1
DESIGNING FOR DRIFT
Planning Ethnographic Qualitative Research
on Groups
Michael G. Pratt and Najung Kim
BOSTON COLLEGE
Ethnographic research is guided as much from drift as design, and is perhaps the source of far more failures than successes.
(Van Maanen, 1979, p. 539)
As the epigraph suggests, ethnographic research may be the source of more failures than success. Clearly it would not be unique in this regard; especially if you consider that the rejection rate by top-tier organizational studies/management journals can exceed 90 percent. But as we discuss below, ethnographic research has its own unique challenges that make it a high-risk/high-reward endeavor. We think it is worth the risks. We challenge the reader to find a more engrossing, energizing, and personally world-changing methodology for studying a group â be it a small team, an organization, an occupation, or a âpeople.â Part of its allure may be in its subject matter â better understanding the cultures of groups. Its fun may similarly derive from the engrossing nature of its methods, such as actual participation in group life and the use of broad, largely unstructured interviews. Or perhaps its appeal is in the challenge in navigating the tension John Van Maanen raises: the pull between design and drift.
The design part is not that difficult to describe. In brief, to examine a group in an ethnographic fashion you have to: (a) select a research question; (b) locate a group to examine this question; (c) design your study; (d) obtain approval for your study through an Institutional Review Board (IRB); (e) gain access; (f) collect data; (g) analyze the data; and (h) write it up. Essentially, it is not that different, in abstract, to how you would approach other examinations of groups. However, as we will discuss below,how these steps are enacted in ethnographies may be unique. For example, engaging in âcommitment actsâ (Feldman, Bell, & Berger, 2003), finding key informants, asking grand tour questions, and the like are central to ethnographic studies. Moreover, what is more difficult to describe, but no less equally important, is the âdriftâ â taking advantage of where the study takes you. To be able to best âcatch the drift,â requires tactics, tips, and training often not found in âhow toâ books on ethnography (e.g., Fetterman, 1998; Spradley, 1979). Our charge is to cover those topics that are the least well covered by existing texts and articles. In this spirit, we discuss the design of an ethnographic group study including gaining access, and preparing for observations and interviews. Before turning to these topics, we first provide an overview of ethnography.
Ethnography: What it is, when you do it, and why
While definitions vary, at their core, ethnographies are the study of a group's culture (Fetterman, 1998; Rosen, 1991; Spradley, 1979; Van Maanen, 1979). These groups may vary in size â from teams (Bechky & Okhuysen, 2011), to occupations and organizations (Kayser-Jones, 2002; Pratt, 2000a; Rosen, 1991; Van Maanen, 1973), to people living in certain areas (Mead, 1928; Venkatesh, 2002). Ethnographies also have specific ways of gathering data, such as participant observation and ethnographic interviews. Some studies are relatively âpureâ ethnographies, which typically involve long periods (e.g., over 6 months) of being âin the fieldâ (Fetterman, 1998). To illustrate, the first author's (Mike's) study of Amway distributors was his most pure ethnography: he spent over 9 months in the field working as a distributor and interviewing other current and former distributors. In organizational research, there are relatively few ethnographies that are pure, but several which contain some ethnographic elements (e.g., Hinds & Cramton, this volume). Mike's work, for example, often uses ethnographic interviews or some limited participant observation, such as rounding with doctors (Pratt, Rockmann, & Kaufmann, 2006) or being involved in a nurse's dress code task force (Pratt & Rafaeli, 1997).
Ethnographies should be used when you are interested in getting the perspective of the cultural participants or âinformants.â The term âinformantâ rather than âsubjectâ (as in a lab study) or a ârespondentâ (as in a survey) is not accidental, but speaks to the nature of the researcher to the researched. Informants, as the name describes, inform you. They are the cultural experts. You are not manipulating the conditions around them as you might for a âsubject;â and ârespondentsâ are often limited to answering questions on the topics and concepts that you deem critical. In the lab and when giving surveys, you are the expert; and you have a fair amount of control over what your subjects do and the kinds of questions they answer. Ethnography puts the researcher, who might have or be getting a PhD, into the unenviable position of often having to appear, or actually be, ignorant. Because you do not know what an informant will say or do, and where the study will âdrift,â you may also feel like you have very little control. For some researchers, these conditions may be enough to dissuade them from doing ethnography!
Moreover, because you are getting the informantsâ perspective, and are not testing concepts or constructs, you have ultimately to translate these understandings into âacademeseâ to get published. This is both a daunting and exhilarating task. Part of the challenge is that you have to be careful not to do too much âviolence to experience,â that is, you need to translate your research in such a way as to do justice to how your informantsâ view the world. And if that is not enough of a challenge, your results will be detailed and specific, but gored on the horn of âgeneralizabilityâ (McGrath, 1982). Thus, it is not clear whether your results will be comparable to any other group. All this and you often have to take hundreds of pages of data and fit them into a research article of roughly 40â50 pages in length.
High probability of failure, relatively low control, putting yourself into a situation where you are not the expert, and being placed between two worlds â that of your informants and that of your academic colleagues â where do I sign up? If you are still reading this chapter, you may have some of the critical characteristics of an ethnographer: persistence, self-confidence, and a desire to learn continuously. And in the spirit of mentioning the âhigh rewards,â at the time of this writing, Mike's ethnography has been the most fruitful â at least in terms of leading to publications (Barnett & Pratt, 2000; Pratt, 2000a, 2000b, 2003; Pratt & Rosa, 2003) â when compared to any other methodology he has used.
Getting back to the issue of when to do ethnography, we have stated that ethnographies are great for exploring the mindsets of individuals within a group â insights that can enrich, support, or even challenge our existing ways of knowing. Like other inductive qualitative approaches, it is good for understanding âwhyâ individuals do things (e.g., their motivations) and âhowâ they do them (e.g., the process). It is not good for describing the prevalence of some attitude or behavior (e.g., how many Fortune 500 employees express high degrees of affective, normative, and calculative commitment towards their organizations). It is good for building theory, but not as good for testing it. It can also elaborate or change how we see existing theories. For example, âorganizational commitmentâ in Amway looked very different from many of the existing models in the literature (Pratt & Rosa, 2003). Finally, ethnographies ground our knowledge and our claims. Most articles suggest that âorganizations are changingâ â but how can we really know this is true without taking the time and energy to figure out how the people we study are viewing the world around them?
First things first? The question and the group
In the beginning there is the research question; sometimes. While it is indeed the case that it is helpful to have a general idea of the kinds of questions you are interested in, sometimes a research question does not crystallize until one finds the appropriate group to study. We'll give you an example. When Mike was completing graduate school, he was struggling to put together a viable dissertation project. He knew he had an abiding interest in organizations that had âstrong culturesâ and conflicting belief systems, but he did not know where to go from there or how to narrow his focus. Fortunately, during this time of struggle, Mike saw his sister who had recently joined an organization he had never heard of: Amway. In their discussion, he became fascinated by the cultural/ideological elements of the organization, and his curiosity was piqued by what appeared to be identity changes in his sister; see Pratt (2003) for details. After reading about the organization, and attending a rally, he knew this was the organization he wanted to study. Once that decision was made, he was able to hone in on a research question around the issues of ideological contradictions, ambivalence and member attachments.
Ethnographic questions are not only of a certain form, they also tend to be of a certain type. As noted above, for qualitative research, the best research questions revolve around questions of âhow?â or âwhy?â rather than âhow many?â or âto what degree?â In addition, these research questions often deal with an issue or problem rather than a specific theory. That is, ethnographic research, like much qualitative research, is often problem centered rather than theory centered. Problem-centered research, as the name implies, is motivated by some conundrum in actual life. In Amway, Mike wondered how people could so strongly attach to a group that appeared to contradict itself in its most fundamental beliefs (e.g., put family before work, but miss kidsâ birthdays to âbuild the businessâ). Because ethnographies attempt to look at the world through the informantsâ eyes â and informants often do not think in theories â it is perhaps natural for ethnographies and problem-centered research to be related. If you start a study with a specific theory in mind, chances are you are simply going to get responses to and from your own perspective. Problem-centered research, however, carries with it at least two unique challenges.
First, problems lie at the intersection of multiple theories. Thus, going back to Mike's dissertation study, conflicting belief systems may be discussed in treatments of ideological conflicts, hybrid organizational identities, and differentiation perspectives on culture, just to name a few organizational examples. It may also require moving beyond the organizational literature. This is challenging when one needs to write up a proposal for an IRB, or possibly for a dissertation committee, because it means that you will be reading widely and integrating multiple theories when writing up your research.
Second, since problem-centered research lies at the intersection of multiple theories, you have to do an extraordinary amount of work, before you even go into the field, to show that the research question you are addressing has not been sufficiently examined before. This often involves writing an introduction to your research (or proposal) by crafting âtheoretical framesâ (Pratt, 2008). Creating theoretical frames involves constructing the argument that multiple theories imply, but do not directly address, the problem or issue with which you are concerned. For example, in Mike's dissertation proposal, he argued that conflicting belief systems should lead to ambivalence among members. However, reactions to ambivalence were that individuals might be more positive or more negative towards their organization, and that they might react by vacillating between beliefs or possibly by being paralyzed. Hence, in his proposal, he showed that extant research had proposed that almost any reaction can follow from ambivalence; but as importantly, he also was able to argue that no one had yet uncovered the conditions under which one response might be more likely than another.
Whether it precedes, follows, or is somehow iteratively involved with the construction of your research question, you must choose a group to study. For an effective ethnographic study, the group you choose is one that you should not be too familiar with. Thus, if you were a consultant before entering into academia, you should not attempt an ethnographic study of consultants. As noted above, you need to get your informantsâ perspective on their culture and âhow they do things.â If you have a fair amount of experience with a group, then (a) you are likely to come into the group with a lot of preconceptions and (b) it is more difficult to get away with asking the types of âdumbâ questions that a newcomer would ask â and those needed to get at the group membersâ world view. That said, it is unlikely that you will be entering a culture that is totally foreign to you. Thus, Rosen (1991, p. 14) writes:
While admitting that working in organizations may not be as exotic as going to a foreign land to study a little-known people, the level of familiarity one has with an organization can vary greatly. For Mike, while Amway was not an entirely foreign culture (e.g., they spoke English and wore recognizable clothing), the culture was nonetheless quite different from what he was used to, and thus he could more easily apply ethnographic techniques.
Other issues that are critical to choosing a group, and which may be more obvious, are finding: (a) groups that will allow you to get the data you need; and (b) groups that will allow you access. With regard to the former, if Mike was interested in examining how office layouts influence group interactions, Amway distributors would be the wrong group to study as they are largely a geographically dispersed group. With regard to the latter, it is helpful before launching into a full project to have some idea of whether or not you will be able to get access to a group, and what kind of access you will get. For both of these criteria, it may help to informally âhang outâ with the group before beginning your official studies. Mike attended an Amway convention to get a sense of what kinds of individuals he was likely to meet, what challenges studying the group might entail, and to help him hone the types of research questions that could be answered by this particular group. He even started building connections with existing distributors.
In general, if you are using ethnography to build theory, you should review the tenets of theoretical sampling, which is different from statistical sampling common to quantitative methods. In brief, theoretical sampling involves choosing a sample for theoretical reasons, such as finding a group that is an extreme example of the phenomena you want to study (and thus where the dynamics you want to study will be clearer), or possibly a prototypical case where a specific group well represents a broader type of group. Mike's ethnography of Amway represented an âextreme caseâ (Pettigrew, 1990). His study of physicians, which was not a pure ethnography, was done because physicians are believed to be a prototypical profession; see Marshall and Rossman (1989), Miles and Huberman (1994), and Patton (1990) for a discussion of different sampling logics in qualitative research. One word to the wis...