INTERPRETACIÓN DE LOS RESULTADOS
D. Las huellas en el libro de Texto de Historia de España
III.2.3. El manual como uso exclusivo ¿Cuaderno de trabajo,
Mark Can you tell me a little bit about your overall use of drink or drugs and what you found good about that?
Karen It stopped the voices.
Mark Okay. What was it you found useful then in terms of stopping the voices?
Karen It give me something else to think about, so…
Microanalysis - It give me ‘It’ – an entity, gave – provided something, me –
personal to Karen. These three words suggest a relationship perhaps between the drug and Karen. It, the drug, appeared to have imparted something to Karen to be grateful for.
Memo – is this another theme or category? That a RELATIONSHIP WITH A DRUG develops and the drug starts to be seen as an entity of its own. If this is so, upon abstinence, an individual may lose the effects of the drug and the relationship. Is this significant? Is there a relationship between participant and drug? If so then what implications for loss, separation, substitution or replacement exist?
Early analysis was therefore a crucial platform that generated categories upon which to base axial and selective coding.
Axial Coding
Axial coding enabled the accumulation of related incidents to be compared. The comparisons and differences of incidents within the same category subsequently revealed the extent of the depth and breadth of a category. Axial coding was the
process that explored linkages and cross-cutting properties. An example of the greater depth axial coding provided is exemplified in Figure 4 below.
Figure 4. Axial Coding: Identifying a Property
Concept – Using drugs to feel good Category – ‘Role of Substance’
Cross-cut or relate dimensions of individual incidents
2 levels of analysis happen here
The words of the participant My conceptualisation of the words
“It like chills me out and takes the edge off” “Self-medicating”
(Self-medicating is a subcategory of the category ‘Role of Substance’)
Where does this happen? Where else does this happen?
Why does this happen? When does this happen?
(The answers started to reveal other incidents in further categories for example ‘Levels of Knowledge of Drugs / Alcohol’)
Figure 5. Axial Coding: Properties and Dimensions Cross-Cutting Matrix “Diagrammatic displays are not just a way of decorating conclusions; they also provide a way of reaching them”. (Dey 1993, Page 192)
A cross over between categories occurred enabling the important connections between categories to emerge. In Figure 5 participants required knowledge about the effects of their drug in order for the role to be performed. Examining the dimension (size and extent) of knowledge revealed potential risks (opiate overdose and respiratory depression) and successes (role). For example experiencing the blocking out of emotional pain by using heroin reinforced the role of the drug whilst lack of knowledge about potential overdose increased mortality risk.
This example illustrated the variable levels of drug knowledge (limited to thorough) and drug role (boredom to emotional pain relief). It superimposed the category ‘Impact on Behaviour and Life’ upon the category ‘Role of Substance’ continuum.
Axial coding was therefore a strategy that used the shared properties of different categories to reveal their connections. Through the identification of shared properties the categories, whilst discreet in a conceptual sense, were exposed and reconstructed as interdependent.
This process was repeatedly applied. Categories were deconstructed into properties and where possible, dimensionally. Deconstructing concepts is a pivotal procedure in grounded theory in order to avoid prior ideas and understanding to dominate concept development. For the practitioner-researcher it is impossible to totally bracket off previous experience as research ‘incidents’ cluster into categories, but the process of deconstruction can help to differentiate the incidental related concepts from previous conceptual understanding developed elsewhere, consciously and subconsciously as a practitioner. Bringing my knowing into the study had substantial value as I discuss in the following pages, however it can be disadvantageous; it was necessary to identify what was known in order to identify new knowns. Chenail and Maione (1997) reiterate the advice of Glaser and Straus (1967) in that ahistorical positioning of the practitioner-researcher is improbable and as such deconstruction of all elements salient in a concept followed by their reconstruction will enable integrity to remain intact. Dervin (1992) couches this in deferent terms which add to the quality of grounded theory research by adding a second dimension, not only does the deconstruction and reconstruction assist the researcher to differentiate their ideas from data generated concepts but it also stimulates the analytical climate by challenging popular or accepted opinion. This Dervin calls ‘sense-making remade’. The challenge of such opinion or position lays at the heart of many a research project; the words of John Stuart Mill (1859, Page 105) come to mind referring to the error inherent in “the deep slumber of a decided opinion” for which much inspiration through this labour intensive analytical procedure was taken.
The reconstruction weaved the categories back together but only after insights, relationships and contradictions had emerged. The context from which they emerged and any cause, effect or intervening properties added to breadth and depth. The deconstruction-reconstruction of categories was therefore a process that encouraged understanding of the subject matter. With such understanding a theory underpinned by
Selective Coding
Axial coding revealed a core category of drug use in mental illness. This was a consistent and overarching activity present in, or directly related to, all categories. The categories related to help or helpfulness were dominant. Major and core categories are discussed in depth later in the Findings and the Discussion chapters, however their identification enabled selective coding to take place. Selective coding consisted of picking up threads that appeared connected to the major categories and subjecting them to further analysis including axial coding. I then searched for incidents that distinctly related to the major category with the aim of moving from conceptual understanding to a practical explanation; a substantive theory.
The analytical procedure of coding incorporates constant comparison of themes. The coding and theoretical sampling took place jointly. Sampling relied upon my sensitivity to incidents or constellations of incidents that would reveal issues relevant to the main categories. Sampling was therefore guided by emerging theory. As the categories grew in depth and breadth but not number I knew I was reaching a point near to data saturation.
The central categories presented and discussed later, contained key issues that encouraged hypothesis led analysis. I was guided by Strauss and Corbin’s (1998, Page 147) criteria for choosing the major categories (helpful and unhelpful incidents) which in summary consisted of the following;
• It / they must be central and relate to all other main categories.
• Must appear frequently
• Must be logical and consistent in property and dimension.
• The label or name given must be sufficiently abstract to be used elsewhere
• Must possess a concept that has grown in depth and is explanatory
• Must be able to explain variation
Thirty-five transcripts in total were made, after the analysis of 26, saturation had occurred. The remaining 9 were listened to and memos written where appropriate in order to (i) discover any new themes, (ii) confirm saturation had occurred and (iii)
identify relevant information relating to the existing themes. Five service user interviews, one practitioner interview and 3 focus groups transcripts were analysed in this ‘scan and focus’ manner (Strauss & Corbin 1998). Whilst new information was not apparent the process ensured that all data had been subjected to inclusion and analysis.