Session 3: Tuesday 2nd June11:00-13:00, room 731     

Grounded Theory

Professor Paul Dowling

The initial phase of the research by Sanderson, discussed in the previous session employed qualitative analysis of focus group discussions to generate 70 statements reflecting attitudes to dance. In the 2008 paper, Sanderson indicates that her analysis was influenced by the work of Glaser and Strauss (1967—she cites a later edition) and Strauss and Corbin (1990; 1998 edn listed here). She makes no explicit reference to Grounded Theory (GT)—the subject of both of these texts—but refers to a central feature of this approach, which is to maintain that theory should emerge from the data rather than being imposed upon it. It is perhaps interesting that no mention is made of the GT texts in the 2000 paper, which is where the analysis leading to the construction of the dance attitude scales is presented. In fact, the analysis of focus group discussions presented in this first paper seems to have been limited, largely, to the identification of groups of statements, which were each represented by one of their number. This might be associated with the open coding phase of GT analysis, but it hardly constitutes theory. On the other hand, the procedure of exploratory factor analysis (EFA) does lead to a conceptualization of the data in terms of the factors that are identified and this is theory. It is probably appropriate to describe the EFA procedure as a transaction between three interactants: the data; a statistical technique; and a sensitivity on the part of the researcher that enables a decision to be taken on appropriate rotation. Surprisingly, perhaps, the application by Jerome McGann (2001) of random Photoshop mutations to a digital copy of Dante Gabriel Rossetti’s The Blessed Demozel in order to reveal hidden structure might be described in the same way: exactly which mutations generate structure worthy of note is in the eye of the mutator. EFA is often described as a technique that enables the revealing of hidden structure, but the technique itself does not tell the analyst when to stop the rotation, or even whether to use orthogonal or oblique rotations; these are decisions to be made by the analyst employing what I am referring to as abductive or pragmatic logic.

I want to describe GT in the same kind of way. The researcher brings to the research setting some sampling, data collection, and data analysis techniques and also what Glaser (1978) describes as theoretical sensitivity. The latter comprises, say, the embodiment of the researcher’s education in the relevant academic field (and, one might say, beyond) and what they have gained from their own reading and previous research. The techniques include: theoretical sampling—sampling decisions being made on the basis of preliminary analysis of data already collected; open coding; conceptualization in the form of a core category; selective coding of data that relates to the core category; memoing (the writing of memos that describe and illustrate codes); constant comparison (between the emerging categories and theory and the data); and theoretical saturation—continuing the process of data collection and analysis until no further developments in the codes/theory emerge.

The advocates of classic GT (cGT)—their principal being Barney Glaser himself—maintain that the method is most appropriately conducted by a researcher who has been mentored in cGT. This, if you think about it, establishes a direct lineage from Barney Glaser—the originator of cGT—to all ‘legitimate’ cGT researchers: a potent strategy of purification (see Dowling, in press). The method—Glaser describes cGT as a distinct method in social research—incorporates some other principles that non-cGT researchers may find surprising. The most familiar of these is the interdiction on the production of a preliminary literature review. The basis for this rule is twofold. Firstly, reviewing extant research on one’s research interest may lead to preconceptualisation, which is a cardinal sin in cGT: theory emerges from the data and is not forced upon it. My own view is that the recognition of theoretical sensitivity also goes some way to admitting preconceptualisation, but that, in any case, we are trained to adopt a critical attitude towards research and so should be able to avoid being unduly or inappropriately biased by our reading of research proximal to our own. The second reason for avoiding the preliminary literature review derives from the principle that the research interest in a cGT study should be determined by that which engages the participants in the research setting and not pre-determined by the researcher: you don’t know in any detail what your research interest is until you get to the field and begin exploring, opening interviews with a ‘grand tour question’—“how’s it going?”, not “can you give me an example of how you have dealt with challenging student behavior in your classroom.” I have some sympathy with this. Certainly ‘let the data speak’—a firm principle of cGT—might reasonably be extended to ‘let the setting speak’. cGT, however, deduces that, if you don’t know what your research question is before you’ve started collecting data, then any preliminary literature review may be focused on what will turn out to be the wrong research question and so your time spent in producing it will have been wasted. My response to this is that the literature review need not be focused so tightly on a specific research question, but rather on the setting more generally. A former  (and successful) doctoral student of mine wanted to research fan communities online, so I initially directed her to make herself an expert on the extant research on online communities and on fan communities, which she did and which enabled her to visualize the potential originality of her study from an early stage and certainly without prejudicing its focus.

It is undoubtedly the case that GT has been the single most influential approach in qualitative educational research, if not social research more generally. A great deal of research makes use of its approach to coding and of constant comparison, for example, without mentioning—perhaps without even recognizing—the origins of these strategies. Very little of this corpus would be recognized, I think, by Barney Glaser as legitimate cGT. This does not matter. What the researcher has to do is make claims about their setting and present an argument in support of these claims. Both the claims and the argument are constructed in the research process, drawing on methodological, theoretical and empirical resources and also, perhaps, the charismatic, compositional facility of the researcher to render a persuasive text. If GT discourse provides resources that can be recruited in this project, then they should be recruited without the necessity of being bound in the nutshell of cGT itself (which will only give you bad dreams!).

GT first appeared as the method deployed in the study, Awareness of Dying by Glaser and Strauss (1965) and was subsequently formulated as their methodological text, The Discovery of Grounded Theory: Strategies for qualitative research (1967). Thereafter, Glaser and Strauss took different paths in developing GT, Glaser continuing to elaborate and develop what he refers to as classic GT and Strauss teaming up with Juliet Corbin (1998 edn) to develop its techniques in ways that Glaser does not always fine consistent with the original idea. Cathy Charmaz and others (see, for example, Nathaniel’s chapter in Martin and Gynnild, 2011) have adopted a more philosophical interpretation of GT: I’m afraid that I tend to side with Glaser here when he declares GT to be a philosophy-free zone.

The paper by Tina Johnston is published in The Grounded Theory Review, which is an open access journal published by Glaser’s publishing house and intended for work of and related to cGT. In reading the paper, it is important to remember that the validity and reliability of the ‘pushing for privileged access’ theory are affirmed by the researcher’s deployment of cGT procedures and in particular, theoretical sampling, constant comparison, and theoretical saturation, and not by the inclusion of large amounts of data in the paper itself. Qualitative research is sometimes criticized for its presumed selective presentation of data, but the point of presenting data in GT is to apprentice the reader into the theory and not to underwrite its validity and reliability; clearly, it is appropriate to choose the best examples for this purpose. Johnston’s theory of parental advocacy, which has emerged from her study, is an example of a substantive GT, which is to say, it is a theory that relates to the particular setting in which she was working. At the end of the paper, however, she considers other contexts in which advocacy might be a visible phenomenon including, for example, sports activities, medical patient care, care of the elderly, and so forth. Johnston’s substantive theory cannot simply be transferred to these other contexts. It can, however, provide the starting point for the development of a formal GT. The procedure for developing formal GT differs from that for a substantive GT in that it generally will not involve the collection of primary data, but rather exploring the extant research on the various other contexts and developing the theory in that way.

It is worth comparing the study by Johnston with that by Kunkwenzu and Reddy (2009) on teacher socialization in Malawi. Both studies were conducted for the purposes of a doctoral thesis (Reddy was Kunkwenzu’s supervisor). The latter study is considerably smaller in scale, but the advantage of the paper is that it does go into some detail in respect of the approach that was adopted to data analysis. I would point out that the data collection and analysis in this latter piece of research does tend to concentrate on problems and coping strategies of beginning teachers and seems to ignore any opportunities that may have arisen in their teaching lives.

A final point to make—which, of course, may not be necessary: ‘grounded theory’ is not itself helpfully thought of as a theory as such, rather, it is a methodological approach or a distinct method (if you follow Glaser); the theory is what its deployment produces.

Preliminary Reading

JOHNSTON, T. L. (2008). ‘Pushing for Privileged Passage: A grounded theory of guardians to middle level mathematics students.’ The Grounded Theory Review. 7(1): 43-60.

Further Reading

CRESWELL, J. (2012). Qualitative Inquiry and Research Design: choosing among five approaches. Third Edition. Thousand Oaks. Sage. Sections on grounded theory and also Appendix D.
DOWLING, P.C. (2012). Being Barney Glaser. Grounded Theory Review. 11(2). no page numbers. [This piece is an essay review of the Martin & Gynnild collection (see below).
DOWLING, P.C. & BROWN, A.J. (2010). Doing Research/Reading Research: re-interrogating education. London. Routledge. c. 7.
GLASER, B.G. (1978). Theoretical Sensitivity:Advances in the methodology of grounded theory. Mill Valley. Sociology Press.
GLASER, B. G. (1992). Basics of Grounded Theory Analysis: Emergence versus forcing. Mill Valley. Sociology Press.
GLASER, B.G. & STRAUSS, A.L. (1965). Awareness of Dying. Chicago IL. Aldine Publishing.
GLASER, B.G. & STRAUSS, A.L. (1967). The discovery of Grounded Theory: strategies for qualitative research. New York. Aldine Publishing Company.
KUNKWENZU, Esthery Dembo & REDDY, Chris. (2009). Using grounded theory to understand teacher socialisation: a research experience. Education as Change. 12(1). 133-149.
MARTIN, V.B & GYNNILD, A. (Eds). (2011). Grounded Theory: the philosophy, method, and work of Barney Glaser. Boca Raton. Brown Walker Press.
McGANN, J. (2001). Radiant Textuality: Literature after the world wide web. New York, Palgrave.
STRAUSS, A. & J. CORBIN (1998). Basics of Qualitative Research: Second Edition: Techniques for Developing Grounded Theory. London, Sage.