In its original sense “grounded theory” stands for a methodology, research program, and method of qualitative research. In a narrower but also more practical sense the term describes a qualitative technique of text analysis comparable to qualitative content analysis and discourse analysis. Barney Glaser and Anselm Strauss introduced grounded theory as a general research program for developing substantial (specific) or formal (general) theory by starting without (too) many assumptions and working immediately with the “data” (mostly text material) – this is what the term “grounded” refers to.
Basic Elements Of Grounded Theory
Three observations are important to understand grounded theory adequately. First, several misunderstandings stem from the fact that the approach uses terms known from quantitative research in a totally different way. For instance, its authors speak of “data” not in the sense of a quantitative dataset, but in the sense of text and even any material (“all is data”). Or authors speak of coding not in the sense of measuring quantitatively but in the sense of summarizing and structuring, e.g., interview transcripts.
Second, Glaser and Strauss have walked different paths since their collaboration and their seminal book, published in 1967. Glaser still favors a strategy of strict induction and emergence – one might call this the “conservative” position of grounded theory. Strauss, however, favors a more “progressive” and practical position. Together with Juliet Corbin, he established grounded theory as a qualitative but nonetheless quite systematic method, e.g., for analyzing media coverage, interview transcripts, protocols of observations, and other material for the purpose of generating theory. Finally, grounded theory is a method of text analysis like qualitative content analysis or discourse analysis, but goes beyond these techniques since it also includes methodological aspects and a micro-sociological action theory. This becomes obvious, e.g., with the coding paradigm in the Strauss version of the approach.
As mentioned already, grounded theory aims at generating theory from data, i.e., mostly from text material such as observational protocols or transcripts of qualitative interviews.
The discovery of a grounded, data-based theory comprises concepts and relations between them (e.g., causes of social action) in a set of hypotheses for a specific sector of reality. But the so-called substantive grounded theory generated by this is a just a first step toward a formal grounded theory, the hypotheses of which claim to be general rules.
The crucial procedure of the approach is called “coding.” Here, one retrieves theoretical concepts (codes) from the text material (data) and provides examples (indicators) also from the text material. These provisional codes as well as indicators are continuously compared to the text material, which also changes as the research process progresses (theoretical sampling; see below). Concepts are classified and categorized step by step, and key concepts, called core variables (e.g., action pattern), emerge. Relations between these core variables are then expressed in hypotheses (e.g., about the causes of action). Strauss describes the principle of coding as neither inductive nor deductive but abductive, i.e., combining both perspectives in a research technique of parallel data sampling, data analysis, and theory development, with inductive and deductive steps supporting each other. This research process comes to an end when the emerging theory does not change any more (theoretical saturation).
There are three types of coding, which take turns in the process of theory development. This can be illustrated for transcripts of interviews: One starts with open coding usually, i.e., with conceptualizing transcripts line by line. In this stage of theory emergence, one conceptualizes a lot of phrases and obtains many codes, which are compared to other material, merged, renamed, and modified. Strauss recommends asking for sociological categories like role or action in order to avoid mere paraphrasing and to find substantive concepts. The second technique is called axial or substantive coding. Here, Strauss recommends using a coding paradigm that is a scheme of social action including categories like causes and consequences. For instance, one searches for phrases in the transcripts (e.g., about job problems) helping to explain the behavior (e.g., alcoholism) mentioned by the interviewee. In doing so, one generates links between concepts and tries to establish hypotheses. Concepts, indicators, and hypotheses are saved in memos, category lists, and diagrams. Computer programs like Atlas/ti were developed especially to support such memos, lists, and diagrams. While coding axially, usually one or two core categories emerge. Then one should continue with the third technique, called selective coding. Here, one looks only at those parts of the interview transcripts relevant from the perspective of the core variable. In addition, category lists and memos are ordered theoretically, concepts are related to the core variable(s), and links between concepts as well as vague concepts are finally specified. This process ends with the above-mentioned theoretical saturation.
Theoretical sampling is a key feature in coding. This term is misunderstood by many researchers who do a selective rather than a theoretical sampling though pretending to do the latter. Selective sampling means defining the criteria of qualitative sampling before one starts to analyze the sample material, which is “fixed” with the sample criteria. Theoretical sampling, however, means continuous sampling and transforming of sampling criteria during the whole research process. Whether to take this or that part of an interview transcript, for instance, depends on the stage of theory development: With open coding in the beginning all material is important at first. But when a core category has been found, new material is chosen with this category in mind.
Another core principle applied in coding is comparable to cluster analysis in quantitative research. In their joint work of 1967 Glaser and Strauss described minimizing and maximizing strategies. The minimizing strategy calls for searching for similar concepts and indicators in the material. The maximizing strategy involves the opposite. This is comparable to homogeneity within a cluster and hetereogeneity between different clusters in quantitative cluster analysis.
Two Versions Of Grounded Theory
With the coding paradigm and several examples for the three types of coding, Strauss has moved grounded theory close to qualitative techniques of text analysis. Thus, his version of the approach is “open-minded,” in the tradition of qualitative research methodology, but also tries to establish a more systematic and at least partly deductive research technique than the Glaser version. Glaser emphasizes more the inductive aspects of grounded theory. He demands that researchers avoid research literature before starting to generate the theory and thus be really “grounded” in the data instead of sticking to preconceptions.
Nevertheless, a literature review is useful at a later stage of theory development. Besides, Glaser considers talking about an emerging theory to be counterproductive before ending the research process, since supporting comments as well as disapproval would end the motivation for writing memos and for comparing and refining categories, concepts, and the emerging theory. In the Glaser version of grounded theory, theoretical memorizing – i.e., writing down all ideas about concepts and relations between them, rewriting them, and sorting them – is the key principle of creativity in developing a substantive or formal theory.
- Glaser, B. G. (1992). Basics of grounded theory analysis. Mill Valley, CA: Sociology Press.
- Glaser, B. G. (1994). More grounded theory methodology: A reader. Mill Valley, CA: Sociology Press.
- Glaser B. G., & Strauss A. (1967). Discovery of grounded theory: Strategies for qualitative research. Mill Valley, CA: Sociology Press.
- Strauss, A. L. (1987). Qualitative analysis for social scientists. Cambridge: Cambridge University Press.
- Strauss, A. L., & Corbin, J. (1990). Basics of qualitative research. Newbury Park, CA: Sage.
- Titscher, S., Meyer, M., Wodak, R., & Vetter, E. (2000). Methods of text and discourse analysis. Thousands Oaks, CA: Sage.