CHAPTER 4: DATA ANALYSIS
3) DEFINING NODES
4.4. Q UERYING THE DATA
TEXT SEARCH QUERY
This enabled the listing of all sources that contain specified text. This provided a method of coding sources by facilitating a search for words and coding the occurrences at a particular node.
WORD FREQUENCY QUERY
This facility listed words and the number of times they occurred in selected items. Collating which words appeared most frequently helped to identify themes and concepts (e.g. the word ‘’different’’ was mentioned 195 times with the majority of references made to how general practice was different to the treatment clinics).
CODING QUERY
This function gathered content based on how it was coded e.g. the content where women talked about methadone and their babies.
MATRIX CODING QUERY
This function created a matrix of nodes based on search criteria. For example, experiences and attitudes about general practice by gender.
COMPOUND QUERY
This facility combined text and coding queries when it was necessary to search for specified text in or near the coded content.
CHARTING
Charts were created to visually represent the coding data and to display matrix query results (Appendix 9).
4.5. RIGOUR
The rigour of the methodology is judged by unique criteria appropriate to the research approach. Rigour in the context of a phenomenological study can be demonstrated by ensuring the methodology is congruent with the philosophy used. In phenomenological research, bracketing, which has been incorporated into intentional focusing on the experience, is one factor that is central to the rigour of the study. In Husserlian phenomenology techniques such as bracketing, form part of the method of transcendental phenomenology. These processes allow an accurate interpretation of the phenomenon in the context of the prevailing philosophy (Seamon 2002b). Credibility, auditability, fittingness, and confirmability have been proposed as reasonable criteria for evaluation (LoBiondoWood & Haber 2002).
4.6. CREDIBILITY
The credibility of qualitative research depends on the use of credible data, a credible analytical process and a credible mode of presentation of the results. To ensure credibility, transcripts of their own interviews were shown to the participants. Each participant was asked to verify and comment on the interview and provide any extra information which they felt was relevant. This ensured that an accurate account of their experience was documented.
The transcripts from three of the interviews were analyzed separately by the supervisory team both in the pilot study and a further three in the main study. Similar themes were extracted and generated from this process and there was close agreement on the basic themes but each analyst may have ’packaged' the themes differently (Armstrong et al 1997). This afforded further credibility as Mays & Pope (1997) caution that research which relies exclusively on observation by a single researcher is limited to the perceptions and introspection of that investigator. Barbour (2001) asserts that the degree of concordance between researchers is not really important but what is valuable is alerting researchers to all potentially competing explanations. This was achieved by discussing both data construction and the analytical process with two supervisors. The use of an assistant researcher was beneficial and insightful to the method development and analysis throughout the pilot study and was used to prevent practitioner/client bias in the main study as it is debatable whether the researcher/ practitioner may have in some way influenced the behaviour and speech that was witnessed in the first pilot interview.
Data consolidation or merging was also facilitated by NVivo to create new variables (not simply transformed ones) that were then investigated or analyzed (Kane et al 2001). This served to protect the narrative
structure of the data so avoiding the problem of decontextualisation or data fragmentation (Mays and Pope 2000).
4.7. AUDITABILITY
The construction, data collection and analysis of data provide an auditable account of the research process. The inferences that have been made in the findings are supported by the decision trail of transformation of the raw data into phenomenological informed expressions and the synthesis of the transformed meaning units into the general structural descriptions. It was possible to present sufficient original evidence in the descriptive account ‘to satisfy the skeptical reader of the relation between the interpretation and the evidence’ (Mays & Pope 1996) by using NVivo as it maintained a visible audit trail (Appendix 9 Screen Dumps & Analysis). All processes and stages of coding
were tracked in such a way as to facilitate the researcher in clearly demonstrating rigour. The descriptions of the participants accurately reflected their experiences and the influence of the researcher practitioner’s bias was laid bare in the bracketing interviews.
Bracketing Interview 1
R Yeah, at first I think you challenged me that I was constructing the data myself and I didn’t like that challenge because I felt, No!, this is a very open experiential type of research but in fact, yes, I had to look at that and realize that I was constructing the datajust by the very framing of the study and just also saying I wanted to ‘tap into experiences’.So I found that very challenging, and in many ways I suppose it was the first line in the sand that I drew in the pilot study as Husserl says ‘suspend the natural attitude’ and come outside the process and look at it in a different way.
The transcription of data is accurate and has been verified both by participants, transcriber, and researcher. In the analysis of the transcriptions, conclusions other than those offered by the researcher
have been discussed in the findings (Polkinghorne 1989:57). The interpretation of a sample of interview transcripts has been undertaken by four independent researchers, namely the principal researcher, academic supervisor, practice based supervisor and the research assistant who was consulted during the pilot phase prior to embarking on interviews for the main study.