So here’s an entry to get me started back in the swing of writing again. Having started interviewing, recording, transcribing and coding, there doesn’t seem to be much time for any thing else. But I thought before I dived in to the data, I’d be well advised to do some preparatory reading about my options in analysing it.
But here’s some of my take home thoughts after reading “Doing Qualitative Research” by David Silverman 4th Edition 2013. I thought I’d better start off with a basic text.
The first thing the book did was to remind me how great Dr Seuss is – especially “Oh the places you’ll go!” as a call for action:
“Oh the things you can find if you don’t stay behind!”
Qualitative data lends itself to “a broadly constructionist approach which focused on how an aspect of the social world is put together by participants”(p.42) That seems to sound like what I am expecting of the interviews – to find what my participants think might be the key concepts of entrepreneurship according to their own perspectives.
Interview responses may be treated as
- giving direct access to ‘experience’ – not so keen on this as I think recall is so vulnerable an unreliable
- actively constructed narratives – this sounds more like it to me
“Constructionist researchers are interested in the practical activities in which persons are continually engaged, moment by moment, to construct, manage and sustain the sense that their social worlds exist as factual and objectively ‘out there’, apart from their own actions.” Holstein & Gubrium 2008b: 375 [p.107]
The ‘facts’ are socially constructed in particular contexts. I thought the threshold concepts might be the ‘facts’ in this instance, but I’m not so sure. I think they might be ‘facts’ which are socially constructed by me as a result of listening to and analysing the interviews. It’s me that is constructing the threshold concepts, I’m not asking my interviewers what the threshold concepts are, I’m deriving what I think they might be from what by subjects are saying. During the Delphi Study stages they might be co-constructed by me and the panel members?
Am I constructing theory – or is it already constructed – when I am attempting to identify Threshold Concepts in Entrepreneurship?
‘In particular if you are a constructionist and treat social reality as constructed in different ways in different contexts, then you cannot appeal to a single ‘phenomenon’ which all your data apparently represent “p.137
Bummer. So does this mean that by attempting to construct a list of threshold concepts (a single phenomena) I am not acting like a constructionist?
- We generalise to theoretical propositions, not to populations (p.145)
- Analytic generalisations (p.145) Robert Yin 2009
- Deductive inference [Giampietro Gobo 2007]
- extrapolation rather than generalisation (Alasuutari 1995 :156-7) each interview could be treated as a ‘case’ or ‘instance’. So again this present the problem of drawing multiple perspectives into one.
I liked this “How you conceive and resolve practical problems in always shaped by your model of how the social world works” p.199
“Interview responses are treated as actively constructed ‘narratives’ involving activities which themselves demand analysis ” p.202 Holstein and Gubrium, 1995.
My interviews are feeing more like a “prompted monologue on behalf of the interviewee” on entrepreneurship. Certainly no one is having a problem thinking of enough to say so far.
The advice is not to ask your research question directly (phew I don’t think I have – or have I?). I have asked people to talk about other people, not themselves and to talk about entrepreneurship passively, although most of them have started to relate personal experiences. It is inevitable that they start to talk about themselves surely – I think however you ask entrepreneurs to identify entrepreneurship it’s hard to not ask the research questions however indirectly – they all know what you’re trying to get at really.
“If respondents are made aware of your interests this can affect their responses” Surely this can have ethical implications? p.206. I am disclosing my interests so ethically I think I’m ok, it’s just how much this affects the responses that I need to consider. I suppose as long as I am aware of this and try and anticipate some of the ways in which this might have affected people’s responses that’s pretty much all I can do.
“Protocol design is less important if you are using a constructionist model and are, therefore, interested in how interviewers and interviewees co-construct a version of ‘reality’. In this approach, questions and answers are made a topic as well as a resource.” p.208 Rapley, 2004) So I think I could argue that I am co-constructing a reality – but not trying to appeal to a single ‘phenomenon’.
Constructionists prefer to talk about the hyphenated phenomena. eg. entrepreneurship-as-narrated-to-an-interviewer
I need to make sure I include something in my thesis about how I relate to my research participants for example as an ex-consultant working in the recruitment sector, and now an academic leading a programme purporting to teach entrepreneurship to undergraduate students.
It will be interesting also to consider how the respondents identify themselves as entrepreneurs. How comfortable are they with it as an identity for themselves? To what extent do they identify as entrepreneurs?
NB Strong’s use of Goffman’s 1974 concept of framing (p.220)
Qualitative research is at its most powerful in exploring things which are everyday and taken for granted (p.235) [at least by the respondent]
The other question which is becoming increasingly pressing for me is “How many interviews is enough?” I’m well aware that the larger your Delphi Study Panel the more quantitative validity your results have but the more interviews I do, the more transcriptions I have to do, and I am already started to feel a sense of saturation creeping in. Perhaps my panel can be larger than the number of people I interview? Or perhaps it doesn’t have to be. It depends to some extent on the numbers of drop outs I get. If I knew everyone was going to stick with me to the end it would be a lot easier. Also if I stick to my qualitative guns and don’t start to see the Delphi Technique as a quantitative approach, I think I can steer a path through this particular sticky patch.
Next question – if I’m not constructing theory in deriving the Threshold Concepts of Entrepreneurship then what exactly am I doing?
In grounded theory, theoretical sampling is used in order to flesh out the properties of a tentative category. Charmaz & Bryant 2011: 292
Constructing Grounded Theory – Kathy Charmaz, SAGE 2006
The theory is that there are such things as Threshold Concepts. So I’m not using Grounded Theory, as I am not constructing theory.
Ground Theory uses theoretical sampling to link concepts to broader theories. http://www.youtube/watch?v=fmWkf5LOmfA Professor Tony Bryant – Introduction to Grounded Theory.
Also www groundedtheory.com Strauss and Corbin
Grounded Theory – the aim is to generate hypotheses that theoretically describe the constructs that arise in every sentence of the stories told (p.239)
“Construed as active, the subject behind the respondent not only holds fact and details of experience, but in the very process of offering them up for response, constructively adds to, takes away from and transforms the facts and details. The respondent can hardly ‘spoil’ what he or she is, in effect, subjectively creating” (1995;117) Holstein and Gubrium
Mixed methods can be justified because ” you have several research questions” p.136 but triangulation doesn’t necessarily have to follow. The risk is that you under analyse one or more of your datasets. I guess this is the risk that I am sensing at the moment – that I under analyse the interview data, it certainly will take a great deal of analysing. But perhaps I don’t need to have completed all the analysis from every angle in order to derive the Threshold Concepts and get started on the Delphi Rounds – that can be going along in parallel. I am slightly (very) concerned about loosing momentum with my research participants if there’s too long a gap between interviewing them and then coming back to them with the first round of Delphi.
It also sounds like I might be adopting an hermeneutic approach.
“A hermeneutic method based on how researcher and subjects interpret the world and attempt to merge their own horizons of meaning” p.45.
I’m starting to think that the data from my interviews might be richer and merit more serious attention that I had thought. There are clearly going to be much more than a quick and dirty way of gleaning threshold concepts in order to proceed rapidly to the next stage – the Delphi Study. There is clearly a whole world of methodological choices that now face me as I get to grips with this first lots of data I am collecting. A lot of analysis to be done.
It may be that the analysis of the interview data will merit a whole section in its own right. In addition to the threshold concepts that I will be able to derive from it. And providing the first step to the Delphi Study.
“Handbook on Methodological Approaches to Handling Interview Data” Gubrium et al. 2012.
Alternatively this part of my research might be treated purely a ‘descriptive study’ rather than anything fancier like hermeneutics or discourse analysis. Although every way of looking at the interview data is potentially very interesting and I feel worth doing. I must keep focused and realistic. The data will stay there for me to come back and look at in many other different ways at a later date.
I need to read around the work of Michael Foucault – especially his discussion of the construction of subjects and disciplines. There is a potential Foucauldian Methodology in Kendall & Wickham, 1998.
CA = Conversation Analysis
DA = Discourse Analysis (Potter & Weatherell, 1987, Potter 2011)