David Smilde wrote:
>
> > I thought the whole point of GT was that the research is exploring issues
> and
> > the 'questions' will emerge from the data. To start with research
> questions
> > will invalidate the inductive nature of the methodology.
>
> Andrew:
>
> *All research starts with pre-existing concepts and questions are simply a
> form of restricted, open-ended concept. Otherwise data collection would be
> impossible since the potential data that exist about the social world are
> infinite. What generally happens when you actually collect and analyze data
> is that the question(s) is/are found to be too vague, general,
> ill-specified, etc.and question-answer sequences, i.e. concepts, are
> specified in different and much greater detail. So the way you started, i.e.
> with a broad list of issues (although I bet it wasn't as broad as you
> suggest) is right on target. Whether or not you put them in question format
> is beside the point. I think you can safely refer to this approach as
> grounded theory in today's climate. Glaser and Strauss have been critiqued
> ad
> nauseum for their supposed tabula rasa empiricism. Some defend them saying
> that they were trying to make an impact in the strongly theoretical,
> deductive type of sociology done in the 50s and 60s, others say Glaser was
> the villian and Strauss had a much more sophisticated pragmatist view
> (probably likely). But I think almost noone who has actually thought about
> the issue thinks such Empiricism is a real possibility and you can just sort
> of acknowledge the criticism, debate, etc.in a footnote and call your study
> "grounded theory"--a label which I think is a very informative.
>
I think David Smilde is one of the smartest people in the business, but
I would like to say a word in favor of saving room for emergent
information.
People do start with questions, and most of the answers they find are
responsive to those questions. But they may also find intriguing,
uncategorized ideas that do not seem to be the answers to any questions
they brought. Those ideas *do* belong in a qualitative study, though in
a more pre-structured study they might seem to be tangents.
When I work with people using Atlas, I suggest that they create a code
called "interesting." The code's purpose is to mark any text that
strikes them as somehow important in a way they cannot yet justify.
Then they can go on coding with the apriori codes that they have brought
into the study with their research questions.
Later, or in moments when the reflective mood is on them, I would
suggest that they create a view - in Atlas this is done with a "code
family" that contains only the code "interesting." Then they can look
at what they marked as important for unknown reasons, and let more
detail emerge about just why it is interesting. That "more detail"
usually comes up as the marking of further, in-vivo, codes, followed by
tentative generalizations. Just what it is about the the data that was
interesting begins to clarify. This process is not so much empirical as
hermeneutic, a conversation with the text in which listening to it
rather than imposing codes on it.
The codes and themes that emerge in this way may or may not end up in
the final write-up. But going through this "i-coding" costs the
researcher almost no extra effort -- and it may reveal very important,
totally unexpected results.
Birrell Walsh
San Francisco
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|