Research Timeline
2.5 Data Analysis: Thematic Analysis
Thematic analysis has been seen as a ‘foundational method’ for qualitative research (Braun and Clarke, 2006). The majority of qualitative research is concerned with thematising meaning. This has led some researchers, to suggest that more specific methods of analysis are required. However Braun and Clarke, writing in 2006 set out a method of thematic analysis which they hoped added rigour to the process. The advantage of using thematic analysis is that it is relatively free of epistemological constraints. For this reason it suits a Critical Realist position, which has separated ontology and epistemology. Thematic Analysis (TA) can be used to identify the reality of participants or to examine discourse around events. Braun and Clarke refer to TA as being able to take a ‘contextualist’ position, which recognises individual meaning making and a broader social context.
Thematic analysis is particularly suited to my research aims, as it allows me to identify themes across the data as a whole, referred to as the ‘data corpus’ and within specific parts of the data, referred to as ‘data sets’. For example the majority of codes appeared across all of the interviews the ‘data corpus’, but some codes were particular to only the CAMHS workers and School staff. The use of Thematic Analysis will allow me to analyse the data corpus, but also to separate the data into sets, defined by the participant’s profession and look for similarities and differences in the themes that are found within the different sets.
Attride-Stirling (2001), writing about an alternative method of Thematic Analysis, recommends that qualitative researchers should be detailed in their description of methods and procedures, to add rigour and transparency to the research
55 process. With this in mind I have provided an account of the six phases of Thematic Analysis, as designed by Braun and Clarke and the choices I made which make the mode of analysis specific to my research question.
Phase 1 – Familiarising yourself with the data
A process of familiarising yourself with the text occurs as a researcher transcribes the data. Thematic analysis does not require transcription to be at such a detailed level as it should be for discourse analysis or content analysis. It does however require verbatim transcription and punctuation that does not change the intended meaning of the participant. Some non–verbal utterances may also have relevance and need recording. It is most important that the transcript retains the information and that it is as ‘true’ to the spoken interview as possible. I have transcribed the participant’s interviews verbatim and checked the transcripts against the original recordings to ensure that the meaning of the participant, as I heard it, is conveyed in the written form.
Phase 2 – Initial Codes
In the second phase of analysis, the transcripts are re-read and the data is coded for repeating patterns or issues of interest. The data extracts that pertain to certain codes are collected together. This process involves identifying the data as belonging to a code and storing this data together. This was done through use of highlighting and notes on transcripts, then the data extracts were collated on file cards.
It is at this point that the researcher needs to identify if they are conducting an ‘Inductive’ or ‘Theoretical’ analysis. A theoretical analysis is driven by the researcher’s theoretical or analytical interests and may be informed by existing literature and theory. I have chosen an inductive method of analysis, whereby the whole of the data set is scrutinised and codes are not pre-determined by the researcher. Braun and Clarke recognise that inductive analysis does not occur in an ‘epistemological vacuum’, as researcher presence is evident in choice of codes. In inductive analysis this is not planned or explicitly sought.
56 At this point coding was applied to as many potential themes as possible and data extracts could belong to several different codes. The data extracts, included some contextual information, so that meaning was not lost.
Phase 3 – Themes
Once the data has been coded and the coded extracts collated, the codes are sorted into potential themes. The researcher analyses the codes and looks for an overarching theme into which the codes fit. Sub-themes can also be identified, should the codes require more definition. At this point the researcher can create a ‘thematic map’, which shows the relationships between the themes, sub-themes and codes.
What constitutes a theme is determined by the researcher but it should have some level of pattern within the data and be relevant to the research question. Prevalence of codes is not as important to the determination of a theme as its relevance and importance, in relation to the research question. This is at the researcher’s discretion and gives them flexibility to follow patterns of interest. I will be using a semantic approach to identifying themes, following what is explicitly expressed in the data and not looking for latent meaning. However the analytic process is a journey from description to interpretation and I will be theorising as to the significance of the patterns in the data and their meanings. Phase 4 – Reviewing the Themes
At this point the researcher is required to review the themes for what is described as internal homogeneity and external heterogeneity (Patton, 1990 in Braun and Clarke 2006). Internal homogeneity is when the data items in the theme cohere and external heterogeneity is when the data within each theme is sufficiently distinct from other identified themes.
It is at this point that researcher can draw up a thematic map which can help to clarify the relationships between themes and sub themes.
57 To check the validity of the themes, Braun and Clark suggest that the researcher return to the original data corpus and ensure that the identified themes hold true to the meanings conveyed in the data. This is also an opportunity for the researcher to code any additional data that has been missed.
Phase 5 – Refine and Define the Themes
The researcher should identify the essence of each theme and be able to state clearly what aspect of the data each theme encompasses. This is not purely description of the theme, but identification of what is of interest and its implications for the research question. The thematic map is refined and finalised and sub themes identified.
Phase 6 – Write up and Analysis
In writing up the analysis the researcher should provide vivid examples of the data to illustrate the prevalence and relevance to the research question. It is at this point that the researcher needs to make an argument for the utility of their findings, analysis and interpretation.