Quick viewing(Text Mode)

What Grounded Theory Is Not Author(S): Roy Suddaby Source: the Academy of Management Journal, Vol

What Grounded Theory Is Not Author(S): Roy Suddaby Source: the Academy of Management Journal, Vol

From the Editors: What Grounded Theory Is Not Author(s): Roy Suddaby Source: The Academy of Journal, Vol. 49, No. 4 (Aug., 2006), pp. 633-642 Published by: Academy of Management Stable URL: http://www.jstor.org/stable/20159789 Accessed: 20-09-2016 15:32 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy of Management Journal

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms ? Academy of Management Journal 2006, Vol. 49, No. 4, 633-642.

FROM THE EDITORS: WHAT GROUNDED THEORY IS NOT ROY SUDDABY University of Alberta

Editor's Note. Three years ago, I invited Robert In the manuscripts I review for AMJ I have seen the (Bob) Gephart to write a "From the Editors" column term "grounded theory" used to describe analysis designed to help authors improve their chances of via correlations, word counts, and pure introspec success when submitting qualitative to tion. I am not suggesting that these techniques can AMJ. Judging from the increasing number of quali not be used in a grounded theory study. But I note, tative studies that have been accepted and pub with some concern, that "grounded theory" is often lished in AMJ since that time, I would like to think used as rhetorical sleight of hand by authors who that his article, " and the are unfamiliar with qualitative research and who Academy of Management Journal," has had a pos wish to avoid close description or illumination of itive impact. their methods. More disturbing, perhaps, is that it Continuing in this tradition, I asked Roy Sud becomes apparent, when one pushes them to de daby?an excellent reviewer (and author) of quali scribe their methods, that many authors hold some tative research?to tackle another "big issue" that serious misconceptions about grounded theory. the editorial team has noticed with respect to qual What are these misconceptions? Before reviewing itative submissions to AMJ: overly generic use of them, I offer you a short description of what grounded the term "grounded theory" and confusion regard theory is. Like most difficult subjects, grounded the ing alternative epistemological approaches to qual ory is best understood historically. The itative research. Like Bob before him, Roy has, I was developed by Glaser and Strauss (1967) as a believe, produced an analysis that will greatly ben reaction against the extreme that had per efit those who are relatively new to qualitative re meated most . They disputed the view search or who have not yet had much success in that the social and natural sciences dealt with the getting their qualitative research published. Hope same type of subject matter. Specifically, Glaser and fully, Roy's analysis will help even more authors to Strauss challenged prevalent assumptions of "grand succeed, thus allowing AMJ and other journals to theory," the notion that the purpose of social research continue to increase the quality of insights pro is to uncover preexisting and universal explanations vided by rich qualitative studies of individual, or of social behavior. In making their challenge, Glaser ganizational, and institutional phenomena. and Strauss looked to the of Charles Sara L. Rynes Saunders Peirce (1839-1914) and early symbolic in teractionists, particularly George Herbert Mead I was not particularly surprised to observe that (1863-1931) and Charles Cooley (1864-1929), each most of the articles identified as "interesting re of whom rejected the notion that scientific truth re search" in a recent AMJ were the product of flects an independent external reality. Instead, they qualitative methods (Bartunek, Rynes, & Ireland, argued that scientific truth results from both the act of 2006). New discoveries are always the result of observation and the emerging consensus within a high-risk expeditions into unknown territory. Dar community of observers as they make sense of what win, Columbus, and Freud, each in different ways, they have observed. In this pragmatic approach to were conducting qualitative inquiries. research, empirical "reality" is seen as I am continually surprised, however, by the pro the ongoing interpretation of meaning produced by found misunderstanding of what constitutes quali individuals engaged in a common project of observation. tative research. Such confusion is most apparent when authors claim to be using "grounded theory." Glaser and Strauss (1967) proposed grounded theory as a practical method for conducting re search that focuses on the interpretive process by Thanks to Sara Rynes for suggesting and commenting analyzing the "the actual production of meanings on this paper. Thanks also to Chet Miller, Jean Bartunek, and concepts used by social actors in real settings" Kristine Fitch, Chris Quinn-Trank, and Marvin Washing (Gephart, 2004: 457). They argued that new theory ton for very helpful comments on early drafts. could be developed by paying careful attention to 633

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 634 Academy of Management Journal August

the contrast between "the daily realities (what is SIX COMMON MISCONCEPTIONS actually going on) of substantive areas" (Glaser & Grounded Theory Is Not an Excuse to Ignore the Strauss, 1967: 239) and the interpretations of those Literature daily realities made by those who participate in them (the "actors"). They also rejected positivist A common misassumption is that grounded the notions of falsification and hypothesis testing and, ory requires a researcher to enter the field without instead, described an organic process of theory any knowledge of prior research. There are several emergence based on how well data fit conceptual variants of this myth, each based on the false categories identified by an observer, by how well premise that the researcher is a blank sheet devoid the categories explain or predict ongoing interpre of experience or knowledge. An extreme variant is tations, and by how relevant the categories are to the notion that not only must the researcher enter the core issues being observed. Most significantly, the field with a blank mind (i.e., without knowl Glaser and Strauss offered a compromise between edge of the literature and absent prior experience), extreme and complete relativism by ar but that she or he must also enter the field with a ticulating a middle ground in which systematic blank agenda (i.e., without a defined research ques data collection could be used to develop theories tion). A less extreme, but more problematic, ver that address the interpretive realities of actors in sion suggests that the researcher must defer reading social settings. existing theory until the data are collected and The method described by Glaser and Strauss analyzed. This notion is reflected in manuscripts (1967) is built upon two key concepts: "constant whose authors avoid any mention of prior literature comparison," in which data are collected and ana until their papers' discussions or concluding sec lyzed simultaneously, and "theoretical sampling," tions. In a similar variant, researchers use grounded in which decisions about which data should be theory to tackle a subject that is in "well-tilled collected next are determined by the theory that is soil"?that is, a subject that has attracted a long and being constructed. Both concepts violate long credible history of empirical research?and use standing positivist assumptions about how the re grounded theory as a justification for ignoring prior search process should work. Constant comparison research in formulating their study. contradicts the myth of a clean separation between Leaving aside the question of whether it is even data collection and analysis. Theoretical sampling possible to disregard one's prior knowledge and violates the ideal of hypothesis testing in that the experience, the idea that reasonable research can be direction of new data collection is determined, not conducted without a clear and by a priori hypotheses, but by ongoing interpreta absent theory simply defies logic. Such research, as tion of data and emerging conceptual categories. Ronald Coase famously observed, is likely to pro Grounded theory, therefore, is a method that is duce a random "mass of descriptive material wait more appropriate for some questions than others. ing for a theory, or a fire" (Coase, 1988: 230). To Clearly, it is most suited to efforts to understand the tally unstructured research produces totally process by which actors construct meaning out of unstructured manuscripts that are unlikely to make intersubjective experience. Grounded theory it past the desk editor at any credible journal of should also be used in a way that is logically con social science. sistent with key assumptions about social reality The notion of using grounded theory as an ex and how that reality is "known." It is less appro cuse to forgo examining extant literature is perhaps priate, for example, to use grounded theory when more problematic because it is often based on a you seek to make knowledge claims about an ob researcher's desire to discover something new. This jective reality, and more appropriate to do so when desire most often manifests when researchers use you want to make knowledge claims about how grounded theory to tackle well-established areas of individuals interpret reality. empirical inquiry?leadership, for example. In Such fine distinctions between abstract catego most cases, the researchers honestly hope to gain ries are perhaps more easily seen through concrete fresh insights by keeping out of the ruts early trav examples. From my reviewing experience for AMJ elers have worn. They thus avoid formal reviews of and other management journals, I have identified relevant literature in their manuscripts to create the six common misconceptions about grounded the impression that their observations and analytic cat ory. Although I make no claim that these categories egorizations of data were not colored by previous are exhaustive, they accurately reflect my experi ideas. ence and offer, I think, a reasonable assessment of Unfortunately, this approach is also based on a common errors researchers make in conducting serious misreading of the seminal texts in grounded and presenting grounded theory research. methodology. Although Glaser and Strauss were

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 2006 Suddaby 635 motivated against grand theory, their formulation Grounded Theory Is Not Presentation of grounded theory was never intended to encour of Raw Data age research that ignored existing empirical knowl I occasionally see papers that start with an ap edge. They distinguished between substantive the propriate and interesting question, are written well, ory, or theory grounded in extant research in a and follow a well-constructed method, but produce particular subject area (e.g., leadership), and findings that are obvious or trite. You've seen such grounded theory, but they observed a direct and papers; they conclude that entrepreneurs are risk necessary link between the two forms of theory: takers, change is difficult, and leaders are charis Substantive theory is a strategic link in the formu matic. The common thread in such research is that lation and generation of grounded formal theory. We it tends to present incomplete or relatively undi believe that although formal theory can be generated gested data. directly from data, it is more desirable, and usually This unfortunate outcome is usually the result of necessary, to start the formal theory from a substan tive one. The latter not only provides a stimulus to a one of three errors in the practice of grounded "good idea" but it also gives an initial direction in theory research. First, there may be some confusion developing relevant categories and properties and in between grounded theory and phenomenology. choosing possible modes of integration. Indeed it is Phenomenological research emphasizes the subjec difficult to find a grounded formal theory that was tive experiences of actors' "lifeworlds" (Husserl, not in some way stimulated by substantive theory. 1969; Schutz, 1972). Methodologically, phenom (Glaser & Strauss, 1967: 79) enologists attempt to capture the rich, if not mun The real danger of prior knowledge in grounded dane, detail of actors' lived experiences. They often theory is not that it will contaminate a researcher's present data in relatively raw form to demonstrate perspective, but rather that it will force the re their authenticity and to permit a holistic interpre searcher into testing hypotheses, either overtly or tation of the subjects' understanding of experience. unconsciously, rather than directly observing. Such data are typically analyzed through some Grounded theory methodologists describe a num what introspective techniques that permit a clear ber of ways to prevent this from happening. One is focus on the relationship between the language to avoid research that adheres too closely to a single used and the objects to which language relates substantive area and, instead, draw from the sev (Moustakas, 1994). eral substantive areas that are frequently reflected Although grounded theory retains some sympa in a given daily reality. Stephen Barley's (1986, thy for phenomenological assumptions and tech 1990) research on technology and structuring, for niques, researchers using grounded theory are less example, succeeds because it draws from (at least) focused on subjective experiences of individual ac two areas of substantive research?technological tors per se and are instead more attentive to how change and structuration theory?that are both of such subjective experiences can be abstracted into ten germane in the same research contexts. Another theoretical statements about causal relations be tween actors. The difference between these two technique is to be continuously aware of the possi bility that you are being influenced by preexisting approaches can be seen in how each uses the tech conceptualizations of your subject area. Particu nique of interviewing. In a phenomenological larly when working through well-tilled soil, study, in-depth are a key means of prob grounded theorists must retain the capacity to ing individuals' subjective experiences (Wimpenny "make the familiar strange" (Spindler & Spindler, & Gass, 2000). The detail and nuance of the stories 1982). A final solution is to try not to overextend interviewees elaborate and the specific words they the objective of grounded theory research. That is, choose comprise the primary unit of analysis. Be researchers may shoot for "the elaboration of exist cause phenomenology is an effort to probe the lived ing theory" rather than untethered "new" theory. experience of subjects without contaminating the None of these approaches justifies ignorance of data (Moustakas, 1994), units of data are often pre existing literature or knowledge. The reality of sented in their raw form. In grounded theory, by grounded theory research is always one of trying contrast, interviews with subjects may start with a to achieve a practical middle ground between a phenomenological interest in subjective under theory-laden view of the world and an unfettered standings, but the primary interest is not in the empiricism. A simple way to seize this middle stories themselves. Rather, they are a means of elic ground is to pay attention to extant theory but iting information on the social situation under ex constantly remind yourself that you are only hu amination. In contrast to phenomenological stud man and that what you observe is a function of ies, grounded theory studies rarely have interviews both who you are and what you hope to see. as their sole form of data collection.

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 636 Academy of Management Journal August

The second grounded theory application error practice should be encouraged?there must be that I see as producing obvious and trite output is some degree of congruence between the research failure to "lift" data to a conceptual level. Again, question (i.e., a researcher's assumptions about the this failure may arise from the epistemological nature of reality and how one might know reality) problem of confusing grounded theory with phe and the methods used to address the question. A nomenology. It is more likely, however, that the realist ontology rests on the assumption that the researcher has simply failed to completely analyze variables of interest exist outside individuals and the data. A key element of grounded theory is iden are, therefore, concrete, objective, and measurable tifying "a slightly higher level of abstraction? (Burrell & Morgan, 1979). An "interpretivist" ontol higher than the data itself" (Martin & Turner, 1983: ogy rests on the contrasting assumption that human 147. The movement from relatively superficial ob beings do not passively react to an external reality servations to more abstract theoretical categories is but, rather, impose their internal perceptions and achieved by the constant interplay between data ideals on the external world and, in so doing, ac collection and analysis that constitutes the con tively create their realities (Morgan & Smircich, stant comparative method. The failure of data to 1980). From this perspective, the key variables of coalesce into definable conceptual structures that interest are internal and subjective. move beyond the obvious may well be the result of Keep in mind that the purpose of grounded the a researcher's failure to thoroughly work between ory is not to make truth statements about reality, data and extant knowledge in an effort to find the but, rather, to elicit fresh understandings about pat best fit or the most plausible explanation for the terned relationships between social actors and how relationships being studied (Locke, 2001). these relationships and interactions actively con Finally, the researcher who has produced sim struct reality (Glaser & Strauss, 1967). Grounded plistic output may simply have stopped collecting theory thus should not be used to test hypotheses data too early. Unlike more traditional, positivist about reality, but, rather, to make statements about research, grounded theory offers no clean break how actors interpret reality. As Martin and Turner between collecting and analyzing data. Rather, a (1986) observed, grounded theory is best used researcher must continue to collect data until no when no explicit hypotheses exist to be tested, or new evidence appears. This process, called "cate when such hypotheses do exist but are too abstract gory saturation," is one of the primary means of to be tested in a logical, deductive manner. This is verification in grounded theory (Strauss & Corbin, where grounded theory is most appropriate? 1998). Premature departure from the field may well where researchers have an interesting phenomenon result in data that are only partly analyzed and without explanation and from which they seek to therefore fail to elevate obvious categorizations to a "discover theory from data" (Glaser & Strauss, more abstract theoretical level. 1967: 1). There are some qualifications to my statement that grounded theory is not theory testing, content Grounded Theory Is Not Theory Testing, , or word counts. Although grounded the Analysis, or Word Counts orists do not use Popperian notions of falsification I regularly see papers in which the authors at as a technique for making statements about reality, tempt to use grounded theory methods to test hy they do "test" their tentative ideas and conceptual potheses. These fall into a larger category of manu structures against ongoing observations. A key scripts whose authors have engaged in the sloppy component of the constant comparative method is practice of methodological slurring (Goulding, such critical evaluation of emerging constructs 2002), using interpretive methods to analyze "real against ongoing observations. Similarly, although ist" assumptions. In most cases the manuscripts grounded theory is not to be confused with content begin with clear sets of positivist assumptions, in analysis or word counting, both techniques can cluding hypotheses, and then proceed to report form part of grounded theory studies. The impor "tests" of the hypotheses with sets of interviews or tant difference is that grounded theory describes an counts of words in relevant publications. In other overall method for systematically gathering and an cases manuscripts will start with interpretive pre alyzing data, but content analysis describes a spe mises, such as the social construction of reputation cific context within which a distinct type of data in the popular business press, and then report word can be gathered and analyzed. Typically, in a counts, with the claim of having performed grounded theory study, content analysis is only one grounded theory. of multiple contexts for acquiring data. Word Although there is nothing wrong with combining counting, which is a subset of content analysis, is qualitative and quantitative methods?in fact, the somewhat more problematic in that it bears posi

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 2006 Suddaby 637 tivist assumptions about the relationship between In framing our introduction and the following con word frequency and meaning, most of which vio ceptual overviews, we have employed the theoreti cal concepts that actually emerged from the study. late the interpretivist assumptions of grounded the In the grounded, interpretive research approach ory (Krippendorff, 2003). used here, the theoretical concepts and framework Another important qualification to my comments are grounded in and emerge from the data and anal about methodological slurring arises from the way ysis that follow (Glaser & Strauss, 1967; Strauss and in which grounded theory tends to be presented in Corbin, 1990), rather than being derived from prior journal articles. Even though grounded theory re theory that guided data collection and analysis. A search is conducted iteratively, by analyzing and "purist" rendition of interpretive research reporting collecting data simultaneously, it is usually pre would, however, entail a lengthy and complex qual itative data presentation before the reader learns sented sequentially. This gap occurs because the what the major theoretical dimensions and contri norms of presentation in management (and other butions are likely to be. In other words, the theory academic) journals have positivist origins and im would normally appear after the data presentation pose discrete and sequential categories of data col (Dact, 1985). We suspend this interpretive reporting lection and analysis on authors trying to present hallmark for the sake of advance clarity, and employ grounded theory research. In pure form, grounded the more traditional presentational strategy of pro theory research would be presented as a jumble of viding a theoretical overview first, to preview the literature consultation, data collection, and analy major findings and resulting model. It is important to keep in mind, however, that these concepts actu sis conducted in ongoing iterations that produce ally emerged from the study itself (along with con many relatively fuzzy categories that, over time, sultations with relevant literature that were guided reduce to fewer, clearer conceptual structures. The by the emerging ). (Anonymous, ory would be presented last. provided by editor) Presenting grounded theory in this pure form, however, would be neither efficient nor compre I like this caveat to the presentation of grounded hensible to the majority of researchers who work in theory because it succinctly avoids the impression the positivist paradigm. The norm that has evolved of methodological slurring by identifying the is to present grounded theory in the traditional messy, nonlinear reality of grounded theory re discrete categories and in the same sequence as search and, simultaneously, providing a practical justification (i.e., comprehensibility) for presenting : theory, data collection, data the data and analysis in a traditional "sanitized" analysis, results. Doing so has the unfortunate con format. sequence of creating the impression of methodolog The key objection to mixing grounded theory ical slurring, even when the constant comparative with hypothesis testing, however, is that doing so method has been used. For those unfamiliar with tends to violate the notion of theoretical emer grounded theory techniques, the mode of presenta gence. That is, when a researcher uses grounded tion may also create the unfortunate impression theory techniques to "test" preconceived notions of that grounded theory methods can be mixed with a what is likely to be observed, chances are he or she positivist research agenda. will "see" the intended categories and overlook There are ways of avoiding this incorrect impres more emergent ones. This form of methodological sion while adhering to the sequential norms of slurring tends toward "forced categories" in the journal presentation. Foremost, the process of data process (Glaser, 1992) and reduces analysis, including coding techniques and category grounded theory technique from its intended pur creation, should be made apparent to the reader. pose of identifying new theory to one of simply Authors can do this in their methods sections and, confirming extant understandings of a social phe in my opinion, such accounts are most effective nomenon as a consequence of the researcher im when the authors provide illustrative examples of posing intentions on the data. coding techniques and the evolution of conceptual categories in a table or appendix. Similarly, authors can note that, although they Grounded Theory Is Not Simply Routine are presenting theoretical concepts in a traditional Application of Formulaic Technique to Data manner (i.e., up front in the study), the concepts Some manuscripts reinforce the myth that did, in fact, emerge from the study. Perhaps one of grounded theory is a relatively mechanical tech the most eloquent versions of this statement came nique. This myth has several manifestations. One is from a manuscript recently reviewed in the Acad to present grounded theory as a series of rigid rules, emy of Management Journal: such as "saturation is achieved when one has con

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 638 Academy of Management Journal August

ducted between 25 and 30 interviews." Another is into what the data are reflecting?because he or she the implication that pouring textual data into a fails to see what they really indicate except in terms software package will yield results. A common of trivial or well known phenomena?then the pub lished findings fail on this criterion. Because there characteristic of most efforts to use grounded the is an interplay between researcher and data, no ory is a neurotic overemphasis on coding. The re method, certainly not grounded theory, can en searcher has diligently followed the rules and gone sure that the interplay will be creative. Creativity from open coding (derived from an initial reading depends on the researcher's analytic ability, the of the data) to more abstract or categorical codes oretical sensitivity, and sensitivity to the subtle and finally to conceptual or theoretical codes but ties of the action/interaction (plus the ability to has made no interpretive effort at any stage of cod convey the findings in writing). A creative inter ing. The result is typically a nice set of conceptual play also depends on the other pole of the re categories that, in the process of routine data anal searcher-data equation, the quality of the data col ysis, become divorced from both the data and the lected or analyzed. An unimaginative analysis original research question. That is, although the may in a technical sense be adequately grounded in the data, yet be insufficiently grounded for the rigid application of grounded theory technique researcher's theoretical purpose. This occurs if might produce passable results, such a mechanical the researcher does not draw on the complete approach usually lacks the spark of creative insight resources of data or fails to push data collection upon which exemplary research is based. far enough. (Corbin & Strauss, 1990: 19). The key issue to remember here is that grounded theory is an interpretive process, not a logico-de Those new to grounded theory research must be ductive one. Positivist models of science encourage come both patient and tolerant of ambiguity, be the notion that researchers stand separate from ob cause it is the ongoing interaction between re jects of inquiry in order to minimize the degree to searcher and data that generates the fundament of which the act of observation interferes with or con successful grounded research. taminates the observation. In interpretive models, this assumption is not so firmly held. The re Grounded Theory Is Not Perfect searcher is considered to be an active element of the research process, and the act of research has a In reading reviews of manuscripts that undertake creative component that cannot be delegated to an grounded theory, I detect a growing fundamental algorithm. Qualitative software programs can be ism in grounded theory research. That is, there useful in organizing and coding data, but they are seems to be a growing gap between those who ac no substitute for the interpretation of data. The tually engage in grounded theory and those who researcher must make key decisions about which write about it. The latter group, unsurprisingly, categories to focus on, where to collect the next tends toward purist idealism and, as a result, they iteration of data and, perhaps most importantly, the repeat and reinforce many of the myths I have meaning to be ascribed to units of data. described above: rigid rules about saturation, me Successful grounded theory research has a clear chanical application of technique to data, and clear creative component. Glaser and Strauss were aware demarcation between theory and data. The gap be of this component and the tension it would create tween pragmatics and purists has become exacer with those who find comfort in trusting an algo bated by a proliferation of how-to manuals and rithm to produce results. Glaser (1978) used the textbooks about simplifying and streamlining term "theoretical sensitivity" to describe the essen grounded theory research. tial tension between the mechanical application of Let me be clear that my intent is not to criticize technique and the importance of interpretive in the emergence of a community of scholars devoted sight. The tension ultimately proved to be a point of to improving grounded theory methodology. This departure between the founders of grounded the has occurred in the context of quantitative research ory, with Glaser favoring creativity and openness to as well. There ought to be a healthy tension be unanticipated interpretations of data while Strauss tween pure methodologists and practitioners. (and coauthor Juliet Corbin) became advocates of Keep in mind, however, the pragmatic core of adherence to formal and prescriptive routines for grounded theory research. It was founded as a prac analyzing data (Locke, 1996). Even so, Corbin and tical approach to help researchers understand com Strauss continued to caution against an overly me plex social processes. It was also designed as a chanical application of method: method that might cccupy a pragmatic middle ground between some slippery epistemological If the researcher simply follows the grounded theory boundaries. Because of this genealogy, grounded procedures/canons without imagination or insight theory techniques are inherently "messy" (Parkhe,

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 2006 Suddaby 639

1993) and require researchers to develop a tacit tion "is the process of forming an explanatory hy knowledge of or feel for when purist admonitions pothesis. It is the only logical operation which in may not be appropriate to their research and may troduces any new idea" (Peirce, 1903: 216). Peirce be ignored. also described abduction as the fallible "flash of This statement is not an excuse for ignorance of insight" that generates new conceptual views of the the epistemological and ontological issues sur empirical world. The notion of abduction has be rounding grounded theory methods. Quite the op come incorporated into grounded theory as "ana posite; it speaks to the need for practitioners to lytic induction," the process by which a researcher know this terrain well, for it is only through under moves between induction and deduction while standing the philosophical underpinnings that one practicing the constant comparative method. Strauss can safely navigate the thorny questions of when, and Corbin noted that induction had been overem for example, one's method is consistent with a par phasized in grounded theory research. They observed ticular view of reality. But being aware of one's that whenever researchers conceptualize data, they epistemological position does not justify dogma are engaging in deduction and that effective grounded tism about conducting grounded theory research. theory requires "an interplay between induction Ultimately, questions of when saturation is achieved, and deduction (as in all science)" (1998: 137). how coding should be done, or when counting is In sum, a healthy tension between methodolo appropriate can be resolved pragmatically. gists and practitioners is desirable, but researchers Perhaps an example will clarify these issues of should try to avoid fundamentalist tendencies in messiness and tacitness. A key point of confusion how they approach and, more importantly, evalu in grounded theory research is the question of ate grounded theory research. knowing when saturation has occurred during data collection. Because grounded theory research uses Grounded Theory Is Not Easy iteration and sets no discrete boundary between data collection and analysis, saturation is not al I often get the impression, when reading some of the more egregious examples of how not to do ways obvious, even to experienced researchers. Yet grounded theory research, that the methodology every submitted manuscript contains a statement that saturation occurred. It is a "box" that must be suffers from its apparent simplicity. That is, much like modern art or the sparse prose of Ernest Hem "checked off" prior to publication. ingway, examples of outstanding grounded theory Deciding saturation has happened takes tacit un raise the immediate reaction, "Hey, I could do derstanding, which is achieved as much through that!" The seamless craft of a well-executed experience as through a priori criteria. The indeter grounded theory study, however, is the product of minate, messy nature of saturation invites a funda considerable experience, hard work, creativity and, mentalist drift toward positivism?the 25 interviews. occasionally, a healthy dose of good luck. But, according to Glaser and Strauss, saturation is a Many of the primary techniques of grounded the practical outcome of a researcher's assessment of ory research are developmental. That is, the quality the quality and rigor of an emerging theoretical of their application improves with experience. The model: "The criteria for determining saturation . . . developmental nature of grounded theory research are a combination of the empirical limits of the derives, in part, from individual differences in re data, the integration and density of the theory and searchers' ability to interpret patterns in qualitative the analyst's theoretical sensitivity" (1967: 62). The data (Turner, 1981). Grounded theory is an inter signals of saturation, which include repetition of pretive process that depends upon the sensitivity of information and confirmation of existing concep a researcher to tacit elements of the data or mean tual categories, are inherently pragmatic and de ings and connotations that may not be apparent pend upon both the empirical context and the re from a mere superficial reading of denotative con searcher's experience and expertise. tent. Many grounded theory researchers describe There are other examples of the pragmatic mid this interpretion as occurring subconsciously, as a dle road of grounded theory. Fundamentalists often result of their constant "immersion" in the da incorrectly describe quantitative approaches as ta?an event that some describe as akin to "drown necessarily deductive and grounded theory as in ing" (Langley, 1999). herently inductive. Practically, neither could ever Although the subjective elements of grounded be true. Pierce recognized that pure induction and theory research confer an advantage on some indi pure deduction are necessarily sterile. New ideas viduals, I do not mean to suggest that this bars result from a combination of these fundamental others from ever conducting grounded theory re approaches, which he termed "abduction." Abduc search. Rather, the advantage some enjoy suggests

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 640 Academy of Management Journal August

that we who are less gifted with intuitive skills or Grounded Theory Is Not an Excuse for the pattern-perceiving abilities require more course Absence of a Methodology work, training, and experience to acquire these sen This essay was motivated by a recurring problem in sitivities. Indeed, the genius of Glaser and Strauss's manuscripts I see, in which researchers claim to have original methodology is that it outlines a procedure performed grounded theory research, support their by which formerly tacit processes are made ex claims with a cursory to Glaser and Strauss plicit. The apparent simplicity of grounded theory (1967), and then offer little, if any, description of their research, however, creates the misperception that it methodology. If a revision is granted and the authors is easy and encourages many without prior experi are pushed to reveal how the data were collected and ence or in grounded theory methodology analyzed, it becomes clear that the term "grounded to "have a go." Not surprisingly, the product of theory" was interpreted to mean "anything goes." such an effort will be similar to the product of a Data may have been collected randomly, coded by researcher untrained in statistical methods "having forced application of preexisting conceptual catego a go" at LISREL. ries, and used to test hypotheses. In sum, the research Exemplary research using grounded theory also either ignored or deliberately violated the core proce requires considerable exposure to the empirical dures and tenets of grounded theory methodology context or subject area of research. Contradicting (Locke, 1996). prevalent ideals of scientific detachment from con When I review a paper containing a claim of text, the constant comparative method implies an grounded theory, I check to ensure that, at a intimate and enduring relationship between re minimum, the authors have described their meth searcher and site. Because of this close and long odology transparently enough to reassure me that standing connection, the personality, experience, they followed core analytic tenets (i.e., theoreti and character of a researcher become important cal sampling, constant comparison) in generating components of the research process and should be the data and that I can reasonably assess how the made an explicit part of the analysis (Strauss & data were used to generate key conceptual cate Corbin, 1998). gories. I'm also interested to see indicia of the Two significant consequences flow from this. researchers' theoretical sensitivity?their open First, in grounded theory, researchers must account ness to new or unexpected interpretations of the for their positions in the research process. That is, data, the skill with which they combine litera they must engage in ongoing self-reflection to en ture, data, and experience, and their attention to sure that they take personal biases, world-views, subtleties of meaning. I look for indications of and assumptions into account while collecting, in consistency between research questions (and their assumptions about the world and how peo terpreting, and analyzing data. Such continuous ple come to understand it) and the methods used reflection is not easy, particularly for those unac customed to a model of science in which the re to answer the questions. Finally, I am very atten tive to the researchers' use of technical language searcher is an important focus of the process. in describing their methodology because I believe Second, because the somewhat artificial bound there is a clear connection between rigor in lan ary between researcher and research subject is re guage and rigor in action. moved, the quality of the contact between re My observations about the lapses I have observed searcher and empirical site and the quality of the in reviewing manuscripts may create the unfortu research produced have a direct relationship. As a nate impression that few exemplary models of result, most high-quality grounded theory research grounded theory research exist. This is, of course, arises from an extensive and ongoing commitment simply not true. There are many and, at the risk of to a line of theoretical research and an empirical offending a large number of exceptional grounded site. Consider as an illustration Karl Weick's com theory researchers, I invite you to read one of my mitment to studying firefighters' "sensemaking" or favorites. Isabella's (1990) study on managers' in Hinings and Greenwood's commitment to studying terpretation of processes of organizational change organizational change in the . Such is an outstanding exemplar of grounded theory illustrative examples are supported by empirical methodology. The methods section states a re research showing that the more time researchers search strategy in which the epistemological link spend on-site in organizational contexts, the between research question and methodology is more they report high levels of "self-learning," clear. The process by which subjects and the more their research is cited by peers were identified and selected is made transparent, (Rynes, McNatt, & Bretz, 1999). as is the mode by which raw data were converted to

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 2006 Suddaby 641

conceptual categories. The author offers sufficient emy of Management Journal. Academy of Manage detail in the data presented to create a sense of ment Journal, 47: 454-462. verisimilitude for the reader but also lifts the data Gersick, C. J. G. 1988. Time and transition in work teams: by weaving ongoing interpretations, experience, Toward a new model in group development. Acad and prior literature into a model of evolving man emy of Management Journal, 31: 9-41. agerial interpretations of organizational change. Glaser, B. G. 1978. Theoretical sensitivity. Mill Valley, The study incorporates all of the key analytic tenets CA: Press. and procedures of grounded theory and, at the Glaser, B. G. 1992. Basics of grounded theory analysis. same time, demonstrates a high degree of theoreti Mill Valley, CA: Sociology Press. cal sensitivity in the researcher. It's also a very Glaser, B. G., & Strauss, A. L. 1967. The discovery of readable study. grounded theory: Strategies for qualitative re Many other exemplars of grounded theory re search. New York: Aldine. search?for instance, Sutton (1987), Gersick (1988), and Browning, Beyer, and Shetler (1995)?share Goulding, C. 2002. Grounded theory: A practical guide for management, business and market research the devotion to transparency and rigor described ers. London: Sage. above. I encourage you to read them, both as how-to Husserl, E. 1969. Cartesian meditations: An introduc manuals and as sources of inspiration and motiva tion to phenomenology [D. Cairsn, trans.]. The tion to engage in grounded theory research. One of Hague: Martinus-Nijhoff. the more satisfying elements of conducting re search in management is that, in contrast to the Isabella, L. A. 1990. Evolving interpretations as change physical sciences, social science research looks at unfolds: How managers construe key organizational events. Academy of Management Journal, 33: how human invention continually generates new 7-41. ways of interaction and organization. Researchers can best understand those new modes of interact Krippendorff, K. 2003. Content analysis: An introduc ing and organizing by using a methodology that is tion to its methodology (2nd ed.). Thousand Oaks, CA: Sage. attentive to issues of interpretation and process and that does not bind one too closely to long-standing Langley, A. 1999. Strategies for theorizing from process assumptions. Fortunately, that's precisely what data. Academy of Management Review, 24: 691 710. grounded theory is. Locke, K. 1996. Rewriting the discovery of grounded theory after 25 years? Journal of Management In REFERENCES quiry, 5: 239-245. Locke, K. 2001. Grounded theory in management re Barley, S. R. 1986. Technology as an occasion for struc search. London: Sage. turing: Evidence from the observations of CT scan ners and the social order of radiology departments. Martin, P. Y., & Turner, B. A. 1986. Grounded theory and Administrative Science Quarterly, 31: 78-109. organizational research. Journal of Applied Behav ioral Science, 22: 141-157. Barley, S. R. 1990. The alignment of technology and structure through roles and networks. Administra Morgan, G., & Smircich, L. 1980. The case for qualitative tive Science Quarterly, 35: 61-104. research. Academy of Management Review, 5: 491-500. Bartunek, J. M., Rynes, S. L., & Ireland, R. D. 2006. What makes interesting research and why does it matter? Moustakas, C. 1994. Phenomenological research meth Academy of Management Journal, 49: 9-15. ods. Thousand Oaks, CA: Sage. Browning, L. D., Beyer, J. M., & Shetler, J. C. 1995. Build Parkhe, A. 1993. "Messy" research, methodological pre ing cooperation in a competitive industry: SEMAT dispositions and theory development in interna ECH and the semiconductor industry. Academy of tional joint ventures. Academy of Management Re Management Journal, 38: 113-151. view, 18: 227-268. Burrell, G., & Morgan, G. 1979. Sociological paradigms Peirce, C. S. 1903. The essential Pierce: Selected phil and organisational analysis: Elements of the soci osophical writings, vol. 2. Bloomington: Indiana ology of corporate life. London: Heinemann. University Press. Coase, R. 1988. The firm, the market, and the law. Rynes, S. L., McNatt, D. B., & Bretz, R. D. 1999. Academic Chicago: University of Chicago Press. research inside organizations: Inputs, processes, and Corbin, J., & Strauss, A. 1990. Grounded theory research: outcomes. Personnel , 52: 869-898. Procedures, canons and evaluative criteria. Qualita Schutz, A. 1972. The phenomenology of the social tive Sociology, 13: 3-21. world. London: Heinemenn Educational Books. Gephart, R. P. 2004. Qualitative research and the Acad Spindler, G., & Spindler, L. 1982. Roger Harker and

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms 642 Academy of Management Journal August

Schonhausen: From the familiar to the strange and Disbanding and reconnecting. Administrative Sci back. In G. Spindler (Ed.), Doing the anthropology ence Quarterly, 32: 542-569. of schooling: 21-46. New York: Holt, Rinehart & Winston. Turner, B. A. 1981. Some practical aspects of qualitative data analysis: One way of organizing the cognitive Strauss, A., & Corbin, J. 1998. Basics of qualitative re processes associated with the generation of grounded search: Techniques and procedures for developing theory. Quality and Quantity, 15: 225-247. grounded theory (2nd ed.). Thousand Oaks, CA: Wimpenny, P., & Gass, J. 2000. Interviewing in phenom Sage. enology and grounded theory: Is there a difference? Sutton, R. I. 1987. The process of organizational death: Journal of Advanced Nursing, 31: 1485-1492.

This content downloaded from 198.137.20.6 on Tue, 20 Sep 2016 15:32:32 UTC All use subject to http://about.jstor.org/terms