the Senate's use of attitudinal screening during its confirmation process. There is now a "widespread view . . . that it is improper or unseemly to inquire into a judicial candidate's substantive legal views" at the appointment stage.

The politicization of judicial selection has been often lamented as divisive and discouraging the nomination of especially able candidates. Bruce Fein contended that it produced a "Court of mediocrity," yielding only unobjectionable judges. Because the confirmation process may screen out the controversial, it allegedly produces justices of lower "stature" than in the past. Senatorial concern for ideology allegedly will detract from concern about quality. Stephen Carter has written extensively on the "mess" that is the contemporary process of confirmation for Supreme Court justices and its adverse effects. While the nomination process has been criticized for various types of attacks on nominee character, a central concern has been their ideological focus. Some argue that Senate confirmation processes are too intrusive into the attitudinal predispositions of nominees for the federal courts. An advocate of judicial independence argued that nominees who declined to respond to Senate questioning "were right to send overly inquisitive senators packing."

One central difficulty with the criticism that confirmation questioning undermines judicial quality is the inability to define the "quality" that we wish to find in our judges. Judicial quality "is said to include temperament, expertise, integrity, intelligence, training, and communication skills." These attributes are obviously amorphous and difficult to define in the abstract, much less apply to particular candidates. None of those who criticize the confirmation process have come up with clearly defined standards for judicial quality or even examples of potential quality judges who failed the nominating process (or low-quality judges who passed the process). Absent such an understanding, there is no basis for the claim that ideological screening has a cost in judicial quality.

Carter argues that questioning judicial candidates on factors relating to their likely future decisions "represents a profound threat to judicial independence." If Senators were seeking some binding advance commitment on judicial votes, that might be contrary to judicial independence, but they are not doing so. They are simply seeking a prediction of how a judge will rule on an issue on which the legal materials are not dispositive. Nothing about judicial independence implies that judges, once appointed, should have unfettered rein to exercise their personal ideological proclivities about the proper path of the law. The criticism erroneously conflates extracting promises on particularized future votes with investigating generalized ideological proclivities relevant to future votes. The two are quite different, and the illegitimacy of the former does not speak to the illegitimacy of the latter.

Indeed, the evidence makes the case for much more specific questioning about particular legal controversies. The consensus view is that nominees may be asked about their "judicial philosophies" but not about particular legal controversies. Unfortunately, this approach is going to yield "virtually no useful information." While it is unreasonable to ask a nominee how he or she would vote on a particular set of case facts, more abstract questioning can still be quite specific. A nominee might well be asked about his or her views on particular legal disputes, such as the constitutional propriety of Roe v. Wade or Buckley v. Valeo. The answer would not commit the life-tenured candidate to any particular future decision but would inform the Senate on how they would inevitably employ their ideological predispositions in future cases.

Carter accurately identifies the rationale for judicial screening when he observes that "[w]e have all been seduced . . . into the vision of the Justices as creators of policy." This seduction can't have required much effort, though, because the justices are quite plainly creators of policy, and their policy choices are typically driven by their attitudinal predispositions. Pinello's research and the research of many others clearly demonstrates this descriptive fact. Carter apparently wishes to pretend the judicial world were otherwise, as many others do, but wishing doesn't make it so. Until judges eschew their attitudinal inclinations (which is extraordinarily unlikely and probably impossible), policy screening for judges is a logical sequela, as Carter realizes.

With respect to concerns for judicial quality, confirmation battles would only have the potential to harm such quality if Presidents were otherwise using an accurate measure of "legal quality" as the standard for their own nominating choices. The historical record does not bear out this condition. Presidents have used numerous criteria in making judicial appointments, of which judicial quality is one but not the paramount criterion. As noted above, Presidents have historically used an ideological screen for nominations. President Clinton confessed to using a belief in privacy rights as a litmus test condition on his nominations. President Bush clearly uses judicial nominations to advance his ideological views. If the President is not using judicial quality as a standard for nomination, but instead is using ideological qualifications, there is no reason why the Senate should be held to a different standard.

The arguments against ideological screening by the Senate and the effect on judicial quality are generally conclusory and, at best, anecdotal. One study has evaluated the effect somewhat more rigorously. The author first surveyed scholars (of law and political science) to produce an evaluation of the relative quality of Supreme Court justices. The level of confirmation battles and associated nominee scrutiny developed only in the late 1960s. There was no difference in estimated quality for the justices appointed after that time, though none of the recent justices approached the standard for "greatness" reached by some of the earlier justices. Perhaps the confirmation process does impede the selection of "greatness," though this finding may have simply been due to the fact that perceptions of greatness appear only after time. In any event, the perceived "greatness" of Justices like Holmes and Brandeis may not be attributable to their traditional legal skills.

Nor is it clear that open attitudinal evaluations of nominees would exacerbate the "confirmation mess." Much of that mess involves scrutiny into personal matters, such as FICA tax payments for nannies, allegations of sexual harassment, or marijuana smoking. Senator Schumer has reported that the "unwillingness to openly examine ideology has sometimes led Senators who oppose a nominee to seek out nonideological disqualifying factors, like small financial improprieties from long ago, to justify their opposition," which led to "an escalating war of 'gotcha' politics" that warped the confirmation process. A more transparent focus on nominee ideology would seem more honest and less messy.

Even if the adversarial Senate confirmation did have the effect of somehow reducing the "legal quality" of the judiciary, that outcome would not delegitimate the inquiries. Pinello's research and ample other research shows that legal quality represents only a fragment of decisionmaking. Legal quality matters least on supreme courts, for which the attitudinal model is the best predictor of outcomes, perhaps because the courts take cases without clear legal answers. Even if intrusive questioning had some negative effect on legal quality, such a result by itself does not make the case against ideological questioning.

The results of Pinello's study are of interest in the context of selecting federal judges. They suggest reason for concern for ethnic and religious diversity on the courts. It would be troubling if religious affiliation or some other attitudinal attribute of a judicial nominee were to be considered disqualifying. It would also be irrational to take this approach. Pinello's data reveal tendencies, not inevitabilities, and some members of every group rendered decisions for gay rights (or against). Pinello himself opposes any use of "litmus tests" for nominees (p. 161). In the case of individual nominees, though, knowledge of Pinello's results might warrant some specific questioning. It would be enormously controversial to question a judge about the role that his or her religion might play in decisionmaking. However, if religion in fact plays such a role, as Pinello finds, the question would seem to be a perfectly legitimate one that should be asked. Judges have no logical claim to be free from scrutiny on matters that may drive their votes.

The importance of attitudinal variables and need for scrutiny of nominees is countered somewhat by Pinello's findings on precedent. He shows that judges will, as they should, vote contrary to their preferences in the face of a clear precedent. However, many cases are not governed by such clear precedents and leave considerable play for extralegal considerations. Moreover, all precedents were at one time the product of a judicial resolution of a case of first impression. As might be expected, Pinello's results show that attitudinal variables are enormously important in such cases. Judicial selection screening is centrally important to determining the path of precedents that will control judicial decisions in the future. His results also show that even ideological judges will follow those clear precedents, once set.

While the conventional wisdom is that the judicial nomination and appointment process is overly politicized, the truth may be the opposite. News is made when the Senate filibusters or refuses to hold hearings for a judicial nominee but not when other nominees breeze through the process. During the six years when Clinton was President and Republicans had a Senatorial majority, over 375 judges were confirmed. In about two years of the Bush Administration, 124 judicial nominees won Senatorial approval. Only a minority of appointments have been blocked for ideological reasons. Pinello's findings demonstrate that not only extreme ideologues on the bench that render attitudinal decisions but that this tendency is commonplace in federal courts, at least on gay rights issues. The results suggest that participants in the process should do more ideological screening, not less. Gay rights is the one area where such screening is appropriate.

B. Federal Courts

The legal academy has seen a longstanding debate over the comparative merits of federal and state courts, especially in the contest of human rights claims. Scholars have often argued that federal courts offer a preferred forum for rights claims. The widely cited classic of the genre is “The Myth of Parity.” This influential article contended that state courts were institutionally less competent or less favorable to claims seeking the vindication of individual rights. The article was influential and continues to be used in support of a federal forum. Most practitioners seem to concur with the article's conclusions.

For many, the U.S. federal courts are a bastion of individual rights protection, defending such liberties from the attempted depredations of the majority, as reflected in the nonjudicial government branches. Federal judges, because they are free from electoral concerns, may be regarded as resistant to the pressures for majoritarian discrimination and both able and willing to protect the interests of disenfranchised minorities. Logically, it might seem that the institutional setting of the federal judiciary better conduces to protecting individual rights.

While theoretically appealing, the historical "record does not support the assertion that judicial review has been a force for protecting individual liberties." Indeed, the "Court frequently has declined to intervene when this paradigm calls for judicial involvement." The Court has shown no systemic interest in protecting the rights of discrete and insular minorities. Under the traditional theory, state judges, who typically must stand for election and many of whom are engaged in partisan elections, would be far inferior to federal courts when it comes to protecting individual liberties. Yet if federal judges have not been so protective of individual liberty as predicted, perhaps the theory is unfair to state judges.

The debate over federal versus state courts has also been influenced by notions of judicial independence. Federal judges, of course, have life tenure and other constitutional provisions, such as salary protection, to ensure their independence. The picture in the states varies considerably, but most state judges serve limited terms and must run in some sort of reelection campaign. The consequent expectation is that federal judges should be more independent of public opinion, campaign contributors, and other politicians. Consequently, they might be expected to provide greater succor to the rights claims of minorities and less receptive to the interests of the state. Justice Brennan argued that independent courts can pass "sober constitutional judgment" at times when elected officials are influenced by the "passions and exigencies of the moment."

As noted above, life-tenured, seemingly independent federal judges have not been consistently employed such sober constitutional judgment in defense of minority rights. Even if independence gives the judiciary the theoretical ability to protect individual rights, it provides no inducement for judges to do so. Nor have elected politicians been consistently unsupportive of minority rights. In reality, elected majoritarian officials have been quite supportive of minority rights. Elected officials must attend to the policy preferences of even their minority constituents, and even the majority of the people may prefer the protection of the rights of minorities. Indeed, the historical record contains ample evidence that elected officials can be more protective of such rights than the courts. Judges subject to election may have an electoral reason to be concerned for gay rights protection.

If elected legislators and executives are protective of minority rights, one might expect elected state judges to be likewise. Some evidence has called into question the purported superiority of federal courts in rights protection. A study of constitutional decisions found relatively little difference in outcomes between state and federal courts. A study of takings claims found similar comparability. The empirical data of this study showed a clear cut advantage for state courts in gay rights claims. The state advantage was all the more remarkable, once case facts were considered. Some case facts are less favorable to the success of gay rights claims, including the family issues. Although "state courts adjudicated appreciably more cases with topics predisposed to lose than federal courts did," the state courts "still substantially outperformed [federal courts] in securing minority rights" (p. 116; emphasis omitted).

Pinello also emphasized the pragmatic importance of filings in state courts, given the issues central to gay rights. Only 22% of the cases that arose in his sample arose in federal courts, and only 35.5% of these issues raised federal questions of any type (p. 101). Some of the most important issues were family matters to be resolved in state court using state standards. Of course, a certain federal ruling under the Equal _Protection Clause could have an enormous impact on state courts and state standards. Given the unlikelihood of such a dramatic ruling, though, the state issues may be paramount, and the state courts have been more receptive to gay rights claims. Legal scholarship on gay rights might benefit from greater attention to state courts.

Pinello's findings have implications beyond gay rights. While gay rights issues are not representative of all litigation, neither are they sui generis. Gay rights claims surely share much in common with claims brought by other minority groups that are disdained by a large portion of the public. While further research is needed on the litigation of other groups, the book's results suggest that state courts might also be a preferred forum for their claims. Pinello concluded that "neither presidential appointment nor life tenure of federal judges necessarily improves the probability that their policy making will be more favorable to disfranchised minorities than that of state counterparts selected by other methods or for shorter terms of office" (p. 113; emphasis omitted).

Justice Brennan's call for the use of state constitutions and courts to protect individual rights is the classic that stands in counterpoint to “The Myth of Parity.” Brennan foretold the fate of individual rights claims in the post-Warren Court era and argued that state courts could be used to enhance the protection of such liberties when the federal forum was unfavorable. There is some generalized evidence that state courts have taken up this role. Professor Erwin Chemerinsky argues that empirical evidence can never conclusively resolve the state versus federal forum dispute. While this is surely correct, empirical information is essential to an informed analysis of the dispute, and Pinello's results add considerable ammunition to those who defend reliance on the state, not the federal, forum for advancing such claims.

C. Precedents

Descriptively, the most important social scientific finding of the book may be the evidence on precedents. Pinello's results show that "rumors of stare decisis's death have been greatly exaggerated" and that integrated models "reveal that stare decisis is statistically significant with consequential impact statistics" (p. 143). He has rigorously proved what law professors were saying all along--legal precedents matter. Pinello also demonstrated, though, that precedents were not the only variable that mattered in judicial decisions. The multiple determinants for judicial votes must be analyzed for practical implications.

The results of Gay Rights and American Law are best viewed as a first step in a long journey to understanding the significance of precedents. These findings are an enormously important first step, because they set us on the path of appreciating the potential significance of precedent, but much remains to be explored. While Pinello's analysis of precedent is a considerable advance over prior quantitative analyses, it is still overly simplistic. He transformed the presence of precedent into a binary variable when in reality the binding effect of precedent lies along a continuum. The coding had a fairly rigorous standard for identifying a precedent and did not enable an answer to questions about its diffusion. For example, does a precedent on gay rights in employment affect decisions on gay rights in family matters? Does a generalized privacy precedent affect subsequent gay rights claims? Pinello's powerful results for precedent may simply be due to his rigorous standard for identifying precedents.

While the quantitative empirical analysis of the book offers considerable advances in understanding the role of precedent, the reader might also like some narratives on precedent to illuminate our understanding. Such narratives might tell us much more about how precedent matters. Considering the narratives (or "the trees") is vital to finding theories for empirical testing and uncovering the best approach to operationalize such testing in order to discern the broader picture ("the forest"). For example, one might wonder why precedents were not followed in some cases. How did the judges in these cases distinguish these precedents? Did the courts of last resort expressly overrule those precedents? Were older precedents less likely to be followed? Did the nature of the language of the precedents have an effect on whether they were followed in subsequent holdings? The answer to these questions may be critical in understanding the operation of the law.

The study of precedents is but a subset of the study of the effect of law on judicial decisionmaking, and Pinello examined other aspects of legal decisionmaking as well. His empirical results also suggest that constitutional text matters in court. While the legal importance of constitutional language is commonly presumed, the reality is that courts have not hewed closely to such language. The Supreme Court has recognized a broad privacy right in the federal constitution, even though such a right is absent from its text. A comparative study of constitutional language is possible at the state level, because some states contain express privacy rights in their constitutions, while others do not. Pinello found that the presence of such an express privacy clause had a positive and significant effect on the success of gay rights claims (p. 114). As with the research on precedent, this finding shows that judging is more than a simply application of attitudinal proclivities and that the law itself matters.

Pinello's research adds to the understanding of legal precedent and how it governs. As noted above, the most remarkable and interesting finding of the research involves the varying power of negative and positive precedents. While Pinello does not discuss this effect, I suspect that the reason for this finding lies in the Bickelian impulses of the judiciary. Because judges are typically regarded as countermajoritarian, at least at some level, they may be abashed in the aggressive pursuit of lawmaking. Life-tenured federal judges might especially feel this constraint. State courts, bearing the democratic imprimatur of elections and consequently possessing their own constituencies, might be less Bickelian in deference and more assertive in rights protection.

While my Bickelian theory is but a hypothesis, it finds some support in Pinello's data. He found that a pro-gay rights judicial decision was less likely in a state that prohibited consensual sodomy but more likely in a state that had passed a gay civil rights law (p. 80). This finding is consistent with the suggestion that judges fear being on the cutting edge of societal change but are more willing to take action once other institutions have signaled that they find such change acceptable. Pinello also found that judges with prior service in nonjudicial elective government were less likely to favor gay rights and "deferred to other branches for innovative government action." This is consistent with the theory that generalized deference plays a role, as is the fact that an elected (and therefore more political) judiciary is somewhat more supportive of gay rights claims. The more aggressive, rights protective role of state judiciaries also provides confirming evidence.

One final implication of Pinello's research involves strategic litigation. The value of the strategically selected "test case" is important to the path of future law. One key to the successful test case is strategic forum shopping, which indicates that litigants should work at the state level and consider the state's cultural environment as well as individual judge attributes before pressing ahead. As precedent is important, a losing claim will have negative externalities on the prospects for success in actions brought by other future litigants. Pinello's model could correctly predict the outcome of over eighty percent of the federal court decisions opposing gay rights (p. 103). Some of these cases surely represented unwise litigation that undermined the prospects for future gay rights litigation. Litigants and particularly interest groups need to be informed about their prospects for successful results and shun cases likely to yield negative precedents, while remaining alert to opportunities for setting a positive precedent (based on institutional and attitudinal variables), which in turn could demonstrably enhance the success of future claims.

III. Lawrence v. Texas

The publication of Pinello's book barely predates the recent federal Supreme Court decision in Lawrence v. Texas, striking down the Texas state law against homosexual sodomy. Given the coincidental timing, it is interesting to examine the Lawrence decision in the context of the book's findings. Some caution is warranted in this analysis. Statistical analyses do not purport to explain every decision but only seek to identify patterns in the data. As a lone decision, Lawrence provides only anecdotal evidence that cannot confirm or refute a much broader quantitative statistical analysis. As a Supreme Court decision, of course, Lawrence is viewed as being far more important than other decisions, but its relative importance will depend in large part on the controlling power of precedent and future cases in which it is cited, in ways analyzed by the book.

At first blush, it may appear that the ruling in Lawrence is contrary to what Pinello might have expected. That decision expressly overruled the recently decided Bowers v. Hardwick, the votes of which were included in Pinello's data, which outcome may seem contrary to the precedential findings of Gay Rights and American Law. The pro-gay rights finding is also contrary to the general pattern of federal court hostility to gay rights claims. The book acknowledges the exceptionalism of the Supreme Court and how its patterns may apply differently in that context (pp. 100-01).

In a number of ways, though, Lawrence is consistent with Pinello's findings and helps illustrate the significance of his conclusions and recommendations. First, the ruling emphasizes the significance of ideology, particularly on courts of last resort. Lawrence was a fundamentally ideological ruling. Although Republican appointees joined the gay rights majority in Lawrence, they were the moderate Republican appointees. The most conservative justices, Scalia, Thomas, and Rehnquist, all dissented. And the moderate Republicans, at least Kennedy and Souter, are on the Court only because of senatorial vigor in screening out more conservative nominees, which illustrates the great significance of the judicial selection discussion above.

Of course, Pinello's book demonstrates that judicial decisionmaking is not "mere politics," in his evidence on the effect of legal precedent. The Lawrence decision is not so contrary to these findings as it might first seem. Recall that Pinello found a relatively weak precedential effect from anti-gay rights rulings (like Bowers) and a strong effect only for rulings in favor of gay rights. He also found that courts of last resort were less bound by precedent. Thus, one would not expect Bowers to necessarily control future decisions, as negative precedent suppressed liberal court of last resort votes only 27.3% of the time in Pinello's study. Conversely, one might expect Lawrence to have considerable precedential power, as positive vertical precedent suppressed conservative votes more than 70% of the time (p. 139). The most important effect of this decision, though, may be in how broadly it can be interpreted. Lawrence has already been invoked in Massachusetts to strike down policies that denied marriage to gays. The precedential power of the Supreme Court case is not automatic, though, and an Arizona court reached the opposite conclusion. The precedential power of Lawrence will vary in future rulings, probably according to the variables that Pinello analyzed. Most gay rights decisions raise no federal issue whatsoever, though, so the state court strategy remains an important one, even after the gay rights success in Lawrence.

The individual justice votes in Lawrence are also roughly consistent with the data. One must be cautious in using broad patterns to project decisions of individual U.S. Supreme Court justices, because the screening process for nomination to the Court is far from a random selection. Justice Thomas, for example, may not be representative of the broader pattern of minority sentiment. Notwithstanding this caveat, the votes are consistent with some of the broad patterns. Justice ideology largely predicted the division of votes. The decision showed an evolution over time in the direction of favoring gay rights. The female justices favored the gay rights claim. The two Jewish justices (Breyer and Ginsburg) both favored the claim. Two of the three Catholic justices (Scalia, Thomas and Kennedy) opposed the gray rights claim. Of the four Protestant justices (O'Connor, Rehnquist, Souter and Stevens), two joined the majority, one concurred and one dissented. The results are roughly consistent with the book's data and expectations.

The data also coincide with some other predictions. The presence of amici on behalf of the gay rights claim may also have been a significant factor. In addition to gay rights groups, the American Psychological Association, the American Bar Association, and other groups, including religious groups, urged that the anti-sodomy law be struck down. The book demonstrates the significance of such amicus support, which surely played a part in the Supreme Court decision. Although Lawrence was but a single data point, its outcome is illuminated by the research of Gay Rights and American Law.

Ultimately, one should not rely on statistical analyses of all courts to make predictions about a decision of the U.S. Supreme Court. Considerable empirical evidence demonstrates that controversial Supreme Court decisions are determined almost exclusively by judicial ideology. The statistical evidence may bear more directly on consideration of the implications of the Lawrence decision. Pinello's findings suggest that Lawrence will be more powerful in support of gay rights claims than Bowers was to the contrary. The Lawrence decision is not invulnerable, though, and its fate will depend in some part on future judicial selections. The precedential power of Lawrence will also depend on the circumstances of the cases and judges who hear future gay rights claims that rely on Lawrence. Pinello's research is a valuable tool for projecting which of those future claims is most likely to succeed.

IV. Coda on Legal Research

One of the most compelling stories surrounding Gay Rights and American Law is its window into legal research practice. Traditional legal research has not made great use of quantitative methods that are common to other academic fields. While such quantitative empirical legal research is increasing, it remains a small fragment of the body of legal research.

Additional empirical research is vital to a full understanding of the operation of the law in America. Other disciplines recognize the value of rigorous quantitative analyses. Where would we be if medicine based its decisions on anecdotal reports of patients whose condition improved following the administration of a drug and failed to conduct sophisticated quantitative analyses of drug safety and effectiveness? Would we want our economic policies to be based on reports of isolated companies' success rather than an overall quantitative picture of economic patterns? The law is every bit as important as medicine and the economy, and it deserves the same caliber of research method.

A. The Importance of Quantitative Empirical Research on Gay Rights

Recent years have seen considerable, high quality conventional legal research on gay rights law. Much of this research has dwelt on the debate over the proper legal "hook" for the recognition of gay rights, e.g., use of privacy principles, equal protection principles, or otherwise. In this vein, Andrew Koppelman has argued that constitutional privacy claims are a weak basis for gay rights claims, so that advocates should look to other legal arguments. Others have questioned the power of the sex discrimination basis for such claims. This debate is very well reasoned, from a traditional legal perspective, but cries out for supplementation in the form of empirical analyses of these claims. The goal of all these authors is to find an effective hook for gay rights claims, and the proof of their arguments must lie in the empirical pudding. Pinello's analysis informs the discussion of these claims.

Patricia Cain has written a long narrative story of gay rights litigation. Her story covers fifty years of litigation in support of such rights and clearly informs our understanding of the legal situation for gay rights claims. However, without the rigor of empirical analysis, such an approach may mislead. Her focus on a few big cases, such as Bowers v. Hardwick, paints a picture that has been described as "bleak." Pinello's broader empirical research suggests that the reality is not so unpromising. As discussed above, any such reliance on narratives risks undue emphasis on particular trees, while losing sight of the broader forest.

Pinello is not the only researcher examining the empirical success of gay rights claims. William Rubenstein examined the extent to which gay rights laws produced state employment discrimination claims. He found that the laws were used with some frequency and had some effect, contrary to the suspicions of some. This empirical information is essential to a rational and effective analysis of the effects of litigation. Quantitative analyses on gay rights litigation can clarify the relative effectiveness of various strategic paths.

While additional research could inform legal argumentation in support of gay rights, one inevitably suspects that the fate of such rights lies more in the composition of the judiciary than in creative legal arguments. As Koppelman concedes, "[t]he courts have enormous discretion in deciding whether gays are protected by the right to privacy, the right to marry, or the right against animus," and "[e]ven the strongest legal arguments do not guarantee success." Pinello's book demonstrates the factual truth of this conclusion. However, Koppelman may go on to go too far when he suggests that the "thought of homosexuality can still drive the judicial mind into desperate confusion." While there are certainly examples of such confusion, the evidence presented by Pinello suggests that this is not commonplace. Rather, the success of gay rights claims depends on the concatenation of attitudinal, institutional, and traditional legal factors. The circumstances in which gay rights claims prevail are fairly predictable, and Pinello concludes that "lesbian and gay reformers reasonably can invest hope in a litigative struggle" (p. 151).

B. The Strangely Negative Reaction to Quantitative Empirical Research

Given the importance of quantitative empirical research, the resistance to such research by the legal academy is both odd and unfortunate. In virtually every other field of academic research, quantitative measures are the gold standard for the discovery of descriptive truth. For legal academics, truth is commonly tested through case analyses. While these may be informative, reliance on the approach runs into obvious problems that are alleviated through quantitative methods.

While this is not the place for a full discourse on statistics, one aspect of quantitative empirical study needs to be understood. Just like more typical legal research, statistics is about drawing inferences from data. Unlike the amorphous standards of typical legal research, though, statistics has some logical rules that give much greater confidence that the inferences are valid. First, accepted quantitative practice requires a certain minimum number of data points. Just as one would not draw conclusions about Barry Bonds's batting based on his success in a single game or series, one should not draw conclusions about the law based on a very limited number of cases. Second, accepted quantitative practice demands that the data points be chosen randomly (from whatever larger set is relevant). Much legal research, including some of the best legal research, openly involves subjective case selection in argumentation. The cases selected may not be representative and raise the suspicion that they are influenced, consciously or unconsciously, by the researcher's own biases. Different authors commonly reach diametrically opposite conclusions from the same basic data. While this is understandable in the context of normative arguments grounded in different values, it is hard to justify in the context of descriptive research. Third, accepted quantitative practice demands transparent and disciplined procedures for analyzing that data. Without such transparency and rigor, the reader cannot know if the researcher considered all the relevant factors and in a reasonable manner.

The legal academy's undue reliance on unrepresentative anecdotes and failure to adhere to principled rules of inference has arisen in the judicial selection debate. Professor Rotunda, for example, emphasizes Scalia votes in favor of a Fourth Amendment plaintiff and on flag burning as evidence that the justices are not ideological. Yet a point is not proved by choosing two votes out of the hundreds cast by the justice. Such reliance on isolated anecdotes look like a desperate effort to cling to a fantasy, in light of the much more extensive empirical studies demonstrating that, on balance, judicial ideology makes a difference. Dean Cass maintains that the "evidence supporting the ideology argument is remarkably weak." Given the dozens of studies demonstrating some statistically significant effect for ideology, this claim is remarkable. None of those studies are referenced, though, by Dean Cass, who does not even acknowledge their existence. It is fair to debate the conclusions of those studies but it is hard to justify ignoring them. Professor Chemerinsky has taken the more empirically defensible position that "ideology should be considered because ideology matters" and "a person's ideology influences how he or she will vote on important issues." His claim could be more persuasive, though, had he used or at least referenced the considerable body of data that demonstrates the point.

Despite the value of quantitative empirical methods, legal academics have shied away from their use. Deborah Hensler has observed that lawyers, academics, and judges have all shown some level of "indifference or hostility to empirical research on ADR." Perhaps this is attributable to the "basic distaste to be found among many students of law for any other than their own traditional language." The negative reaction has even produced a "tendency to exclude scholars with divergent perspectives from scholarly conferences and other academic interchanges." Those of the law have "generally reacted with hostility to [empirical] attitudinal literature" on the grounds that it reflects "an overly-crude picture of judicial decisionmaking." While there is something to this criticism, it should not provoke hostility, it should instead provoke law professors to undertake their own searching and rigorous analyses to refine the literature and identify whatever measure of truth it may contain. Quantitative research has its own inevitable limitations in legal analysis and can be overly reductionist. It must be supplemented by more qualitative analyses, but quantitative research remains a vital tool for understanding the law.

The preparation of Gay Rights and American Law includes a valuable experience in the conduct of empirical research and the reactions of law professors that is described in the book. Much of the information necessary for coding judges' backgrounds is readily available from public sources, but Pinello was unable to identify the religious affiliation for some of the judges in his database. He sent a survey to judges to ascertain information about their background, including their religious affiliations, and received answers from around two-thirds of the judges, a very good return rate for such research (p. 214)

In order to ascertain information on religious affiliation for other judges, he posted an email message to a listserve for law professors (LAWPROF), asking members if they could give him this information. His request produced a measure of criticism, with law professors responding that this was "a private matter," or "none of your business" (pp. 156-57). There was an air of hostility in the reaction by law professors. Although Pinello had taken precautions to protect the privacy of individual judges and there was no reason to believe that his inquiry invaded reasonable expectations of privacy, the privacy concern was repeatedly raised. Perhaps the expressed privacy concerns were in fact masking a broader concern about the research.

No one responding to the inquiry expressly objected to the empirical analysis of the determinants of judicial decisions in principle, but many of the law professors seemed distinctly uneasy about the process. One poster particularly disagreed with the notion "that we ought to discover what we can about the privately held religious beliefs of judges so we can consider whether the statistical tendency of persons with similar beliefs to rule for or against groups with which we sympathize should lead us to favor or oppose their selection" (p. 160). Unfortunately, this disagreement came with no further reasoning, so one is left wondering why we ought not discover this fact. I suspect that it is either (a) a general discomfort with empirical analysis among legal academics or (b) a fear of discovering facts that, while true, are discomfiting. Many of us would prefer to pretend that background characteristics, especially religion, are not determinants of judicial decisions.

While legal academics appear somewhat leery of quantitative empirical research on courts, some federal judges have been downright hostile to the methodology. Judge Edwards has referred to some such research as "absurd," "odd, if not bizarre," "embarrassing[]," and "shockingly ill-reasoned." Judge Wald was less vituperative but equally aggressive in her attack on this research. Relatively few judges objected to the conduct of Pinello's research, as evidenced by his high return rate. A few, however, characterized his research as "irrelevant" or "offensive." One New York judge went further, writing to the author:

I fail to understand how the race, religious affiliation, or political party affiliation can possibly have any relevance or bearing upon what you characterize as a "scholarly journal article involving one or more decisions by our court." I am equally troubled by your inclusion of race and religion as part of the biography of a judge.

After Pinello wrote this judge a follow up letter, containing references to the published literature on decisionmaking, the judge responded:

I respond to your letter of ___________ and in so doing, reaffirm my disagreement and displeasure with your premise that the decisions of judges are related to and influenced by "political party affiliation, gender, religion, race, generational cohort and so forth . . . . Notwithstanding all of the sundry sources to which you make reference, the judges in this State, and particularly in this region, have by their oath of office sworn to uphold the constitutions of the United States and the State of New York, and in practice, dedicated to equal and impartial justice for all. Your obvious bent that judges in our courts favor those of the same faith, national origin, race, or political persuasion is offensive to every concept of equality, fairness and justice which we are sworn to uphold and strive to provide.

While the aspirations of this response are noble, two of its features are vivid. First is the cavalier dismissal of the "sundry sources," which were empirical articles, without any attempt to grapple on the merits with the result. Second is the refusal to participate in a study question. Pinello's research, after all, might have demonstrated that attitudinal variables like race and religion did not impact judicial decisionmaking. The judge in question perhaps feared that Pinello would find results such as those he did.

One should not make too much of this single response, because the vast majority of judges participated in Pinello's survey without objection. The judiciary is not generally attentive to or concerned about political science research, though. It is when such studies are produced in the legal literature by legal academics that judges have become exercised. It was law review articles to which Judges Wald and Edwards vigorously responded. The foundation of the judicial concern about the empirical research is obvious, as Judge Edwards worried about the "contemporary climate of skepticism towards courts." Judge Edwards' fear is that the research may "mislead the unsuspecting . . . into thinking that judges are . . . influenced more by personal ideology than legal principles."

The judicial concern for appearances is certainly understandable, but it cannot deny descriptive reality. For litigants, including gay rights litigants, it is the reality of outcomes that matters rather than appearances. Furthermore, appearances inevitably follow reality at some distance, as the truth can be obscured only so long. In fact, empirical research will not be so demeaning to the judicial enterprise as may be feared. While some political scientists have suggested that judicial decisionmaking is utterly political and that the law is meaningless, much of the best recent research demonstrates that the law does have a profound effect on judicial decisions, which Pinello's research confirms. This research should bolster our appreciation for the role of the law, though it is unlikely to support the traditional formalistic vision of the law.

It should be clear from the existing research that law's effect is not so simple and straightforward as may be taught in first-year law school classes. Studies like Pinello's are necessary to illuminate how, when, where and why law matters in judicial decisions. The determining role of law cannot be merely assumed. Relying on such assumptions about the law and the power of precedent merely leaves legal academics without persuasive ammunition when other disciplines present rigorous empirical analyses that appear to demonstrate that law does not matter. Quantitative analyses, such as that of Pinello, show that in fact the law does matter, but not in the simple formalistic manner that some legal academics assume. Empirical legal research is growing but remains at an embryonic level.

Much vital research on judicial decisionmaking remains to be done, and it should be performed by, or in collaboration with, law professors and published in legal journals. I have suggested above some intriguing questions posed but not answered by Pinello's data. One huge lacuna in the existing empirical research involves the significance of the language of opinions. The existing empirical research has focused overwhelmingly on judicial outcomes alone (e.g., which party prevailed), without examining the judicial language that set precedents for future rulings. This may be because outcomes are much more easily captured as a quantitative variable or perhaps because many of the political scientists doing the research don't believe that such precedential language matters. However, reliance on simple outcome coding can miss crucial legal developments. Outcome coding would treat Roe and Casey as identical, because both were rulings in favor of abortion rights. Lawyers of course know that the two decisions were not at all identical in their real-world implications. If "what judges say is even more important than how they vote," then research needs to move beyond simple outcome coding and attend to the importance of the content of opinions. By capturing such differences, law professors could greatly enhance the understanding of law. To have real persuasive impact, though, this research would require analytical rigor, such as that offered by the methods of quantitative empiricism. Simply citing anecdotal cases isn't enough.


The results of Pinello's empirical analysis are both intriguing and important. They illuminate and inform key legal controversies. There is a dispute in the legal academy between the significance of legal realism and legal formalism as determinants of judicial outcomes. Social scientists have weighed in strongly on the side of the realists. Both sides of this controversy have made exaggerated claims about the irrelevance of the others' claims.

Pinello's results provide no succor to either extreme. He finds clear evidence that both the law and politics matter, in substantial measure. Those who wish to entirely isolate the law from politics, eschewing political considerations in judicial selection in hopes of pure professionalization of the judiciary, are pursuing a vain aim. Judicial backgrounds, including political affiliations, are an important part of decisionmaking outcomes and it is willfully blind to attempt to ignore or evade that well-established descriptive fact. However, attitudinal politics is not destiny, and Pinello provides considerable evidence that the law itself matters greatly in outcomes. Judges will moderate or even negate their personal preferences in the presence of contrary legal commands. Rather than conflicting, the realism and legal frameworks interact with one another cybernetically, a relationship that requires much further research. Only when this association is understood can we truly understand the workings of the legal system.

Gay Rights and American Law contains relatively little normative content (aside from the author's occasional asides) and focuses on the descriptive determinants of decisions, but results such as these have substantial normative implications. The recognition that attitudes and institutional factors influence decisions is critical to judicial selection and the institutions that influence judicial decisionmaking. The Pinello results might also be important to the state/federal court question. They demonstrate the superiority of state courts for gay rights litigants and provide some basis for an implication that this superiority may be more general. Although the Lawrence decision may alter this conclusion, it bears noting that the U.S. Supreme Court's decision in that case was much slower to come than many comparable state court decisions in favor of gay rights. The results should illuminate the efforts of advocates to extend and apply the ruling in Lawrence.

Finally, Gay Rights and American Law is important to all further research on the law, as conducted by both legal academics and by social scientists. For academics, the most significant findings may be those regarding the significance of precedent. The book substantially informs our understanding of judicial decisionmaking but leaves many other questions unanswered. While traditional legal research may shed light on the answers to these questions, quantitative empirical investigation is necessary for a confident understanding of judicial behavior. Legal academics should not shun such analyses but should embrace them as central to the mission of legal research. While the use of quantitative empirical methods in legal research is growing, far too much research ignores such methods and the results of prior quantitative empirical studies. Professors of law should strive to improve, not evade, quantitative empirical legal studies.