wk 6

deefer

qual-data-analysis-using-a-dialogical-approach.pdf

Home >Applied Sciences homework help >wk 6

Data Preparation and Analysis

In: Qualitative Data Analysis Using a Dialogical Approach

By: Paul Sullivan

Pub. Date: 2012

Access Date: July 14, 2021

Publishing Company: SAGE Publications Ltd

City: London

Print ISBN: 9781849206105

Online ISBN: 9781446268391

DOI: https://dx.doi.org/10.4135/9781446268391

Print pages: 64-80

This PDF has been generated from SAGE Research Methods. Please note that the pagination of the

online version will vary from the pagination of the print book.

https://dx.doi.org/10.4135/9781446268391

Data Preparation and Analysis

In this chapter, I give some guidelines on data preparation and analysis. I do this within a broader discussion

of the value of guidelines in qualitative research. So, I first of all outline how qualitative research in general

involves a dynamic engagement with the text along both ‘bureaucratic’ (guideline-following) and ‘charismatic’

(stylistic) lines. I then examine in more detail how ‘bureaucracy’ and ‘charisma’ are principles relevant to a

dialogical approach to qualitative analysis. The examination of bureaucracy will involve outlining how the data

can be prepared, including principles of transcription and the selection of units of analysis. It will also involve

examining how the data can be analysed and the coding scheme that can be employed. The examination of

charisma will involve outlining how the analyst(s) bring their own style into the analysis.

‘Bureaucracy’ and ‘charisma’ are terms borrowed from Max Weber (1947). He uses these terms to refer to

the kinds of authority that permeate different organisations. Here, I apply these terms to qualitative analysis in

order to illustrate how the authority of analysis resides both in the bureaucratic procedures that are followed

and the charismatic style of the analyst. As we will see, both ‘bureaucracy’ and ‘charisma’ offer the researcher

different and complementary possibilities for analysing the data set.

Bureaucracy and Charisma in Qualitative Research

This section explains a bureaucratic and charismatic approach to qualitative data analysis, adapted from

Weber (1947), for our purposes here. In bureaucracy, there are the following features: (1) authority lies in

rules and procedures; (2) these can be verified; (3) it is systematic/exhaustive; and (4) value resides in the

impersonal. In the case of data analysis, this means that there is a procedure to be followed, this can be left

as an ‘audit-trail', the procedure systematically processes all of the data and the findings can be corroborated

or at least given independent value on the basis of the procedures followed.

I should note, however, that Weber argued that such ‘ideal’ types are rarely found in reality. These are the

values that underpin a bureaucracy but they may not always be realised. In qualitative analysis, for instance,

the following may ‘corrupt’ the ideal sense of the bureaucracy: (1) ungrounded but provocative interpretations;

(2) mixing the data with theory in an unpredictable way; (3) mixing the analysis with personal musings and

reminiscence; and (4) using aesthetically pleasing language to give weight to an interpretation. This latter

tendency is particular salient in many key authors in the social sciences, including Freud, Goffman and Weber.

Their style is non-systematic, unpredictable and often dependent on the personality of the analyst for its

claims.

These characteristics, which disrupt the bureaucratic approach, involve a charismatic element coming into the

analysis. According to Weber (1947), a charismatic style of organisation gets its authority from the personality

of the leader, who is often seen to have magical or exceptional properties. This kind of leadership is often

revolutionary but may subsequently become bureaucratic, as in the case of Christ and the Christian Church.

SAGE

SAGE Research Methods

Page 2 of 16 Data Preparation and Analysis

In research, the authority of the leader arises from the writing style of the analyst.

The example below illustrates this charisma. Goffman (1981) discusses Richard Nixon's treatment of a female

journalist–when he asked her to change from wearing pants to a skirt, to keep her husband happy, and the

audience laughed:

Example of a charismatic analysis

I surmise that although his audience may have laughed loudly, they may have seen his gesture as

forced, wooden and artificial; separating him from them by a behavioural veil of design and self-

consciousness. All of that would have to be understood to gain any close sense of what Nixon

was projecting, of his alignment to those present, of his footing, and I believe structural linguistics

provides us with the cues and markers through which such footings become manifest, helping us to

find our way to a structural basis for analyzing them.

(1981: 157)

Here, although Goffman recommends structural linguistics, his own style of interpreting the interaction is

characteristically charismatic. He surmises the audience's reaction although he is not aware of what it

is; he uses lovely metaphorical adjectives and tautology–‘forced, wooden and artificial’–as well as vivid

images–‘behavioural veil of design'. His unique writing style enters into his interpretations and he relates

these to the other theory that he has read. An audit trail is difficult here.

I am using the ‘bureaucratic’ and ‘charismatic’ distinction not to signify that one is better than the other, but

rather to highlight these as different tendencies in qualitative research. These tendencies are apparent in

every stage of the research. Even a transcription of the data defies total bureaucracy. Instead, the analyst

needs to make choices around what symbols to use and how relevant they are (Gee, 2005). Nonetheless,

there is a transcription scheme available, to help make it as bureaucratic and systematic as possible.

When it comes to the analysis of data itself, often there are guidelines as to what to follow to make the

analysis systematic and impersonal. In the appropriation of these guidelines however, the personal style

and charisma of the researcher(s) makes a significant difference. Madill, Jordan and Shirley (2000), for

example, have demonstrated that the personal style of the analyst plays a significant part in the analysis of

grounded theory. In their study, two different analysts followed the general procedures of grounded theory in

the analysis of the same data set. It emerged that one analyst had a ‘serialist’ style, following the process

step by step–generating themes from clusters of lower level categories. The other analyst was a ‘holist’

learner, generating higher-order categories before moving on to examine lower-level divisions within these

categories. These two styles were quite complementary, leading to a rich analysis. Their personal styles and

the guidelines worked well in the analysis of the data.

In Table 4.1, I list more generally the bureaucratic and charismatic elements of various schools of qualitative

analysis. The aim of doing this is to show how both bureaucracy and charisma are important elements of

SAGE

SAGE Research Methods

Page 3 of 16 Data Preparation and Analysis

qualitative research, in general. I do not elaborate on Table 4.1 as it would take us back into other qualitative

methodologies, an introduction to which can be found in Chapter 1. The bureaucratic and charismatic features

of qualitative methods listed in Table 4.1 are not exhaustive. It is only an indicative list.

Now that I have outlined the dynamic interplay between bureaucracy and charisma in the analytic

engagement with the text, in the next section I will examine how ‘bureaucracy’ and ‘charisma’ can underpin a

dialogical approach to qualitative analysis. For ease of organisation, I will discuss methods of data preparation

and analysis as ‘bureaucratic'. I discuss impersonal procedures that can be followed in doing an analysis.

I also discuss ‘charisma', particularly as part of a write-up, but in much less detail, as there are no real

guidelines in a charismatic approach. As we will see, however, some charismatic elements (e.g., making

choices) come into bureaucratic procedures and some bureaucracy can come into the write-up through

following guidelines. In this way, a good analysis should depend on its authority from both bureaucracy and

charisma.

Table 4.1 Bureaucratic and charismatic elements in qualitative research

Bureaucracy in a Dialogical Approach

In this section, I outline the bureaucratic procedures that can be involved with a dialogical approach. This is

SAGE

SAGE Research Methods

Page 4 of 16 Data Preparation and Analysis

further broken down into two subheadings–data preparation and data analysis. In ‘data preparation', I discuss

how the data can be selected and transcribed in a dialogical approach. In ‘data analysis', I discuss how the

data can be coded. In both these cases, I use concrete examples from empirical research to highlight the

procedures involved. It is also important to note, however, that the judgement as to whether or not these exact

procedures should be followed does depend on the charisma, aims and goals of the researcher.

Data Preparation

Qualitative data includes interviews, focus groups, diaries, images and video. Researchers are often faced

with a daunting data set after engaging in a qualitative project. How, then, should one prepare a systematic

analysis under a dialogical framework? In what follows, I will offer some thoughts on transcription, using a

concrete example. I will also describe how Anna Madill and I prepared our data for a dialogical analysis in a

recent collaboration (Madill and Sullivan, 2010).

How should the Data be Transcribed?

In this sub-section, I briefly introduce transcription notation and outline some of the dilemmas that are

associated with transcription. I use a concrete example to illustrate some of these issues.

There are some disputes in the literature (e.g., O'Connell and Kowall, 1995) as to which transcription symbols

should be used, and how to denote the features of talk that go beyond words, such as the prosodic (including

intonation and stress), paralinguistic (e.g., whether the words are said in a joking manner) and the extra

linguistic (e.g., gestures). Part of the problem is that these symbols are interpretations rather than reflective

of objective reality. Eliot Mishler (1991), for instance, has shown how the same extract can be transcribed in

different ways reflecting different purposes. Moreover, what these sounds signify (e.g., does emphasis signify

sarcasm or enthusiasm) is also a matter of interpretation and judgement.

The dialogical approach has been inspired by the literary analysis of Bakhtin. As such, it is worth briefly

speculating on what he would advise in terms of including these extra-discursive markers when transcribing

interviews and focus group data. As we saw in previous chapters, for Bakhtin (1981), the genre, the use of

direct and indirect speech and types of discourse are the key markers of emotional intonation. As well as this

he relies on the narrator/author to give the extra-discursive detail in a character's speech. We saw that in his

own analysis, he re-writes written speech with capitals to denote emphasis even though it is missing in the

text that he takes.

There is also some degree of ambivalence in Bakhtin's writings when it comes to the potential of audible

speech to reflect the emotional register. He argues that verbal speech can sometimes distort the emotional

loud and living intonation excessively monologizes discourse and cannot do justice to the other

person's voice present in it.

SAGE

SAGE Research Methods

Page 5 of 16 Data Preparation and Analysis

(Bakhtin, 1984 [1929]: 198)

I am not sure why Bakhtin believes this to be the case exactly as it is a rather cryptic comment. It is a

charismatic claim without any evidence. Perhaps he makes the claim because, when spoken out loud, our

own inner doubts may be steamrolled away by the ‘position’ we assume in relation to the other–e.g., a position

of righteousness. Possibly an emphasis on oral speech brings too much attention to the self–other dialogue at

the expense of the self–self dialogue. As Billig (1987) points out, the absence of an outer critic can sometimes

allow the ‘inner-critic’ to be foregrounded.

Either way, I disagree up to a point with Bakhtin's reading of verbal versus written speech. For the purposes of

examining subjectivity and emotion, some discursive markers may be useful guides to the emotional register.

In particular, I use the following aspects of the well-known Jefferson system, named after Gail Jefferson who

developed it (see Atkinson and Heritage, 1984, for a more complete description):

[] Square brackets mark the start and end of overlapping speech.

((swallow)) Additional comments from the transcriber in double parentheses, e.g., about features of context or

delivery.

CAPITALS Capitals mark speech that is emphatic.

() Empty parentheses signify inaudible talk.

– Underlined words signify stress in tone.

I hasten to point out that there are many other symbols in the Jefferson system, such as markers that give

in-breaths and out-breaths and markers that time pauses. However, I prefer not to use these as I find that

they detract away from the readability of a transcript and foreground attention on the sounds of the words,

for the reader, at the expense of what is being said and the format used to say them–e.g., genre, direct and

indirect speech and types of discourse. As Gee (2005) point out, the validity of an analysis does not depend

on how much detail is in the transcript, but rather on how the transcript works together with the other features

of the analysis and the goals of the study.

Even for the conventions outlined above, I prefer to use them only minimally. In this sense, the transcription

system here is what one may call ‘broad’ or less detailed as opposed to ‘narrow’ or very detailed. When

transcribing, judgements need to be made around the context of the discursive marker. For instance,

emphasis in speech may be important in some contexts but, as I exemplify in a moment, in other contexts it

may not be relevant to the argument being made or the features of the talk that are foregrounded.

Transcriptions build on each other as the significance of what is being said comes into focus. Initially,

interviews, for instance, may be transcribed without extra-discursive features. Langdridge and Hagger-

Johnson (2009) refer to this lack as ‘simple transcription'. For many researchers, this can be done by a typist.

From reading this transcript (further familiarisation), the researcher may focus on some key areas of the

interviews/focus groups and particularly important interactions. These can be given further transcription

SAGE

SAGE Research Methods

Page 6 of 16 Data Preparation and Analysis

attention. Although this iterative process may build further detail into the text, in a dialogical approach, the

judgement on whether or not to use them (e.g., mark emphasis) depends on how relevant they are to the

emotional register of what is being said.

An example may help clarify some of these issues around transcription (the same example will be used later

to illustrate data analysis). The transcribed data below is taken from an artist instructing students on how to

make glass. It is reported in McCarthy, Sullivan and Wright (2006) and Sullivan, Smith and Matusov (2009).

The artist's real name is Donna but the students’ names (Jane and Zoe) are pseudonyms. The transcription

presents a number of interesting features which may help clarify the level of appropriate transcription in a

dialogical approach:

Transcription example

Jane: can I use this ruler

Donna: sure honey, how you doing Zoe: grand

Donna: OKAY ((said in funny/joking voice))

Zoe: are we able to do uhm, like, I have rods upstairs and uhmm, will I just bring them down and we'll

leave them together and stuff and ()

Donna: sure they'll be gorgeous, yeah, wiRED up [on a panel

Zoe: just wired]

Donna: you know like some of those pictures you have, they'd be simple, bring down anything like that

that you LOVE

Zoe: o.k.

Donna: because if you love it you know, it will all happen and come together then in the piece, you know

Zoe: o.k., so you don't, so it doesn't have to be like, well I will do this () as well anyway

Donna: so like the little journey around America

Here, Donna's emphasis on ‘love’ is capitalised because her emphatic speech also reflects the value she

places on this concept. In her diaries and interviews she also talks about love and art. Putting it in capitals

brings out her investment in the word, in the broader context of her life. On the other hand, perhaps it was

a mistake to put wiRED in caps because it does not add much to the analysis. I could have used another

Jefferson convention of underlining the word to signify a rise in tone, but it was difficult to discern whether it

was a rise in tone or an emphasis. Indeed, it is hard to know why she put stress on this word in any case.

There are some other notations that are included that are worth mentioning. Putting contextual information

in brackets–e.g., (said in a funny/joking voice)–helps us to understand not only the kind of atmosphere that

Donna is feeling and/or trying to create, but also that Donna places value on this joking atmosphere. On the

SAGE

SAGE Research Methods

Page 7 of 16 Data Preparation and Analysis

other hand, perhaps I should have directly used notation to try to follow the sounds of the voice rather than

giving the contextual information of it being said in a joking voice. I definitely read it as a joking voice but there

was no laughter in the voice. Finally, I marked the overlap using square brackets which could also indicate

that there is some interruption between the speakers, but in the overall context of the extract it is difficult to

interpret what this signifies. As such, maybe I ought not to have marked the overlap.

It should also be clear that in this particular case, only some transcription conventions were used. I see this

as both a weakness and strength. It is a weakness because there may well be very important signs around

the talk that would aid my interpretation of what Donna is doing. By not transcribing these, I am ignoring

data. On the other hand, as Gee (2005) points out, even the most detailed transcription systems also miss

data. Perhaps more importantly, however, I want to do an analysis that will direct the reader towards the most

significant details of what is being said for the argument that I am making. There is a risk in transcribing other

details that the reader will consider these as very significant, because they are present in the transcription,

when in the overall context of understanding the emotional register, these may be insignificant details. This is

a judgement call that the researcher needs to make–and as such militates against the bureaucratic dimension

of a transcription convention.

Despite my preference for the bare minimum in terms of transcription convention, it is clear from discourse

and conversation analysis that detailed transcription systems are very effective in drawing attention to what

initially appears insignificant and giving it significance in subsequent interpretation. However, as we saw in

Chapter 2, this kind of discourse analysis often has quite different aims from a dialogical analysis–particularly

in terms of foregrounding the game-playing subject and/or understanding the rules that structure talk.

Key Moments

Most approaches to qualitative data involve an initial familiarisation with the entire data set. However,

reduction is necessary for an interpretation and analysis to take place. In thematic-led approaches, this tends

to be done by systematically coding all of the data from the bottom-up and then putting these in hierarchical

categories. Hierarchy is a key part of a bureaucratic approach in general.

However, another approach that I favour, is to focus on the ‘key moments’ or ‘key extracts’ from the data

set to achieve a reduction this way. The use of ‘key moments’ emerged out of a recent collaboration with

Anna Madill (Madill and Sullivan, 2010). ‘Key moments’ are an ‘utterance’ of significance. An utterance is a

significant unit of meaning, different from the sentence or the line and is defined by its readiness for a reply/

reaction. As a unit of meaning, it can be of variable length.

The Donna example above, for instance, is a ‘key moment’ in so far as it captures a significant aspect of

her interaction with Zoe and Jane (it shows the emphasis she places on love, her informal, chatty style with

the students and Zoe's questions). Later in the chapter, I do an analysis of this ‘key moment’ along with an

analysis of an interview done with another artist. This other interview extract is also a key moment but longer

than Donna's example. Again, however, it captures a significant aspect of what the experience of making art

SAGE

SAGE Research Methods

Page 8 of 16 Data Preparation and Analysis

involves for her, over a number of conversational turns.

Originally, Anna Madill and I decided to use ‘key moments’ as part of managing a large data set of transcribed

interviews with eleven different medical students (each interviewed at least twice) taking a year out to do

psychology. We were interested in their experiences of doing medicine. After going through the entire data

set, we reduced all of the interviews to forty ‘key moments'. This involved an iterative process. We first

selected a wide variety of different extracts. These were a mix of what we found interesting and what seemed

most relevant to the research question. In the second stage, the more peripheral of these were excluded on

the basis of relevance and we were left with forty ‘key moments'.

It is possible to decide on criteria for selecting parts of the data as ‘key moments'. These criteria may differ

across studies, but to give a sense of the kinds of criteria that are possible, I will outline some general criteria

that Madill and I decided on in the study of intercalating medical students.

In terms of content, what counted as a key moment, in this particular study, included the following: (1) an

anecdote around a difficult or interesting teaching or medical situation; (2) Managing interpersonal relations in

order to secure medical knowledge, such as taking patient histories and taking blood; (3) their expectations as

to what the branches of medicine would involve–e.g., psychiatry offers different possibilities for dealing with

people from surgery; and (4) their reflection on their relationships to other staff and students in the hospital.

In terms of form, a key moment was given a boundary by a move in the interview to relating: (1) an anecdote

relating to self or someone else who had a personal impact; (2) a significant belief or expectation around

medicine; and (3) a reflection on their own identity and practice in medicine. These often overlapped–e.g., an

anecdote was intertwined with a personal reflection. Generally speaking, these moves were prompted by an

interviewer question, but occasionally the interviewees spontaneously moved the interview in this direction.

Because these ‘key moments’ are given their boundaries by the content and form, they are of unequal length.

Some are a few hundred words long, while others are up to a thousand words long. Finding the precise place

where the ‘key moment’ begins was sometimes easy to identify as the participants moved into and out of

an anecdote/belief/reflection to something quite different. Sometimes, however, the interview meandered in

different directions and here the key moment had fuzzy boundaries. In these cases, the key moment begins

and ends at what we considered to be the outer periphery of the personal significance that the interviewees

attached to the moment.

Moreover, while we tried to find at least one ‘key moment’ in each interview, some participants had far more

reflections, anecdotes and experiences to relate than others and across different interviews as well. This

means that some participants are more represented than others in the list of ‘key moments', although each

participant has at least one key moment and most have more. We gave each ‘key moment’ a label to help

identify what its referent was–e.g., ‘facial injuries', ‘psychiatry', ‘dressing up'.

While these key moments were extremely useful in our study, there are no definite rules in a dialogical

approach. For some research questions it may be better to package the data into relevant ‘key moments';

SAGE

SAGE Research Methods

Page 9 of 16 Data Preparation and Analysis

for others, it may be better to count everything as relevant but still divide the data set into ‘key moments’ or

utterances; for others still, it may be better to dispense with key moments and put separate parts of the text

together under ‘themes’ and then analyse these along dialogical lines.

Sound Bites

Regardless of whether one uses ‘key moments’ or not, it will not be possible to use all of the data in the

final report or the article. This can create some difficulties. For example, the participants may use a particular

expression or put things in an unusual way that one wants to bring into the analysis. However, this interesting

nugget or ‘sound bite’ may not be in one of the quotations you selected–generally because within a longer

quotation it may lose its impact. Often, in qualitative data, this difficulty is overcome informally by saying:

‘Elsewhere, this participant also said…'.

I suggest formalising this habit in qualitative analysis. In other words, while going through the data set and

the ‘key moments', it is possible to take a separate note of all the particular expressions or ‘sound bites’ that

seem to articulate an experience in an interesting way. Below, I list some ‘sound bites’ I wrote down in my

analysis of a ‘mental health’ team who participated in a focus group discussing their place in the organisation.

The details of this study are published as Sullivan and McCarthy (2008) and explicated further in Chapter 6.

For now, these are just examples of the kinds of sound bites that can be written down:

We haven't had enough change.

Often you have people making plans who have no concept of the impact on the ground of what they

are talking about, they have no clue whatsoever.

The prospectus report was all put together without a single conversation with anyone in the health

board anywhere. If that's not detached I don't know what is.

And yet at the same time big areas seemed to be unclear and grey. And still are unclear and grey

and may be getting greyer, do you know what I mean.

I would say a lot of it is the bureaucratic system. Em, the protect your back system, the don't make

a mistake, because again nobody will take responsibility.

These sound bites are useful because they can be explicitly brought into an analysis of a broader data set

to substantiate it. We will see in Chapter 6 how ‘sound bites’ can be presented in a table. In Chapter 6, we

will examine how ‘sound bites’ from different areas can be used to create a dialogue among participants and

points of view that would not otherwise be in direct dialogue.

Data Analysis

Much of data preparation also involves data analysis in so far as one is continually making interpretations

around what to transcribe, what to focus on, what arguments to make and what to leave out.

SAGE

SAGE Research Methods

Page 10 of 16 Data Preparation and Analysis

In the bureaucratic side of the analysis, however, there are a number of further steps that are possible. One

can identify: (1) the genre and the type of discourse; (2) the affect or the emotion; (3) the time–space that is

being used (more technically–‘the chronotope’); and (4) The context of where it is happening or what is being

said. As I discussed in Chapter 2, these are all significant aspects of the data.

This way of organising the data emerged through the same collaboration with Madill (Madill and Sullivan,

2010) and it is worth reading this study for some further detail. Here, I analyse the interaction between the

artist, Donna, and her students (data also presented on p. 70 as an example of transcribing data). This

assumes that this interaction is a ‘key moment'.

In Table 4.2, Donna's interactions with Jane and Zoe are organised according to the genre and discourse, the

emotional register, the time–space elaboration of this and the context.

Table 4.2 An analysis of a key moment in the case of the artist, Donna

As we can see, Table 4.2 shows one method of coding the data. It involves outlining various aspects of

the talk, such as the genre, discourse and chronotope (or time–space). It is also possible to include short

quotations, alongside a label, to remind oneself of what exactly one was referring to with a coding label–e.g.,

‘Jane authored as “honey”'. Although not done here, if there are other key moments that are also similar to

the kind of genre and discourse evident in this particular extract, the reference to these can also be included

under ‘extract’–e.g., (Key moment 12, 49).

Below, I give another example of this kind of initial data analysis in order to illustrate the variety of contexts

in which the coding scheme can be used and the variety of ways in which the genres and chronotope can be

employed. While the data above is taken from a study of a glass-making workshop, the data below is taken

from an interview with an artist, Christine (pseudonym) who discusses her experience of different levels of

consciousness in making art. I then present the analysis of this in terms of the coding scheme above.

SAGE

SAGE Research Methods

Page 11 of 16 Data Preparation and Analysis

Extract from Interview with Christine

C: Christine;

R: Researcher

R: o.k.

C: I suppose its like medit., meditating, you're, you, you let go of that, of the surface

R: yeah

C: right, the surface world, that the

R: yeah

C: that the noise is from the radio, the [people

R: yeah]

C: walking outside the, you know when I'm picking up now

R: yeah

C: you, you sort of sink into a different conscious level where you are just making, you are just doing

it and your hand is reaching out for blue because it knows blue is the right thing to do or and its like

driv…, you know its like all those unconscious levels where you're just submerged and you're not

thinking uhm ‘why yellow'?

R: yeah

C: you, you don't, because if if you actually want pull yourself away from that

R: all right

C: Level of consciousness well I think, well, I mean, I mean suddenly as you come out of it then,

when you feel, you know like you might look at your watch or something, something jolts you like,

you say ‘ohh, I must pack up'

R: yeah

C: you suddenly have realised that you have made something or started HALF of this might have

been completed and that all those decisions were quite sort of (short pause) there was a flow going

that you can't really explain

R: yeah

SAGE

SAGE Research Methods

Page 12 of 16 Data Preparation and Analysis

C: but, there there, its not there because of uhm, its not there in a, on a TOTal intuitive level, its there

on a, a, a sort of uhm EDUCated intuitive level

This ‘key moment’ is coded like the previous one, in Table 4.3, according to genre, discourse, emotional

‘key moment’ has fuzzy boundaries and is a little longer than the extract of Donna teaching art.

The analysis in Table 4.3 gives a starting point for coding Christine's ‘key moment'. Some of these codes may

appear vague. Other descriptions may also be possible. For example, one may have included ‘exposition’

or ‘confession’ in the coding of genre. Hopefully, this shows that although there are guidelines, like other

qualitative approaches, the actual choice of codes can reflect the personal style of the analyst(s) and what

their particular interests are. In Chapter 8, I outline how, bearing this charisma in mind, it is still possible to

evaluate a dialogical approach.

Table 4.3 An analysis of a key moment in the case of the artist, Christine

So far, in this section, we have looked at the value and importance of a bureaucratic-led approach to data

analysis in dialogue. In summary, it allows one to be systematic with the data set. What this means is that

the authority for the interpretation does not depend on the magical properties of the individual making the

interpretation, but instead depends on the procedures that were employed in arriving at the interpretation.

At the same time, however, in qualitative research, following procedures rigorously is still not enough to

guarantee a good analysis. As we saw already, many individual judgements are possible. Choices around

SAGE

SAGE Research Methods

Page 13 of 16 Data Preparation and Analysis

what codes to use can be an exercise in creativity in some cases. There are also a variety of charismatic

styles that are possible in the appropriation of procedures. This has not been demonstrated here as the codes

I used are my own, but Madill, Jordan and Shirley (2000) do give an excellent example of different styles in

following procedures in qualitative analysis.

The Charismatic Side of Dialogical Research

In this section, I outline how a ‘charismatic’ side of data analysis is relevant to a dialogical approach. I do this

through using an example of Bakhtin's (1984) analysis of data.

In a dialogical approach, the strengths of both bureaucracy and charisma should intertwine. Bureaucratic

procedures and the charismatic engagement with the data may wax and wane as the analysis proceeds.

This is particularly so in the write-up. In the write-up, perhaps a vivid image may be useful here, a theoretical

reflection there, a pointed comment somewhere else. The point, here, is that writing an analysis involves

taking ownership of it through one's own style. In this way, the authority of the interpretation depends on

more than just the capacity of the analyst to rigorously follow procedures. Rigour is important but it is not the

only quality that is necessary in a qualitative analysis. Authority also lies in the charismatic capacity of the

individual to actualise procedures.

One final example may suffice to illustrate this charismatic-led interpretation, but one that intertwines with

some generic and traceable bureaucratic steps. It is taken again from Bakhtin's analysis of Devushkin in

Dostoevsky's novel ‘Poor Folk'. We examined some of this analysis in the previous chapter. Here, however,

Bakhtin does not add capitals to denote what he considers to be the internal amplification of a sound-value

in ‘copying', but rather he takes Devushkin's speech and re-writes it as a dialogue between him and an

anonymous ‘other':

Example of a Re-Written Dialogue

THE OTHER: One must know how to earn a lot of money. One shouldn't be a burden to anyone. But you are

a burden to others.

MAKAR

DEVUSHKIN: I'm not a burden to anyone. I've got my own piece of bread.

THE OTHER: But what a piece of bread it is! Today it's there, and tomorrow it's gone. And it's probably dry

one, at that!

MAKAR

DEVUSHKIN:

It is true it is a plain crust of bread, at times a dry one, but there it is, earned by my toil and put

to lawful and irreproachable use.

(Bakhtin, 1984: 210)

Bakhtin, here, re-writes a monologue as if it were a dialogue. He invents the interrogatives of THE OTHER

SAGE

SAGE Research Methods

Page 14 of 16 Data Preparation and Analysis

and thereby puts a context to Devushkin's polemical certainty. In this context, far from being certain,

Devushkin is fighting his own doubts, embodied in the mocking voice of a generalised other. Bakhtin (1984)

is unable to prove this charismatic claim as he makes up the interrogatives of THE OTHER, but it certainly

allows a provocative interpretation. Bakhtin (1984) brings to life the dialogical quality of speech through this

charismatic analysis.

In Chapter 6, I use a similar technique with the ‘sound bites'. I put various sound bites, which embody

particular viewpoints, in dialogue with each other. This can help to shed light on how the different participants

in an organisation anticipate each other's judgements and attempt to answer these back.

There is more to this kind of analysis, however, than just a charismatic writing style alone. In the example

above, for instance, Bakhtin also illustrates, and elsewhere comments on, the emotional register (Devushkin's

defiant despair), the genre (a confession), the struggle between self and other (a hidden dialogue) and the

time–space (a threshold moment). These are all generic, even impersonal features, central to his analysis

and emerge from the data set that he seeks to describe.

His analysis is charismatically led but underpinned by sound bureaucratic requirements. He uses quotations

to back up his claims, he deals with alternative interpretations of his ‘data’ and he compares it to other data

sets (other Dostoevsky novels). As such, somebody else may reveal these features of discourse as well but

perhaps not do it in such a dramatic or charismatic way. Consistent with Bakhtin's (1984) own theory, the

ideas are given life and take on colour in their unique embodiment in the speaking personality.

Discussion

I am suggesting that a dialogical analysis can be used both in a bureaucratic way–leaving an audit trail–but

with charismatic elements nonetheless–e.g., choosing extracts that reflect personal interest; judging what

transcription symbols to use; judging how the audience will react to particular directions in the analysis;

judging how to present the argument and how the data fits in with this.

In my own research, I sometimes endeavour to write bureaucratically and sometimes charismatically,

depending on the audience. Needless to say, the approach needs to meet the expected requirements of the

audience. In the case of journals, some expect a lot of quotations and an exhaustive analysis, while others

are more concerned with the theoretical issues at stake and a few exemplars may suffice here.

Nonetheless, there is a lot of scope for the researcher to combine these approaches and use them

flexibly. There is a lot of strength in dynamically moving between bureaucratically-led and charismatically-led

approaches. Perhaps in the earlier stages, a bureaucratic-led approach may dominate, while later, in writing-

up, a charismatic-led approach may dominate or even intertwine with the bureaucratic approach.

In the next chapter, I will bring further attention to the interplay between the ‘bureaucratic’ and ‘charismatic’

approaches. I do this by giving some general guidelines on how to write up a dialogical analysis. These

guidelines are exemplified through an analysis of three artists making art.

SAGE

SAGE Research Methods

Page 15 of 16 Data Preparation and Analysis