Thematic analysis in research

Thematic analysis is a technique used by qualitative researchers to understand the material they gain through interviews, focus groups, texts and multiple types of media.  Thematic analysis aims to understand the key themes that occur in a transcript of these interactions.

A theme is a pattern or idea which runs throughout and across a set of data which can summarise or at least minimise that data and often interpret it.  It can be essentialist or constructive, descriptive of the data (without interpretation) or conceptual (with interpretation) but that depends on the interpreter and whichever approach is taken it must be supported by evidence in the data itself.  A theme should have importance, and just because a particular concept occurs more frequently, it does not imply that it is of importance and therefore should be considered as a theme (Braun & Clarke, 2006).  Accordingly, there is no standard level of occurrence with which a theme should occur in the data in order to be considered a theme (Braun & Clarke, 2006).  The data within a theme may not make sense on its own but when it is all pulled together it will form a cohesive idea or trend.  Examples of a trend might be ‘Society make judgements about fat people’, and evidence to back this up may be sentences such as ‘People assume I’m lazy; it’s not that I’m greedy, I know that’s what they’re thinking’.  Another example of a theme might be ‘Being healthy is an active choice’ and evidence to back this up might be sentences such as ‘It’s not easy; every morning I drag myself to the gym and I know if I eat fruit instead of biscuits that it’s good for my health, but it’s a struggle sometimes’.

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research In Psychology, 3(2), 77-101. doi:10.1191/1478088706qp063oa

The main steps involved in performing a thematic analysis are as follows:

  • During data collection, start to note down patterns which grab your attention
  • Familiarise yourself with all the data, transcribe it using one of the recognised conventions (Edwards & Lampert, 1993, cited in Braun & Clarke, 2006) and read it multiple times taking note of patterns
  • Some researchers develop a ‘code manual’ which undergoes testing using additional data, serving as a guide for data interpretation and adding rigour to their process. (Fereday & Muir-Cochrane, 2006)
  • Code data by asking “what is going on?”, “what’s being described here?” and then summarise succinctly. Allocate a different number for each code and analyse all of the text.
  • Identify themes by looking for patterns and over-arching ideas. This will involve sorting the ideas into similar groups, and a table, brain-storm or diagram can be used to aid this process (Braun & Clarke, 2006)
  • Review and refine themes perhaps applying Patton’s dual criteria (1990), checking that the theme applies to both the extracts and the overall data set, avoiding repetition, grouping into logical sections. Ignore or merge similar and smaller ones together
  • Define and name the theme by identifying the story and understanding how it fits within the bigger picture; make sure it still covers you working title. This is an ongoing process and my not occur at one chronological point in time.
  • Produce your report where you write up your themes using evidence; referring to previous literature and using diagrams if necessary. Use the most powerful or typical extracts to illustrate your point.

Fereday, J., & Muir-Cochrane, E. (2006). Demonstrating Rigor Using Thematic Analysis: A Hybrid Approach of Inductive and Deductive Coding and Theme Development. International Journal Of Qualitative Methods, 5(1), 1-11.

The validity of a qualitative/thematic analysis be improved in multiple ways:

  • Follow guidelines set out by the BPS
  • Ensure you have quality data, collected in a fair and unbiased way and transcribed in detail; have someone check your transcription for accuracy
  • Get to know your data well before embarking on coding so that you can understand the whole of the text / discourse
  • Note down any strong reactions, or biases or hunches that you might have during the process which may hinder your interpretation as you go; your relation to the discourse is important and needs to be analysed, but avoid unnecessary bias.
  • Work from the data only at coding stage; don’t interpret too early
  • Be thorough, go through every section so that you don’t miss anything, ensuring each item gets a similar amount of consideration and that the most salient parts of the text aren’t the only points noticed
  • If unsure about what a part of the data is saying, consult other psychologists’ opinions
  • Group your coding and themes logically, so that their relation with each other is clear
  • Avoid duplication and complication of themes; keep it simple, by cross referencing each theme with each other, making they are distinct and understandable
  • Keep to about 5 themes
  • Take your time at every stage of the process so as to produce quality interpretations
  • Maintain a focus on your working title, so that findings are relevant and that you achieve what you said you would
  • Ensure your analysis and interpretation clarifies and makes sense of the subject matter, avoiding jargon.

Leave a Reply