Thread by @timbocop

Thoughts on themes & results in #ThematicAnalysis. There are many ways to think about this, here's mine (currently).

Themes don't emerge, but also themes are not your results. They're ways of organising data that help you tell a story about some aspects of your research area.

Themes aren't right or wrong, they're just more or less interesting. Themes like "barriers" and "drivers" aren't *wrong*, they're just usually not conducive to interesting stories. Maybe they suggest a need to keep looking for themes that will help tell a richer story.

Look for themes that you wouldn't have thought of before you started analysing the data. Look for themes that surprise you. That's where the really good stories are.

Your results are really a story or stories about your data. In writing results: when you put an extract or make a point under a theme, the relevance of that extract or point to the theme needs to be clear. This is where explicit explanation via interpretative text comes in.

If linking between a point and its theme is proving tricky, the point/extract might be in the wrong place or the theme might need adjusting. Or you might need a new theme.

Or you might need to leave out the point or extract because it's not part of the story you're telling with this paper. Don't shove it in where it doesn't belong: that muddies the story and the story matters more than any single quote or point, insightful as it may be.

Judgement and familiarity with the data and research area / question / context are needed to test whether a story reflects the dataset, even if it is necessarily selective and emphasises some aspects more than others.

Reflexivity is not an add on; a section of writing put in at the end. It should permeate this storying process, testing your unfolding ideas against your position, power, prejudice and subjectivity, and against the wider dataset.