_________________________________________

Routledge, 11 New Fetter Lane, London EC4P 4EE

1993, 285 pp, ISBN 0-415-05851-1 (hbk)

_________________________________________

The author wrote this book base on two basic assumptions: (1) learning how to analyze qualitative data by computer can be fun, (2) it is better to introduce qualitative analysis in the students’ terms rather than other persons’ terms (like a question of Mohammed coming to the mountain, rather than the mountain coming to Mohammed). There is a tendency that a new generation of undergraduates and postgraduates expects to handle qualitative data using the new technology. For whatever reasons, these students will not give qualitative analysis the same attention and commitment as quantitative analysis, if only the later is computer-based. This book is written primarily for them, and for other researchers that are new to qualitative analysis and the use the computer for this purpose for the first time.

Through the book Dey wants to provide a practical and unpretentious discussion of the main procedures for analyzing qualitative data by computer, with most of its examples taken from humor or everyday life. It examines ways in which computer can contribute to greater rigor and creativity, as well as greater efficiency in analysis.

The methods presented in the book assume the use of specialist software, but there is no one software application that supports the whole range of procedures in qualitative data analysis. The researcher can choose any available software for qualitative data analysis (e.g. the software packages that distributed by Renata Tesc, or Hypersoft that develops by the author)) to support a particular configuration of procedures.

To know the book more closely, let us look at the content chapter by chapter.

This chapter provides us the general line of the book that contains sixteen chapters. There are three introduction chapters (chapter 2-4) about qualitative data, qualitative analysis and computer, eleven chapters about procedures of qualitative data analysis (chapter 5–15), and the last chapter for conclusion. About the procedures themselves are drawn like the schema in page 2.

In this part the author describes the nature of qualitative data. Comprehensive comparison between qualitative and quantitative is made in order to reach deep understanding and to build strong foundation for qualitative data analysis. Important aspect that needs to be underlined here is that qualitative data deals with meanings, and quantitative data deals with number. This includes the meanings of all objects that we are researching. The meanings can be expressed through actions as well as texts (or images).

ANALYSIS

15 Producing an account

14 Corroborating evidence

13 Using maps & matrices

12 Associating &linking 12-13 Connecting categories

11 Linking Data

8-10 Categorizing data

7 Reading and annotating 10 Splitting & splicing

9 Assigning categories

6 Managing data

5 Finding a focus 8 Creating categories

DATA

The core of qualitative analysis lies on three related processes: describing phenomena, classifying it and seeing how the concepts interconnect. Dey draws these as a circular process (p. 31) to show that they interconnect each other. But because qualitative analysis is iterative process, he also represents them by iterative spiral (p. 53).

The first step in qualitative analysis is to develop thorough and comprehensive description of the phenomenon under study. Geerz (1973) and Denzin (1978) call this as ‘thick’ description. If ‘thin’ description merely states ‘facts’, a ‘thick’ description includes information about the context of an act, the intentions and meanings that organize action, and its subsequent evolution (Denzin, 1978). Thus description encompasses the contexts of action, the intentions of actor, and the process in which action is embedded.

Classification is the second process in qualitative data analysis. Without classifying the data, we have no way of knowing what it is that we are analyzing. We also cannot make meaningful comparisons between different bits of data. So, classifying the data is an integral part of the analysis. Moreover, the conceptual foundations upon which interpretation and explanation are based lay on it.

Alvin Toffler (in Coveney and Highfield, 1991) said that, we are so good at dissecting data that we often forget how to put the pieces back together again. This problem will not arise if description and classification are not ends in themselves but must serve an overriding purpose, that is to produce an account for analysis. For that purpose we need to make connections among building block of concepts of our analysis. Here the author offers graphic representation as a useful tool in analyzing concepts and their connections.

The author used this chapter to introduce computer as a main tool for qualitative data analysis. Here we can find several aspects in which a computer can help us to analyze our data. Detail applications of these aspects will be found separately in the next chapters. Fortunately, explanations that was given by the author in this chapter are not due to particular research method, but they can be applied for any research method that produce qualitative data.

The computer has provided some solutions for qualitative analyst, particularly with regard to managing and coding data efficiently. It also provides a new set of tools in the form of facilities for searching and linking data. There are two things in which computer supports us: computer enhancements and computer transformations. The former related to computer support in recording and storing data, filling and indexing data, coding and retrieving data. The later deals with searching and interrogating data, electronic links between data and auditing analysis.

Finding a focus is the first step in our analysis. It is not something we consider as afterthought, once we have embarked on our research and already produces our data. It is a process initiated in the very moment we first conceive of a research project. In finding a focus Zuka (in Praverand, 1984) suggest us to be an empty cup, not full with our opinions and speculations.

To give a direction in finding a focus, we can use the questions such as what kind of data do we want to analyze, how can we characterize this data, what are our analytic objectives, why have we selected this data, how is the data representative/exceptional, who wants to know and what do they want to know. These questions are not a sequence that should be followed in logical order. So, the researchers are free to use them, base on their intention priority. Besides that, we also can use resources such as personal experience, general culture, and academic literature to help find a focus.

Good analysis requires efficient management of one’s data. Because of that we should record data in format which facilities analysis. Here the computer plays an important role in order to realize this purpose. The computer has a capacity to locate and retrieve information that is remarkable by human standards. The computer can also improve our efficiency in managing data. Here we file information only once, and then obtain access to it as required. In interview, if we file information about different speakers then we can reference the data more economically and retrieve the full references whenever required. If using questionnaire, the questions can be filed once, and then it is sufficient to record a brief reference for the data. The full question can be displayed on the screen as required.

How well we read our data may determine how well we analyze it. Related to this, Dey gives place for ‘reading’ in qualitative data analysis. The aim of reading is to prepare the ground for analysis. Reading itself is not passive but interactive. How do we read data in interactive way? Here the author mentions some techniques: (1) the interrogative quintet by asking the question Who? What? When? Where? Why? These questions can lead all sorts directions, opening up interesting avenues to explore in the data, (2) the substantive checklist, (3) transposing data, (4) making comparisons, etc.

In qualitative research much of our data may take the form of notes. Annotating data involves making notes about the notes (p.88). To distinguish the two, Dey call the notes about notes ‘memos’. In principles, we need to record (notes) data as soon as possible.

In order to analyze our data, we must be able to identify bits of data. One way to do that is by grouping the data (the author also calls it creating categories). Here we put all the bits of data which seem similar or related into separate piles, and then compare the bits within each pile. We also can divide up the items in a pile into separate sub-piles if the data merits further differentiation.

Grouping data involves developing a set of criteria in term of which to distinguish observations as similar or related. This is done by developing a set of categories, with each category expressing a criterion (or a set of criteria) for distinguishing some observation from others, as similar or related in some particular respect(s) (p. 96). Developing categories usually involves looking forwards towards the overall results of the analysis as well as looking backwards towards the data. So, the process is one of continues refinement, until we think that we don’t need to do it more.

In practice we don’t need to separate these activities with the previous one. But to make clearer, the author considers them as distinct activities. At a practical level, this activity involves the transfer of bits of data from one context (the original data) to another (the data assigned to the category). The bits of data are not actually transferred: they are copied, and the copy filed under appropriate category (p.113). So, the process so simple: copying and filling. The computer software has been design to facilitate this task.

There are some general and specific decisions in assigning categories (we can find them in p.125). After this, we are challenged to make further decisions such as: "Should we assign other categories? Should we create a new category?

After creating and assigning categories, now we should consider ways of refining or focusing our analysis. Tesch (1990) call this process recontextualization of the data, in which we view the data in the context of our own categories rather than in its original context. In the previous process maybe we produce a (probably very large) number data bits which have been assigned to one or more of the various categories used in analysis. Therefore we can now split up it into a number of subcategories. There are some issues in splitting categories into sub categories data bits such as: Do the they make sense conceptually? Are they useful practically? Do they look useful analytically? (p.137).

Splicing is a process of identifying formal connections among categories. Here we concentrate our effort on central categories emerging from the preliminary analysis. Then, we tray to look in detail some aspects in categories such as: how are they distinguished conceptually, how do they interrelate, are they of the same status or super/subordinate, etc. (p.151).

In breaking up the data, we lose information about relationships between different parts of data. We also lose our sense of process- of how things interact or ‘hang together’. To capture this information we need to link data as well as categorize it. Again we can use the computer software to create various links (single or multiple). For the better results we need to label the links, use a list for clarity and consistency, ground links conceptually an empirically and use a limited links list to reduce complexity (p.157).

Firs of all, it is better for us to see difference between link and connection as it is shown in the scheme below.

Link

Connection

Here we use the link to establish a substantive connection between two bits of data. But in making connection we connect two categories base on our observation and experience of links and how they operate. So, the links are an empirical basis for connecting categories.

There are two ways in making connection: connecting through association and connecting with linked data. In the first one we identify correlation between categories, even in the second one we identify the nature of link between data bits.

The relationships between categories of our data are frequently very complex. To handle that problem, the author uses two diagrams: matrices and maps (p.193). The matrices are used for making comparisons across cases, and maps for representing the shape and scope of concepts and connection in the analysis (our computer will help us to do these). If we use maps, we also can give particular signs to the lines that connect the shapes. For example: length of line for type of relationship, arrows for direction of the relationships, positive and negative sign for the value of the relationships, line thickness for the empirical scope of the relationship.

Corroborating evidence is procedure in which we think critically about the quality of the data (Becker & Geer, 1982). We try to collect evidence to check the quality (include validity and reliability) of data (p. 224). Here the computer can help us to do this task. For example, it helps a bit by making it easy to look for counter evidence. Instead of retrieving only those data bits which support our analysis, we can also retrieve those which are inconsistent or contradiction. Other tasks that we do in this step are encou-raging confrontation with the data (p.234) and choosing between rival explanations.

The author starts this chapter with interesting statement: what you cannot explain to others, you don’t understand yourself" (p.237). It means that producing an account is not just something we do for the audience, but also for ourselves. Through the challenge of explaining ourselves to others, we can clarify and integrate the concepts and relationships we have identified in analysis.

The techniques of producing an account are drawing diagrams, tabulating tables and writing texts. To produce an account, we have to incorporate these disparate elements into a coherent whole. As the ultimate product of the analytic process, it provides the overall framework for our analysis.

In the last part of this chapter we can find the issue ‘generalization’. There are two aspects of generalization, inference and application. Here the author mentions that qualitative analysis often provides a better basis for inferring generalizations than for applying them.

All qualitative analysis procedures in this book are presented as a logical sequence of steps. This sequence reflects the logical relationships between different phases. But in practice we rarely proceed in a direct line from the first encounters with the data to the conclusions. Because of that, it is more realistic to imagine qualitative data analysis as a series of spirals as we loop back and forth through various phases within the broader progress of the analysis (p.264).

The author also remain us that the computer software only provides a set of procedures that can replace or facilitate the mechanical tasks involved in analyzing data, but not the creative and conceptual tasks that this requires. So, the important and crucial tasks still left on the analyst.

There are many books about qualitative data analysis, but not so much of them that use computer as a tool in order to help us in analyzing our data. Because of that, I think it is not wrong if we give a positive value for this book. ‘Abundance of data in term of texts’ and ‘workloads’ that sometimes make people allergic with qualitative data, now can be something fun, because the computer software can make our task more effective and efficient.

Procedures for qualitative data analysis that are proposed by Dey in this book are complete enough, because he give us a guidance to do all steps that we should do in qualitative data analysis (from finding a focus and managing data until producing an account). By following these procedures the researchers can finish their tasks in analyzing qualitative data. If we look the book of Miles & Huberman (Qualitative Data Analysis, 1994), we will find similar procedures, as it is shown in the next table.

The way they present the procedures is also the same, that is in the form of a logical sequence of steps, as it was described in introduction chapter. But in practice both authors prefer to employ an iterative model because qualitative data analysis tends to be an iterative process (Miles & Huberman call it Interactive Model (p.12), even Dey uses the term Iterative Spiral (p.53)).

As it is mentioned before, Dey wants to provide an accessible and practical guide for the researchers in his book. Unfortunately, it is not easy for us to find these things in the book, because the content in each chapter is not organized very well (there is no pattern). Most contents of the book are presented in narrative way. Every chapter starts with long explanation about the reason ‘why we should do this or that’. The practical steps (about ‘how to do’) are also presented in form of long narration. Because of this, it is difficult to capture the main ideas that are presented, and the impression of book is theoretical rather than practical. As a practical book I thing it is better to present the procedures in form ‘we have to do those or these’ by giving them number or bullet. To compensate these weaknesses Dey always makes summary (after long narration) in each chapter. But again, as a practical book I guess the researchers will put their hope more than this on the book (we should remember that this book is also provided for the researchers who use the computer for the first time in analyzing qualitative data).

As comparison, the book of Miles & Huberman presents very clear structure. Every method/ procedure was presented in the same pattern. It was begun with the name of method/procedure, analysis problem, brief description, illustration, until time required, so we can follow and understand the procedure easily. Each method/procedure of data display and analysis is described and illustrated in detail, with practical suggestions for user for adaptation and use.

About the language of Dey’s book is rather difficult, because (1) the author tends to present every concept in long narration, (2) most examples that are used frequently integrated in the long texts and not familiar for me (the author rarely use the examples from an education setting). So, I can say that the way he presents the contents sometimes make me rather difficult to understand the book.

Maybe the author understands that long texts can make the reader feel boring and not interesting in the book. To reduce these things he trays to present a large number of examples. Sometimes they are presented in form of texts, and the others in form of pictures, schemas or graphics. But again, it is little bit difficult to find and understand examples in form of texts because they are frequently integrating in other texts.

The quality of examples are good in general, because most of them taken from humor (created by Victoria Wood and comedian Woody Allen) and everyday life. Sometimes they can help the reader become more interest and easier to understand the book. Especially for pictures, schemas and graphics, they are clear enough and well presented. I think the experiences of the author as a researcher (had worked in a variety of qualitative methods), as a lecturer (in research methodology) and as a software developer (Hypersoft) help him in creating various examples. But especially for me (that come from an ‘eastern’ country) sometimes it is not easy to understand the humor from the western country. Besides that, as a reference (or tool) for the researchers, it is better to presents more examples that are taken from others research (such as from educational settings), because it is more applicable for the reader (or user).

In writing this book the author uses many references. A few books came from 60’s and 70’s and the rests is from 80’s and earlier 90’s. But among these references, only three books that discuss about ‘using computer for qualitative data analysis’ (Pfaffenberger (1988), Fielding (1991), Tesch (1990)). So most contents of the book create by the author himself base on his experiences as a researcher, lecturer and software developer. It can be understood because such kinds of references are still rare.

Although there are some weaknesses as I mention before, this book stills a good reference (or tool) for qualitative data analysis. It is not only addressing a particular research method (or research question), such as case study or survey, but it can used to analyze our qualitative data that is collected through any research methods. Moreover, analyzing qualitative data by using the computer that was offered in this book could be something that very promising.

Finally, for one who interesting in this book can find it in Toegepaste Onderwisjkunde (TO) Library, University of Twente, Netherlands (register number: TO 3:167d39), British Library (England), Library of Congress (USA) or can buy it through internet (http://www.amazon.com) or the publisher.

In conducting one of various types of developmental research (see articles about Developmental Research from Akker & Plomp (1993), Nieveen (1997), Richey (1997), Visscher, Gustafson & Plomp (1998)) we frequently employ research methods/approaches/ techniques (such as: survey, case study, observation, interview, etc) that produce qualitative data. We can use the procedures that discuss in Dey’s book to analyze the data, no matter what kinds of research methods/ approaches/techniques that we use to collect it. For example, we can analyze qualitative data that was collected through questionnaire survey that use open question, or qualitative data from case study about teacher creativity in geometry instruction.

Related to my research project (to develop and implement geometry curriculum for Indonesian elementary school), I think this book is very useful, especially on aspects in which computer can help me to analyze my qualitative data. In conducting the developmental research (developing and implementing several prototypes of geometry curriculum) I will do many times interviews (e.g with experts, teachers or students) and observations. All these things will produce a large number of qualitative data. By following and practicing the procedures in both Dey’s and Miles & Huberman’s book I hope I can analyze my data in the right and the good way.