Qualitative Data Analysis

A User Friendly Guide for Social Scientists

Ian Dey

• Publisher: Routledge
• Year: 2003
• eBook ISBN: 9781134931453
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

First Published in 2004. Learning how to analyze qualitative data by computer can be fun. That is one assumption underpinning this introduction to qualitative analysis, which takes account of how computing techniques have enhanced and transformed the field. The author provides a practical discussion of the main procedures for analyzing qualitative data by computer, with most of its examples taken from humour or everyday life. He examines ways in which computers can contribute to greater rigour and creativity, as well as greater efficiency in analysis. He discusses some of the pitfalls and paradoxes as well as the practicalities of computer-based qualitative analysis. The perspective of "Qualitative Data Analysis" is pragmatic rather than prescriptive, introducing different possibilities without advocating one particular approach. The result is a largely discipline-neutral text, which is suitable for arts and social science students and first-time qualitative analysts.

Chapter 1
Introduction

Q. What colour is snow?

A. White.

To most of us, the answer ‘white’ may seem satisfactory, but to an Eskimo it would seem a joke: Eskimos distinguish between a wide variety of ‘whites’ because they need to differentiate between different conditions of ice and snow. So it is with qualitative data analysis: in a recent review of the field, Tesch (1990) distinguishes over forty types of qualitative research (Illustration 1.1). Just as the Eskimos distinguish varieties of white, so researchers distinguish varieties of qualitative analysis. There is no one kind of qualitative data analysis, but rather a variety of approaches, related to the different perspectives and purposes of researchers. To distinguish and assess these different perspectives fully would be a formidable and perhaps rather fruitless task, particularly as the boundaries between different approaches and their relation to what researchers actually do when analysing data is far from clear. But is there a basic core to qualitative research, as there is a basic colour ‘white’, from which these different varieties are derivative?

Different researchers do have different purposes, and to achieve these may pursue different types of analysis. Take a study of the classroom, for example. An ethnographer might want to describe the social and cultural aspects of classroom behaviour; a policy analyst might want to evaluate the impact of new teaching methods; a sociologist might be most interested in explaining differences in classroom discipline or pupil achievement—and so on. Different preoccupations may lead to emphasis on different aspects of analysis. Our ethnographer may be more interested in describing social processes, our policy analyst in evaluating results, our sociologist in explaining them. This plurality of perspectives is perfectly reasonable, remembering that social science is a social and collaborative process (even at its most competitive), in which (for example) descriptive work in one project may inspire interpretive or explanatory work in another (and vice versa).

ILLUSTRATION 1.1
DIFFERENT APPROACHES TO QUALITATIVE RESEARCH

---

action research	ethnographic content analysis	interpretive interactionism
case study	interpretive human studies
clinical research	ethnography	life history study
cognitive anthropology	ethnography of communication	naturalistic inquiry
collaborative enquiry	oral history
content analysis	ethnomethodology	panel research
dialogical research	ethnoscience	participant observation
conversation analysis	experiential psychology	participative research
Delphi study	field study	phenomenography
descriptive research	focus group research	phenomenology
direct research	grounded theory	qualitative evaluation
discourse analysis	hermeneutics	structural ethnography
document study	heuristic research	symbolic interactionism
ecological psychology	holistic enthnography	transcendental realism
educational	imaginal psychology	transformative research
connoisseurship and criticism	intensive evaluation
educational ethnography

---

Source Tesch 1990:58

Given the multiplicity of qualitative research traditions, one might reasonably wonder whether there is sufficient common ground between the wide range of research traditions to permit the identification of anything like a common core to analysing qualitative data. On the other hand, the very notion of ‘qualitative’ data analysis implies, if not uniformity, then at least some kind of family kinship across a range of different methods. Is it possible to identify a range of procedures characteristic of qualitative analysis and capable of satisfying a variety of research purposes, whether ethnographic description, explanation or policy evaluation is the order of the day? The relevance and applicability of any particular procedure will, of course, depend entirely on the data to be analysed and the particular purposes and predilections of the individual researcher.

Having identified a multiplicity of perspectives, Tesch manages to reduce these to three basic orientations (1991:17–25). First, she identifies ‘language-oriented’ approaches, interested in the use of language and the meaning of words—in how people communicate and make sense of their interactions. Second, she identifies ‘descriptive/interpretive’ approaches, which are oriented to providing thorough descriptions and interpretations of social phenomena, including its meaning to those who experience it. Lastly, there are ‘theory-building’ approaches which are orientated to identifying connections between social phenomena—for example, how events are structured or influenced by how actors define situations. These distinctions are not water-tight, as Tesch herself acknowledges, and her classification is certainly contestable. No one likes to be pigeon-holed (by some one else), and nothing is more likely to irritate a social scientist than to be described as atheoretical! However, Tesch does suggest a strong family resemblance between these different research orientations, in their emphasis on the meaningful character of social phenomena, and the need to take this into account in describing, interpreting or explaining communication, cultures or social action.

Thus encouraged, we can look for a basic core of qualitative data analysis, though not in some consensus about research perspectives and purposes, but rather in the type of data we produce and the way that we analyse it. Is there something about qualitative data which distinguishes it from quantitative data? And if qualitative data does have distinctive characteristics, does this also imply distinctive methods of analysis? My answer to both these questions is a qualified ‘yes’. In Chapter 2 I distinguish between qualitative and quantitative data in terms of the difference between meanings and numbers. Qualitative data deals with meanings, whereas quantitative data deals with numbers. This does have implications for analysis, for the way we analyse meanings is through conceptualization, whereas the way we analyse numbers is through statistics and mathematics. In Chapter 3, I look at how we conceptualize qualitative data, including both the articulation of concepts through description and classification, and the analysis of relationships through the connections we can establish between them.

I said my answers were qualified, for though we can distinguish qualitative from quantitative data, and qualitative from quantitative analysis, these distinctions are not the whole story. We can learn as much from how meanings and numbers relate as we can from distinguishing them. In social science, number depends on meaning, and meaning is informed by number. Enumeration depends upon adequate conceptualization, and adequate conceptualization cannot ignore enumeration. These are points I take up in Chapters 2 and 3. My aim is to introduce the objects and methods of qualitative analysis, as a basis for the subsequent discussion of procedures and practice.

It is easy to exaggerate the differences between qualitative and quantitative analysis, and indeed to counterpose one against the other. This stems in part from the evolution of social science, most notably in its efforts to emulate the success of the natural sciences through the adoption of quantitative techniques. The fascination with number has sometimes been at the expense of meaning, through uncritical conceptualizations of the objects of study. Nowhere is this more apparent than in the concepts-indicators approach, where specifying the meaning of concepts is reduced to identifying a set of indicators which allow observation and measurement to take place—as though observations and measurement were not themselves ‘concept-laden’ (Sayer 1992). The growing sophistication of social science in terms of statistical and mathematical manipulation has not been matched by comparable growth in the clarity and consistency of its conceptualizations.

Action breeds reaction. In response to the perceived predominance of quantitative methods, a strong undercurrent of qualitative research has emerged to challenge the establishment orthodoxy. In place of the strong stress on survey techniques characteristic of quantitative methods, qualitative researchers have employed a range of techniques including discourse analysis, documentary analysis, oral and life histories, ethnography, and participant observation. Nevertheless, qualitative research is often cast in the role of the junior partner in the research enterprise, and many of its exponents feel it should have more clout and more credit. This encourages a posture which tends to be at once defensive of qualitative methods and dismissive of the role of the supposedly senior partner, quantitative research.

Beneath these rivalries, there is growing recognition that research requires a partnership and there is much to be gained from collaboration rather than competition between the different partners (cf. Fielding and Fielding 1986). In practice, it is difficult to draw as sharp a division between qualitative and quantitative methods as that which sometimes seems to exist between qualitative and quantitative researchers. In my view, these methods complement each other, and there is no reason to exclude quantitative methods, such as enumeration and statistical analysis, from the qualitative toolkit.

Reconciliation between qualitative and quantitative methods will undoubtedly be encouraged by the growing role of computers in qualitative analysis. The technical emphasis in software innovation has also encouraged a more flexible and pragmatic approach to developing and applying qualitative methods, relatively free from some of the more ideological and epistemological preoccupations and predilictions dominating earlier discussions. The development of software packages for analysing qualitative data has also stimulated reflection on the processes involved, and how these can be reproduced, enhanced or transformed using the computer. The development of computing therefore provides an opportune moment to consider some of the main principles and procedures involved in qualitative analysis. I outline the general contribution of the computer to qualitative analysis in Chapter 4. In doing so, I take account of how computers can enhance or transform qualitative methods. This is a topic I address explicitly in Chapter 4, but it also forms a recurrent theme throughout the discussion of analytic procedures in the rest of the book.

On the other hand, software development has also provoked concerns about the potentially damaging implications of new technological forms for traditional methods of analysis. Some developers have emphasized the potential danger of the software they themselves have produced in facilitating more mechanical approaches to analysing qualitative data, displacing traditional analytic skills. This concern has highlighted the need to teach computing techniques within a pedagogic framework informed by documented analytic principles and procedures. Paradoxically, however, existing accounts of qualitative methodology and research are notoriously deficient in precisely this area. Burgess (1982), for example, in his review of field research, complains that there are relatively few accounts from practitioners of the actual process of data analysis or from methodologists on how data analysis can be done. The literature is littered with such complaints about the lack of clear accounts of analytic principles and procedures and how these have been applied in social research. Perhaps part of the problem has been that analytic procedures seem deceptively simple. The conceptual aspects of analysis seem frustratingly elusive, while the mechanical aspects seem embarrassingly obvious. Thus Jones suggests that qualitative data analysis involves processes of interpretation and creativity that are difficult to make explicit; on the other hand, ‘a great deal of qualitative data analysis is rather less mysterious than hard, sometimes, tedious, slog’ (Jones 1985:56).

The low status and marginality of qualitative research generally have fostered defensive posturing which emphasizes (and perhaps exaggerates) the subtleties and complexities involved in qualitative analysis. It has also led to a heavy emphasis on rigorous analysis. The resulting analytic requirements can seem quite intimidating, even to the experienced practitioner. There has also been a tendency to dress methodological issues in ideological guise, stressing the supposedly distinctive virtues and requirements of qualitative analysis, by contrast with quantitative methods, for example in apprehending meaning or in generating theory. At its worst, this aspires to a form of methodological imperialism which claims that qualitative analysis can only proceed down one particular road. As Bryman (1988) argues, more heat than light has been generated by the promulgation of epistemological canons that bear only a tenuous relation to what practitioners actually do. To borrow an apt analogy, we need to focus on what makes the car run, rather than the design and performance of particular models (Richards and Richards 1991).

This lacuna has been made good to some extent in recent years (e.g. Patton 1980, Bliss et al. 1983, Miles and Huberman 1984, Strauss 1987, Strauss and Corbin 1990), though not always in ways accessible to the firsttime practitioner. This book is one more attempt to help plug the pedagogical gap referred to above. The focus is on the engine rather than on any particular model. My assumption is that the practical problems of conceptualizing meanings are common to a range of different perspectives. For example, the interpretive approach of Patton (1980) emphasizes the role of patterns, categories and basic descriptive units; the network approach of Bliss and her colleagues (1983) focuses on categorization; the quasi-statistical approach of Miles and Huberman (1984) emphasizes a procedure they call ‘pattern coding’; and the ‘grounded ...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Figures, Illustrations and Tables
• Preface
• Acknowledgements
• Chapter 1: Introduction
• Chapter 2: What Is Qualitative Data?
• Chapter 3: What Is Qualitative Analysis?
• Chapter 4: Introducing Computers
• Chapter 5: Finding a Focus
• Chapter 6: Managing Data
• Chapter 7: Reading and Annotating
• Chapter 8: Creating Categories
• Chapter 9: Assigning Categories
• Chapter 10: Splitting and Splicing
• Chapter 11: Linking Data
• Chapter 12: Making Connections
• Chapter 13: Of Maps and Matrices
• Chapter 14: Corroborating Evidence
• Chapter 15: Producing an Account
• Chapter 16: Conclusion
• Appendix 1: ‘If the Impressionists Had Been Dentists’ (Woody Allen 1978)
• Appendix 2: Software
• Glossary
• References