Further Adventures in NVivo

I’ve finally read all the control group autobiographies and listened to all the control group oral histories, so now I can concentrate on something I personally find a lot more fun and stimulating - actually analysing my data!  In the last few days I’ve been using Nvivo to code the last of the autobiographies and put together my Social History Society conference paper.  Some people like it for the charts and graphics you can produce with it, but I prefer to use it as a practical digital alternative to highlighters and sticky labels.  So what have I been doing with Nvivo and the autobiographies?

image

Better image here.
I love the modelling feature on Nvivo.  Not only is it a handy way to visualise the questions I want to answer using my data, but I’ve also used it to make diagrams for other things when I’ve needed to - it’s much easier than using something like Word, where you insert a text box and you can never quite get it the same size as the other boxes and you move one thing just for everything else to do skewiff.  This is my initial map of the questions I wanted to ask of my life story sources.

image

Better image here.
And this is what those questions looked like as nodes (not that they look like questions when you can’t use question marks!).  Assigning bits of text to the relevant node from the original source is very easy - you highlight the bit of text, then there are a variety of ways to relate it to a node.  My favourite is to open up a dialog box of all the nodes and tick the ones I need, but other options are very much available.

image

Better image here.
After I’d coded all the sources, I did even more coding, as having all the responses to questions like ‘did they enjoy school?’ in one place is a bit unwieldy to say the least.  So now I’ve separated out positive, negative, and mixed responses, which is much easier.  I must admit, I do like to use a bit of pen and paper to note down how many of which categories of people (only and non-only me and women) said certain things, and their reasons for doing so.

image

Better image here.
So this is an example of bit of relevant text grouped together under a node - in this case, negative experiences of sports at school.  Yes, I’ve colour-coded the original text so I know at a glance who’s only and non-only, male and female (strong pink = only child woman, weak blue - non-only-child man).  Yes, I know pink and blue are arbitrarily assigned colours for gender stereotyping - but I may as well take advantage of my social conditioning and know what a colour’s referring to without having to think.

image

Better image here.

Another feature of Nvivo I find really useful is the classification table.  As I said the other week, I find it hard to compress biographical details into a quantifiable format, hence I only have a few categories.  I don’t have geographical region as a category, for example, because some people moved around a lot.

image

Better image here.
From the classification table comes a classification report, so you can see how many of each characteristic you have.  This is a breakdown of why people were only children - 31 were unassigned because they weren’t only children, while 41 only children gave no indication of why they had no siblings.  As you can see, though, parental health was a big reason to stop at one child.  I noticed from further down this report that I have WAY more men than women in my study, and it’s a good thing I saw that so I can alter my conference paper accordingly!

image

Better image here.
‘But what about visualisation?!’ I hear you ask.  Well, alright then, even though I’m not a huge user of it myself, apart from the models of course.  This is one of my favourite sources, the autobiography of James Kirkup (though he doesn’t do only children any favours by fitting the stereotype so neatly!).  The coloured lines on the right indicate which nodes a particular part of the text has been coded at, so you can see which questions a particular source addressed most frequently.

image

Better image here.
More visualisation, this time of word frequency.  There are some boring connecting words in there, but also some nice words like ‘mum’, ‘dad’, and ‘home’, which indicate popular topics in the autobiographies.  Actually, the top word as ‘p’, as in ‘p. 146’, so it’s probably a good idea to filter out very short words!

So that’s what I’ve been up to, and will probably be up to for a while yet (though hopefully I won’t let my secondary reading fall by the wayside now I’m actually doing something enjoyable!).  It might be that by not using the visualisation tools much, I’m not using Nvivo to its full potential, but I’m using it in a way that works for me.  Considering it’s more aimed at social scientists, and I actually have to dedicate part of my thesis to justifying using it so that more traditional historians don’t go into paroxysms, I think that’s alright.

Share: Twitter Facebook

About Alice Violett

Reader of books, editor of web content, haver of PhD

Colchester, UK https://www.draliceviolett.com