BERKELEY, 1983: LAW AND SOCIOLOGY IN RELUCTANT CONVERSATION

Well into her graduate studies in sociology at Stanford, Laurie decided to go to law school. Her decision was not motivated by doubts about her prospects as a sociologist, but rather because she was serious about understanding law, not only as a sociologist but also as a lawyer. So, in the fall of 1983, Laurie began law school at Berkeley, where she and Catherine were classmates.

Although the Jurisprudence and Social Policy program at Berkeley provided a rich environment for sociolegal study in 1983, Berkeley's law school had a different focus. Especially in the first-year subjects, most faculty members at that time were firmly committed to the study of legal doctrine, simpliciter.

Laurie was curious and willing to listen, but she resisted unquestioning devotion to the established 1L project of crunching doctrine, juggling hypotheticals, and mastering the blackletter law. In Berkeley's fine tradition of student rebelliousness, Laurie persistently brought sociological perspectives to the study of cases. Law students often do not realize that the point of a Socratic dialogue is for students to learn from one another. But as Laurie's classmate, Catherine was delighted by her resistance to the usual monocular focus on legal doctrine, and frequently found Laurie's expansion of what was relevant and interesting about the cases to be the most edifying and inspiring part of class.

In many ways, Laurie's experience of law school solidified her identity as a sociologist. She would often state, with a sly look and a wry smile, that she was “not a lawyer.” And indeed, throughout her career, she used sociological methods to investigate research questions generated by sociological theories. But as the years progressed, the collaborative research groups she assembled increasingly included lawyers. She did this as she came to understand that the increasingly sophisticated research questions she wanted to investigate required using elements of legal doctrine as independent variables in her statistical analyses. In this, her 1996 move from Wisconsin to Berkeley proved pivotal.

BERKELEY, 1996: BREAKING THROUGH DISCIPLINARY BOUNDARIES

In 1996, Laurie and Linda went out to lunch at Berkeley, where they had both just joined the faculty, to discuss the prospect of working together. Linda asked Laurie to describe her work on socio-legal endogeneity (which Linda could barely pronounce) in employment discrimination law (about which Linda knew quite a bit). As Laurie talked, Linda sketched a counter-clockwise circular model of Laurie's socio-legal endogeneity theory onto a napkin, beginning with the passage of ambiguous antidiscrimination laws and the emergence of organizational complaint handlers (Reference Edelman, Petterson, Chambliss and ErlangerEdelman et al., 1991), and proceeding through the generation of symbolic compliance structures (Reference EdelmanEdelman, 1992), the inflation of liability risk (Reference Edelman, Abraham and ErlangerEdelman et al., 1992), the generation of rational myths (Reference Edelman, Uggen and ErlangerEdelman et al., 1999), and employers' use of compliance structures in EEO disputes (Reference Suchman and EdelmanSuchman & Edelman, 1996).

But the model stopped shy of 12 o'clock. The last step in Laurie's legal endogeneity research agenda—showing that symbolic indicia of compliance and their underlying managerial logics actually did get incorporated into the judges' decisions about the legal meaning of “discrimination”—had yet to be attempted. Laurie described what such a study would entail, including the assembly of a sufficiently large sample of cases, coding procedures, and the types of variables she contemplated including in regression models. Ever the lawyer, Linda found it troubling that none of the independent variables pertained to employment discrimination doctrine, and she wondered how one would operationalize “judicial deference” to institutionalized employment practices, which Laurie posited as the study's main dependent variable.

As they talked through these issues, disciplinary differences quickly manifested, vexing communication. Consider terminology. “Deference” has a particular meaning in law, one having to do with legal rules about the weight judges should accord administrative agency regulations in statutory interpretation. Laurie was using the term as a shorthand way of describing a certain pattern of judicial inference. It took Laurie and Linda much of their lunchtime together just to figure out that when they spoke of “deference,” they were talking about two different things.

More fundamental were differences in their initial ideas about what variables needed to be controlled to convincingly demonstrate that it was the presence or absence of symbolic compliance structures, and no other variables, that predicted the “deference” phenomenon. For Linda, the absence of independent variables relating to aspects of legal doctrine was a serious problem. To test the last step in the legal endogeneity process, Linda suggested, one would at least need to code each case for the legal theory asserted (disparate treatment, disparate impact, hostile work environment harassment, etc.) and the bases of discrimination claimed (race, sex, national origin, age, etc.).

With respect to legal theory, in some types of cases, it would be remarkable if judges did not use the presence of compliance structures in their analyses, because for some, but not all, legal theories, the applicable legal doctrine made their presence or absence relevant. For example, Linda explained, in hostile work environment cases, a lawyer would expect to see judges taking account of institutionalized employment practices like the existence of anti-harassment policies and internal grievance procedures. This is because legal doctrines made their presence or absence relevant. To determine whether patterns of judicial inference can be predicted by the presence or absence of symbolic compliance structures, disparate treatment cases—where the presence or absence of institutionalized practices is not relevant to the doctrinal analysis—would need to be separated from hostile work environment cases. Coding for legal theories would be equally important in disentangling the disadvantages plaintiffs face based on their demographic characteristics (e.g., a plaintiff who is Black and female) and their legal claims (e.g., a plaintiff who alleges discrimination based on her status as a Black woman).

The resulting research group did code for legal theories, and it operationalized intersectionality with respect to both legal theories and plaintiff characteristics. With respect to intersectionality, data analysis yielded the first-ever empirically rigorous quantitative corroboration of Reference CrenshawKimberlé Crenshaw's (1989) assertion that federal judges were less amenable to intersectional claims than to single basis claims (Reference Best, Edelman, Krieger and ElliasonBest et al., 2011). And with respect to the coding of legal theory more generally, the group ultimately found the endogeneity effect in the disparate treatment cases, the very place where lawyers would least expect it, and did not find the effect in hostile work environment cases, where lawyers would (Reference Edelman, Krieger, Elliason, Albiston and MellemaEdelman et al., 2011).

What is important here is not only that Laurie was right (which always delighted her), but also that, without the interdisciplinary focus and the lawyer's expertise, the study might have been undertaken without separating the cases by legal theory. If that had happened, the effect visible in the intentional discrimination cases—where deference was not legally justifiable but present—could have been lost in the absence of an effect from the hostile work environment cases—where deference was legally relevant but absent. Even if an effect had been found, without controlling for legal theory, lawyers would have readily dismissed the finding as an artifact of faulty research design. In similar fashion, without recognizing and coding for claim intersectionality as something distinct from demographic intersectionality, there would have been no way to test, and find quantitative empirical support for, Professor Crenshaw's theory.

The judicial deference research group did more than just generate important research findings. It successfully flight-tested a new, thick method of interdisciplinary collaboration capable of identifying and solving research problems that would otherwise be invisible or insoluble.

BERKELEY, ANN ARBOR, AND MĀNOA, 2017: INTERDISCIPLINARITY THROUGH AND THROUGH

In 2017, Laurie began to recruit another diverse team of researchers for a project on judicial reasoning in disability discrimination cases. Not only is the group interdisciplinary, it is intergenerational. Three of us (Laurie, Linda and Catherine) are full professors who have been in legal academia since the early 1990s. Rachel is a sociology professor at Michigan who was a graduate student when she began working with Laurie in 2005. Diana and Yan are advanced PhD students in Berkeley's Jurisprudence and Social Policy Program. Todd joined the team during the COVID-19 pandemic, and is the last in a long line of graduate students who had the good fortune to be advised by Laurie. Laurie leveraged her position to facilitate intergenerational scholarly collaboration, quite literally bargaining in her own retention negotiation for tuition and fee waivers for graduate research assistants.

Our project reflects Laurie's understanding of the power of combining different types of expertise. Some of us are lawyers with years of experience reading cases, thinking about how the procedural posture of a case affects the reasoning, and why judges craft opinions in certain ways. Some of us know how to access or analyze social science data and theorize health, medicine, and disability. Some of us are experts on disability and employment discrimination law. Some of us knew little about disability discrimination law before joining this team and approached coding cases with the curiosity and keenness of a novice. Working on Zoom across six time zones, we developed and implemented a coding scheme to investigate how judges reason about and decide employment discrimination cases involving different types of disabilities.

Laurie was drawn to disability law in part because it promised to shed new light on legal endogeneity. In addition to prohibiting discrimination, disability law requires “reasonable accommodations” to put people with disabilities on a level playing field with other members of society. Are judges' understandings of which accommodations are “reasonable,” and of employer defenses of undue hardship and direct threat, influenced by organizational practices and institutionalized ideas about business rationality?

Laurie also realized that when studied over time, disability law could provide a chance to test her hypothesis that legal ambiguity prompts legal endogeneity (Reference EdelmanEdelman, 1992, Reference Edelman2016). In the Americans with Disabilities Act (ADA) of 1990, the definitions of many key terms were highly ambiguous. The ADA Amendments Act (ADAAA) of 2008 clarified many of the terms in the original act, including “disability.” Laurie hoped to test whether this decrease in legal ambiguity shaped judges' interpretations of disability and accommodations and their responses to employers' practices.

Extending insights about intersectionality gleaned from our earlier work (Reference Edelman, Krieger, Elliason, Albiston and MellemaBest, Edelman et al., 2011), studying disability cases let Laurie further theorize how cultural ideas about plaintiffs' characteristics shape litigation outcomes. How do judicial interpretations of ‘disability’ relate to social, organizational, or medical constructions of disability? Do judges reason differently, or require different types of documentation, or come to different conclusions, depending on whether disabilities are visible or invisible, constant or intermittent, widely accepted or relatively unknown, medically legitimated or not, and stigmatized or unstigmatized?

To answer these research questions, Laurie knew that we would need to develop a comprehensive coding scheme to capture the nuances of societal, organizational, and judicial views of and assumptions about disabilities and accommodations, and to control for the many factors—including those relating to legal doctrine—that might influence judicial reasoning and case outcomes. And to identify subtle, probabilistic patterns in judicial reasoning and case outcomes, we would need to code a relatively large number of cases (n = 1200).

Coding at the level of nuance that Laurie's research questions demanded is complex, particularly for a law like the ADA, whose analyses involve burden-shifting and similar-sounding legal standards that judges often appeared not to appreciate. To carry out this complex coding, we used an interactive, iterative process that prioritized intercoder reliability. The heart of this process were regular Zoom meetings to discuss disagreements in coding. When we disagreed, we took turns explaining why we decided to code a particular way. Sometimes, we persuaded others that our coding was “correct” for the purposes of our research questions and coding scheme. Other times, we allowed ourselves to be persuaded. Sometimes we found that our coding instructions—and sometimes, even broader research questions—needed to be clarified. Regularly and systematically discussing differences was essential in cultivating a shared understanding of how to code how courts reason.

This process also let the members of our team learn from each other, helping the interdisciplinary team become more than the sum of its parts. For example, Linda and Catherine often shared their extensive knowledge of ADA law, helping others work their way through cases in which the legal reasoning seemed particularly muddled. Diana and Yan created new processes for collecting and managing data, while drawing from their recent years of legal practice to flesh out hypotheses. Todd mastered the coding of judicial politics. Rachel answered questions about sampling, coding, independent and dependent variables, and how to combine the qualitative and the quantitative analyses. Laurie asked many questions about the law and about statistical analysis. Her willingness to admit what she did not know created a safe space for others to ask questions and learn from colleagues.

As we each read hundreds of cases, we developed a greater sense of the patterns and injustices in disability cases. Laurie pushed us to find ways to track the patterns we were noticing so that we would be able to describe them quantitatively. Again and again, we read about people who worked a job for decades, became injured on the job, and then were terminated because they were now too disabled to perform the job. Noticing this terrible pattern, Laurie encouraged us to go back through the cases we had coded, adding a variable noting when the plaintiff's disability resulted from an on-the-job injury. This variable will allow us to track an ongoing injustice and a hole in the social safety net. As we coded, we also noticed that judges, like many others, appeared to have difficulty recognizing invisible disabilities, particularly the debilitating effects of pain. We therefore developed a set of variables tracking how judges write and reason about conditions involving pain. Having experienced pain, Laurie worked to understand it sociologically with the goal of trying to make our workplaces more humane.

In addition to her analytical insights, Laurie's contributions reflected her irrepressible playfulness, a joie de vivre that helped counter-balance both the time-consuming precision of our scientific method and the somber realities of the cases we were reading. In one case, an employee with PTSD sought to bring a trained service dog to work. Catherine coded this case as mentioning “accommodations involving people” but raised it with the team because the codebook only mentioned assistance from human beings. Laurie, a lover of dogs, weighed in: “I agree (especially if the dog is cute). Including service dogs seems to fit the spirit if not the language of the variable.”

LEARNING FROM LAURIE

For several of us, Laurie was our employer in addition to our collaborator. And as a supervisor, she made real the political commitments that motivated much of her scholarship into the possibilities and realities of employment relationships. She gave us what she wanted for every plaintiff in every one of the 1200+ cases we coded. Laurie ensured that we each had the opportunities and support needed to do our best work, notwithstanding our many limitations, and with the accommodations our bodies, relationships, and humanity required. Consistent with Laurie's theoretical and empirical expertise on institutions and social change, she did this through institutional design, not just individual kindness. Laurie secured funding for graduate student researchers' tuition, fees, and salaries from Berkeley, which supported Diana and Yan, as well as Rachel when she was a graduate student. Even more, through sharing the work among an integrated team of researchers, Laurie ensured that we could each alternatingly step up and step back as needed, and that the work would go on. Through these accommodations, the project has already survived cancer, surgery, chronic migraines, hospitalization, COVID, the birth of one daughter and one son, family and friends in crisis, family and friends in joyful celebration, tenure, and the largest strike in the history of higher education. And now, it survives Laurie's death, too.

Laurie died just as we were wrapping up 3 years of coding and beginning work on data analysis and writing papers. Having convened us for this project, she left us to carry it on without her. Among the many heart-breaking things about Laurie's untimely death is that she will miss the excitement of analyzing the data and finding results.

Whatever we find when we turn to data analysis, we will owe Laurie a debt of gratitude for spending the last years of her life conceptualizing this project and bringing us together to pursue it. Here, we have tried to share some of what made Laurie such an extraordinary scholar, collaborator, mentor, and friend. But we will continue memorializing her, perhaps in the manner she would prefer, in the project itself, in which she will be a co-author for years to come.