Chapter 3. Research Methodology
3.5 Process of conducting the study
3.5.5 Role of the Pilot Study
In order to maximise the benefits inherent in conducting a small-scale study ahead of the main study (Sampson, 2004), it was decided to attempt to analyse the findings of the study from two perspectives: the effectiveness of the research strategy and interview protocol, and the outcomes in terms of research findings. This section will cover a critical analysis of both areas, starting with the research protocol testing, followed by an assessment of the research findings.
3.5.5.1 Assessments of methods.
The results of the pilot study in terms of the research protocol: data collection through semi-structured interviews and subsequent data analysis its key variables were summarised as follows:
Variable Assessment
Research Protocol Both respondents were enthusiastic about the process, and were fully engaged.
Both agreed that more detail might be needed ahead of the interview, in order that the most suitable person within the organisation was approached.
Sampling may need refinement at Individual level.
Interview Face-to-face ‘worked better’ for both respondent and researcher.
Timings Assumed lead in times and interview timings worked well;
however analysis and write up timings were grossly inaccurate and need revision
. Generation of
Information
Thematic approach, with limited set questions allowing the respondent to expand on key points and express himself or herself. Rich and thick data were generated, however there was some repetition across theme areas with the initial introductory exchanges.
Only one potential new theme emerged from the 2 interviews.
Introduction of bias a potential issue for researcher.
Analysis With limited data from only 2 interviews no major revisions emerged from the initial template and coding structure. Likely that this will not be the case for broader, less familiar interviewees.
Transcription attempted initially by researcher, but reverted to secretarial support due to time constraints.
Issue emerged with technology support and suitability of NVIVO and/or TAMS.
Table 10 Pilot Study - Research process assessment.
[Source: Author]
During interviews the thematic boundaries allowed the respondent to focus on their experiences and what they felt was important, with little prompting from the researcher. The key issue arising from the process was that a face-to-face interview seemed to allow both respondent and researcher to engage more effectively, as was agreed by both at the feedback session at the end of the interviews. As a result during the main study face-to-face interviews only were conducted. Whilst the overall protocol seemed to meet the needs of the study at this stage, the area that garnered the most learning, in terms of process rigour and coherence with the phenomenological approach, was within the data preparation and analysis stages.
3.5.5.2 Pilot study: Data preparation and analysis.
In order to understand the issues arising from the data analysis for the pilot study it is necessary to outline the procedure adopted. The interviews conducted were recorded, with the participant’s consent. Flick (2009) offers that the use of recording devices ensures the documentation of data independent of perspectives: both of the researcher and of the interviewee and as such it seemed appropriate to apply this approach to ensure the most complete record possible. The first step on analysis of the data was the transcription of the interviews conducted. Whilst there is no adopted academic standard of transcription and little guidance on the process (Poland, 1995;
Wellard and McKenna, 2001), as with all qualitative research it is suggested that it is vital where possible to strive for reliability (Silverman, 2010). In order to attempt to ensure this a second check of the transcript was conducted, both to ensure accuracy and to enable the anonymisation of the data and compliance with any confidentiality guarantee where required. Both respondents appreciated the opportunity to revisit their thoughts, and although there were few changes, some specific operational details were deleted at the request of one of the respondents.
The interview transcripts for the pilot study were then subjected to coding, initially attempted through the use of NVivo software (Welsh, 2002). As noted in the review of the research protocol on this particular occasion, for a variety of technical compatibility reasons encountered whilst travelling during the analysis stage, NVivo could not effectively be employed. After consulting with supervisors and taking advice from the wider group an alternative was found in the TAMS Analyser programme, which proved suitable for the purposes of the small-scale study. Although drawn from a limited sample, it was felt that some insight was gained from the findings, however throughout the process a sense
generation of further nodes - was effective. However, the use of the electronic analysis programme encouraged a personal sense of detachment, not only from the data, but from the ability to be able to be able to represent the essence of the findings.
Thus an alternative data analysis plan was sought and the six-step approach offered by Creswell (2007) was selected. The plan was built on the work of Moustakas (1994) and the data analysis procedures as outlined by van Kaam (1966) and Colaizzi (1978). The plan is adapted, in that it was originally offered as a vehicle for multiple-researcher analysis, however in this study there is a single researcher. Notwithstanding, the approach remained appropriate for this study in that it was deeply rooted in phenomenological methodology (Creswell, 2007; Moustakas, 1994). The next section will outline how this more coherent process was used in the main study.