6.4 Grounded theory

Grounded theory is very much associted with qualitative research and was developed in the 1960s as an antidote to more quantative approaches to research. Key to grounded theory is the notion that the emergent theory is grounded in the data, rather than existing separately. Despite its name, grounded theory is as much about coding data as it is about generating theory and indeed the coding of the data is the "pivotal link between collecting data and the emergent theory to explain these data" (Charmaz 2006, p. 46).

Comparison is also an important component of grounded theory. This can mean a comparison of data collected by different methods, such as interviews and observations. It can also mean data collected at different times in the research, so observations of the same research setting at the start of the research compared with observations conducted at a later stage.

This section on grounded theory owes much to Kathy Charmaz' (2006) book, Constructing Grounded Theory. Details of this and Glaser and Strauss' original work on grounded theory can be found in the key references section.

Initial coding

Initial coding is the first phase of coding in grounded theory and involves coding data as actions or from the participants' own words rather than imposing pre-defined categories. In particular, it is advised that gerunds are used to code actions, so 'receiving news' or 'experiencing pain' as they are more descriptive and meaningful than simple topic words such as 'news' or 'pain'.

The participants' own words can be used as code, referred to as in vivo codes. Theses are not necessarily words which are used ost often but which which imply a particular meaning or experience or which are specific to a particular group.

Initial coding can involves different units of analysis, so at the level of the word, the line or by incident.

Word-by-word coding
At this level of coding images and meanings are particularly important. It may be particularly helpful when coding documents and websites.

Line-by-line coding
This is probably the most common starting point to coding in grounded theory. The idea is that by looking at the data line by line you define what is happening in the data as it is happening and so can identify "implicit concerns as well as explicit statements" (Charmaz, 2006, p. 50).

Incident by incident coding
This is the type of coding most suited to observational data because you are comparing each scene rather than particular words or lines of text (Although, of course a lot of observational data may well be in written form).

Focussed coding

Focussed coding is the second phase in coding. This comes after line by line and incident by incident coding and is designed to synthesize and explain larger segments of data. It involves using the most significant or frequently occuring of the earlier codes to examine your data as a whole. When comparing data to data it may be that a code used in one context may shed light on another. This is focussed coding. Alternatively, it may be that by comparing new data with the earlier codes, some of the earlier codes no longer make sense and do have to be re-examined. This is the process of refining the data.

Axial coding

The purpose of axial coding is to sort, synethsize and organize large amounts of data and reassemble them in new ways after initial coding.  Axial coding is therefore the stage at which the data are brought back together in a coherent whole.   So, rather than seeing the data as individual and distinct codes, categories and subcategories are linked together and the relationship between them is examined and explained.  It is this coding which answers questions such as "when, where, why, who, how, and with what consequences" (Strauss and Corbin quoted in Charmaz, 2006. p. 60).

Moving from coding to theory

There are two stages in the journey from coding to theory in grounded theory: the first is memo writing and the second is theoretical sampling.  Memo writing helps you to capture your thoughts and the comparisons and connections you make as a result of your initial coding.  As you write successive memos you can start to increase the level of abstraction in your analysis which may, in turn, develop into theoretical categories.  Memo writing can also be a means of fine-tuning any further data collection.  In some respects, therefore, memo writng resembles the notes a practitioner researcher might make in a personal research journal and can also be a starting point in the process of writing up your research.

Coding and memo writing can help identify and develop theoretical categories from your data. Theoretical sampling takes the process one step further by actively seeking and collecting pertinent data to elaborate and refine categories which constitute your emergent theory.  The idea is that this theoretical sampling continues until no new properties relating to your categories emerge.  So, rather than just sampling at the start of the research project, to address particular research questions or to reflect particular groups of people, theoretical sampling allows you to further elaborate and refine the theories until yuor categories are 'saturated' with data.  Grounded theory still uses initial sampling but the purpose is different.  Initial sampling is where you start whereas theoretical sampoling directs where you go.    

Examples of research which has used grounded theory

A clear account of grounded theory coding is provided in Philip Osteen's paper, 'Motivations, Values, and Conflict Resolution: Motivations, Values and Conflict Resolution: Students' integration of personal and professional identities', based on the experience of Social Work students.  It is published in the Journal of Social Work Education, Fall2011, 47(3), pages 423-444

Last modified: Tuesday, 4 Mar 2014, 16:40