I have knocked my 21 interviews into more-or-less usable form. I now have about 27,000 words of interview response data which is a fair amount to work my way through.

To give me some initial pointers, I have made it all into tag clouds using the very user-friendly site tagcrowd.com

The picture below shows a tag cloud for all my interview data, including the frequency of the words in the tag cloud. Ignore the highlighting (Snag It put that in as it was doing the screen grab) it’s the size of the words which is important.

‘Group’ and ‘work’ are obviously key words, and the fact that this is an ‘online’ ‘course’. Moving beyond the obvious, though, I’m interested in the words which relate to constructing knowledge together: answers, asked,chat, collaborative, communication, discussion, experience….

Nothing’s ever straightforward, is it?

I realised that I asked different questions of the tutors and the students, so ‘question 1’ won’t mean the same thing in every place. So now I have question 1 (the tutor’s question) and question 1s.

I have an interview which intersperses my questions with the answers, so Ive added another style to pick out my questions.

Then the interviewee introduced two tables to her answer. And very useful tables they are. I don’t think they’ll import into NVivo, though, so I have had to rearrange as text. And then she attached a PDF document – and then a Word document. I’ve imported the text of both – but not the formatting, and then I’ve had to add in notes to remind me what I’ve done at each point.

So, I had my eyes tested and realised why I’ve not been getting on with my literature review – I can’t see to read the literature! New glasses on order, so I’m making a start on preparing my data for analysis.

You’d think epistolary interviews would be straightforward to input, but there are decisions to be made, even so. Because somewhere on the back burner I have an article about epistolary interviewing, I’m temted to retain a lot of things which are irrelevant to my doctoral research.

Type size, face and colour, how quickly they responded, whether they interspersed their answers amonst my question, whether they started a new email or hit ‘Reply’ to mine. I’ve decided all that is irrelevant at this point, so everything is being styled ‘Normal’. Same font, same font size, same spacing. This gets problematic when they have mixed their question with my answer, but I’m sure I’ll think of something.

I’m going to put all this in NVivo eventually, so I want it as NVivo friendly as possible, which means thinking about the styling. I’ve used four heading levels to style the copy. That means I can pick out their name, their group, whether they are student or tutor and which question they are answering. I hope these prove sufficient – it’s going to be so frustrating if I find there were other categories I should have added at this stage.

Oh, and I have to pick everyone a pseudonym. Must remember to make them noticeably different this time. Last time I had Carol and Karen and Caroline and it thoroughly confused me.