Facing up to issues in grounded theory research --preconceived frameworks and verification

"The very attraction of grounded theory may lie in the way it obliges us--because of its commitment to theory--to face up to some fairly basic issues about the nature of social research. If we accept the elementary (but awkward) principle that to do research requires reflection on what we are doing and how we do it, at the very least we should try to confront and clarify these issues." Grounding Grounded Theory (1999), Ian Dey p. 24

This is the first of a number of posts I will make discussing grounded theory and my own interpretation of the methodology. As Dey points out, there are a "plurality of 'interpretations'" of grounded theory that seem to fall into three camps: 1) the orthodoxy of Glaser, 2) the safe schemas of Strauss and Corbin, or 3) the doctrine of dimensionality of Schatzman. I have increasingly developed a sense of what grounded theory is through reading, but as I do more reading in preparation actually to perform the methods of grounded theory analysis, I feel that I need to coalesce some of my thinking. That's what these posts will be, and I anticipate that they will continue through the researching process.

I want to talk about two questions that Dey raises at the end of his first chapter in the book cited above. Here's the first one:

--How much scope does grounded theory allow for adopting preconceived frameworks as an aid to analysis?

This question is, of course, the source of the big break between Glaser and Strauss (and Corbin). Glaser adheres to the principle that the researcher should lay aside all preconceptions and theories as they analyze data. This question gets at the scope that will be allowed for adopting preconceived conceptual frameworks. This doesn't necessarily mean a particular theory but simply the use of ANY frameworks ahead of time for understanding data.

It seems to me that the question is moot since it is impossible NOT to bring conceptual frameworks and even theories with us as we examine data. I suppose we could get too worked up over this point, but I very much have hinged my study on faulting previous theory building for creating that theory through deductive analysis of data based on outside theories. Theories to understand data. I want to go the other way--data to form a theory to understand the data. Thus, I think I need to side more in the camp with Glaer. One comment made in chapter 2 about categories seems relevant here: "Glaser and Strauss (1967, pp. 240-41) present categories as 'sensitizing' concepts that related meaningfully to the realities of interaction (as perceived by participants)" (28). I think the last part is significant. I need to key into as much as possible how my participants are seeing this interaction and representing it. The important thing is not my interpretation of the interaction, but their interpretation (or my interpretation of their interpretation). The data itself will indicate the concepts that will emerge.

Strauss and Corbin developed a particular method for "integrative coding" rather than allowing for broader possibilities for coding. Strauss and Corbin's "coding paradigm" followed conditions, context, action/interaction strategies, and consequences. This is what is referred to as "axiel coding." They believed the use of this coding paradigm "allowed data to be related systematically in complex ways" (Dey 11). I look at this paradigm and I am immediately reminded of Aristotle's Topoi. The topics from which the rhetor selects his or her arguments. I, personally, don't see a huge problem with applying these analytical heuristics to the data to see what they reveal. It seems that Strauss and Corbin are creating a more detailed procedure to guide coding data and analyzing it. Glaser would be more open an improvisational in his coding.

A bigger question regards how much will I bring in theories and preconceived ideas regarding rhetorical reflection into my analysis. Strauss and Corbin allow a number of different options for possibly using theory. My stance, I think, will be this:
--define my preconceived theories, ideas, and biases
--set them aside
--code with an open mind
--see what happens

I think in my initial pilot coding I will try to be "objective" and let the data speak. When I have what I hear the data saying I may then turn to my theory to give it a name or an explanation. I'll just have to feel out what I will do. From my reading of Strauss and Corbin, I believe they adopt this approach which is flexible and allows the use of theory if it seem appropriate. My thinking right now may reflect that I have read more from Strauss and Corbin than Glaser (other than the 1967 book). I have ordered three Glaser books to look at, so I may learn more about his approach and change my mind. For now, Strauss and Corbin seem to outline a better procedure to follow for doing the coding successfully (and systematically).

Second question:
What place has verification in grounded theory?

This question has significance for me because the current research design I will follow involves grounded theory analysis to develop a coding instrument and then a test to see how well this coding instrument is useful for analysis through a large scale content/rhetorical analysis of student reflective texts.

So I would generate and then verify--all in one dissertation study.

The issue of verification of theory for grounded theory is complex. Glaser's position is that verification has no place in grounded theory--it is a different methodology. Grounded theory takes two positions toward verification. One position is that verification is left to others researchers at a different time. Grounded theory's job is not verification, so it would go against this purpose of grounded theory to have a study that involved both generation and verification of theory. The second position of grounded theory regarding verification is that the theory generated from grounded theory contains implicit verification within it because it fits the data. There is no need to check the theory because these ideas are induced from the data--to verify the theory we need only look at the data. I believe I am oversimplifying things--as Dey mentions there is "ambivalence in grounded theory about the status of discovery and verification" (38). It is clear to me, as well, that "verification" has two different meanings in these two positions.

I am reluctant to make my own study a two part study where I generate and then verify. Not only would it possibly make my study very large and time consuming, but it would violate some of the principles of grounded theory methodology (it seems to me). Also, I conceive of my study as the initial stage of theory building where a more homogeneous set of data is examined. The second stage is to maximize differences in one's theoretical sampling to see how the theory becomes refined.