Melissa Hardy & Alan Bryman (Eds.) (2004). Handbook of Data Analysis. London, Thousand Oaks, and New Delhi: Sage Publications, xviii + 704 pages, ISBN 0761966528 (hardback), $130.00, £85.00, EUR 141,90

Abstract: Editors HARDY and BRYMAN present qualitative and quantitative analysis methods in the same book. The individual chapters tend to deal with one or the other approach, though, and there may have been some opportunities missed in terms of discussing both together. The chapters are of high quality and present the current state of knowledge in a wide range of methods. Researchers and students will find the book useful and informative, but this is likely to be a book for libraries rather than individuals to purchase.

The title of this book is ambitious in its lack of qualifiers—this is a handbook, not of qualitative or quantitative analysis, not of discourse or statistical or social-network or ethnographic analysis, but of "data analysis". The ambition is not unfounded: the book brings together discussion of the current state of a wide range of approaches to and methods of (social) data analysis. [1]

In their introductory chapter the book's editors, Melissa HARDY and Alan BRYMAN, recall their "early decision to include both 'quantitative' and 'qualitative' techniques in a single volume" (p.1). They discuss the similarities between quantitative and qualitative data analysis under a series of subheadings, recognising that both are concerned with data reduction, with answering research questions, with relating data analysis to the research literature, and with variation; both treat frequency as a springboard for analysis, both seek to ensure that deliberate distortion does not occur, both argue the importance of transparency, and both must address the question of error. HARDY and BRYMAN argue that it does not particularly matter whether a given piece of research is classed as qualitative or quantitative, as long as the procedures are systematic and the conclusions sound. They suggest that over-specialisation in the social and behavioural sciences can be a barrier to useful intellectual exchanges, whether that specialisation relates to disciplines and fields or to analytic techniques. [2]

HARDY and BRYMAN's introductory essay is argued cogently but, despite the line that they take, the articles in their book mostly fall on one side or other of the qualitative-quantitative divide. Topics that could have been usefully related to both qualitative and quantitative approaches tend to relate to one or the other rather than both. The editors point out that "quantitative" researchers must make judgements about the boundaries of classifications and that "qualitative" researchers do not all avoid statistical approaches. They ask, "Does the fact that a researcher calculates a correlation coefficient or bases a conclusion on differences in the counts of events suddenly toss the research into the quantitative camp?" (p.2). This is a valid point, but in this book the discourses employed in different chapters make it clear whether qualitative or quantitative approaches are being dealt with. [3]

The book is divided into five parts: Part I, Foundations; Part II, The General Linear Model and Extensions; Part III, Longitudinal Models; Part IV, New Developments in Modeling; Part V, Analysing Qualitative Data. As the word "model" suggests, the middle three of these five parts are all focused on quantitative analysis. The first part is mixed, with two articles on qualitative and four on quantitative topics; the final part is qualitative. Almost all chapters are about quantitative or qualitative analysis but not about both. For example, Chapter 4, "Inference", by Lawrence HAZELRIGG, starts out with a general discussion of inference as "a process of arriving at a conclusion" (p.65) but proceeds to a discussion focused exclusively on statistical inference. A related chapter, Chapter 21, "Causal Inference in Sociological Studies" by Christopher WINSHIP and Michael SOBEL, begins with an overview of philosophical theories of causality and ends with a list of the contributions to empirical research that can be made by "the counterfactual model of causal inference", e.g., "matching provides a powerful nonparametric alternative to regression" (p.499). Inference, at least in the terms of arriving at a conclusion, is of relevance to non-numerically as well as numerically based forms of data analysis, but only the second type is dealt with here. Both of these articles are clear and well argued, but both are clearly unrelated to qualitative analysis. [4]

Following the editors' introductory remarks, this lack of crossover between discussion of qualitative and quantitative studies is disappointing. For example, in Chapter 6, "Feminist Issues in Data Analysis," Mary MAYNARD deals only with qualitative forms of analysis. MAYNARD explains that "a pragmatic approach towards mixing methods and even developing quantitative ones has developed [… but] qualitative studies still remain the preferred research approach for many feminists, with the consequence that it is in this area that they have most to say about data analysis" (p.133). The association between feminism and qualitative methods has been observed elsewhere (PLATT & HOPPER 1997, p.285), and MAYNARD cites, but does not discuss, work by Ann OAKLEY on feminist research that could have opened out this discussion to include both qualitative and quantitative approaches. This chapter is excellent as an overview of feminist issues in qualitative data analysis, but in the context of this book I would have been pleased to see some ideas developed about what a feminist approach to quantitative analysis might involve. [5]

All discussion of a "quantitative-qualitative divide", or even of bridging this divide, tends to reinforce the perception that there is a divide. It seems likely that if we did not tell our students that there was a divide, presenting them with forms of analysis as equal but different rather than (perhaps unintentionally) socialising them into being qualitative or quantitative researchers, then we would all make a lot more progress. Since that is not likely to happen (and I acknowledge the irony of making such a comment in the pages of a journal dedicated to qualitative journal), we must explore ways and look for means by which qualitative and quantitative analysis can proceed hand-in-hand rather than moving along separate paths. An example of this is discussed in another review in this issue of FQS, in Leen BEYERS' review of Jane ELLIOTT's new book. While remaining appropriately critical, we must welcome such efforts. [6]

Throughout the book the writing and presentation are excellent. About ten of the book's thirty chapters deal with qualitative analysis and are likely to be of greatest interest to FQS readers. For reference, these articles are, by topic and authors: on tools, of all types, used in qualitative analysis, by Raymond M. LEE and Nigel G. FIELDING; on content analysis, and particularly on analyzing large amounts of textual data, by Roberto P. FRANZOSI; on semiotics and data analysis, by Peter K. MANNING; on conversation analysis, by Steven E. CLAYMAN and Virginia Teas GILL; on discourse analysis, by Jonathan POTTER; on grounded theory, by Nick PIDGEON and Karen HENWOOD; on narrative in social science research, by Barbara CZARNIAWSKA; on the post-modern turn in qualitative research, by Sara DELAMONT and Paul ATKINSON; on historical analysis, by Dennis SMITH; and, as mentioned, on feminist issues in data analysis, by Mary MAYNARD. Those interested in these methods may recognise that the authors are all prominent in their fields; the book is the better for having contributors of such good pedigree. [7]

Of the qualitatively oriented chapters, two stood out for me. The one I found most insightful was Chapter 22, "Tools for Qualitative Data Analysis" by Raymond M. LEE and Nigel G. FIELDING. LEE and FIELDING have written on the use of computers in qualitative analysis and summarise some of that work here. They do so in the context of describing the use of three different types of tools in the conduct of qualitative research: literal tools, which are physical devices such as computers; conceptual tools, which are methodological or theoretical approaches used in research and which range from simple techniques to developed theories; and operative tools, which are procedures such as those used in handling and coding data. This article provides a very useful overview of the tools, computer-related and otherwise, currently available to qualitative researchers. For more on using technological tools in qualitative research interested readers might like to consult FQS 3(2). [8]

Another contribution worth reading is Chapter 7, "Historical Analysis" by Dennis SMITH. SMITH brings together a number of approaches, including ideas on time and periodisation put forward by ISAAC and GRIFFITHS, PADGETT and ANSELL's work on historical social networks, FRANZOSI's work on semantic grammars and their analysis, RAGIN's work on qualitative comparative analysis, and Barrington MOORE's comparative analyses. (Of these, network analysis is covered in Chapter 22 of this book by BREIGER, and FRANZOSI has a chapter of his own on content analysis). SMITH draws on the work of these writers to discuss techniques useful in collecting and manipulating historical data, placing this discussion in the broader context of historical inquiry. Drawing on MOORE, he makes an important point: issues of methodology are, or at least should be, secondary to the questions that began the research process. SMITH's piece feels like one in which the author has run out of space to say all that he would like to, but it does well in covering a number of stimulating approaches to the topic. [9]

It is notable that this book contains no chapters on, for example, analysis of ethnographic or visual data. This is perhaps because such data tend to involve analysis that is less formal and more idiosyncratic, and hence require approaches that may not fit with the editors' statement that "regardless of the form of the original data, analysis requires researchers to develop coding schemes, classification protocols, definitional rules, and procedures for ensuring reliability in the application of all of these tools" (p.vii). [10]

An issue not raised in the book relates to the connection between different aspects of the research process. Qualitative researchers, far more than their quantitative colleagues, tend to approach and conduct their research as a holistic process in which data collection, handling, analysis and dissemination all play parts. Quantitative researchers often conduct secondary data analysis removed from the collection process, but qualitative researchers seldom do so. Recent activities that encourage the secondary analysis of qualitative data, such as the guest-edited issues of FQS (see http://www.qualitative-research.net/fqs/fqs-e/inhalt1-05-e.htm, http://www.qualitative-research.net/fqs/fqs-e/inhalt2-05-e.htm and http://www.qualitative-research.net/fqs/fqs-e/inhalt3-00-e.htm), reinforce rather than contradict this assertion: quantitative researchers need no such prompting to conduct secondary analysis. This difference between typical qualitative and quantitative practice is likely to make a difference in terms of how researchers analyse data, but this difference is not discussed here. [11]

Those who approach this book with an interest in qualitative methods of analysis will find useful and stimulating articles, and editing and writing which are both of a high standard. Overall, the content of this volume emphasises that there remain differences between qualitative and quantitative analysis. These differences, despite HARDY and BRYMAN's comments about the similarities between them, remain to be explored and developed. Buying "Handbook of Data Analysis" is likely to be worthwhile only if you are interested in both quantitative and qualitative methods (and only if it comes out in a cheaper paperback edition); it's more likely to be the kind of thing you'd should ask your friendly local librarian to stock for you. [12]

Platt, Jennifer & Hopper, Simon (1987). Fragmentation, Social Theory, and Feminism: Sociology in the United Kingdom. Contemporary Sociology, 26(3), 283-285.

Iain LANG is an Honorary Research Fellow in the Public Health & Epidemiology Group in the Peninsula Medical School in Exeter, UK, and an NHS Academic Specialist Trainee in Public Health. His research interests are in health & ageing, volunteering and care-giving, social network analysis, and social research methods. He is on the editorial team of FQS and has previously contributed two reviews: Readme First for a User's Guide to Qualitative Methods (by MORSE & RICHARDS 2002) and Stretching" EXERCISES for Qualitative Researchers (JANESICK 2004).

Public Health & Epidemiology Group
Peninsula Medical School
Barrack Road
Exeter, UK. EX2 5DW

Lang, Iain (2006). Review: Melissa Hardy & Alan Bryman (Eds.) (2004). Handbook of Data Analysis [12 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 7(2), Art. 20, http://nbn-resolving.de/urn:nbn:de:0114-fqs0602200.

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

ISSN 1438-5627

Creative Commons Attribution 4.0 International License