David Silverman (2006). Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction (Third edition). London: Sage, XV + 428 pages, ISBN 9781412922449 (Hardcover) £70.00, ISBN 9781412922456 (Paperback) £22.99

Abstract: This edition of SILVERMAN's well-known book offers a wide-ranging introduction to the problems facing any qualitative researcher, especially as concerns the design of qualitative projects and the analysis of qualitative data. It is in many ways a personal book, often referring to the author's own experience and reflecting his own intellectual development. He is clear about his preferences and doubts, but offers good arguments for both. While it is presented as a textbook for undergraduates, it may be considered too demanding intellectually in some cases. The review offers an extensive overview of the book's contents, in order to facilitate a teacher's choice of it as a course book, but it is recommended without reservation to any serious qualitative researcher.

Key words: qualitative research, qualitative data analysis, naturally occurring data, ethnomethodology, conversation analysis, constructionism, semiotics, emotionalism

Even if we could imagine a textbook freed from authorial prejudice, it would be a pretty dull affair—rather like those awful book reviews which do little more than list the titles of each chapter(SILVERMAN, p.377)

David SILVERMAN is both one of the most successful writers on "qualitative methods" and also a major representative of what may be called the sceptical trend in qualitative research. The fact that the book under review is in its third edition, the previous ones being published in 1993 and 2001, already testifies to the first point. It is one of a set of three. There is another book authored by SILVERMAN, called Doing Qualitative Research: A Practical Handbook (second edition 2005), which offers "a guide to the business of conducting a research project at the graduate level" (p.XIV) and Qualitative Research: Theory, Method and Practice (second edition 2004), which is an edited collection of essays on a range of particular approaches written by specialists. [1]

SILVERMAN's scepticism is clearly expressed right in the preface. He opposes those kinds of qualitative research that are effectively a-theoretical, are mostly based on unstructured interviewing, stressing "authenticity" and the expression of "experience" rather than reliability and validity, and have the analysis of data being determined by a particular moral or political position. Or more positively, SILVERMAN stresses the importance and even the unavoidability of theory, the use of naturally occurring data, the issue of credibility in qualitative research, while he totally rejects partisanship as a basis for assessing research findings or judging the relevance of research topics (p.XIII). In short, he can be seen as a "contrast figure" to the trend promoted by Norman DENZIN and Yvonna LINCOLN in their journal Qualitative Inquiry, and also strongly represented in the successive editions of their Handbook of Qualitative Research (DENZIN & LINCOLN, 1994, 2000, 2005). [2]

I do sympathise with SILVERMAN's position, so if I have a "prejudice" as a reviewer, it is a positive one. On the other hand, I have a special interest in the choices that seem to underlie the book, as I have myself produced one that, in large part, covers similar topics (TEN HAVE, 2004). The previous edition of SILVERMAN's book has been reviewed already in FQS (KALEKIN‑FISHMAN, 2001), which allows some interesting comparisons, between the second and third editions and between her impressions and mine. [3]

Comparing the three successive editions of Interpreting qualitative data, a first and obvious observation is that each next edition is bigger than the previous one. The first had 224 pages, all included, the second 325, while the present one has 428. In his "Preface", SILVERMAN refers to a change in the teaching of qualitative research to the effect that courses in the field now tend to focus on students actually doing research, often in the form of a small "training project". This edition offers, therefore, a wider range of "support devices" in addition to the basic text. As before, there are many "exercises", inviting students to consider particular research problems, or gather and/or analyse actual data. The idea is that the practical skills needed for doing research are best learned by actual experience in research activities, rather than by abstract "prescription". New are explicit statements of learning objectives at the start of each chapter as well as many "boxed" case studies, student tips and links to relevant Internet sites throughout the text. [4]

There are, of course, many ways in which a book on qualitative research can be organised in "parts", "chapters" and "sections". As the title, Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction, makes clear the overall focus is on the analysis of data, rather than their "production" or "collection". There are some scattered remarks on data collection when this seems to have an impact on the analysis or as an extra suggestion. Throughout the text, SILVERMAN stresses that the analysis depends on the character of research questions, which in turn, depend on a chosen theoretical framework, while the researcher has, at the same time, to be "open" to what the data suggest by themselves. [5]

The first part, "Theory and Method", functions as a general introduction. Besides some general warnings to students and a contrast analysis of quantitative and qualitative research, it offers typologies of "methods" and "models". The four major methods are, according to SILVERMAN, "observation", "analysing texts and documents", "interviews and focus groups", and "audio and video recording". It is on the basis of this typology that part 2, which presents the core of the book, is organised, although the second, "method", is covered in two separate chapters—"texts" and "visual images". The (theoretical) models are also differentiated into four types: "naturalism", "ethnomethodology", "emotionalism" and "postmodernism". These different models do not return as chapter headings, but reemerge in various guises in the chapters dealing with methods. [6]

In the first of these, on "Ethnography and observation", SILVERMAN presents an overview of varieties of ethnography, with short vignettes of some examples. The major argument is that any observation is "theory-laden", even if this is not expressed explicitly by the observer. He opposes a style of "just telling it like it is", which he calls "naturalism", and displays affinity with (a mild version of) ethnomethodology's invocation to treat common sense as topic rather than resource, as well as its stress on the details of social practices, rather than, say, "perspectives". [7]

The next chapter in part 2, "Interviews", offers a clear display of what I called above SILVERMAN's skepticism. Early on, he discusses the questions "Why interview?", and "why should we ever depart from naturally occurring data and use contrivances like interviews and focus groups?" (p.113). The largest part of the chapter offers an elaborate discussion of the sense of interview data from the perspective of three different approaches which he calls "positivism", which is interested in facts about behaviour and attitudes, "emotionalism", which is after authentic expressions of personal experiences and feelings, and "constructionism" which treats interview talk itself as the topic of analytic attention. It is the latter, of course, that he favours, offering summary discussions of some inspiring examples. When we compare these three approaches with the "models" mentioned in part one, and the two styles in the previous chapter, we must conclude that both the labels and the characterisations are not very stable, but seem to have been adapted to the "local arguments" in an ad hoc fashion. [8]

Third in this part is a chapter on "Texts". While observations and interviews are rather obvious kinds of materials in qualitative research, the analysis of texts is less common. This chapter offers discussions of four ways in which these can be analysed. The first of these is Content Analysis, in which "researchers establish a set of categories and then count the number of instances that fall into each category". The problem with this mainly quantitative method is how to choose one's categories. In the chapter under discussion, this approach seems to function as a contrast to the other three. The second one analyses narrative structures. It has its origin in structural linguistics and searches for internal structures in textual materials. Third come from "ethnography" which looks at the ways in which various kinds of documents—such as files, statistical records, official proceedings and those available on the Internet—actually function in society. For the final one, ethnomethodology, SILVERMAN has selected Membership Categorization Analysis (MCA), as conceived by Harvey SACKS, as an example. All four are amply illustrated with concrete examples and supplied with useful suggestions. As in the previous chapter, the order of the presentations seems to suggest an order of preference. Although each gets a "fair" treatment, the reader gets the most solid introduction to MCA. [9]

The fourth chapter in part two is called "Naturally occurring talk". It starts with a preference for "naturally occurring data" in contrast to "researcher-provoked" ones, and continues with "why work with tapes?" and the transcription of audio tapes. This sets the scene for the major research style using tapes and transcripts, Conversation Analysis, followed by a treatment of Discourse Analysis, as used in Discursive Psychology, and a discussion that compares these two. [10]

The topic of the final chapter in part two is "Visual images". It is presented as a compensation, of sorts for the neglect of visual aspects in previous chapters. SILVERMAN differentiates kinds of visual data and research strategies, mentioning quasi-experiments, visuals as supplements to researcher-provoked data and naturally occurring data. The bulk of the chapter, however, treats "three widely used ways of analysing visual images": content analysis, semiotics and workplace studies. Both content analysis and semiotics have already been discussed in the chapter on texts, while workplace studies figure as an extension of conversation analysis. [11]

I earlier characterised SILVERMAN's "position" on many aspects of qualitative research as one of scepticism. While this attitude was inferable in many places in part two, it is much more outspokenly expressed in the first chapter, "Credible qualitative research", or part three, "Research practice". He again argues strongly against trends that focus on "experience", "subjectivity", and/or "authenticity", or that use a researcher's value position or politics as a major criterion for the value of his or her research report. "If qualitative research is to be judged by whether it produces valid knowledge, then we should properly ask highly critical questions about any piece of research" (p.275). What SILVERMAN argues for, instead, is to treat qualitative research as one way to do social science, which is not basically different from other scientific pursuits, including qualitative social science and natural science. He is not afraid, therefore, to use concepts like falsification, reliability, validity and generalisability. In fact, the latter three concepts are used in the heading of this chapter's major sections. [12]

In quantitative research, reliability refers to stability of measurement. In qualitative research, it is often argued, this is not relevant, as any observation or interpretation is unique, being circumstantial, while its objects are in constant flux. SILVERMAN, however, maintains that in qualitative research reliability is, in a sense, adapted to this type of investigation, a serious consideration. The idea is to make one's conclusions less dependent on particular circumstances and personal preferences and more open to an outsider's inspection. It involves making the research process as systematic and transparent as possible and facilitating the reader’s access to the original data, for instance by using "low-inference descriptors", providing the context for quotes, and using standardised and detailed transcriptions. [13]

Validity has to do with the relation between an account, concept or description, and what it purports to represent, its object. SILVERMAN first offers a rather critical discussion of two methods for checking validity which are sometimes recommended as particularly fitting for qualitative research: triangulation and respondent validation. Although these methods can be useful as a way to produce additional insights or data, they do not seem to produce their proposed function because they are based on rather debatable assumptions of respectively a fixed underlying reality and an epistemological privilege of participants' common sense knowledge. What he recommends, instead, is a series of connected efforts to combat "anecdotalism" by systematising one's research process through analytic induction, the constant comparative method, deviant-case analysis, comprehensive data treatment, and tabulations when appropriate. The chapter is concluded with a discussion of sampling issues, stressing well-considered purpose and/or theoretical sampling to increase the relevance and credibility of the research. [14]

The other two chapters of this part on "Research practice" offer very sensible but much less uniquely "Silvermanian" considerations and practical suggestions. In the one on "Research ethics", he shows, referring to quite vivid examples, that there are no easy and general solutions to ethical problems. Therefore, the researcher has to develop a "contextual awareness" to balance research interests and subjects’ concerns when dealing with subject identifications, privacy negotiations, etc. "Writing your report" offers a short, 10-page summary of rather obvious points of attention. [15]

The fourth and final part of the book is called "Implications". Its first chapter deals with "The relevance of qualitative research". It offers a succinct discussion of the general problems of the "practical application" of social research in general and qualitative research in particular. SILVERMAN in critical of some conventional "roles" of the researcher vis à vis practical matters, such as "scholar", "state counsellor" or "partisan" and argues for the need to differentiate the various audiences for research results: academic colleagues, policy makers, practitioners and the general public. He is rather sceptical of a dialogue with policy-makers, who are seen to prefer quantitative results. Concerned segments of the general public and especially practitioners seem to be more promising partners for a dialogue with research. For the researcher, this would require a respectful attitude towards their partners concerns and competencies as well as a realistic assessment of local constraints. Offering a kind of mirror view of actual practices as a stimulus for a reconsideration of those practices seems to be the most fruitful way to engage in a dialogue with practitioners such as counsellors and other service agents. [16]

The final chapter is called, "The potential of qualitative research: eight reminders". In it, SILVERMAN refers to his own experiences as a researcher, followed by eight sections which summarise major points from the preceding chapters:

Note that the "hyphenated" character of phenomena refers to the idea that an apparently stable phenomenon takes on different meanings depending on the context in which it is used. The book ends with the usual appendices as well as a glossary. [18]

Apart from its explicit content, it is also important to consider a writer's "style"—his way of locally organising what he wants to say. SILVERMAN clearly prefers an active voice. He argues in a personal way, takes clear stands on major issues and often refers to his own experience as a researcher. On the other hand, he also very often refers to the writings of others, summarising and quoting from their texts. Sequentially speaking, such referencing and quoting can be done in two ways, in first or second position. In the latter case, the author starts an argument in his own voice and afterwards uses a reference or a quote as a support for it. This seems to be the usual or default way. SILVERMAN, however, quite often puts other writers in first position, as in "Michael Agar (1986) has described a 'received view' of science", which is the first sentence in chapter 3 (p.65). In a similar fashion, many of the (numbered or bulleted) lists scattered in the text are "adapted" from other writers' publications. It strikes me that, at least in many of the quotes, such a textual dependence on other writers seems unnecessary, in the sense that the ideas represented in this way do not seem to be very original or special. In any way, the aspect of his style does not contribute to a fluent reading. [19]

As noted at the start of this review, SILVERMAN presents this book as oriented to an undergraduate audience. At times he does indeed directly addresses students about to work on an undergraduate assignment, sometime is a bit patronising way. Most of the time, however, the way he discusses important research topics seems equally valid for graduates and PhD students, or even researchers who want to reflect on their research practices. In other words, the book may be useful for a wider (higher?) audience than the officially intended one. On the other hand, undergrads may get impatient at some of the more extended "theoretical" discussions. Indeed, the book's over 400 pages take quite some time to read. [20]

As I noted in my overview, the chapters in the core part of the book are based on different "methods", or rather, data types, while within the chapters the treatment of those data types is often differentiated according to the theoretical and methodological approaches or "models" used. I also noted that the characterisation of those "models" seems to be adapted to local argumentative considerations, while the explanations are scattered across the "methods" chapters. For students new to these approaches, this may be quite confusing. [21]

Taking the book as a core text for an introductory undergraduate course in qualitative methods may free a teacher from quite a lot of theoretical and methodological explications, but it may be difficult to coordinate the reading work with a weekly schedule of concrete activities and discussions of practical problems. Except for the "Exercises" at the end of each chapter, the book does not give any suggestions for the organisation of such a course. A teacher, therefore, still has a lot of managerial and design work to do. [22]

In her review of the second edition, Devorah KALEKIN‑FISHMAN (2001) made some remarks that I want to take up here again, with reference to the third edition. While praising the book as a resource for teachers and advanced researchers, she suggests that it is less useful as a course book, because of what she calls its "overload": "The very comprehensiveness of the book is likely to undermine its usefulness for potential researchers—students, who are looking for plain guidance". She specifically mentions the extensive discussions of exemplary studies, which at times divert the reader's attention from the topic of the chapter in which they are cited. Although SILVERMAN has made an effort to make the third edition more relevant to students, I am afraid she would not be convinced by the effort, as this edition is still very comprehensive. Indeed, goes quite deeply into many of the exemplary studies. I have not checked the two editions systematically, but my impression is that in the present edition few exemplary studies and other quotes have been deleted and quite a lot has been added, often stemming from recent collections (co)edited by SILVERMAN (2004; SEALE, GOBO, GUBRIUM & SILVERMAN, 2004). For my taste, the additions are very useful to bring the book up-to-date, and (as I know from experience) deleting sensible parts of one's text just to limit the overall size is hard to do. [23]

A more principled objection she has concerns the chapter on the quality of qualitative research. She writes: "The chapter on validity, reliability and generalisation makes some odd proposals" (KALEKIN‑FISHMAN, 2001, para.7). And while she concedes that "there is no doubt but that qualitative research must clarify the bases for deciding what constitutes good research and what an acceptable interpretation of findings should be. This does not, however, justify sticking to problematic terminology". For her, concepts like validity and reliability carry such strong positivistic connotations that it is confusing to use them in an interpretative context. She is, apparently, unconvinced by SILVERMAN's effort to adapt these to a qualitative context. I must say that I rather liked this "adaptation" or "bending" of these conventional holy cows of methodology. [24]

Again, the third edition of Interpreting qualitative data does not provide "plain guidance", but rather offers "food for thought". So whether the book fits a teacher's preferences will depend on the latter's didactic conceptions and methodological taste, as well as his or her particular student cohort's overall intellectual level and interest. [25]

With this third edition, David SILVERMAN has again produced a high quality text. It is informative and covers a wide range of important aspects of qualitative research with a focus on data analysis. The selection of topics and the overall direction of the argument is personal and experience-based, but, at the same time, not limited to the author's own work. In my judgement, SILVERMAN covers the most important aspects of qualitative research in a way that is both challenging and inspiring. He rightly stresses that qualitative research is far from easy, but that it can be accomplished. For each case, however, a researcher has to find out how is should be done this time, in these particular circumstances. It is not a "cookbook", therefore. There are no foolproof recipes, only recommendations, or as SILVERMAN notes, "reminders". [26]

The position taken by SILVERMAN is quite marked. It is strongly influenced by ethnomethodology and conversation analysis, but he also keeps some distance from these "schools", aligning with a broader trend of "constructionism" and occasionally referring to semiotics. At the same time, he I quite critical of what he calls "emotionalism", with its focus on "experience" as voiced in interviews and focus groups. Adherents of that kind of research may not be impressed by his arguments and may not be inclined to follow his preference for "naturally occurring data". [27]

Whether this book is the best choice for an undergraduate course requires careful local consideration, but I would recommend it without hesitation to any serious graduate student or researcher about to design and execute a qualitative research project. As I said: food for thought. [28]

Denzin, Norman K. & Lincoln, Yvonna S. (Eds.) (1994). Handbook of qualitative research. Thousand Oaks: Sage. [2nd ed. 2000, 3rd ed. 2005]

Have, Paul ten (2004). Understanding qualitative research and ethnomethodology. London: Sage.

Kalekin‑Fishman, Devorah (2001, August). Review : David Silverman (2001). Interpreting qualitative data: Methods for analysing talk, text and interaction [11 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 2(3), Art. 6, http://www.qualitative-research.net/fqs-texte/3-01/3-01review-kalekin-e.htm [Date of access: June, 6, 2007].

Seale, Clive; Gobo, Giampietro; Gubrium, Jay & Silverman, David (Eds.) (2004). Qualitative research practice. London: Sage.

Silverman, David (1993). Interpreting qualitative data: Methods for analysing talk, text and interaction. London: Sage.

Silverman, David (2001) Interpreting qualitative data: Methods for analysing talk, text and interaction (second edition). London: Sage.

Silverman, David (Ed.) (2004) Qualitative research: theory, method and practice (second edition). London: Sage

Silverman, David (2005). Doing qualitative research: a practical handbook (2nd ed.). London: Sage.

Paul TEN HAVE was, until his retirement in 2002, associate professor in sociology at the University of Amsterdam, where he taught qualitative research methods. His last book is Understanding Qualitative Research and Ethnomethodology (Sage, 2004). A second edition of his previous book, Doing Conversation Analysis: A Practical Guide will be published by Sage in the fall of 2007. In previous issues of FQS Paul TEN HAVE has reviewed Conversation Analysis and Discourse Analysis: A Comparative and Critical Introduction (by WOOFFITT 2005).

Have, Paul ten (2007). Review: David Silverman (2006). Interpreting Qualitative Data: Methods for Analysing Talk, Text and Interaction (Third edition) [28 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 9(1), Art. 16, http://nbn-resolving.de/urn:nbn:de:0114-fqs0801160.

Forum Qualitative Sozialforschung / Forum: Qualitative Social Research (FQS)

ISSN 1438-5627

Creative Commons Attribution 4.0 International License