Qualitative Research and Theory Development

Mystery as Method

Mats Alvesson, Dan Karreman

• Publisher: SAGE Publications Ltd
• Year: 2011
• Print ISBN: 9780857023247, 9780857023230
• eBook ISBN: 9781446259887
• Edition: 1
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

Empirical data is one of the cornerstones of knowledge in the social sciences, and yet the researcher often takes it for granted, reserving his or her imaginative faculties for finding a theory that ?fits the data?. This revealing account of the theory-data relationship calls this faith in data into question and establishes a reflexive framework and vocabulary to explore the creative, political and philosophical elements of data production.

Rather than thinking about the theory-data ?fit?, Alvesson and Karreman will encourage you to consider the research process as one of theory-data interplay, asking if creative empirical material can challenge established theory and inspire new lines of development, and if breakdowns and mysteries encountered in research can be a constructive rather than destructive process.

They will encourage you to think critically about empirical data in terms of construction rather than verification, and most importantly they will encourage you to develop theory that is interesting and novel, rather than naive or irrelevant, making this title essential reading for those who often find the traditional vocabulary and frameworks of social science research obvious or simplistic.

1

---

THE USE OF EMPIRICAL MATERIAL FOR THEORY DEVELOPMENT

---

For social scientists, empirical work is one – if not the only – core activity. At the same time, theoretical knowledge is often seen as the most interesting, valuable and prestigious part of a scientific study. It is also broadly seen as the most difficult element to add. The academic reader of a PhD or a reviewer of a research paper will often find the lack of a theoretical contribution one of the greatest shortcomings. It is not so difficult to produce a description of what people do and say through interviews, observations and other methods, but to continue beyond that and suggest insights, concepts, explanations and other ‘deeper’ aspects offering a more abstract theoretical understanding that goes beyond the relevance of a particular case or sample studied is not so easy. The empirical and theoretical elements are not always engaged in a productive interplay. This is the starting point for this book and we hope to offer some ideas on how this interplay can be accomplished in a creative, challenging, and novel way.

How can empirical studies contribute to the development of theory? According to the conventional wisdom in social science, two basic approaches are available: deduction and induction. Most researchers would still probably adopt a nomothetic approach, thus emphasizing the importance of the deduction of theoretical ideas from earlier knowledge, the formulation of hypotheses, and the testing of these as key ingredients (c.f. Freese, 1980; see also Popper, 1963; 1972). Empirical tests are used either to verify theories, as neo-positivists would put it, or to refute them, as Popper and his followers frame it. Inductivist approaches such as grounded theory (Glaser & Strauss, 1967; Strauss & Corbin, 1994) and many versions of so-called case study research (Eisenhardt, 1989; Yin, 1984) would emphasize the building of theory based on data. Either way, both inductivist and deductivist approaches share a belief in a clear separation of theory and data and a deep seated trust in the capacity of data to inform and correct theory building. They also share a strong belief in premeditated process and both downplay the subjectivity of the researcher.

The case study approach could serve as an illustration for our point. Although Robert Yin is by far the most well-known advocate of the case study approach we will pay particular attention to Kathleen Eisenhardt’s (1989) take on the capacity for theory development in the case study approach. Eisenhardt suggests that theory development here proceeds through distinct steps: a tentative formulation of research questions (with as little theoretical baggage included as possible); the selection of cases; the crafting of instruments and protocols; an engagement with the field; analyzing within-case data; searching for cross-case patterns (if multiple cases are researched); shaping hypotheses; enfolding literature; and reaching closure. She also stresses that these moves typically follow one another in an iterative fashion. One would expect at lot of travelling back and forth between stages during research and that this movement is critical for the development of new ideas and insights. In addition, Eisenhardt stresses the inductive character of the theory development process. Theory emerges through intimate contact with empirical materials, and through the frictions and tension between and within various data sets.

She argues that this also leads to one of the biggest strengths with theory development through the case study approach. Since a case study typically leads to rich and messy data sets, these data sets are rife with contradiction and paradox. This makes it possible to juxtapose conflicting evidence, thus freeing up the curious mind to rethink the relationships between the data points. Eisenhardt argues that this increases the potential for new and creative theory:

That is, attempts to reconcile evidence across cases, types of data, and different investigators, and between cases and literature increase the likelihood of creative reframing into a new theoretical vision. Although a myth surrounding theory building from case studies is that the process is limited by investigators’ preconceptions, in fact, just the opposite is true. This constant juxtaposition of conflicting realities tends to ‘unfreeze’ thinking, and so the process has the potential to generate theory with less researcher bias than theory built from incremental studies or arm-chair, axiomatic deduction. (Eisenhardt, 1989: 546)

She also claims that theory generated from case studies is likely to lead to theory that can be measured and be proven false. The idea here is that since theory emerges from empirical settings it always is rooted in concrete realities and eventualities, which means that it in a sense it is already operationalized and also is less likely to be immunized from falsification. Data will keep the speculative mind in check. Eisenhardt believes as well that theory generated from case studies is likely to be empirically valid. Again, the intimate relationship between data and theory almost guarantees that theory reflects underlying realities. Data, or evidence, will provide a compass that can keep the theory generation process on course.

Yin and Eisenhardt make a strong case for using case studies to generate theory. We would agree with this. Case studies, when properly designed, will be helpful for theory-building purposes for a variety of reasons, but most importantly because they provide a strong potential for a certain thickness of description. Properly executed case studies generate an abundance of empirical material that is almost certain to challenge established assumptions and perspectives. We also think that Yin and Eisenhardt are mostly right about the advantages of case studies. However, we would beg to differ on how case studies can facilitate theory generation. It is clear that both Yin and Eisenhardt have a strong belief in the robust nature of data. This, they both claim, is the big advantage of theory generation from case studies. Data will navigate the process and provide well-grounded and robust theory that has a strong empirical validity. Theory will provide an insight into the complexities and intricacies of empirical reality. In this sense, Yin and Eisenhardt use theory to resolve data, hence the resumed lack of application range.

In this respect inductivists like Eisenhardt and Yin (and Glaser and Strauss), who would claim theory is to be developed through sifting through data, are no different from deductivists who would see theory emerging through the accumulation of verified (or corroborated) hypothesizing. These views of social science are in many ways different, but both rely on data as the central element in social research. Theory is supposed to ‘fit’ data – either by design, where a lack of fit should lead to rejections or revisions of a theory, or by default, where theory is understood as emerging from data. Theory and data are thus seen as ‘external’, two different entities that can and should be related while still being recognized as separate.

In this book we shall suggest adopting a different approach. In particular, we would wish to highlight the usefulness of empirical material for theory development through recognizing the fusion of theory and empirical material in the research construction process. We would emphasize the potential of empirical material as a resource for developing theoretical ideas through the active mobilization and problematization of existing frameworks. In particular, we shall point to the ways empirical material can be used to facilitate and encourage critical reflection: to enhance our ability to challenge, rethink and illustrate theory. This approach recognizes the constructed nature of empirical material and ‘proofs’ (Astley, 1985; Denzin & Lincoln, 2005; Shotter, 1993; Shotter & Gergen, 1994; Steier, 1991). It assumes that something is going on out there in reality and there may be better or worse ways of addressing this reality that can be more or less backed up by what might appear to be evidence.

However, it also takes seriously the view that frameworks, pre-understandings and vocabularies are central in producing particular versions of the world. ‘Data’ in social science are seldom so strong or clear-cut that a researcher can claim to have produced unproblematic knowledge about how complex social reality looks or operates. This is not an excuse for not taking empirical material seriously, but perhaps often to do so in an open-minded and humble way. We would propose a relaxation of the emphasis on ‘data’ and a greater interest in the contribution of how ‘data’ are constructed for the benefit of theoretical reasoning (c.f. Sutton & Staw, 1995).

Some time ago, ‘empirical’ research frequently meant that one could assume an independent reality out there which could be perceived and measured through indications of this reality, i.e. data. Through the careful design of procedures, the collection and processing of data based on this design and the subsequent analysis of these, empirical research could say yes or no to various hypotheses about the chunk of reality targeted for study. Nowadays, it would be seen in many social science camps as old-fashioned, intolerant, and theoretically and philosophically unsophisticated to favour this idea. The label of positivism – as currently broadly defined (or not defined, but used) – invites all sorts of pejorative comments. During recent times, there have been more varied views on what constitutes empirical research, making the meaning of this activity quite vague. Reading texts of all kinds, for example, could constitute undertaking empirical research for some people.

But typically, ‘empirical research’ refers to taking a strong interest in gathering or constructing empirical material that says something about what goes on out there – in the social life existing outside of the research practices of academics or available texts. Even the increasing number of people in social science who are skeptical to the possibility of the ‘collection’ and processing of data in order to say yes or no to various hypothesis and theories will often take an interest in empirical work. In many forms of qualitative studies (e.g. in grounded theory) the assumption is that data, carefully processed, can guide the researcher to understand specific phenomena and to develop theory (Glaser & Strauss, 1967; Strauss & Corbin, 1994).¹ In interpretive work it is assumed that we can access and study social reality through indications of the meanings and symbolic interactions that are viewed as crucial elements in social communities. Even though postmodernism give strong reasons for being more careful and modest about such enterprises than previously, it would be remiss of us not to be interested in what we can learn from empirical work.

The key point of this book is to suggest a framework and vocabulary for thinking about theory development that is inspired by empirical studies and different from conventional views of building on data associated with grounded theory and other ‘dataistic’ approaches. We would thus emphasize the creative and imaginative constructions of empirical material. Rather than assuming that ‘data’, like a signpost, point in a specific direction, ‘data’ read as empirical material make a variety of readings possible and may also make different knowledge results possible. Rather than asking and checking if there is a data-theory fit, we ask and explore if empirical material can encourage the challenging and rethinking of established theory and thus inspire novel lines of theory development.

Questioning the Faith in Data

This great faith in data and empirical inquiry as a cornerstone in knowledge development has been challenged by a multitude of intellectual streams during recent years. A powerful example is what may be referred to as ‘non-objective’ interpretivist perspectives. These put an emphasis on how pre-understanding, paradigm and metaphor can pre-structure our basic conceptualizations of what we want to study. Our approach to, perceptions of, and interpretations of what we experience are filtered through a web of assumptions, expectations and vocabularies that will then guide entire projects and be crucial for the results we arrive at (e.g. Brown, 1977).

Somewhat more far-reaching critiques have been raised by feminists pointing at how male domination and masculine standards have influenced the dominant epistemology and methodology in social science (Jaggar, 1989). Male domination has produced a masculine social science built around ideals such as objectivity, neutrality, distance, control, rationality, and abstraction. Alternative ideals such as commitment, empathy, closeness, cooperation, intuition, and specificity have thereby been marginalized. Scientific rationality is thus expressing male domination, rather than superior reason. If one looks at the psychology of researchers and conflicts between different groups, the idea of the distanced and neutral scholar who is rationally oriented towards objective truth becomes peculiar (Bärmark, 1999; Popper, 1976). Researchers can often be very committed to their research and emotionally attached to theories and results. Critique and counter-evidence can then lead to defensiveness rather than a willingness to radically revise a position.

A related point of view has been expressed by critical theorists who would emphasize the political, interest- and value-laden nature of social enquiry (Alvesson & Deetz, 2000; Delanty, 2005; Kincheloe & McLaren, 1994). It is argued that knowledge development is grounded in human interests (Habermas, 1972). In social science, it is impossible to say anything of social significance without having some implications for the formation of society – social science is notoriously and inevitably political. Neither the researcher nor the other actors involved in influencing a research process and its outcome (research foundations, research leaders, editors and reviewers, the people studied and the mass media, and others who would guide them in how to think and how to express themselves) can exist in an ideological vacuum. It is seldom possible to identify and sort out the ideological from non-ideological elements when studying families, gender issues, the socialization of children, consumption, our care of elders, voting behavior, ethnicity, etc. The vocabulary is, for example, not neutral, even though commonly used language will often give the impression of being so. That human interests and cultural, gendered and political ideals can put their imprint on methodological ideals as well as on research practices and results makes it very difficult to see science as a pure activity – neutral and objective in relation to the reproduction or challenging of social ideologies, institutions and interests.

Even more profound are the views of discursivists and (other language-focused) constructivists which would deny science any privileged access to the objective truth about the social world outside language and language use (Potter & Wetherell, 1987; Steier, 1991). Language constructs rather than mirrors phenomena, making representation and thus empirical work a basically problematic enterprise, or so it is argued (Denzin & Lincoln, 2005; Gergen & Gergen, 1991). What (possibly) exists ‘out there’ (e.g. behaviors) or ‘in there’ (e.g. feelings or motives) is complex and ambiguous and can never simply be captured, but given the perspective, the vocabulary and the chosen interpretation, ‘reality’ can emerge in a particular way. Any claim of truth then says as much or more about the researcher’s convictions and language use than about the object of study. Foucault (1980), one of the most influential social theorist (broadly defined) at present, claimed that social scientific knowledge was closely associated with power (the regulation of social reality through arrangements and ordering devices) and less with exploring or distorting truth than creating it.

To sum up, it is increasingly common to claim

that there is no clear window into the inner life of an individual. Any gaze is always filtered through the lenses of language, gender, social class, race, and ethnicity. There are no objective observations, only observations socially situated in the worlds of the observer and the observed. (Denzin & Lincoln, 1994: 12)

The critique of positivism and neo-positivism is massive. However, this does not stop the majority of researchers from doing normal science more or less as if nothing has happened. Questionnaire researchers still assume that the X’s put in small squares by respondents make it possible to determine what goes on in the social world. Qualitative researchers still present interview statements as if they were pathways to the interiors of those being interviewed and that observational data via codification and categories will mirror social practice, although it is broadly recognized – also amongst positivists – that data need to be interpreted to say anything. And that the process always involves the (selective and contestable) construction of data, as well as any use of it.

One problem with the critique of approaches having a strong faith in data is that it is perceived as categorical, provocative and destructive, and therefore is neglected. Another problem is that much of the critique addresses philosophical and epistemological issues, while the craft of doing research – for example fieldwork – has received much less attention. This is largely viewed – at least in most textbooks and also in research reports – as a technical matter, separated from theoretical and philosophical ideas about knowledge production, although some change is on the way here. Method (the action-related principles and ideas on how to produce and make sense of empirical material) still largely remains comparatively unaffected by all the work that has tremendous relevance for our understanding of methodological practices. The wealth of insights into problems of developing knowledge and the limitations to social science as a rational project need to be connected to research practices. Many researchers feel that all this philosophy of science and associated critique of traditional research are of limited relevance (e.g. Melia, 1997). We think the challenge is to try to incorporate parts of these into research practice. This rather heterogeneous but rapidly expanding critique of social research and its uncertain relevance for specific methods for doing fieldwork, interpreting and writing poses one context for this book.

The ambition then is to work with empirical material and to take it seriously without giving it a non-motivated robust status, as well as to treat it as if it offers a strong authority for forms of knowledge that are based on claims of being grounded in data and revealing reality. The ambition is also to put critical ideas about dataism into constructive use, where the possibilities – rather than the problems and impossibilities of empirical research – can be emphasized.

An Honest Account of How Empirical Material is Created and Processed

The aforementioned, more theoretically- and philosophically-based critique of empirical results as a solid building block in the accumulation of knowledge can be supplemented with some consideration of the practical problems of mirroring reality out there in research texts.

A major issue with the (limited) reliability of empirical studies concerns the many (more or less) coincidental and arbitrary in...

Table of contents

• Cover Page
• Title
• Copyright
• Contents
• Preface
• 1 The use of empirical material for theory development
• 2 The role of constructions and language in empirical research
• 3 Key methodological principles for detecting mysteries
• 4 A methodology of sorts for theorizing from empirical material
• 5 Illustrating the development and resolution of mysteries
• 6 Fieldwork techniques and mystery creation
• 7 Research guided or assisted by mystery?
• References
• Subject Index
• Author index