Thematic Analysis
Poorly demarcated, widely used 
Should be seen as foundational method for qualitative analysis and the first one to be learned.
This article argues that Thematic Analysis shouldn’t be seen as a process within other methods, but a method in its own right. 
One key benefit of TA is its flexibility. 

QA methods can be divided in two camps: 

• 1. Are tied to or stemming from a particular theoretical position: some of them only operate within that frameworks (conversation analysis, interpretative phenomenological analysis) , others widely applicable (grounded theory, discourse analysis, narrative analysis)
• 2. They’re independent of theoretical framework and can be applied across different approaches: thematic analysis: compatible with essentialist and constructionist paradigms within psychology 

Vocabulary:
Data corpus = all data collected during research project 
Data Set = data from corpus that is used for particular analysis; 1^st:  Could be many / all individual data items within data (for example only interviews); 2^nd: could be identified by a particular analytic interest in some topic in the data, data set is every time topic is being referred too. IN MY CASE: AFFECTS
Data item = each individual piece of data collected (e.g. an interview) 

What is TA? 
Method for identifying, analysing, and reporting patterns within data 
Sometimes is interprets various aspects of research topic 
Questions to be considered BEFORE analysis:

• What counts as a theme? (what ‘size’ does a theme need to be?) – retain flexibility, overall themes and subthemes will start appearing 
• A rich description of a data set, or a detailed account of one particular aspect. 1^st choice useful when researching and under-researched area. 2^nd choice if you want to provide a more detailed account 
• Inductive versus theoretical thematic analysis: inductive = coding data without trying to fit it into a pre-existing coding frame; theoretical thematic analysis = more analyst driven, focuses on an aspect of data 
• Semantic or latent themes: usually TA focuses on what is explicitly said and nothing beyond. Ideally, analytic process involved progression from description (data being organised etc.) to interpretation (attempt to theorize the significance of patterns + broader meanings). ‘Latent level’ TA identifies underlying assumptions, ideas etc. that are shaping and informing the semantic content of the data. This involves interpretative work; analysis is not just description but is already theorised. 
• Epistemology: essentialist/realist vs. constructionist thematic analysis: with essentialist approach you can theorise motivations, experience straight forward, because it’s assumed that there is a unidirectional relationship between meaning and experience. Constructionist perspective: meaning and experience are socially produced and reproduced rather than inhering within individuals. Therefore, you can’t focus on motivation or individual experience, but theorize sociocultural context and structural conditions that enable the individual accounts that are provided. 
• The many questions of qualitative research” research questions can be refined as project progresses. Questions asked during interview are not analysis! 

Doing TA: A step-by-step guide

Phase 1: Familiarising yourself with your data: 

• Read data repeatedly, immerse yourself, search for meanings, patterns, etc. 
• Depending on if you’re looking for detailed / overall analysis, latent / semantic themes, data / or theory driven will inform how you read Mine: probably overall, semantic, theory driven
• Start making notes, marking ideas for coding 
• Transcription of verbal data: time is not wasted, as it informs early stages of analysing, you will develop more thorough understanding of data 

Phase 2: Generating initial codes: 

• Codes identify key feature of data (semantic content / latent) 
• ‘The most basic segment of the raw data that can be assessed in a meaningful way regarding the phenomenon’
• Process of coding is part of analysis
• Coded data is different from units of analysis (themes) which are often broader 
• Themes (developed in phase 3) are where interpretative analysis of the data occurs and in relation to which arguments about the phenomenon being examined are made
• Code for as many potential themes / patterns as possible 
• Code extracts of data inclusively (keep a little of the surrounding data)
• Individual extracts of data can be coded in as many different ‘themes’ as they fit to 
• Data sets can have contradictions – don’t smooth them out

Phase 3: Searching for themes: 

• Involves sorting the different codes into potential themes and collating all the relevant coded data extracts within identified themes
• Use visual representations to sort codes into themes (or mind maps etc.) 
• Think about relationship between codes, between themes, between different levels of themes 
• Have main themes, sub-themes, some can be discarded, you can have ‘miscellaneous’
• Don’t abandon anything yet without looking at all extracts in details (next phase), as it is uncertain if the themes will hold as they are or will be changed again 

Phase 4: Reviewing themes:

• Here you will realise that some candidate themes are not themes (data too diverse, not enough data etc) and others might collapse into each other (two apparently separate themes might form one theme) 
• 1. Read all collated texts and check if they form a coherent pattern
• 2. Consider validity of individual themes in relation to the data set, but also if candidate thematic map ‘accurately’ reflects the meaning evident in the data as a whole
• Reread entire data set 1. To see if themes ‘work’ in relation to data set; 2. To code additional data that have been missed in earlier coding stages. (This is to be expected) 
• Stop when coding doesn’t anything substantial 
• At the end you should have good idea of what your different themes are and how they fit together and the overall story they tell about your data 

Phase 5: Defining and naming themes

• Identify the ‘essence’ of what each theme is about 
• Don’t just paraphrase the content of data, but identify what’s of interest’ 
• For each individual theme write a detailed analysis 
• Consider themes themselves, but also in relation to the others (no overlaps)
• At the end you need to be able to clearly define your themes. Describe scope and content of each theme in a couple of sentences 
• Names need to be concise & punchy and give an immediate idea what the themes is about 

Phase 6: Producing the report 

• Final analysis and write up of the report 
• Tell complicated story of your data in a way that is convincing the reader of its merit 
• Analysis should be concise, coherent, logical, non-repetitive
• Extracts must be embedded within analytic narrative, that illustrate compellingly
• Analysis must go beyond description 
• Make an argument in relation to research question 

Questions you need to ask towards the end of analysis: 

• What are the assumptions underpinning it? 
• What are the implications of this theme? 
• What conditions are likely to have given rise to it? 
• Why do people talk about this thing in this particular way? 
• What is the overall story the different themes reveal about this topic? 

Potential pitfalls to avoid when doing TA 

• Failure to analyse data: TA is not a collection of extracts; extracts are illustrative of the analytic points 
• Using data collection questions as ‘themes’ = no analytics works has been carried out 
• Themes do not appear to work / too much overlap / themes not internally coherent (can stem from failure to provide adequate examples). Reader must be persuaded, to avoid ‘anecdotalism’ 
• Mismatch between data and analytic claims
• Mismatch between theory and analytic claims (experiential framework wouldn’t make claims about social construction of research topic; constructionist thematic analysis wouldn’t treat people’s talk of experience as a transparent window on their world)

What makes a good TA 

Checklist: 

Advantages and disadvantages of TA
Flexibility means that range of things that can be said about data is too broad: can be potentially paralysing for the researchers 
Limited interpretative power beyond mere description if it’s not used within existing theoretical framework that anchors analytics claims that are made 

from:

Virginia Braun & Victoria Clarke (2006) Using thematic analysis in psychology,

Qualitative Research in Psychology, 3:2, 77-101, DOI: 10.1191/1478088706qp063oa