1. When you start seeing patterns, don't pretend you can't see them, but start actively looking for alternative explanations and patterns that would invalidate these initial insights (p437)
2. Once all the data is in, we have two main ways to organise the analysis: the questions from the design stage, and the analytic insights that emerged during data collection.
3. You can go back to interviewees for more, to enrich or clarify.
4. "Perfectionism breeds imperfections" - "perfectionist and forced analysis ... undermines the authenticity of inductive, qualitative analysis". Analysis will find patterns. But also "vagaries, uncertainties and ambiguities".
5. Description comes first. (p438) Before you can start to answer "why?" you have to describe what you have. p439 lists ways of organizing and reporting data. Maybe for me, it is most sensible to simply organise around the questions i asked (in TMA02 I said i did it this way to make data analysis easier, so why fight it?). Other options, are round sensitizing concepts (community, practice, ...?) or round people (would descriptions of the people be useful?) I think i will group the data around the questions and then see. If too thin, one option is to send the thin data out to those involved for comment. A second one is to describe the individual responses as mini case studies and see how that compares with the other cut. Patton calls this an "analytical framework approach" and says (p 440) "an interview guide, if it has been carefully conceived, actually constitutes a descriptive analytical framework for analysis.