Abstract: Numbers apparently talk. With few numbers, qualitative researchers appear to rely on examples or instances to support their analysis. Hence research reports routinely display data extracts which serve as telling instances of some claimed phenomenon. However, the use of such an evidential base rightly provokes the charge of (possible) anecdotalism, i.e. choosing just those extracts which support your argument.

I suggest that this methodological problem is best addressed by returning to those features of our theoretical roots which tend to distinguish what we do from the work of quantitative social scientists. Although SAUSSURE is most cited in linguistics and structural anthropology, he provides a simple rule that applies to us all. In a rebuke to our reliance on instances, SAUSSURE tells us "no meaning exists in a single item". Everything depends upon how single items (elements) are articulated.

One everyday activity in which the social world is articulated is through the construction of sequences. Just as participants attend to the sequential placing of interactional "events", so should social scientists. Using examples drawn from focus groups, fieldnotes and audiotapes, I argue that the identification of such sequences rather than the citing of instances should constitute a prime test for the adequacy of any claim about qualitative data.

Key words: methodology, ethnography, focus groups, language, conversation analysis

The modern thinker who instructs us to deal in sequences is, most obviously, Harvey SACKS (1992). Through SACKS's pioneering work on the structures of everyday talk in the 1960s and 1970s, the systematic method of conversation analysis (CA) was first established (see SILVERMAN, 1998). [1]

However, for the purposes of this paper, I want to show that the relevance of sequential organisation extends far beyond those who want to use CA. To achieve that end, I will refer to an earlier founding figure in social science—Ferdinand de SAUSSURE. Following that path, I will suggest that SAUSSURE's focus on the articulation of different elements lays the basis for research based on sequences rather than instances. [2]

SAUSSURE (1974) asks us to study the ways in which relationships and differences are articulated within sign systems. He rejects a substantive view of language—concerned with the correspondence between individual words and their meanings—in favour of a relational view, stressing the system of relations between words as the source of meaning. According to this view, signs are not autonomous entities but derive their meaning only from the place within a sign-system. What constitutes a linguistic sign is only its difference from other signs (so the colour red is only something which is not green, blue, orange, etc.). For instance, the status of any train arises from its place in a timetable. So, if the 10.30 from Zurich to Geneva does not leave at 11.00 it is still the 10.30 train. [3]

Signs can be put together through two main paths. First, there are combinationary possibilities (e.g. the order of a religious service or the prefixes and suffixes that can be attached to a noun—for example, "friend" can become "boyfriend", "friendship", "friendly" etc.). SAUSSURE calls these patterns of combinations syntagmatic relations. Second, there are contrastive properties (e.g. choosing one hymn rather than another in a church service; saying "yes" or "no"). Here the choice of one term necessarily excludes the other. SAUSSURE calls these mutually-exclusive relations paradigmatic oppositions. As in the railway timetable, signs derive their meaning only from their relations with and differences from other signs. [4]

SAUSSURE's argument seems mainly to have been taken to heart in the analysis of textual or visual data—perhaps because such data is almost self-evidently articulated. However, there are more widely applicable gains for data analysis in SAUSSURE's approach. Once we recognise that "no meaning resides in a single element", we need to think twice about searching data for individual instances or examples. In interpreting any instance, we cannot neglect the sequence in which it is embedded. So, for instance, analysis based on a single answer by interviewees will usually be inadequate. Thorough analysis must usually be based on an extended sequence of interviewer-interviewee talk (see RAPLEY, 2004). [5]

However, this example pushes SAUSSURE much further than he was prepared to go. SAUSSURE's concern for articulation is always at the level of structures and systems rather than interaction ("langue" not "parole"). Yet articulation occurs not only at the level of impersonal systems. As we shall see in the data extracts that follow, participants are themselves deeply attendant to relations between different activities. This is shown, for example, in the complex ways we have of delivering and receiving invitations. So, following an invitation, the inviter may treat a pause as indicating some problem and provide an "excuse" ["or perhaps you are too busy right now"] (see HERITAGE, 1984, pp.241-4). [6]

Given that we all demonstrably work with and upon sequences of actions, this means that syntagmatic relations are much more local than SAUSSURE was prepared to acknowledge. In the examples that follow, I will show how data analysis can pay close attention to the local embeddedness of interaction while drawing inspiration from SAUSSURE's emphasis on the articulation of the relation between elements. [7]

I begin with a discussion of how to analyse focus group data. I will then draw out some implications for how we might analyse other kinds of qualitative data. [8]

Phil MACNAGHTEN and Greg MYERS (2004) were interested in how the scientific debate about genetically-modified (GM) food was reflected in popular feelings about the subject. Through focus groups, they sought to elicit "the different ways people relate to animals and … the ways their beliefs and values about animals relate to implicit beliefs about what is natural" (2004, p.67). [9]

Below is one extract from their data. It begins with a leading question from the moderator:

Extract 1: MACNAGHTEN and MYERS 2004, p.75, adapted [M=moderator; X and Y=participants] [10]

For my present purposes, I am only concerned with the issue of how we might analyse such data. Very helpfully, MACNAGHTEN and MYERS discuss two different strategies based, in part, on practical contingencies. Working to a tight time-scale, MACNAGHTEN paid more attention to setting up the focus groups than to data analysis. His strategy involved the following three simple steps:

The authors note that this simple method offers a rapid way of sorting out data to bear on a particular research topic. When we begin data analysis, we may be in an unknown terrain. In this sense, MACNAGHTEN's method allows us, as they put it, to "map the woods". [12]

The kind of rapid answers that can arise through "mapping the woods" undoubtedly have an appeal to social problem oriented research. However, this method of identifying repeated themes overlooks the fact that the focus group participants are not isolated individuals but are engaged in a conversation. To understand the conversational character of the data, MYERS suggests that we need to look at how meaning gets constructed in the interactions between moderator and participants and between the participants themselves. In Extract 1, he notes:

This detailed analysis, the authors suggest, is more like "chopping up trees" than "mapping the woods". Unlike the latter approach, it rejects the assumption that there is a one-to-one link between utterances in focus groups and people's "views" on animals and GM research. Instead, it shows how: "a focus group transcript is a way of recovering, as far as is now possible, a moment-to-moment situation, and the shifting relations of people in that situation" (MACNAGHTEN & MYERS, 2004, p.75). [14]

However, like any method of data analysis, "chopping up trees" presents potential problems. First, it is clearly a much slower method than if we proceed by identifying "key passages". Second, its linguistic approach may run the risk of losing sight of the research problem with which we began. In this example, critics may justly argue that MYERS' sequential analysis has little bearing on the debate about genetically-modified food. [15]

Any qualitative researcher will recognise that the alternative approaches posed by MACNAGHTEN and MYERS exemplify two widely used (and very different) methods of analysing our data. "Chopping up trees", with its fine-grained sequential analysis, seems a more soundly-grounded research method than the scatter-gun approach of simply quoting favourable instances. However, at least "mapping the woods", whatever its limitations, tells us something about a substantive phenomenon. Can a concern with articulation and sequential organisation tell us about more than conversational structures? To obtain an initial answer to this question, I will discuss a second focus group study in which such substantive matters are more clearly highlighted. [16]

Sue WILKINSON and Celia KITZINGER (2000) were interested in the way in which both laypeople and many medical staff assume that "positive thinking" helps you cope better with cancer. They point out that most of the evidence for this belief derives from questionnaires in which people tick a box or circle a number. By contrast, WILKINSON and KITZINGER prefer to treat statements about "thinking positive" as actions and to understand their functions in particular sequences of talk. Put simply, they seek to insert "scare marks" around "positive thinking" and to examine when and how it is used. [17]

Let us look at one data extract that they use from a focus group of women with breast cancer:

In this extract, Hetty's account of feeling "devastated" by a cancer diagnosis is met by appeals to a "positive attitude" by both Yvonne and Betty. On the surface, then, Extract 2 seems to support the idea that "positive thinking" is an internal, cognitive state of people with cancer. However:

"this overlooks the extent to which these women are discussing 'thinking positive' not as a natural reaction to having cancer (the natural reaction [mentioned by Hetty] is that, 'obviously you're devastated because it's a dreadful thing'), but rather as a moral imperative: 'you've got to have a positive attitude'" (WILKINSON & KITZINGER, 2000, pp.806-7). [19]

So WILKINSON and KITZINGER's analysis suggests two different ways in which these women formulate their situation:

This distinction shows the value of looking at looking at how talk is organised and not just treating it "as providing a transparent 'window' on underlying cognitive processes" (2000, p.809). By contrast, WILKINSON and KITZINGER's focus on sequences of talk allows us to get a quite different, more processual grasp of the phenomenon. [21]

Moreover, in this version of "chopping up trees", we do not lose sight of the substantive phenomenon. Unlike questionnaire studies, which usually simply confirm lay or medical beliefs about the usefulness of certain mental responses to life-threatening illness, this research reveals that expressions of "positive thinking" may have more to do with public displays of one's moral position than with how people actually respond to their illness. [22]

Such a conclusion provides new insights of potentially great value to both patients and health workers. Moreover, this finding is simply not available from answers to questionnaires or from conventional qualitative analysis of the data which, no doubt, would find multiple instances of "positive thinking" within these women's talk. [23]

WILKINSON and KITZINGER show how close analysis of sequential organisation can be practically relevant. The next two examples, drawn from my own tape-recorded data, seek to underline this point. [24]

John MCLEOD has reminded us that "almost all counselling and psychotherapy research has been carried out from the discipline of psychology" (1994, p.190). Given the dominance of experimental and/or statistical methods favoured in psychology, one consequence has been a focus on quantitative studies which apply outcome measures to individuals. [25]

Psychological issues have also been to the fore in interview studies. Although these have sometimes been based upon open-ended questions and qualitative data analysis, the main concern has been to elicit changes in perception and knowledge. This focus upon individuals has meant that what people say has been treated as a more or less transparent window on their world(s). [26]

In designing my research on HIV-test counselling, I used a two-prong strategy to provide a different view of the phenomenon. First, instead of using measures of outcome, I chose to study how HIV counselling worked in actual counsellor-client interviews. Second, rather than looking at the client in isolation, I examined sequences of counsellor-client talk (see SILVERMAN,1997 and, for a shorter version, SILVERMAN, 2005, pp.113-9). [27]

At all the HIV-testing centres I studied, counsellors tried not to make assumptions about why people had come for an HIV-test. So pre-test counselling usually began with a question about why the client was there. Extract 3 below starts in just this way. It is at the very beginning of a counselling session held at the sexually-transmitted disease department of a hospital in a provincial British city. When asked by the counsellor (C) about why he wants an HIV-test, this male patient (P) tells a story about what happened on his girlfriend's holiday:

As you read P's account, notice how he displays massive attention to how his story of his girlfriend's holiday will be heard. "With her friend" (turn 2) tells us that his "girlfriend" had not gone away on her own, where going away on your own may be heard as implying a problem with a relationship. "Her friend" does not tell us the gender of the "friend". However we know that, if that gender had been male, it would have massive implications for the story that is being told and, therefore, P would have been obliged to tell us. Given that he doesn't, we must assume that "her friend" is "female". Moreover, we can also assume, for the same reason, that it is not a sexual relationship. [29]

But P also leaves a question hanging about why he had not accompanied his girlfriend given that "going on holiday together" can be heard as appropriate to the relationship girlfriend/boyfriend. "I didn't go because I was busy" (turn 4) attends to this question. He shows that this "not going" is accountable and provides its warrant: "because I was busy". P makes accountable that he did not accompany his girlfriend on her holiday thereby invoking the routine character of described events and, thereby, constituting his described behaviour as morally acceptable business as usual". [30]

P's account also provides a description of an event that may be heard in terms of other moral issues. "She was on holiday" (Turn 2) conjures up the category "holidaymaker" which can be heard to imply innocent enjoyment but may also be associated with other activities e.g. holiday "romances", holiday "flings". Because we know that holidays may be a time when moral inhibitions may be temporarily lifted, the upcoming description of potentially "promiscuous" behaviour is potentially downgraded or at least made comprehensible. [31]

"She's just told me (.) that she had sex with (.) a [Xian] when she was out there" (Turn 4) consists of a series of highly implicative descriptions of activities. Having "sex" with a third party implies "being unfaithful". Although the earlier description "on holiday" (confirmed by the place-locater "when she was out there") may make this description understandable, it may not make it excusable. As we shall see, P engages in considerable interpretive work to preserve the moral status of his girlfriend in a way that does not threaten his own status as a "reasonable" person. [32]

"Well not actually had sex with" (Turn 4): here the damaging description "having sex (with a third party)" is immediately repaired by B. Thus we have to suspend the implied category "unfaithful girlfriend". But this repaired description is ambiguous. For instance, are we to hear "not actually sex" as a physical or social description of the activity? [33]

"She said that this guy (0.2) this is what she told me this guy had (.) forced herself (.) hisself upon her you know" (Turn 4). It is clear from his next utterance that P is attending to this ambiguity as something in need of further explication. If "he forced ... hisself upon her", then we are given a description which implies the categories rapist/victim where "victim" implies the activity of not giving consent. [34]

So P reworks his original category "having sex", with its damaging implications, by positing the absence of consent and thus a withdrawal of the warrant of the charge "unfaithful girlfriend" and a return to a description of the events without any charge. [35]

However, there is a further nice feature embedded in P's description. It arises in its preface: "she said that this guy (0.2) this is what she told me". P's story of these events is thus doubly embedded (both in "she said" and in "this is what she told me"). How does "this is what she told me" serve to repair "she said"? [36]

We can unpick the nature of this repair by recognising that, when somebody offers an account, the upshot of which puts them in an unfavourable light, we may suspect that they have organised their description in order to put themselves in a more favourable light. So, if P had simply reported what his "girlfriend" had said about this incident, then, although he would be implying that he was a "trusting partner", he could be seen as "too trusting", i.e. as a dope. [37]

Now we see that "this is what she told me" makes him into an astute witness by drawing attention to the potential credibility problem about his girlfriend's account. However, note that, unlike this comment, P is not directly stating that his girlfriend is to be disbelieved. Rather her story is offered just as that—as her story without implying that P knows it to be true or false. [38]

The beauty of P's repair into "this is what she told me" is that it puts him in a favourable light (as an astute observer), while not making a direct charge against his girlfriend's veracity (an activity which would allow us to see him as a "disloyal partner"). This allows a hearer of his story to believe or disbelieve his girlfriend's account and permits him to go along with either conclusion. [39]

Is this the story of an unfaithful girlfriend or of someone who has been shamefully assaulted? However, we decide, P fits himself into the descriptor "loyal partner" and so is in the clear. P's elegantly crafted story leaves it up to the hearer to decide which story best describes these "events". Extract 4 shows how C chooses to hear P's account:

C elects to ignore the ambiguities brilliantly embedded in P's story. Note how her account of the upshot of what she has just heard is prefaced by "so obviously" (Turn 8). But her response is not simple-minded. By electing to hear this as a story of a rape (rather an instance of promiscuity), C is attending to the nature of her task which is, after all, to do with her client's perception of risk rather than the moral status of his sexual partner. [41]

Notice how P speedily (Turn 9) agrees with C's choice of one version of "what happened", even underlining and explaining C's version (Turn 11). Having set up his story as ambiguous, P would put himself in a difficult situation if he did not follow how C hears it. To persist in an explanation which has so obviously been rejected by C (his girlfriend's possible promiscuity), might now define himself as a disloyal partner. [42]

Through C's response, P discovers what he meant all along [his girlfriend was raped]. This reminds us that sequential organisation is not just an abstract matter dealt with by obscure social scientists but is, rather, something attended to in great detail by societal members. Members are deeply and skilfully involved in analysis of the upshot of their own and other's actions. [43]

So far, so good. But what is the practical relevance of all this? First, such attempts to understand the skills of the parties involved, as displayed in situ, provide a much more adequate basis for training practitioners than normative instruction or even role-plays (see SILVERMAN, 1997). Second, it makes us realise that supposedly reliable measures of "outcome" look problematic once we see that what "really happened" is a puzzle upon which participants themselves work in real time. [44]

My final two examples, drawn from audiotapes and fieldnotes, focus upon this feature even more directly, showing how the question "really?" may be used as a charge against the truthfulness of a participant's account. [45]

These clinics treat children born with hare lips and/or cleft palates. Cleft-palates can stop babies feeding and so are usually repaired in the first few months of life. A hare lip is treatable by routine, low-risk, cosmetic surgery usually carried out when the patient is in the teens. The rationale for delaying cosmetic surgery in the cleft-palate clinic is that, since appearance is a matter of personal judgement, it is best left until somebody is of an age when they can decide for themselves rather than be influenced by the surgeon or their parents. In practice, this reasonable assumption meant that the doctor (D) would ask the young person concerned a question in the format shown in Turn 1 of Extract 5 below taken from an English clinic:

Barry's answer was common at this clinic. Short of a later self-correction or a persuasive parental intervention (both difficult to engineer), it meant that many such patients did not get cosmetic surgery. [47]

Drawing upon evidence of this kind, I argued that questioning such young people about their looks set up the consultation as a psychological interrogation likely to lead to non-intervention. This was strengthened by the fact that, later in the consultation, it became clear that Barry, after all, did want cosmetic surgery. Barry's case and that of others showed that these adolescent patients had far less difficulty when they were simply asked whether they wanted an operation rather than being asked to assess how they felt about their appearance. So here we can immediately see a practical outcome from such detailed "chopping up trees". However, a visit to a clinic in Brisbane, Australia, provided me with the deviant case shown in Extract 6:

In one leap, Simon (S) seems to have overcome the communication difficulties that a question about your appearance usually generates in these clinics. He freely admits that he "notices" and "worries" about his looks and, consequently, would "like to get it done". What are we to make of this apparently deviant case? [49]

The first thing to report is that, at eighteen years of age, Simon is considerably older than Barry and the other children seen in my English clinic. So reticence to discuss one's appearance may be age-related and different medical strategies may be applied to different age-groups. [50]

However, there was something more interesting about Simon's case. This was how his reports about his worries were treated by doctors in his clinic. Extract 7 below is a continuation of 6:

What is going on in Extract 7? Why is Simon's apparently straightforward response subject to further questioning? To answer these questions, I noted comments made by a doctor before Simon had entered the room. These are shown in Extract 8:

We see from Extract 8 that, even before Simon enters the room, his "degree of maturity" will be an issue. We are advised that Simon's answers should not stand alone as expression of his wishes but should be judged as mature or immature and, perhaps, discarded or reinterpreted. [53]

After Simon leaves, this doctor worries some more about what Simon's answers "really" mean:

Eventually, this doctor concludes that Simon's relaxed manner is merely "a cover-up" for his self-consciousness about his appearance. Although this is rather an odd conclusion since Simon has freely admitted that he is concerned about his appearance, it generates general consent and all the doctors present agree that Simon is "motivated" and should have his operation. [55]

This deviant case considerably added to my understanding of the mechanics of decision-making in the cleft-palate clinic. The English data had suggested that asking young people about their appearance tended to set them problems which could lead away from the cosmetic surgery they might want. The Australian data showed that, even where a patient confidently reported his concern about his appearance, this created a further complication. In this case, the doctors worried about how someone so concerned could present themselves in such a confident (or "sunny") manner. [56]

A Catch-22 situation was now revealed. The doctors' practical reasoning unintentionally resulted in the following impasse:

The impasse derived from the coupling of the doctors' understandable desire to elicit their patients' own views with psychological versions of the meaning of what their patients actually said. This underlines the importance of understanding the versions that participants actually employ in their interactions and avoiding the search for a stable mental state "behind" someone's talk. [58]

Once again, by locating single utterances within a sequence of talk, we are able to see the process through which they take on meaning. It remains to explore how far this process is purely interactional. To do this, I will take one further case which, although it is from a very different setting, seems to be very similar to what we have seen in the Australian cleft-palate clinic. [59]

Around the time, I was observing the cleft-palate clinics, GUBRIUM (1988) was doing an ethnographic study of Cedarview, a U.S. residential treatment centre for emotionally disturbed children. Extract 10 below involves three boys (aged 9-10) who are talking in their dormitory room. GUBRIUM reports that he overheard this conversation from an adjacent room while reading comics with other boys.

In this extract, Gary and Bill are challenging Tom about his access to firecrackers. Compare what is said to what we have just seen in the Australian cleft-palate clinic:

Despite two very different settings and participants (a peer group and a professional-client interview) note how participants systematically search for what "really" is the case, using that term to frame questions and to provide answers. In formal terms, both extracts look like the kind of charge-rebuttal sequences that are common in courts of law. Is it appropriate to say that we are dealing with a single phenomenon which happens to be located in a variety of contexts? [62]

Yes and no. An analysis of the features of charge-rebuttal sequences is indeed a useful exercise since it can identify the various strategies available to people to make or rebut charges. However, we must not exclude the different agendas the participants bring to different contexts and the resources they can draw upon in, say, medical clinics, peer group interaction and law courts. Without this further step, our analysis runs the danger of becoming purely formalistic and, thereby, likely to lack the kind of practical relevance in which I am interested. [63]

GUBRIUM (1988) suggests how we can reframe this argument to mark out the limits of two different kinds of ethnography. Structural ethnography simply aims to understand participants' subjective meanings. It makes great use of open-ended interviews and, as such, is the most common approach. By contrast, articulative ethnography seeks to locate the formal structures of interaction. It is usually based on audio- or videotapes of naturally-occurring interaction and identifies sequential structures like charge-rebuttal sequences or preference organisation. GUBRIUM argues that, although both these kinds of ethnography answer important questions, they cannot, even in combination, define the whole of the ethnographic enterprise. To do this, we need to understand the context in which the parties generate their meanings and interactions. [64]

To meet this goal, practical ethnography recognises that members' interpretations are neither limitless nor purely formal. For example, in GUBRIUM's residential home, staff members would construct particular versions of children in different contexts e.g. a treatment review team versus a meeting with the child's family. Again, charge-rebuttal sequences may look very different in children's talk versus a clinic or courtroom. Such actions are, as GUBRIUM (1988) puts it, "organisationally embedded", i.e. different settings may provide the participants with differing meanings and interactional resources. GUBRIUM's argument is set out in Table 1 below:

The options shown in Table 1 are not so much alternatives as complementary questions which need to be answered in a particular sequence. As I shall show, this is because the main strength of qualitative research is its ability to study phenomena which are simply unavailable elsewhere. [66]

Quantitative researchers are rightly concerned to establish correlations between variables. However, while their approach can tell us a lot about inputs and outputs to some phenomenon (e.g. counselling), it has to be satisfied with a purely "operational" definition of the phenomenon and does not have the resources to describe how that phenomenon is locally constituted (see Figure 1). As a result, its contribution to social problems is necessarily lopsided and limited.

Moreover, when qualitative researchers use open-ended interviews to try to tap the perceptions of individuals, they too make unavailable the situations and contexts to which their subjects' refer (see Figure 2).

The real strength of qualitative research is that it can use naturally-occurring data to locate the interactional sequences ("how") in which participants' meanings ("what") are deployed. Having established the character of some phenomenon, it can then (but only then) move on to answer "why" questions by examining how that phenomenon is organisationally embedded (see Figure 3).

The kind of research outlined in Figure 3 can answer "why" questions by locating the circumstantial limits of usage. This paper has offered several examples of these limits: e.g. cleft-palate surgeons need to find out their patients' "true" feelings about their appearance; how HIV counsellors seek to establish their clients' risk-status and not to explore the morality of their partners' actions; the ways in which "positive thinking" is a culturally-approved way for cancer patients to define their outlook and the resources available to members of a children's group. [70]

A decade ago, in a paper with GUBRIUM (SILVERMAN & GUBRIUM, 1994), I argued that, unlike quantitative research, we delay "why" questions in favour of "what" and "how" (for later thoughts by GUBRIUM, see HOLSTEIN & GUBRIUM, 2004). In this paper, I have sought to develop that argument. Research which simply describes instances of perceptions is perhaps best left to the correlations available in survey research. By contrast, qualitative research can address the "whats" and "hows" of interaction. These "whats" and hows" are to be found by studying the local management by participants of sequences of interaction which are themselves organisationally embedded. [71]

Gubrium, Jaber (1988). Analyzing Field Reality, Qualitative Research Methods Series No.8. Newbury Park: Sage.

Holstein, James & Gubrium, Jaber (2004). Context:working it up, down and across. In Clive Seale, Giampietro Gobo, Jaber F Gubrium & David Silverman (Eds.), Qualitative Research Practice (pp.297-311). London: Sage.

Macnaghten, Phil & Myers, Greg (2004). Focus group. In Clive Seale, Giampietro Gobo, Jaber F Gubrium & David Silverman (Eds.), Qualitative Research Practice (pp.65-79). London: Sage.

Rapley, Tim (2004). Interviews. In Clive Seale, Giampietro Gobo, Jaber F Gubrium & David Silverman (Eds.), Qualitative Research Practice (pp.15-33). London: Sage.

Sacks, Harvey (1992). Lectures on Conversation, edited by Gail Jefferson with an Introduction by Emmanuel Schegloff (2 volumes). Oxford: Blackwell. Oxford.

Silverman, David (1997). Discourses of Counselling: HIV counselling as social interaction. London: Sage.

Silverman, David (1998). Harvey Sacks and Conversation Analysis, Polity Key Contemporary Thinkers Series. Cambridge: Polity Press; New York: Oxford University Press.

Silverman, David (2005). Doing Qualitative Research: A Practical Handbook (second edition). London: Sage.

Silverman, David & Gubrium, Jaber (1994). Competing strategies for analyzing the contexts of social interaction. Sociological Inquiry, 64(2), 179-88.

Wilkinson, Sue & Kitzinger, Celia (2000). Thinking differently about thinking positive: a discursive approach to cancer patients' talk. Social Science and Medicine, 50, 797-811.

David SILVERMAN is Emeritus Professor of Sociology at Goldsmiths' College and Visiting Professor in the Management Department, King's College, University of London. His interests are in qualitative methodology, medical sociology and language. He is the author, co-author or editor of 14 books including: Interpreting Qualitative Data (Second Edition, 2001), Doing Qualitative Research (Second Edition, 2005) and with Clive SEALE, Giampietro GOBO and Jaber GUBRIUM, co-editor of Qualitative Research Practice (2004).

King’s College
Work Interaction and Technology
150 Stamford Street
London SE1 9NH
England

Silverman, David (2005). Instances or Sequences? Improving the State of the Art of Qualitative Research [71 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 6(3), Art. 30, http://nbn-resolving.de/urn:nbn:de:0114-fqs0503301.

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

ISSN 1438-5627

Creative Commons Attribution 4.0 International License