Grounded Theory & Constructivist Grounded Theory
WARNING: Although this blog post starts out quite wittily light-hearted, it gets progressively more dense and heavy… like me after a few drinks, in fact. But to some up, it’s all about this:
A ‘traditional’ approach to research might well be that you observe something, state a hypothesis, make a prediction, test it, and write up your methods and results, though hopefully not with the explosive ending seen in that clip above. But what about if you’re in a human research jungle (spurious link to jungle themed YouTube vid’ here: https://www.youtube.com/watch?v=0cD9cBEaNBc) and you don’t really know where you are going and actually you’ve learned that making predictions is a form of bias based on your own frame of reference? Why would you intentionally bring a set of prejudices to a situation that might actually be very far removed from how your own misconceptions are colouring it? Surely you would want to understand the meaning of a given phenomenon in the context it arises?
It could also be that your research is in an exploratory phase, where you have to develop some ideas based on a given situation first. This exploratory style would require you collecting data as you go. The data would then have to provide some evidence of a picture you were painting in the context of your journey. Ideally you wouldn’t want to colour your own judgements by bringing too many preconceived ideas with you to the field, not that you can escape them entirely, but you’d at least want to aim for a fresh starting point. So why predict anything? Surely the journey is of value here, not necessarily at the expense of, but definitely in addition to the destination, which isn’t predictable. I’ll repeat and rephrase that: it’s NOT about prediction, folks!!!
It could also be that you do know very well where you are going (context-wise) and you need to record your research journey, a research journey that may well be unique due to the time and place you are embarking on it. You could well be engaging in ethnography, or participant observation – going native? So what might be a useful process for getting going in such a manner? You’ll be pleased to know there is one. Actually, not just one, but a few, that have been developed over the years to become less and less positivistic in nature.
Firstly, one important research approach, that also loosely inspired the way I carried out the research for my PhD, was grounded theory (Glaser and Strauss, 1967; Partington, 2002; Charmaz, 2003; Goulding, 2005; Selden, 2005). When its first proponents, sociologists Glaser and Strauss, came up with the idea of grounded theory in the 1960’s, they defined it loosely as ‘the discovery of theory from data’. Meaning: collect your data first, and then work out what it’s telling you theoretically. It was supposed to be theory developed from the ground up. Hence, grounded theory.
In technical speak these two grounded theory gurus described their basic position that generating grounded theory was ‘a way of arriving at theory suited to its supposed uses’ in the place that it came from (1967:3), or, theory derived from and for a specific context, for a specific reason. This method sits nicely with those of the periods referred to in a previous blog post of mine as ‘crisis of representation’ and ‘postmodern experimental’, born out of the ‘blurred genres’ period where greater importance started to be placed on gaining understanding from the local ethnographic context – with a view to understanding what went on between people in different places and settings. Yes, this would produce a ‘relative’ understanding of any alleged meaning arrived at, but that could then be extrapolated out to compare and contrast with other cases or similar phenomena reported on elsewhere. It wouldn’t always be necessary in such cases to have control groups as the research questions you were trying to find answers to might only be relevant in the context you were looking at there and then.
In this vein, though, Glaser and Strauss advocated the use of comparative analysis as a strategic method for generating theory applied to social units of any size, type or whereabouts. So that could apply for example to a household, a friendship group, a class, a school, a local community, village or other social group. And of course these could be anywhere in the world: from the high-rises of Hong Kong to the Gorbals of Glasgow.
A method Glaser and Strauss suggested to discover theory from this starting point, was through the generation of theory from conceptual categories derived from evidence, or data. So you’d gather your data, analyse it and ascribe conceptual categories to it – or code and theme it, and then develop your context specific theory from there. This was established as somehow ‘factual’ by means of comparative analysis with other social units or within the one in question. Internal validity would therefore be enough, at least in the first instance, to give credence to your findings. You could establish through double- and/or triple-checking, through a variety of methods, whether a finding was really meaningful in the way your research suggested.
The use of empirical generalisations was seen as necessary to delimit a grounded theory’s boundaries of applicability, as was the specifying of a unit of analysis from which to elaborate the particularity of the case (i.e. saying clearly to what and where the findings apply – scoping). Subsequent theory was to be verified through acquisition of more data, from which original theories may be modified. Neither accurate evidence, nor evidence type, nor number of cases, were held as crucial for generating theory. Glaser and Strauss (1967:30) explained:
“A single case can indicate a general conceptual category or property; a few more cases can confirm the indication… generation by comparative analysis requires a multitude of carefully selected cases, but the pressure is not on the sociologist to ‘know the whole field’ or to have all the facts ‘from a careful random sample’. His job is not to provide a perfect description of an area, but to develop a theory that accounts for much of the relevant behaviour… to… generate general categories and properties for general and specific situations and problems.”
The theory generated by this approach is described by Glaser and Strauss in terms of two elements: 1) conceptual categories and their conceptual properties, and, 2) hypotheses or generalised relations among the categories and their properties (Glaser and Strauss, 1967:35). Both categories and properties are described as concepts indicated by the data, and Glaser and Strauss asserted that while a category stood by itself as a conceptual element of the theory, a property, in turn, was a conceptual aspect or element of a category. They advocated the deriving of categories from the data in question, not borrowing categories from other data in an attempt to ‘force round data into square categories’, and that the generation of theory should aim at achieving diversity in emergent categories, synthesised at as many levels of conceptual and hypothetical generalisation as possible. This synthesis, they said, would provide apparent connections between data and lower and higher level conceptual abstractions of categories and properties. Concepts were required to be analytic – ‘sufficiently generalised to designate characteristics of concrete entities’, and sensitising – to ‘yield a ‘meaningful’ picture, enhanced by illustrations that would communicate the meaning in terms of understandable experience. The purpose of this was to enable the reader to ‘see and hear vividly the people’ in the area under study, which would in turn facilitate understanding of the subsequent theory developed for the area.
For Glaser and Strauss, hypotheses were then constructed from the relations generated among the categories arising from the comparison of differences and similarities among groups or units studied (1967:39-41):
“In the beginning, one’s hypotheses may seem unrelated, but as categories and properties emerge, develop in abstraction, and become related, their accumulating interrelations form an integrated central theoretical framework – the core of the emerging theory. The core becomes a theoretical guide to the further collection and analysis of data. … Integration of the theory is best when it emerges, like the concepts. The theory should never just be put together, nor should a formal-theory model be applied to it until one is sure it will fit, and will not force the data. … The truly emergent integrating framework, which encompasses the fullest possible diversity of categories and properties, becomes an open-ended scheme… because, as new categories or properties are generated and related, there seems always to be a place for them in the scheme.”
While Glaser and Strauss offer a clear vision of how to implement grounded theory as an emergent research approach, Selden (2005) and Charmaz (2003) critique their position as nevertheless being underpinned by positivism and objectivist epistemology. In Charmaz’s view, Glaser’s (1978, 1992) position is accompanied with assumptions of an objective, external reality, a neutral observer who discovers data, reductionist inquiry of manageable research problems, and objectivist rendering of data. Similarly, Charmaz understands the position of Strauss, with Corbin (1990, 1998), to assume an objective external reality, aiming towards unbiased data collection, proposing technical procedures and the need for verification. Charmaz goes on to assert that Strauss and Corbin move into a position of post-positivism through, for example, the proposition of ‘giving voice to their respondents’. In view of this understanding Charmaz presents her own approach, dubbed ‘constructivist grounded theory’.
As a starting point for understanding this constructivist view more fully, it is useful to consider that Collin (1997) refers to the broader level of ‘construction’, and ascribes it to interpretation, analysis and evaluation grounded in the view that human thought, discourse, agreement, or concepts generate the social world in a non-causal sense.
Under this umbrella of construction there are several sub-categories, including the above-mentioned constructivism, social constructivism, and the social constructionist position. ‘Constructivism’ is said to focus on the way in which the individual mind constructs what is taken to be reality, with or without any systematic relationship to the external world (Gergen, 1999; Piaget, 1954; von Glasersveld, 1991; Stacey, 2001). In ‘social constructivism’ individual minds are understood to construct reality in a way significantly formed by relationships (Vygotsky, 1962; Bruner, 1990; Stacey, 2001). While the ‘social constructionist’ approach views the social as the process of articulating individual selves and the world, placing the social relationship as prior and primary over the concept of the individual (Stacey, 2001). In this social constructionist perspective, blame and responsibility are not attributed to the individual since actions are understood as joint, because individuals are understood to shape each other’s actions – “as part of a living ‘we’”, ‘forming while being formed’ (McNamee and Gergen, 1999; Shotter, 1993; Stacey, 2001).
Regarding, ‘forming while being formed’, I’ve always loved this Escher drawing to illustrate that point:
As for how this impacted on the way I did research for my own PhD, while I accepted the social constructionist view in general, I found that the constructivist view described more accurately the way in which I collected data and the way I derived my findings.
The implications of this in terms of Charmaz’s constructivist view of grounded theory can be seen as follows:
“Constructivist grounded theory celebrates firsthand knowledge of empirical worlds, takes a middle ground between postmodernism and positivism, and offers accessible methods for taking qualitative research into the 21st century. Constructivism assumes the relativism of multiple social realities, recognises the mutual creation of knowledge by the viewer and the viewed, and aims towards interpretive understanding of subjects’ meaning. The power of grounded theory lies in its tools for understanding empirical worlds. We can reclaim these tools from their positivist underpinnings to form a revised, more open-ended practice of grounded theory that stresses its emergent, constructivist elements. We can use grounded theory methods as flexible, heuristic strategies rather than as formulaic procedures” (Charmaz 2003:250-1).
In view of Charmaz’s critique of grounded theory and the more generally stated proposition of the constructivist view above, the research I carried out for my PhD was is grounded in an integration of a grounded theory and a constructivist grounded theory approach.
The way I did this was to assess a variety of methods of data collection in order to fulfil the criteria of grounded theory as I’ve attempted to outline above. I then used data analysis techniques that were complementary to the grounded theory approach as well.
I realise that while this blog post started off fairly lightweight it has in its second half descended into academic-lingo. I make no apologies because those working at this level as researchers will need to get their head round this. However, If you have any questions or wish to discuss further, please leave a comment below or tweet me at @cazzwebbo