Glaser, Barney G. and Anselm L. Strauss. 1967. The Discovery of Grounded Theory: Strategies for Qualitative Research. Chicago: Aldine.
Looks at how theory can be generated by empirical data, instead of data pursued by theory. Grounded theory based in comparative analysis. Sociology aims to advance theory, predict and explain behavior, be applicable to scholar and laymen alike, and to guide a style of research on areas of behavior. “Theory based on data can usually not be completely refuted by more data or replaced by another theory” (4), as it is too closely related to the data – therefore, theories made here are often long-lasting and responsive to social change. GT is tangible, applicable to everyday life through statistics, experiences, etc. Generating theory in GT is derived from doing research in the area, instead of a priori assumptions that may not be representative of the current social phenomena. At time of publication, G+S complained that many sociologists were testing out “great man” theories (Durkheim, Marx, etc.), but not questioning the theories themselves.
**** “The sociologist with theoretical generation as his major aim need not know the concrete situation better than the people involved in it (an impossible task anyway)” (30). “Theory as process” (32) – theory as ever-developing, never complete or perfect – works to understand realities of social interaction and its place within structures.
Developing substantive (empirical) or formal (conceptual) theory at middle-range levels. In this, forming multiple theories, instead of grand ones promoted by logico-deductive traditions.
Elements of a theory – things indicated by the data, NOT data itself!
- Categories – standalone conceptual element of the theory (sub-theory)
- Properties – conceptual element of a category (sub-sub-theory)
Notes – categories can be borrowed from pre-existing theory, but this limits the outcomes of emergent ones and properties. “An effective strategy is, at first, literally to ignore the literature of theory and fact on the area under study, in order to assure that the emergence of categories will not be contaminated by concepts more suited to different areas. Similarities and convergences with the literature can be established after the analytic core of categories has emerged” (37).
Concepts should be analytic (generalized to describe concrete ideas, but not the entities themselves), sensitizing (offering meaningful depictions of the ideas). Hypotheses, here, are suggested relationships between categories, their properties, each other. These are NOT tested. Multiple hypotheses are concurrently researched.
Sampling – based on theoretical purpose and relevance, not of structural circumstance. — which groups are selected — why, and how?
Use of ongoing inclusion of groups – to be included, you must have “enough features in common” with other groups, to be excluded you have to show a “fundamental difference” (50). Cannot randomize sample, as it is not required to discover relationships between concepts or even their existence.
What technique is best to get your data? Whatever one works for you to have the best access to it, ethically.
“Anecdotal comparison” (67) – “Through his own experiences, general knowledge, or reading, and the stories of others, the sociologist can gain data on other groups that offer useful comparisons” — lived experience is an acceptable way to construct sensitizing concepts.
Comparison groups control over conceptual level and population scope, help to maximize/minimize differences and similarities of data. This is necessary for developing categories, properties, and emergent theory. (I maximize data – demonstrating similarities between largely diverse groups, over large areas of space — and I do large!)
Theoretical saturation comes when “no additional data are being found whereby the sociologist can develop properties of the category” (61). Depth of a theoretical category is the amount of data collected on a group/category.
In generating formal theory, take out the specifics. Instead of looking at gender at fests, look at gender.
Examine change, transition, counterstatements, exceptions, themes. GT good for keeping in tune with on-ground phenomena, helping modify and form new theories – particularly in areas where there are no or few data/theories to begin with.
When you think you’ve seen a new idea or notice a theme during coding, stop and write a memo.
Code while you gather data, or in short spurts after data collection. Form categories for which codes can belong; when you can gather no new information about an area, move on to the next. Write theory when you have enough of a glimpse of several categories and their properties —- and how they relate to each other.
Are you more interested in forming or verifying theory? Substantive or formal? How dense is concepts and properties? What kinds of data are you using? How is theory integrated, and how clear is its use/presence?
Chapter 7 – meh.
Chapter 8 – More meh.
People criticize GT for not fitting the requisites of credibility, as their standards for credibility are based off of sampling, hypothesis construction, presentation of evidence in relation to a pre-formed theory. However, G+S note that these standards shouldn’t really apply, as GT has different metrics – instead, we form our credibility in how we collect, code, analyze, and present theory in relation to data.
Closing the research – when you think you’re knowledgeable enough about the phenomena studied. Not an arbitrary judgment, but when you examine your data (which can always be mined further), and you’re not getting new developments, this signals a close.
Research is immersed enough to know what he’s talking about, but is still somehow detached which “served also to protect him from ‘going native’ while still passing as a native to a large extent, when the people whom he is studying either have temporarily forgotten his outsider status or have never recognized it” (226).
Vivid descriptions of the field are necessary!
***Good GT will be understandable by the layperson, and attempt to closely reflect their realities. Categories should “fit” the data – and its potential applications. Theories should be generalizable – abstract to make a theory multi-conditional and flexible to accommodate change reflected in the everyday world, but also specific enough to be attune and sensitized to the phenomena studied. Hence, “to achieve a theory general enough to be applicable to the total picture, it is more important to accumulate a vast number of diverse qualitative ‘facts’ on many different situations in the area” (243- BOOM!). Theory’s application should offer a sense of control to the user, to produce and predict change in the area, and be flexible in its application. Controllable variables (not static, but those that fit, generalize) offer those who come after you to structure their research. Access variables give that control to the researchers – that is, what relationships or actors stand in the way of making this a static interaction?
Limitations of GT (a la Merton)
- Data can be forced to fit the theory, no matter what they “actually” mean.
- As data can be fitted to theory, the theory is seldom threatened. If qualifying data cannot be found, it’s not the theory’s fault
- These data cannot be fully tested against standardized measures.
- There’s too many variables that can’t actually be tested, and you’re consistently adding more.
- “Real science” is about testing variables that are involved in the phenomena.
BUT! G+S: “Generating grounded theory is what most of us end up doing, even if we start out to fit an existing theory to our data” (262).