In this article, we begin to respond to the deceptively simple question: How do cause lawyers decide when and where to litigate on behalf of their cause? And, to begin to unpack this question, we utilize methods that are relatively novel in the current literature on cause lawyers and cause lawyering.
We situate, with a few obvious temporal and institutional constraints, the locus of agency for this decision squarely with the cause lawyers themselves (see also Reference Barclay, Fisher, Sarat and ScheingoldBarclay and Fisher 2006; Reference Barclay, Marshall, Sarat and ScheingoldBarclay and Marshall 2005; Reference Levitsky, Sarat and ScheingoldLevitsky 2006; Reference Marshall, Sarat and ScheingoldMarshall 2006). This positioning of agency with the cause lawyer is in contrast to much of the existing literature, in which the locus of decision making for when and where to legally act has often been located external to the lawyers: it is a choice adopted by the larger social movement to which they are bound (e.g., Reference Sarat, Scheingold, Sarat and ScheingoldSarat & Scheingold 2006); or simply a reflection of the available legal opportunities in a defined jurisdiction (e.g., Reference AndersenAndersen 2005; Reference Ellmann, Sarat and ScheingoldEllmann 1998; Reference Michalowski, Sarat and ScheingoldMichalowski 1998) or some combination of both of these elements.
Alternatively, the decision of when and where to act is occasionally identified as an almost incidental by-product of the initial decision of lawyers to both act as a cause lawyer and for whom to act in this capacity (e.g., Reference HilbinkHilbink 2004; Reference Menkel-Meadow, Sarat and ScheingoldMenkel-Meadow 1998; Reference Sarat, Sarat and ScheingoldSarat 1998; Reference Shamir, Chinski, Sarat and ScheingoldShamir and Chinski 1998). Once the decision to act and for whom is made, it is implied that the choice of venue and its timing will follow automatically from this initial choice; a logical presumption if one is discussing cause lawyers engaging local welfare bureaucracies or city-based housing courts on behalf of poor persons or displaced tenants (e.g., Reference Kilwein, Sarat and ScheingoldKilwein 1998; Reference Scheingold, Sarat and ScheingoldScheingold 1998), but far less determinative if one is considering cause lawyers acting in coordination with emerging social movements challenging social norms that are ubiquitous in policy (e.g., Reference Meili, Sarat and ScheingoldMeili 2006).
By returning agency to the cause lawyers, we create a context in which the decision of when and where to act becomes a choice among alternative locations and potential time periods. A choice that is contingent upon some limited set of selection criteria that are invoked similarly by different cause lawyers in different time periods. And, the consideration of contingent choices with assumed commonalities across subjects, locations, and time is an area in which statistical models can offer important insights. Accordingly, we turn in this article to statistical models to model this behavior in order to unpack some of the factors that might lead cause lawyers to pursue a case in one location at a set time while rejecting opportunities at other locations.
Although the use of statistical models is not explicitly eschewed by scholars investigating cause lawyers and cause lawyering, it has been largely absent as an analytical tool in the existing literature. This dearth of statistical models reflects a scholarly recognition that, because cause lawyering, especially on behalf of social movements, often occurs in dynamic legal and political environments, the circumstances are intricate and varying in ways that do not always invite simple generalizabilty across different lawyers, locations, and time. As Reference Sarat, Scheingold, Sarat and ScheingoldSarat and Scheingold (2005: 12–13) noted, “What lawyers can do to serve their cause is shaped by a variety of factors, for example, the goals of the cause or movement, the resources that it can make available or that lawyers can mobilize, the possibilities at the practice site, the lawyer's own experience, skills, and understandings, the lawyer's social capital and networks, the nature of existing social, political, and legal arrangements, and so forth.” Nonetheless, even while acknowledging the constraints mentioned earlier, we will demonstrate the intellectual value of beginning to incorporate such statistical models to gain leverage on the nature of the commonalities underpinning the choices pursued by an important subset of cause lawyers, those who pursue high visibility litigation on behalf of social movements for marginalized groups.
The story we will offer in answering the question of how do these cause lawyers decide when and where to litigate on behalf of their cause is, at its heart, a story about the power of law to facilitate political and social transformation. As Reference Abel, Sarat and ScheingoldAbel (1998: 69) notes: “Because law constitutes the state, law can reconfigure state power. Because the state usually acts through law, the state can be constrained by law.” For their part, cause lawyers attempt to selectively and strategically wield this inherent duality in the relationship of law to state action; often in pursuit of larger social movement objectives. Given law's singular importance in projecting state power and its related ability to exemplify often unquestioned social norms, the choice of the location and timing of legal challenges by cause lawyers can represent defining moments in the life of a social movement and its possibilities for success (e.g., Reference CainCain 2000). The choice both manifests and subsumes the prior decision of the cause lawyer to act and to act through law at that moment rather than through alternative forms or institutions.
To consider this question, we use evidence from the timing and actions of cause lawyers in the marriage equality cases in the United States from January 1990 through December 2004. It is an apt example. Since 1995, cause lawyers have been an integral component of the movement for marriage equality. And, reflecting the fragmented nature of the movement for lesbian and gay rights generally (see e.g., Reference EskridgeEskridge 2002; Reference StoneStone 2012), many cause lawyers in a wide variety of states were faced with the decision of whether to litigate in their location in support of the emerging movement. Although there existed every appearance of similar legal opportunities and active political debates around the issue in a wide number of states, cause lawyers directly initiated litigation on marriage equality in only 13 of the 50 states in the period from 1990 through 2004 (Reference BarclayBarclay 2010).
Legal Opportunities
As a prerequisite to the initiation of litigation, cause lawyers require a suitable legal environment; one that permits them to utilize existing opportunities in the legal structures to engage, through the mechanisms of law, with the state to achieve desired policy goals (e.g., Reference Abel, Sarat and ScheingoldAbel 1998; Reference HilbinkHilbink 2004; Reference LevitskyLevitsky 2008; Reference Marshall, Sarat and ScheingoldMarshall 2006; Reference Scheingold, Sarat and ScheingoldScheingold 1998). Without a legal environment that includes such opportunities, the ability to initiate legal actions on behalf of a cause is precluded in that particular location (e.g., Reference Ellmann, Sarat and ScheingoldEllmann 1998; Reference Michalowski, Sarat and ScheingoldMichalowski 1998), even if they might remain available in other forums (e.g., Reference Abel, Sarat and ScheingoldAbel 1998). And, even if such legal opportunities exist in terms of formal accessibility, the goals of any litigation that can be pursued are severely constrained in the absence of a somewhat receptive legal environment—one with a willingness to engage openly, but necessarily supportively, with the ideas proffered by the cause lawyers (e.g., Reference Bisharat, Sarat and ScheingoldBisharat 1998; Reference McCann, Silverstein, Sarat and ScheingoldMcCann and Silverstein 1998; Reference Shamir, Sarat and ScheingoldShamir 2005).
Ideally, cause lawyers require the flexibility offered by the positive confluence between existing legal precedents and the evolving legal norms associated with such precedents (e.g., Reference Krishnan, Sarat and ScheingoldKrishnan 2005, 354–357). But, an appropriate legal environment is certainly not synonymous with a legal environment in which cause lawyers can expect to automatically prevail in their legal claim (e.g., Reference AndersenAndersen 2005). And, even legal claims that are identified as legally unsuccessful at the time can generate long-term positive results for a social movement (e.g., Reference NeJaimeNeJaime 2011). For example, the litigation efforts themselves can still advance important movement goals given even marginal receptivity to the claim from the courts (e.g., Reference Israël, Sarat and ScheingoldIsraël 2005). Reference Barclay, Fisher, Sarat and ScheingoldBarclay and Fisher (2006), writing about the early marriage equality movement, noted that initial cases allowed the nascent movement to reclaim the same-sex marriage issue as their own from other movements that had appropriated it previously. Similarly, Reference McCann, Silverstein, Sarat and ScheingoldMcCann and Silverstein (1998: 269), writing about the animal rights movement, note that “lawyers and non-lawyers alike noted the significant contribution litigation made to political education and publicity, and, in turn, to movement building.”
As the earlier discussion demonstrates, even the determination of an appropriate legal environment—an aspect that speaks directly to the skill sets shared by all cause lawyers, as in their formal legal training—is neither as simple nor as obvious as it might first appear. This complexity is evident in relation in our current example: the same-sex marriage cases initiated by cause lawyers.
Between January 1, 1990 and August 25, 2004, cause lawyers initiated same-sex marriage cases in 13 states: Alaska, Arizona, California, Connecticut, Hawaii, Indiana, Maryland, Massachusetts, New Jersey, New York, Oregon, Vermont, and Washington. At the time the respective case was initiated, not a single one of these 13 states had a state constitutional amendment proscribing same-sex marriage. In fact, at the time that the last of these cases is first filed with the relevant state court—in Connecticut on August 25, 2004—only three of the 50 states in total (Alaska, Nebraska, and Nevada) had such a state constitutional prohibition in place, even if many states would subsequently introduce such constitutional amendments, including an additional 13 states in early November 2004.
Similarly, although most states would eventually pass statutory restrictions on the celebration and recognition of same-sex marriage, not all states had done so by August 2004 and not all states would go on to do so. Furthermore, the presence or not of a statutory restriction in a state does not seem by itself to be singularly determinative of the choice of cause lawyers to litigate in a given location. Of the 13 states with cases filed by cause lawyers in this period, six states possessed statutory restrictions at the time and seven states did not (Reference BarclayBarclay 2010).
Based on the evidence from these same 13 states, a state's current policy position on other lesbian and gay rights' issues also does not seem to have been defining of an appropriate legal environment. For example, in 1974, the state courts in Massachusetts—one of these 13 states noted earlier—had curtailed the application of its statutory prohibition on sodomy. But, at the time that the cause lawyers first initiated the same-sex marriage case in Massachusetts on April 11, 2001, state police and prosecutors still actively and publicly enforced the statute against gay men. In fact, the same cause lawyers who initiated the marriage case in Massachusetts would contemporaneously initiate a legal challenge, in GLAD v. Attorney General (2002), to the application of the state's statute prohibiting sodomy. And, by April 2001, when that litigation in Massachusetts was filed by cause lawyers, state legislatures in 27 states had already repealed their existing statutory prohibition and state courts in an additional eight states had struck down their respective state's prohibition. And, obviously, by the time the last of these 13 cases was filed by cause lawyers in Connecticut in August 2004, the U.S. Supreme Court's decision in Lawrence v. Texas (2003) had created a similar legal environment around the issue of sodomy in all 50 states.
Clearly, as this evidence from the same-sex marriage cases reveals, cause lawyers were left with an array of possible choices for locations with an appropriate legal environments in which to initiate cases. This is true even if we begin by acknowledging that some states—such as those with clear, recent constitutional restrictions precluding same-sex marriage and/or states with a recent legal history revealing marked non-receptiveness to other issues in lesbian and gay rights—are less likely to appear to cause lawyers as having the necessary legal opportunities at the time of consideration.
For example, cause lawyers in selecting suitable legal environments although December 2004 might choose to eschew:
1. locations with state constitutional restrictions to same-sex marriage passed, or eventually to be passed, by December 2004Footnote 1 —Alaska, Arkansas, Georgia, Kentucky, Louisiana, Michigan, Mississippi, Missouri, Montana, Nebraska, Nevada, North Dakota, Ohio, Oklahoma, Oregon, and Utah; and,
2. locations where proscriptions on consensual sodomy among adults were removed by the US Supreme Court in Lawrence v. Texas (2003) rather than via earlier legislative or judicial action by the state itself—Alabama, Florida, Idaho, Kansas, Louisiana, Michigan, Mississippi, Missouri, North Carolina, Oklahoma, South Carolina, Texas, Utah, and Virginia.
Even if they avoided litigating in these two types of locations, cause lawyers would still be left with 26 states that apparently meet the criteria of a suitable legal environment. As it happened, cause lawyers would initiate litigation in only 13 of these remaining 26 states through December 2004.
This pattern of selection among very similar legal environments is particularly highlighted if we focus on one set of cause lawyers, the lawyers at the Gay and Lesbian Advocates and Defenders (GLAD), operating in only one geographic region, New England. The GLAD lawyers were involved in the same-sex marriage litigation in Vermont, which began in July 1997. And, the same GLAD lawyers initiated litigation on marriage in Massachusetts in April 2001 and in Connecticut in August 2004. Yet, notwithstanding marked similarities in legal environments, these lawyers never initiated litigation in nearby New England states of New Hampshire nor in Rhode Island. New Hampshire would legislatively introduce civil unions in 2007 and same-sex marriage in 2009. Rhode Island would legislatively introduce civil unions in 2011 and same-sex marriage in 2013.
Similarly, the immediate evidence from the GLAD example highlights that the “home” jurisdiction of cause lawyers also need not be a definitive indicator of the likelihood to sue in a particular location: GLAD's lawyers initiated the case in Vermont four years prior to initiating similar litigation in Massachusetts, notwithstanding the offices of GLAD being located in Boston.
Such a finding of an array of locations with the appearance of an appropriate legal environment should not be surprising. When acting in coordination with social movements representing marginalized groups, cause lawyers are often faced with many locations whose policies evince the requisite set of injuries that offer the subsequent basis for a legal claim. For example, the fact that the laws of many U.S. states in the 1930s and 1940s embedded a wide variation of segregationist policies allowed the cause lawyers challenging those policies to target selected states from among an array of possible options in pursuing their planned litigation (e.g., Reference TushnetTushnet 2005). The small subset of states chosen in the litigation in Brown were apparently chosen for very strategic reasons by the cause lawyers at National Association for the Advancement of Colored People (NAACP) from a much larger possible set of states reflecting such segregationist policies (e.g., Reference TushnetTushnet 1995).
In the marriage equality cases, the potential need to select from a larger set of states is reinforced by the fact that several of dominant legal organizations, such as GLAD, American Civil Liberties Union (ACLU), Lambda Legal, and the National Center for Lesbian Rights, operate organizationally with a defined regional purview.
Beyond Legal Opportunities
If an appropriate legal environment in the location is only the first component, rather than the sole component, in the selection of timing and location of litigation by causes, it raises the question as to what other factors might drive such lawyers to choose one location over another and one time over another time.
It is worth noting that these additional factors need not be operating solely in the realm of law. Once the legal environment is viewed as appropriate, there is no obvious additional legal “prompt” to signal if and when a cause lawyer should act to engage with this new environment. Therefore, the likely factors are presumably outside of the area of immediate expertise of the cause lawyer as a person trained in maneuvering within the formal law. In fact, given that similar legal opportunities exist across several locations, the question turns to why are cause lawyers not pursuing litigation in each of those apparently receptive legal jurisdictions. Consequently, from this point forward in the article, we frame these additional factors from the perspective of “prompts” to action by cause lawyers—these are the factors that act to actually propel cause lawyers to initiate litigation in one of the locations that they have previously determined to have a suitable legal environments.
In the section later, we outline and eventually statistically test several factors, beyond simply the presence of an appropriate legal environment, that are likely to influence the selection of a location and timing of planned litigation by cause lawyers.
Government Action
One simple answer to the question of factors shaping this choice by cause lawyers of location and timing is that this choice may not be entirely under the control of the cause lawyer. In some cases, action by government officials may preempt the normal determination of an appropriate location and timing by cause lawyers as they instead find themselves defending the policy actions of others in a place and on a timetable set by others. For example, when local government officials in California and New York, including Mayor Newsom in San Francisco and Mayor West in New Paltz, introduced same-sex marriage through local government action in February and March 2004, respectively, cause lawyers in those states were rapidly drawn into pursuing litigation—litigation that they themselves had not previously initiated up until that point in either California or New York—as they sought to defend these unexpected policy gains.
Importantly, the impetus for action by elected government officials can occur for very different political and legal reasons than the factors that might prompt cause lawyers to invoke the law on behalf of social movement goals. Clearly, like social movements, elected government officials are likely to be aware of mounting pressure for action generated by individual constituent requests and/or social movement activity around the particular policy issue. And, as we will see below, the timing of these government actions seem to be consistent with rising trends in general public opinion in the locations in questions. Yet, the timing of government action need not necessarily be reflective simply of attitudinal shift among the general population or elite sub-groups generated to-date by social movement action. For example, it can also reflect a complex political calculation by individual elected officials involving the precipitous shifting of bureaucratic, coalitional and/or electoral possibilities that could possibly be achieved through publicly pursuing a policy shift at that chosen time. As such, it could be about a policy actor seeking increasing institutional influence, publicly staking a claim to distinguish their position in a fluid political coalition, or electorally conscious officials preemptively seeking support from a previously unincorporated, but increasingly influential, set of constituents.
Public Opinion
It appears counterintuitive to consider the level of public opinion when addressing the question of the prompts to legal action. Courts presumably operate from a different basis than their more electorally focused institutional counterparts. After all, it is this very presumption of a different basis for acting that is often thought to motivate social movements to contemplate litigation rather than lobbying the legislature, particularly when pressing the claims of politically marginalized groups (e.g., Reference VoseVose 1958, cf. Reference OlsonOlson 1990). Nevertheless, there are still reasons to believe that attitude shift among the general population impacts the determination of location and timing by cause lawyers.
There is obviously some limited correlation directly between the trends in public support and the direction of subsequent judicial action (e.g., Reference BarnumBarnum 1985; Reference Mishler and SheehanMishler and Sheehan 1993). If nothing else, judges are members of a society and are likely to reflect many of the same social understandings, including social stereotypes about marginalized groups, as predominate generally at that time (e.g., Reference RichmanRichman 2009). And one might expect cause lawyers to pursue cases in locations where judges are at least minimally predisposed to actually engage the legal claim; an aspect that might be most easily evidenced by the presence of a growing trend of support for the claim among the general public (e.g., Reference BarnumBarnum 1985). Yet, this approach is not our present basis for considering the state of public opinion in reference to decision making by cause lawyers.
Beth Robinson, one of the three cause lawyers who in July 1997 initiated the same-sex marriage case in Vermont, noted about the relationship of legally pursued causes and public opinion:
We can craft the best arguments in the world in court, but if we're not also standing in booths at the local county fair, or speaking to a nearby church congregation on a Wednesday night, we risk losing in the court of public opinion our well-deserved gains in the court of law (Reference RobinsonRobinson 2001: 243; see also Reference Barclay, Marshall, Sarat and ScheingoldBarclay and Marshall 2005: 187).
Within the quote, Robinson clearly distinguishes the reception of legal arguments by judges from the expected reaction of the population to any resultant court decision. According to this idea, the cause lawyer's assessment of public opinion arises from a defensive posture—it seeks to avoid popular opinion generating a legislative backlash to judicial action in ways that counter any policy advances won in court (see also see Casillas, Enns & Reference Casillas, Enns and WohlfarthWohlfarth 2011). It effectively captures the complex reasoning that is likely operating behind a cause lawyer's consideration of public opinion in determining the timing and location of any litigation.
Yet, there are practical problems that arise with proposing that cause lawyers might be utilizing the consideration of state-level, public opinion data to determine the appropriate location and timing for initiating a legal claim on behalf of a cause. The first problem is one of temporal directionality. Public opinion on an issue is often not sampled by polling organizations until well after a number of cases are brought by cause lawyers—it appears that the initiation of litigation is more likely to prompt the surveying of popular opinion on the subject rather than vice versa. For example, the Gallup Organization did not ask questions on same-sex marriage until March 1996; 25 years after the first same-sex marriage cases were filed by cause lawyers (see Reference Barclay, Fisher, Sarat and ScheingoldBarclay and Fisher 2006), five years after cause lawyers initiated the Hawaii case in December 1991 (Reference EskridgeEskridge 2002), and three years after the Hawaii Supreme Court first ruled on those legal claims in Baehr v. Lewin 1993. Similarly, the first questions to assess the general public's position on interracial marriage were only asked by the Gallup Organization in September 1958; over a decade after cause lawyers in California initiated the first modern case, Perez v. Sharp 1948, challenging state prohibitions on interracial marriage.
The second problem arises around the meaning reflected in responses to polling questions. Cause lawyers are prone to need information on new approaches to existing social norms. Opinions expressed by individuals on many new or controversial issues are not always stable enough to be reliable as indicators to the subsequent pursuit of action. Even on policy issues, which appear to dominate the political discussion of the day, a substantial portion of any survey population expresses what public opinion scholars subtly call “nonattitudes” (e.g., Reference Pierce and RosePierce and Rose 1974). More disconcerting for those who seek to use the results of public opinion data, the apparent lack of knowledge on a topic appears to be no barrier to an individual's passionate or forceful support of an opinion when one is solicited by a public opinion survey. “Most people know little about politics but are nonetheless willing to give opinions on even the most esoteric policy issues when asked to do so by survey researchers” (Reference AlthausAlthaus 2003: 2). This leads to great instability in opinion. According to Gallup polls, public opinion on the issue of same-sex marriage apparently moved as much as eleven percent within a single month; usually without even a contemporaneous political event to prompt this shift. For example, in March 2005, only 28 percent of the surveyed population supported the recognition of same-sex marriages whereas a month later in April 2005, 39 percent of the same population apparently now endorsed same-sex marriage when prompted by the exact same question.
The third problem is the appropriate concept to be captured. It is not always clear what is the correct measure to observe in making an assessment of the population's actual position around a new approach to existing social norms. For example, in July 1968, only 19.8 percent of respondents in Gallup polls approved of interracial marriage—a new high to that point in time. At first glance, the level of public support would appear to dissuade a cause lawyer from filing a case in this area at that time. Yet, this admittedly diminished level of support occurred five years after a cause lawyer initiated the claim and a year after the U.S. Supreme Court ruled unconstitutional such state prohibitions in Loving v. Virginia (388 U.S. 1). One possible reason for not directly using the perceived level of popular support only on interracial marriage in determining this choice could be that anti-miscegenation claims were already identified as part of a larger set of legal and political claims associated with the transition in race relations in general rather than solely classified based on the specific issue of interracial marriage. A similar claim could be made about the second wave of same-sex marriage litigation, which occurs as lesbian and gay individuals are enjoying unprecedented support for removal of employment discrimination. According to the Gallup Organization's survey, 80 percent of the surveyed population supports “equal rights in terms of job opportunities” by April 1993 and the level of support continues to climb throughout the next decade.
Finally, public opinion data is not always effectively differentiated down to the state-level, but instead is usually targeting shifts in national public opinion around emerging policy issues. Even in the best case scenarios, reliable public opinion is collected only sporadically at the state-level and usually by media sources that do maintain consistency in the focus or wording of questions over time. And, we could postulate that most cause lawyers lack both the resources and expertise to fund their own public opinion polling, even if they had the desire to do so. And, they are certainly unlikely to be able to do so across several locations and multiple time periods as would be required to accurately assess changes in the general opinion of the population.
Local Print Media
Cause lawyers might also be prompted into action by changes over time in the level of support for their cause demonstrated by the local print media. This factor has the advantage of being both visible and easily accessible to cause lawyers as they contemplate when and where to pursue possible legal action. In fact, it is this very aspect of ease of accessibility that potentially places the current factor as a supplement to the consideration of the level of public opinion noted earlier. Given the difficulty of obtaining reliable public opinion data across locations and time, the position of the dominant local newspaper in each location might be considered an apt substitute by which to assess the current position of that jurisdiction's population. After all, newspapers are often thought to reflect their readers' values in part to ensure good continued circulation.
Cause lawyers, eager for signs that a location that evinced a suitable legal environment might also offer other signs of its appropriateness for litigation, could parse the changing levels of support offered by local print media for their cause. The prompt to action presumably would be evidenced by a substantial and positive upswing in support over time; that is, a difference over an extended period between the historical levels of support and the most recent level of support for the cause demonstrated in the same newspaper. Such a marked difference signals that the cause is now front and center on the political agenda in that location as far as the print media is concerned.
Returning to the defensive theory raised in Beth Robinson's quote and noted earlier in relation to public opinion, we might expect that the fact of being highlighted as part of the current political agenda by the print media in that location means that the generation of sufficient political support for any resultant legal action have been met. Conversely, it might also be taken as evidence that the issue has generated sufficient political interest that government action is soon likely on it. Under our current approach to explaining the choice of timing and location, either perspective is likely to influence cause lawyers in their determination, as each element relates to a factor that we have already noted earlier as being potentially influential in this selection.
Testing the Source of Cause Lawyers' Impetus to Action
To evaluate the factors that prompt cause lawyers into action, we constructed a survival model (also known as an event history model or duration model) that considered the decision to initiate a legal claim over same-sex marriage between January 1, 1990 and December 31, 2004. As noted earlier, cause lawyers initiated litigation directly on the issue of the state constitutionality of same-sex marriage in 13 states during this time period. Those 13 states were (in chronological order): Hawaii, New York, Alaska, Vermont, Arizona, Massachusetts, Oregon, Washington, California, Indiana, Maryland, New Jersey, and Connecticut.Footnote 2 The initial litigation by cause lawyers on this issue began in Hawaii on May 1, 1991, and the last case was filed in Connecticut on August 25, 2004.
In the analysis later, we focus on a small subset of four of those 13 states where cause lawyers litigated and compared them with a small subset of seven of the remaining 37 states where cause lawyers did not litigate during this time period. These limitations on the analysis occur primarily because of inherent limitations in access to data about each location (as we outline later). Consequently, the analysis in the survival model focuses only upon the following 11 states (in alphabetical order): California, Colorado, Connecticut, Florida, Illinois, Kansas, Massachusetts, New York, Pennsylvania, Tennessee, and Utah.
Nevertheless, these 11 states incorporate an array of divergent characteristics around the question of whether and where to initiate litigation on behalf of same-sex marriage. In four of these states—Massachusetts, California, Connecticut, and New York—cause lawyers initiated litigation over same-sex marriage. In two of these four states—Massachusetts and Connecticut—the timing of the filing was selected solely by the cause lawyers. In the other two states—California and New York—cause lawyers filed cases contemporaneously with local government action on same-sex marriage. In California, the local government action was the decision of the Mayor of San Francisco to begin celebrating marriages in February 2004. In New York, the local government action was the decision of the Mayor of New Paltz and other municipalities to begin celebrating marriages in March 2004.
In the remaining seven states—Colorado, Florida, Illinois, Kansas, Pennsylvania, Tennessee, and Utah—cause lawyers did not pursue action. In one of these seven states—Utah—the population introduced a state constitutional amendment at the very end of the period under consideration (November 2, 2004), and in another state, Kansas, they did so shortly thereafter (April 5, 2005). In contrast, Illinois would eventually legalize civil unions in June 2011 and pass marriage equality in November 2013; both legal changes occurring through legislative action. Colorado would introduce Civil Unions in March 2013 through legislative action. As of the time of writing, cause lawyers are litigating in the federal courts against prohibitions on same-sex marriage presently in place in Pennsylvania, Tennessee, and Utah. Finally, it is worth noting that statutory restrictions on same-sex marriage operated in both a state in which cause lawyers initiated litigation, California, as well as in other states where they did not initiate litigation.
Missing from the selection of 11 states are the two states that were the site of the very first cases on this cause in the 1990s: Hawaii and Alaska. The decision to exclude these states was purposeful. The earliest litigation initiated in pursuit of a cause can expect, by definition, to be occurring in a political environment that is mostly hostile to the proffered claim; if such hostility did not exist to the position advocated by the cause lawyer, they would not be required to engage in the current litigation (Reference Abel, Sarat and ScheingoldAbel 1998; Reference Scheingold, Sarat and ScheingoldScheingold 1998; cf. Reference Southworth, Sarat and ScheingoldSouthworth 2005). Thus, the decision to initiate legal action at that point in time is likely based on other incentives, such as issue claiming and issue framing (e.g., Reference Barclay, Fisher, Sarat and ScheingoldBarclay and Fisher 2006).
Measuring Government Action
We created a simple dichotomous measure that captured when government officials acted and denoted the immediate time period after that action. As noted earlier, mayors in both California and New York changed standing policy on the issue of same-sex marriage in February and March 2004, respectively.
Measuring Public Opinion
To generate a reliable measure of public opinion on lesbian and gay rights, an index score was created from the responses to three separate questions asked in each of the 11 states in the American National Election Studies (ANES) throughout the period, 1990 through 2004. The first element was the percentage of positive support for gays and lesbians on a feeling thermometer (v232). The second element was the percentage of positive support on the question (v876): “Do you favor or oppose laws to protect homosexuals against job discrimination?” The third element in the index was the percentage of positive support on the question (v878): “Do you think gay or lesbian couples, in other words, homosexual couples, should be legally permitted to adopt children?” We used linear interpolation to fill the missing data associated with each of these three elements. Subsequently, we divided by three the combined scores for each state in each surveyed year on these three elements to form an index score that could only range in the positive from 0 to 100.
The data from ANES offers the advantage that the questions were consistently asked throughout the time period and across all 11 states. This allows us to effectively compare changes in public support over time and location. This element resolves some of potential problems associated with using public opinion data. Although this ANES data may not have been immediately accessible to cause lawyers in making their assessments (as there is an almost 2-year lag until release), we can presume that less scientific surveys conducted by the media in these locations followed the same general trend over time and reported similar results. Using an index score that includes a question related to general levels of acceptance of lesbian and gay individuals (the feeling thermometer), a question related to employment discrimination (the job discrimination question), and a question related to issues of family (the adoption question), this measure of public support responds to the potential issue that same-sex marriage is simply viewed by most individuals as part of a more comprehensive consideration of lesbian and gay rights.
Measuring Print Media Support
To measure local print media support, we collected data through two databases. LexisNexis provides data for some leading national newspapers included in this study, but it covers only a small and unrepresentative sample of local and regional newspapers. In contrast, NewsBank's Access World News database includes a much larger number of local and regional newspapers. Both sources were used in this study. In two of the 11 states, Utah and Florida, we collected information from more than one newspaper as there was more than one dominant newspaper operating in each of these locations.
Using these two databases, we created a measure of demonstrated print media opinion on same-sex marriage through the use of several search terms, particularly “gay marriage.” And, we focused only on editorial opinion—official editorials and columns—to reduce quantitative variation across the newspapers because of institutional differences in resources, reporters, news space, and breadth of coverage. Opinion pieces provide relatively clear expressions of policy preferences, and the official editorials are likely to reflect the institutional preferences of the newspapers, and may be mirrored by more subtle, but similar biases in the news pages (Reference BreedBreed 1955; Reference ChomskyChomsky 1999). We included all opinion pieces that took definable positions on same-sex marriages, domestic partnerships, and/or civil unions. Most took explicit positions; some eschewed explicit statements of preference, but leaned in one direction, and a position was attributed to them. Editorials and opinion pieces that did not take any identifiable position were excluded from the sample on the grounds that if we, as highly interested researchers, could not discern a position on the issue from the editorial or opinion piece, we believed that it was unlikely that any cause lawyers reading the article would also be able to discern one.
We counted the total number of official editorials and columns in each of the chosen newspapers in each month between January 1990 and December 2004 that contained a discernable position on same-sex marriage or domestic partnerships and civil unions. This created three separate measures depending on the position adopted by the newspaper in any editorial or column. Editorials and columns could be pro–same-sex marriage, pro-civil union/domestic partnership, or anti–same-sex marriage. And, finally, to create a measure of the print media's level of support in the locations in question, we simply created a score that accumulated and summed the number of pro–same-sex marriage editorials and opinion pieces for the 18 months prior.
Indications from the Survival Model
The survival model considered the likelihood of cause lawyers initiating in any location based on three criteria: (1) the presence of government action in support of same-sex marriage as noted by a dichotomous variable; (2) the level of public opinion on lesbian and gay rights in each location in each month using the created public opinion index score; and (3) the number of pro–same-sex marriage editorials and opinion pieces for the 18-month period in each location in each month. The model was constructed around an assumption that the likelihood for initiating a case was changing throughout the period—it presumably became slightly simpler as cause lawyers garnered more experience, litigating the issue in state courts and as the social norm they sought to challenge become increasingly de-naturalized through the ongoing social discourse that the earlier judicial and legislative actions were generating. This expectation was incorporated into the survival model using the parametric form of a Weibull distribution to constrain and model the data. The results are reported in Table 1.
Table 1. Parametric Survival Model (Using a Weibull Distribution) on the Initiation of a Case by Cause Lawyers in 11 U.S. States, January 1990–2004
Notes: *Significant at the 0.05 level.
N = 2382; subjects = 13; failures = 4.
Likelihood ratio (LR) χ2 (3) = 177.97 p = 0.000.
Obviously, 13 papers is an extremely small N for statistical purposes. Maximum likelihood estimation models, including this parameterized form, tend to remain relatively robust even in small-N situations. And by considering each time period per location separately, the present statistical technique mitigates to some extent the potentially biasing effects by its treatment of the sample. Nevertheless, this limitation in the degrees of freedom imposes a limitation upon the amount of factors introduced into the model; we introduce only the three variables listed earlier and constrain the constant. And, given these restrictions, we are primarily utilizing the model's results as initial indicators of expected influences rather than as solid evidence of strength of effect of any or all of these influences.
The results of the model indicates that, when government officials move of their own accord on the issue of same-sex marriage as occurred with mayors in California and New York in 2004, cause lawyers can be quickly expected to follow with litigation. Such a response makes sense as an attempt to defend the actions of these government actors who have set the agenda on this issue in that location. It also makes sense as a means for cause lawyers, who have long been involved in the issue, to reclaim the issue in ways that allow them to shepherd the strongest possible arguments into the legal and political arena.
Cause lawyers in California and New York litigated quickly in pursuit of both of these goals. In California, cause lawyers were involved in Lockyer v. City and County of San Francisco (2004) in defense of Mayor Newsom's actions in San Francisco. And, they filed the claims initiating Woo v. Lockyer (2005 and 2006) and In Re Marriage Cases (2005, 2006 and 2008) directly seeking same-sex marriage in the state. In New York, cause lawyers were involved in People v. West (2004) in defense of Mayor West's actions in New Paltz. And, they filed the claims initiating Hernandez v. Robles (2006, 2005b and 2006), which was directly seeking same-sex marriage in the state.
Similarly, the results of the model indicates that the level of state public opinion in support of lesbian and gay rights generally appears to influence the decision by cause lawyers to initiate legal action. Interestingly, the sign for this prompt is negative, signifying that states with higher level of public support for lesbian and gay rights are not necessarily the most likely to prompt action by cause lawyers. Although this appears counterintuitive, it is consistent with the evidence: cause lawyers in two of the locations with generally supportive local populations, California and New York, did not initiate litigation on this issue until prompted by government action, as noted above. Moreover, it is worth noting that the current model is considering the influence of public opinion in relation to both government action and local newspaper support; each of these latter factors integrate indirect mechanisms for incorporating the general attitudes of the local populace on issues.
Finally, the current model indicates that the presence of a substantial and positive upswing in the level of support for their cause in the dominant, local newspaper also prompts cause lawyers to initiate legal action in that location at that time. This result is consistent with the change in levels of support for same-sex marriage offered by the dominant, local newspaper in Massachusetts, New York, and Connecticut around the time of the litigation by cause lawyers in those locations.
The Boston Globe, the dominant, local paper in Massachusetts, averaged 0.17 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 6 months) in the period from January 1990 through September 1999. In contrast, in the 18 months prior to April 11, 2001, the date when the cause lawyers at GLAD initiated Goodridge v. Dept of Public Health (2003), the Boston Globe averaged 0.44 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 2 months). The difference in these averages was statistically significant (t = −2.24, p = 0.027). This relationship is well represented in Figure 1 where the dashed red line symbolizes the average in the pre-prompt period and the dotted blue line symbolizes the average in the 18-month prompt period. And the impact of the effect can be evidenced in the distance between these two lines—the greater the distance, the greater the difference in the stance of the newspaper in the 18-month prompt period from its prior behavior.
Figure 1. Same-Sex Marriage and Civil Union Editorials in Boston Globe.
Similarly, the Hartford Courant, the dominant, local paper in Connecticut, averaged 0.09 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 11 months) in the period from January 1990 through January 2003. In contrast, in the 18 months prior to August 25, 2004, the date when the cause lawyers at GLAD initiated Kerrigan v. Dept of Public Health (2006), the Hartford Courant averaged 1.16 pro–same-sex marriage editorials and opinion pieces in any month (or at least one such editorial or opinion piece every single month). The difference in these averages was statistically significant (t = −9.33, p = 0.000).
Finally, the New York Times averaged 0.09 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 11 months) in the period from January 1990 through August 2002. In contrast, in the 18 months prior to March 2004, the date when the cause lawyers initiated Hernandez v. Robles (2005), the New York Times averaged 1.28 pro–same-sex marriage editorials and opinion pieces in any month (or at least five such editorial or opinion piece every four months). The difference in the averages between the prompt period and the preceding period was statistically significant (t = −6.55, p = 0.000).
This finding that cause lawyers responded to a substantial and positive upswing in the level of demonstrated support apparent in the local newspaper might also help explain those incongruous examples where cause lawyers failed to act despite an apparently favorable legal environment and a supportive local populace. For example, notwithstanding an increasingly supportive level of public opinion, the current measure of print media support offers insights into the reluctance of cause lawyers to become involved in the first New York same-sex marriage case of the 1990s, Storrs v. Holcomb (1996). That case was filed pro se on May 18, 1995. Cause lawyers associated with Lambda Legal Defense Fund as well as other lesbian and gay rights' legal groups were reluctant to become involved in the claim at that time and apparently actively discouraged the case from being filed (see Reference AndersenAndersen 2005; Reference Barclay and FisherBarclay and Fisher 2003). This decision by cause lawyers in New York not to initiate litigation was somewhat surprising. Same-sex marriages cases were being pursued contemporaneously in other locations, such as Hawaii and the District of Columbia. And, cause lawyers had in fact joined existing litigation in the Hawaii case despite their initial reticence (Reference EskridgeEskridge 2002). Further, New York State at that time lacked any clear statutory prohibition on same-sex marriage and its marriage statutes were not even “gendered” in relation to the parties (e.g., Reference HalliganHalligan 2004). Finally, New York had been a major site in the history and development of lesbian and gay rights, including being the site of the Stonewall Riots in July 1969. In 1995, it was the home to several influential lesbian and gay rights' groups, including Lambda Legal Defense Fund (Reference AndersenAndersen 2005).
Yet, assuming that cause lawyers for these groups were using even a simplified version of the current measure of support by the local print media, they would have determined that the location was not ready for them to initiate at that time. This result is confirmed statistically (t = 0.622, p = 0.535). The New York Times averaged 0.02 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 50 months) in the period from January 1990 through April 1995. And, it averaged approximately zero pro–same-sex marriage editorials and opinion pieces in any month during the 18-month prompt period. Notwithstanding occasional editorials and opinion pieces, the New York Times was not demonstrating a marked increase in the level of demonstrated support for same-sex marriage during this period. As noted earlier, this is in clear contradiction to the period immediately preceding the filing of the claim in the later New York same-sex marriage case, Hernandez v. Robles (2006).
A similar pattern emerges in the slightly more complex case offered by the decision in February 2004 of San Francisco's Mayor Gavin Newsom to issue marriage licenses to same-sex couples. Here, the reticence of the cause lawyers to litigate on this issue prior to Mayor Newsom's actions is in question, rather than their response once this government official had acted. Their reticence appears more logical if we assume that the cause lawyers were responding to the cues set by the local print media. In addition, the California example also highlights the important distinction between simply identifying locations with demonstrated ongoing support for the cause and those locations in which there is a substantial and positive upswing in the level of such support in the local media.
For example, the dominant local paper, the San Francisco Chronicle, had a strong history of support for same-sex marriage. Yet, when we consider the lack of any recent substantial and positive upswing in its demonstrated support for the issue, it might explain why cause lawyers initially assumed that the location was not conducive to such a claim at that time; at least until the Mayor acted. This result is confirmed statistically (t = −0.327, p = 0.744). The San Francisco Chronicle averaged 0.06 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 17 months) in the period from January 1990 through July 2002. But, the San Francisco Chronicle only differed slightly in their demonstrated support in the period from August 2002 through January 2004. During this latter period, it averaged 0.11 pro–same-sex marriage editorials and opinion pieces in any month (or approximately one such editorial or opinion piece every 9 months). The slight difference between these averages is well represented in Figure 2.
Figure 2. Same-Sex Marriage and Civil Union Editorials in San Francisco Chronicle.
Conclusion
In coordination with social movements, cause lawyers seek to transform existing social norms and current policy. Although a key component of their activities in pursuit of these goals involves the strategic use of litigation, cause lawyers' power to redefine prevailing social norms often arises as much from the public, performative aspects afforded by litigation as from any resultant legal ruling by a court (e.g., Reference McCann, Silverstein, Sarat and ScheingoldMcCann and Silverstein 1998). Beyond promoting the issue for both potential members and the larger public, litigation facilitates the documentation of new scientific evidence, the marshaling of persuasive arguments for improving the social positioning of the group, and the ability to situate through analogy the proposed change in policy (e.g., Reference McCannMcCann 1992). These aspects of litigation allow cause lawyers to speak directly to the social meaning attributed to social norms and address the authority these norms are accorded in current social practice. As much as they are targeting the removal or reinterpretation of the formal legal component of law, cause lawyers for social movements are primarily interested in using litigation and the forum offered by the court case to reconstitute the social construction of identity, power, and social practices that are embedded within those formal laws and their subsequent enactment by the state. Consequently, we should not be surprised that the current research endorses the notion that the public's attitude in a location—as evidenced by both public opinion and the local media's position—are as important as the legal opportunities that operate in that same location.
Moreover, the selection of the location and the timing of litigation generate their own substantial contribution to the path dependency that emanates from planned litigation. Planned litigation of the sort pursued by cause lawyers on behalf of movements involves building over time and across cases. Beyond precedential value, the legal development of an issue is shaped by the choices made about timing and location as much as it is influenced by framing of the issues incorporated into the filed briefs.
For example, the choice by cause lawyers to litigate same-sex marriage claims in Vermont in the late 1990s refocused the movement, which had become fragmented and unfocused from fighting ballot initiatives on a state-by-state basis (Reference StoneStone 2012). The briefs filed by the cause lawyers redefined the salience of certain legal arguments about marriage, procreation, and parenting in ways that redounded generally in the larger public discourse. And, the process used by the cause lawyers for inviting amici on key concepts in those arguments created an organizational template for future cases pursued by the larger movement. Yet, to the movement's obvious displeasure, the actual policy outcome in that Vermont case single-handedly legitimized the concept of “civil unions” as a viable state alternative to the recognition of marriage equality. As this shows, the movement's advancement was shaped by the timing and location of the litigation, even as the actual advances on the policy front was more mixed.
Finally, in the dynamic dance between state legislatures and state courts, which denote the path to policy making adopted by most states on the issue of marriage equality (e.g., Reference BarclayBarclay 2010), the timing and location of litigation by cause lawyers may speak as much to legislative possibilities as legal opportunities. Cause lawyers may be litigating to engage in legislative agenda setting, assist in the framing of legislative debates, as well as to encourage legislative, rather than judicial, resolution on the policy issue.
For example, in New Jersey, cause lawyers from Lambda Legal filed the initial briefs in that state's marriage equality case, Lewis v. Harris, in July 2002 and the very presence of this ongoing litigation prompted the state legislature to introduce domestic partnerships in 2004. Using this legislative action as a basis for the litigation, cause lawyers then refiled to prompt the state legislature into transitioning from domestic partnerships to civil unions, which were introduced by the legislature in December 2006. With legislative backing, cause lawyers again filed litigation in Garden State Equality v. Dow 2013 to challenge successfully the existing use of civil unions by New Jersey and to push for the introduction of marriage equality against gubernatorial resistance. A key component in this activity was the very publicness of that litigation, which eventually prompted the governer to withdraw his legal challenge. As is obvious from this example, the timing and location of litigation by cause lawyers in New Jersey was much more about playing to political dynamics than it was about legal outcomes in the particular legal case.
