Data Analysis in Qualitative Research

Theorizing with Abductive Analysis

Stefan Timmermans,Iddo Tavory

• Editorial: University of Chicago Press
• Año: 2022
• ISBN:
  9780226817729
• Categoría: Sozialwissenschaften
• Categoría: Soziologie

From two experts in the field comes an accessible, how-to guide that will help researchers think more productively about the relation between theory and data at every stage of their work. In Data Analysis in Qualitative Research, Iddo Tavory and Stefan Timmermans provide a how-to guide filled with tricks of the trade for researchers who hope to take excellent qualitative data and transform it into powerful scholarship. In their previous book, Abductive Analysis: Theorizing Qualitative Research, Timmermans and Tavory offered a toolkit for innovative theorizing in the social sciences. In this companion, they go one step further to show how to uncover the surprising revelations that lie waiting in qualitative data—in sociology and beyond.In this book, they lay out a series of tools designed to help both novice and expert scholars see and understand their data in surprising ways. Timmermans and Tavory show researchers how to "stack the deck" of qualitative research in favor of locating surprising findings that may lead to theoretical breakthroughs, whether by engaging with theory, discussing research strategies, or walking the reader through the process of coding data. From beginning to end of a research project, Data Analysis in Qualitative Research helps social scientists pinpoint the most promising paths to take in their approach.

1: Surprise

Surprises1 are the fuel that powers research engines. Drawing upon data of millions of research papers, patents, and scientific careers, sociologists and computer scientists Feng Shi and James Evans concluded that “science and technology advance through surprise.”2 Decades earlier, sociologist of science Robert Merton already noted “the serendipity pattern” in empirical research whenever an “unanticipated, anomalous, and strategic” observation becomes the basis for developing a new theory. Merton’s formulation was, as usual, insightful. While we may think about surprises as anomalous, we typically don’t think of them as “strategic.” But the observer needs to be able to recognize the theoretical implications of a serendipitous finding strategically. Merton added that this “requires a theoretically sensitized observer to detect the universal in the particular.”3 How can we systematically cultivate surprising findings? How do you sensitize yourself to detect emerging surprises? This book helps you do just that.

Our first book, Abductive Analysis: Theorizing Qualitative Research,4 staked out a claim in an increasingly unrealistic field of qualitative data analysis. Our major concern was that the conventional approaches to data analysis impoverish the potential of qualitative research. At best, most qualitative researchers pay lip service to an epistemic model that they don’t follow in practice. The methods sections of research articles—both inductive and deductive—are most often formulaic fantasies divorced from the real-life drama of analyzing qualitative research data.

Joining other scholars who had mined the work of the early pragmatist Charles S. Peirce,5 we argued that abduction—the mode of inference drawing from surprise—is an underappreciated aspect of the research act, especially in the context of theory discovery. Focusing on abductive reasoning is a productive and logically sound way to rethink qualitative research and data analysis. Abductive analysis requires in-depth familiarity with a broad array of social theories and an intensive engagement with observations in order to develop theoretical contributions. Most methods books in qualitative research are either about coming up with something from scratch, or about testing, verifying, and generalizing theory. Instead, abduction in abductive analysis refers to the process of crafting theory—figuring out both what surprising observations are a theoretical case of, and where the theoretical case and the findings diverge in interesting ways.

While not quite philosophically precise, a cooking metaphor is useful to explain the difference. If you open Yotam Ottolenghi’s bestselling cookbook Plenty, you will find his famous recipe for poached baby vegetables with caper mayonnaise. The recipe begins: “To make the mayonnaise, place the garlic, egg yolk, vinegar, mustard, salt, and lemon juice in the bowls of a food processer. Start blending and then very slowly dribble in the oil until you get a thick mayonnaise. Fold in the capers and lemon zest and set aside.”6 If you follow the instructions, using the given ingredients, you should be getting the meal Ottolenghi describes. That’s a deductive logic. You have a good idea of what you’re going to eat; it’s pretty delicious, but is often estranged from the local foodstuffs. Even though Ottolenghi is an Israeli-English chef with several delis and restaurants in London, his recipes remain the same whether you live near the Park Slope food co-op in Brooklyn or have access to the farmer’s markets from Los Angeles. By contrast, a purely inductive logic would probably have us grab something from the fridge; fry, roast, or boil it; and hope for the best. Perhaps, in the grounded theory tradition, we would compare the flavors of different food combinations for patterns.7 If the resulting dish grows on us, we may decide to include it in the family rotation of meals, following the recipe that we “discovered.” While deduction starts with a theory and makes empirical predictions, induction generalizes from observations.

An abductive logic is different. You are craving those poached vegetables with caper mayonnaise. You have memorized Ottolenghi’s book pretty much by heart, with grease spots on the pages of your favorite recipes. His book is tucked in among other cookbooks on your kitchen shelf. You open the fridge and scan the food in front of you, push the cauliflower aside to see what’s behind it, notice that you are out of garlic and that the baby carrots are way beyond their prime. But look, there is a basket of kumquats you purchased impulsively during last Saturday’s shopping trip. What a pleasant surprise and interesting cooking challenge! You don’t have the ingredients for Ottolenghi’s dish but maybe you can make something novel that incorporates the sweetness and tartness of the kumquats. What would be the best way of using these grape-size fruits? You’ll skip the caper mayonnaise and maybe try a citrus syrup instead. The flavor palate of the dish is going to be completely different from Ottolenghi’s poached baby vegetables. You invent your own dish. Abduction, then, is a creative moment where you take what you know and go beyond it to try out something different. It occurs when you have a surprising observation and formulate the most promising hypothesis. Of course, the original recipe inspired the new dish, as did the different cookbooks on your shelf, but the moment of realization of the local possibilities is exhilarating.

Abduction in scientific research is a logic that organizes the process of coming up with a new hypothesis based on surprising research findings. Abduction intervenes in the tension between knowing what you are interested in by chasing usual suspects and remaining open to new unexpected findings for which a usual-suspects explanation is inadequate. It requires in-depth familiarity with a broad range of theories that help you navigate familiar landscapes and point the way when you run into alien terrain. The logic works best when you gather data in an innovative way—a new site, a new set of questions, an overlooked group of stakeholders. You also need to have enough data to examine variation across analytical dimensions. In short, theoretical surprises do not grow spontaneously but are seeded by how you read your theoretical peers and how you conduct your research.

The purpose of Abductive Analysis was to carve out a new epistemological position. We developed the pragmatist semiotics of meaning-making, explained why the defamiliarization generated by coding procedures is helpful, and emphasized the critical importance of writing with an eye on a community of inquiry. We included some examples, but it was more a rethink than a how-to book. Observers pointed out that some researchers were already doing something that resembled abductive analysis. In that sense, the book met its promise: it found an audience across disciplines and most references to abductive analysis are from researchers relieved that they can report what they did as a legitimate methodology. However, we were and are more ambitious than repackaging current practice. Our purpose was to stimulate theorizing that plays to the strengths of qualitative research. But, we admit, you would not know how exactly to do this from the first book.

The challenge is that we don’t—and indeed can’t—have a simple foolproof protocol for surprise. Think back to the cooking example. While a deductive approach gives you a recipe to follow, and an inductive approach might tell you what a meal could look like if you combine the stuff in your fridge, the flash of abductive insight that we depicted requires careful attention to what’s in the fridge, a deep knowledge of different cookbooks, and insight into how to coordinate cooking time for various ingredients in order to come up with a new dish. How can something like this be ever taught in a book format? This is the kind of knowledge that comes from years of apprenticing in different kitchens. You do not become a chef by memorizing recipes. Yet while there are no abductive culinary shortcuts, we can still tell you something about how to go about organizing your shopping trips, what to look for in your fridge and pantry, and how to read cookbooks.

The current book aims to translate our ideas about abduction into concrete research practices. We contend that abductive analysis means doing qualitative research differently, not just relabeling what you were already doing. Abductive analysis means designing research explicitly to maximize the chance of data surprises; engaging the literature in a different way; approaching data coding with a mindset aimed at defamiliarizing the familiar; and writing research up deliberately and reflexively to evoke abductive insights. We show in the next chapters how working abductively permeates a research project from the moment you muse about doing research, through collecting the first observations, through the final write-up. It demands an analytical sensibility that immerses you in the literature, guides the collection of empirical materials, and requires working with these observations as they come in. Surprising findings do not fall out of thin air: you will need to cultivate them. This book shows you how to discover surprises by thinking differently about your research.

What this book is, and what it isn’t

Our kind of qualitative research intertwines empirical and theoretical narratives. We agree with Loïc Wacquant when he wrote that “far from being antithetical, vivid ethnography and powerful theory are complementary and that the best strategy to strengthen the former is to bolster the latter.”8 We presume that you share our goal of theorizing based on qualitative research; why else would you be reading this book? But maybe we shouldn’t assume a shared purpose; maybe we should make the case first that theorizing is a worthwhile endeavor. There are, after all, other compelling ways of working with empirical materials where theory isn’t really at the center of the work.

In many of the hotspots of qualitative inquiry, researchers don’t primarily theorize but engage in applied research. Plenty of qualitative health researchers, for instance, work towards better outcomes within a public health, medical, or health services framework. They examine why patients are noncompliant with medication regimes, or how direct-to-consumer advertising sells pharmaceuticals.9 Other qualitative researchers apply their skills for technology companies to make products that will find a market. For more than twenty years, Xerox had a research division that included ethnographers to study human-computer interaction in order to design copiers that would be intuitive to use. While some of this work, especially by the principal investigator Lucy Suchman,10 engaged in theorizing, the purpose of the unit was to help engineers design user-friendly machines. Globally, many, probably most, qualitative researchers conduct contract research, write bids for government projects, apply to foundations, and submit proposals to granting agencies. Since such agencies are rarely interested in advancing a theoretical agenda, this renders a qualitative researcher much more of a consultant or expert working towards an institutional portfolio. Grant agencies often pay for data gathering but provide few funds for analysis, and the resulting publications provide mainly descriptive insights.

Some superb research also doesn’t aim for novel theorizations but convincingly proves what many people already suspected. The research of Devah Pager on the racial profiling of job seekers using experimental methods is such an example. She found that employers were more likely to hire a White man with a felony conviction than a Black man without a criminal record.11 This research is not theoretically innovative, but it is critically important to uncover institutional racial biases in the labor market, and Pager’s work meaningfully intervenes in an important political struggle. At Pager’s memorial service, the writer Ta-Nehisi Coates talked about how her research validated the experiences of people of color—giving them a decisive argument against people who said discrimination was “all in their heads.”12

Still other researchers are attracted by the humanistic side of qualitative inquiry rather than by the goal of abstracting and theorizing. Whatever else we do, qualitative researchers closely chronicle the human condition. In that capacity, we are not alone but part of a crowded field of journalists and documentary makers, as well as fiction writers. Chronicling is done by people from the outside looking in, but also by people from the inside looking out. They tell their own and each other’s stories on social media, in interviews, blogs, documentaries, news articles, op-ed pieces, reflections, diaries, internet groups, videos, and memoirs.

While humanistic, exemplifying, and applied qualitative projects constitute alternative goals for conducting qualitative research, they also benefit from theoretical engagement, even if the goal is not developing novel theories. Maybe they should be more theoretically adventurous. In fact, some of the strongest exemplars of these research traditions engage implicit theoretical agendas.

It is no accident that Suchman’s work for Xerox was so influential; it was due, in large part, to her deep engagement in sociological interactionism and ethnomethodology. Applied studies that sidestep theories risk shortening their reach. Similarly, some of the best studies that seem to only “exemplify” theory actually do more than that. Thus, for example, Annette Lareau’s excellent Unequal Childhoods13 can be read as an exemplification of Pierre Bourdieu’s theory of habitus formation and cultural capital. Lareau’s book is notable for its innovative methodology: she inserted herself into the homes of working- and middle-class parents to observe class-based socialization in action. While methodology is surely part of what makes this book such compelling reading, Unequal Childhoods specifies, empirically and theoretically, the mechanisms through which classed socialization occurs in the home and at schools.

As for humanistic research, a common warrant for qualitative research is that the work itself reveals and bridges a discrepancy between what an audience thinks about a community and how the members of a community behave.14 The researcher shows through the richness of observations that those ostracized as deviant are human after all, and those in power or exalted in the media are actually more flawed than we imagined. This kind of interview study or ethnography folds the extremes of behavior into a universal humanism, while still mapping cultural idiosyncrasies. It is difficult to pull this off without engaging theories and concepts, even if they are banished to the endnotes. Some of the most popular works of qualitative research—Mitch Duneier’s Sidewalk, Matt Desmond’s Evicted, or Tressie McMillan Cottom’s Thick15—draw extensively from social theory but you need to know this literature to recognize who they are in conversation with in the text. Again, not just inserting but also developing a theoretical dimension in such a humanistic project may expand its reach beyond those already interested in a substantive research area, as well as help distinguish it from the journalistic human interest stories.

While there are many warrants for qualitative research and all of them benefit from a strong theoretical backbone, the path we advocate fully embraces the theoretical craft of academics as scholars. The theoretical contribution of qualitative research is the endeavor’s protection against irrelevance. Theory is how research travels beyond a small substantive niche. The warrant for our kind of research is that we find something counterintuitive, unexpected, or surprising and develop theory from it. We aim to foster a particular craftsmanship of working with qualitative data in light of existing bodies of scholarship in order to make a theoretical and conceptual contribution. These concepts and theorizations not only abstract relevant experiences but, if successful, can open new perspectives and produce new research foci. If theory shows how a pattern found in one place illuminates what happens in another, then there is something deeply humanistic about it.

To take one example we both admire and that we return to later on in this book, Leigh Star and James Griesemer coined the notion of “boundary object” when they studied the acquisition of bird specimens for a natural history museum. They conceptualized the boundary object as a negotiation strategy for situations where parties with different interests do not need or want to confront each other. A boundary object is unified enough that it has a common meaning but also sufficiently plastic that it may have vastly different meanings in different communities.16 The concept of boundary object exemplifies its own power as a boundary object. The notion traveled and was used to mediate very different situations. This concept caught on in dozens of studies. Researchers incorporated boundary objects to examine how the same issue is able to mean different things for different groups ...

Índice

• Cover
• Title Page
• Copyright Page
• Contents
• 1:  Surprise
• 2:  Beginnings
• 3:  Theories
• 4:  Collecting
• 5:  Open Coding
• 6:  Focused Coding
• 7:  Clues about Clues
• 8:  Writing It Down, Writing It Up
• 9:  Abduction in Action
• Appendix: From Qualitative to Computational Abductive Analysis
• Acknowledgments
• Notes
• Bibliography
• Index