‘Themes’ are features of participants’ accounts characterising particular perceptions and/or experiences that the researcher sees as relevant to the research question.
‘Coding’ is the process of identifying themes in accounts and attaching labels (codes) to index them.
Researchers will generally choose to define features as themes where they recur several times in the data set, within and/or across transcripts. This is not, however, a hard and fast rule. If a single comment made by one participant is particularly helpful in elucidating their account, you may want to devise a theme that encapsulates it and include it in your template.
It is important to recognise that themes in qualitative research are not hiding in the data, waiting to be ‘discovered’ by the researcher. Rather, they arise from the engagement of a particular researcher with the text, as he or she attempts to address a particular research question. As such, they are pragmatic tools to help the researcher produce their account of the data. When deciding whether and how to define themes, keep this pragmatic intent in mind, ask yourself the question, ‘if I code the text in this way, is it likely to help me build my understanding of the data?’
For a discussion of the philosophical issues regarding the relationship between text, analysis, and the participant’s experience, visit the what is Template Analysis? section.
For a discussion about how to judge the quality of thematic coding, visit the quality checks and reflexivity section.
In template analysis it is common to identify some themes in advance, usually referred to as ‘a priori’ themes. Usually this is because a research project has started with the assumption that certain aspects of the phenomena under investigation should be focused on. A recent example from Nigel’s research is a qualitative evaluation of the Gold Standards Framework (GSF) for community palliative care. This framework, known as the ‘GSF’, specifies seven key issues regarding the organisation and delivery of care that practitioners need to address. It therefore made sense to use those seven issues as a priori themes when analysing GPs’ and District Nurses’ accounts of their experiences with the scheme (King, Bell, Martin and Farrell, 2003).
Another justification for using a priori themes is that the importance of certain issues in relation to the topic being researched is so well-established that one can safely expect them to arise in the data. For example, a researcher investigating patient experiences of chronic illness may feel that ‘uncertainty’ may be safely used as an a priori theme, given its prominence in the literature.
The main benefit of using a priori themes is that they can help to accelerate the initial coding phase of analysis, which is normally very time-consuming. There are also some important dangers associated with their use, which you need to bear in mind. Firstly, by focusing on data that fit the a priori themes, you may overlook material that does not relate to them. Secondly, you may fail to recognise when an a priori theme is not proving to be the most effective way of characterising the data. To prevent these pitfalls, it is crucial to recognise a priori themes as tentative, equally subject to redefinition or removal as any other theme. In the GSF study, mentioned above, two of the original seven a priori top-level themes were removed and included along with others under a new top-level theme. You should also try to restrict the number of a priori themes as far as possible, if you start with much of the initial template already defined, the danger of it having a blinkering effect on your analysis will be considerable.