lecture 3 - qualitative data analysis Flashcards
what do you need to decide before starting analysis
Before you start the analysis (preferably before you even start collecting data) you need to decide:
-What’s your epistemology? (e.g, looking at “employability” in a positivist, contextualist or constructivist way)
-What counts as a theme?
-A rich description of the whole data set, or a detailed account of one aspect?
-Inductive or deductive TA?
-Semantic or latent themes?-
Braun and Clarke’s six steps
1-familiarise yourself with data
2-generating initial codes
3-generating themes
4 reviewing themes
5 defining and naming themes
6 producing the report
Familiarize yourself with the data
Doing the interviews (what was your impression of the interview?)
Transcribing the interviews (re-familiarize yourself with what was said and how it was said)
-Read each transcript carefully
Reading all of the transcripts together (if you have more than one interview; this allows you to get a sense of the data as a whole)
Make notes about anything that you find interesting in this stage
e.g., all my participants had difficulty answering the question ‘who are you’, and took a long time to respond
what is a code
A code is a brief phrase that ‘summarizes’ what a particular
section of the transcript ‘is about’
It’s a descriptive label
generating initial codes
You code segments of text that cannot be broken down further into
meaningful units.
- Code everything (don’t assume some things are unimportant)
- Coding is usually done several times – initially the coding is fairly semantic, staying on the surface, using participants’ own words. In a second round of coding, more interpretation (and collapsing of codes) can be done.
- When you’re done, write out your full list of all the codes
- You can code by hand (using pen and paper, or Word), or NVivo
- generating themes
‘Sorting’ the codes into groups (themes)
You’re starting to analyse your data here, and are considering how certain codes might group together to form an overarching theme
Not all codes fit into a group (i.e., into a theme) – it’s ok to have a theme called ‘other’ – maybe you can find a home for these codes later
Something else you can do at this stage Some themes themselves can group together into a new theme – in that case the original themes themselves will now be called ‘subthemes’
Note: you are not ‘searching’ for themes
- reviewing themes
Check that each theme has enough data to support it
->if not, it may not really be a theme in its own right, you may have to discard it
- Check that the codes for each theme are not too divers
->if so, you may have to split the theme into two (or more) themes - Make sure each theme is distinct (no overlap between themes)
- Sometimes several themes can become subthemes within one overarching theme
- After you’ve done all this, think about whether your themes accurately capture
what your participants have said (+ context, + how they have said it, if relevant)
Reviewing themes may take some time.
Try to leave it for a day and then come back to it: you may get fresh insights.
- Defining and naming themes
For each theme, write a summary of ‘what the theme is about’, its essence.
- You do this by looking at the quotes from participants on which the theme is based.
- Paraphrase what participants said and tell the reader what is interesting about these things.
- Use a few literal quotes as examples for each theme
- Make sure the names of each theme are clear: short phrases are best, avoid single-word names!
- Then, think about how each theme relates to other themes
- Can you write a story about your data that connects all of your themes?
- producing the report
Writing up means ‘telling the story of your data’:
* You need to provide enough evidence (quotes) that your themes
adequately capture the essence of your data
* Do not just paraphrase your participants; you need to tell a story,
convince the reader that your research question should be answered
in a certain way, move beyond description to interpretation
