CHAPTER 2 Methodology
3.10 Data analysis
This section reviews the theoretical and methodological aspects of data analysis, followed by a description of how the obtained data was transcribed and analysed.
Data analysis aims to create meaning (Dibley 2011. Ellis, 2010. Boeije, 2010. Jorgensen 1989) and develop situational context (Ellis, 2010) from the potentially lengthy transcripts obtained through interpretative
phenomenological interviews (Dibley, 2011. Riessman, 1993). In addition, the resultant data must address the pre-determined research questions (Boeije, 2010).
An interpretative phenomenological methodology, exploring the views of the participants, requires a data analysis method that, in addition to describing a situation, also interprets the data to increase comprehension and provide clarity, yet allowing the participant to remain central to the process through the use of direct quotations (Wagstaff & Williams, 2014. Larkin & Thompson, 2012. Pringle et al, 2011a. Backstrom & Sundin, 2007. Van der Zalm & Begrum, 2000). IPA analysis allows “rigorous exploration of
idiographic, subjective experiences and more specifically social cognitions”
(Biggerstaff & Thompson, 2008; p219). This detailed, first-person account is a distinctive characteristic of this approach.
Pringle et al (2011a, p 21) questioned “whether it is possible to describe
something without adding an interpretation at the same time? “. In order to
describe the phenomenon only, in this context, the presented data would report key characteristics of emergency preparedness. In contrast, this study aims to understand experiences and concepts in additional to defining characteristics.
However, this study specifically aims to explore how paramedics experience emergency preparedness. A conclusive list cannot be generated as each person’s experience and opinion will differ dependent on numerous factors such as their past experience, area of interest and personal opinion. To generate this depth of understanding, the researcher must take a central role in interpreting and questioning the content of the interview transcripts
to generate this new knowledge (Wagstaff et al, 2014. Pringle et al 2011a. Smith et al, 2009. Biggerstaff & Thompson, 2008). As a result, a degree of subjectivity exists in the data analysis role, but can be minimised by choosing a method that has structure, is auditable and enables the researcher to justify their interpretation and thematic analysis (Pereira, 2012. Ellis, 2010. Jootun et al, 2009. Whittemore et al, 2001).
The advantages of IPA data analysis are its adaptable nature, allowing it to be used in unfamiliar situations or where the opinions of participants are challenging to obtain (Pringle et al, 2011a). In addition, it has a clear,
auditable structure, which can be applied to numerous disciplines (Smith et al, 2009. Giorgi and Giorgi, 2003). Consequently, as it follows an
interpretative philosophy, the generated themes can be developed as subsequent analysis occurs (Tuohy et al, 2013. Larkin & Thompson, 2012. Smith, 1999b), thus enabling an evolution of results as the study progresses. This is particularly important when examining a developing area such as emergency preparedness, as only limited literature is available to establish the interview questions. In contrast, Giorgi (2000) acknowledges that the structured nature of analysis challenges the academic freedom of the researcher, although Smith et al (2009) notes that the process is only a recommendation, and can be modified as needed.
The literature highlights that the suggested process of IPA data analysis is a guideline and commonly follows the following stages:-
Read transcript through several times, making notes and comments.
Identify and label emerging themes and meanings within the text.
Relate back and link themes to quotes in text, using a cyclical process.
Look for potential links between themes that may lead to master- superordinate themes.
Repeat the process with subsequent transcripts.
Connect/cluster the themes from the texts into super-ordinate themes, with related subthemes.
Examine texts more closely for greater depth of meaning and interpretation.
Produce a summary table of theme for the group, and a detailed interpretative reflexive written account.
Smith et al (2009) in Pringle et al (2011b), p15.
This structured approach is not definitive, and although there are
communalities in approach and process, the input from the researcher must show imagination and creativity (Wagstaff et al, 2014. Larkin & Thompson, 2012. Smith et al, 2009. Eatough & Smith, 2008). Due to the interpretative freedom that this process allows, the analysis process must be transparent to the reader (Smith et al, 2009), so it is clear how the interpretations are derived and also that the narrative account remains focused at the
idiographic level, which is central to IPA, resulting in the participants quotes telling each individual story (Pringle et al, 2011a. Smith et al, 2009).
The practicalities of this process will now be reviewed in the context of this PhD. Each stage was reflected upon in my research diary and photographic evidence obtained to demonstrate how each stage was carried out, to aid with transparency and create an audit and decision trail.
Firstly, each script was read thoroughly, multiple times, to enable familiarity with the text and provide an immersive view into each interview (Wagstaff et al, 2014. Larkin et al, 2006). In transcribing each interview, I was able to ‘hear’ the participant’s voice which reminded me of additional aspects of each interview, which were not captured on the digital recording. In
addition, the research diary, with the initial notes that were made about the interview and the interaction between the researcher and participant, was reviewed alongside each transcript. This process acted as an additional contextual reminder of the interview. This stage enabled active engagement with the data and the emphasis on each individual experience within the transcript.
Notes were then developed at an exploratory level, and initial manual coding commenced on a line-by-line basis (appendix 16). This ‘open coding’ consists of identifying apparent areas of interest including language, events, places and areas of concerns (Smith et al, 2009. Larkin et al, 2006) and also considers how the participant discusses, comprehends and considers
emergency preparedness. In addition, descriptive, linguistic and conceptual concepts were highlighted and noted on the text (Smith et al, 2009)
(appendix 17). At the end of this stage, a detailed and holistic review of each interview had occurred, with notes reflecting the phenomenological focus, detailing what it is like to be a participant and what some of the key features of these experiences are and what they mean to the participant. At this stage, emergent themes were noted on index cards for further review. In addition, a brief summary of each interview was added to the interview transcript, with key summary bullet points.
Moving from the descriptive to the interpretative phase required ‘abstract’ thinking, with the researcher noting keywords, patterns, metaphors and imagery (Smith et al, 2009). It is at this stage that the researcher is “trying to
make sense of the participants trying to make sense of their world” (Smith &
Osborn, 2008. p53). A similar pen and paper approach was deployed, initially on one transcript, before moving to the next. To aid in the line-by- line text review, various techniques were utilised to ensure that the required detail was captured. These techniques included reading the text out loud, in small chucks (and covering up non-relevant parts), on the computer and on hard copy and also by reading the document backwards (deconstruction) (Smith et al, 2009). During this process, notes were made of metaphors and other ‘gems’ that stood out that required further review and analysis. These ‘gems’ are infrequent, but their value, in representing a theme, is much greater than the part of the transcript it represents (Smith, 2011c). These gems were highlighted, and used as theme titles or
descriptions, which further roots the analysis in the participant’s words (Pringle et al, 2011a).
These notes were then moved from individual transcripts onto an A3 piece of paper where themes were mapped (and associated subordinate themes were developed), with the aim of patterns developing which provide an insight into each lived experience (appendix 18). It was important at this stage to establish some order and structure to how the results would be presented, so a table detailing key quotes and the theme they represented was devised.
Whilst IPA focuses on the individual account, there is also a need to note patterns across the transcripts. A table was created, from the
themes/subordinate themes developed and a mapping exercise was carried out (appendix 19) noting similarities, differences and interrelations that was significant (Smith et al, 2009. Larkin & Thompson, 2012. Eatough & Smith, 2008). During this stage, additional consideration was given to the extracted quotes to ensure that they, and the developing themes, were illustrative of the participant’s interview.
The process of identifying themes and subordinate themes was challenging and overwhelming and the literature notes that this can feel like ‘drowning
in a deep bowl of spaghetti” (Wagstaff et al, 2014). Multiple methods of
visual organizing occurred, including the use of index cards that were labelled and moved around as themes developed and disappeared, and
mind-maps to represent the concepts and themes as they moved around the central themes (Wagstaff & Williams, 2014. Whiting & Sines, 2012). Using a mind map enabled a visual structure of how the final themes were
connected and structured and resulted in a visual pathway for data analysis (Whiting & Sines, 2012). Through visual mapping and discussion with supervisors, an IPA group and a nursing colleague, the final themes for discussion were determined from the initial themes. This stage of the process was reflective, iterative and cyclical (Wagstaff & Williams, 2014). Through verbal discussion it was apparent that the concluding themes had clear links to the text and were suitable for review and further discussion.
From these visual tools and table, a narrative can then be created (chapters 6, 7 and 8), with a thread created for each theme, displaying pertinent extracts of the transcript to illustrate the key points. A table summarizing key themes was developed, detailing the superordinate and subordinate themes that had been identified.
A significant decision was whether to utilise data analysis software during this phase. Although there are sound reasons for using programs such as QSR NVivo, the IPA literature supports researcher preference on this issue (Wagstaff et al, 2014). As the primary focus of IPA data analysis is actively engaging with the transcripts and coding and analysing them (Smith et al, 2009. Langridge, 2007), this was carried out with paper, pens and index cards enabling intimacy with the data, resulting in greater insight and a richness to the analysis phase.