What’s NVivo? you may ask. It’s a qualitative data analysis software program that I use to keep control of the data that I’m gathering as part of my research. You can read more about NVivo here.
I first came across NVivo in its earlier incarnation as NUD*IST, which was a much more playful and memorable name. I am still using NVivo 7 because my clapped-out old laptop here wouldn’t download NVivo 8, and now I see that there is an NVivo 9. Fortunately I use mine under my university’s site licence (I think that it’s very expensive) and for now the older versions are still available.
NVivo (or NUD*IST in its earlier life) was developed originally by Lyn Richards, a sociologist at La Trobe University, and it reflects many of the time-honoured ways that academics work with data anyway- identifying major themes and points (which often found their way onto index cards); highlighting themes in a particular document with different coloured pens, cutting up documents to group all the themes together etc. NVivo does much the same thing, digitally. And because every project is different, and because people work differently, no two NVivo outcomes would be the same.
First you need to put a document into NVivo. This isn’t a problem for me as I type up notes on the computer as I go. Here’s a shot of all the documents or ‘sources’ that I have about Judge Willis in Melbourne. Some of them are full transcribed documents, others are my notes. If I have the document in hard copy or saved as a PDF elsewhere, I might save it as a ‘proxy document’ with a very stripped down content skeleton, with the expectation that I can go back and look at the full document easily.
When you have a document you identify the themes in it. You call these themes ‘nodes’, and it’s just like tagging, or using a different coloured highlighter pen for each theme. You develop the themes as you go along.
So, in the picture about, I might be reading Paul de Serville’s ‘Port Phillip Gentlemen’, and I might notice that a paragraph is about ‘authority’ or ‘class’ or ‘gentlemanly expectations’. I would highlight the paragraph and select the node on the left hand side, or create a new node if it was something that I hadn’t come across before.
This means that you develop a long list of nodes that you’ve identified across all your documents. You can group related concepts into ‘tree nodes’ or just leave them alphabetical as shown below. Because you have developed the nodes yourself, you get to know what is there and move around it quite quickly. When you’re working with a particular document it also collects the nodes that you’re working on as you go along into a drop-down menu, and as they tend to recur, it means that you’re working with a smaller set. But here is my master list as of today of the Port Phillip nodes I’d developed. If I worked on a new document tomorrow and identified new nodes, they would be added to the list.
So if you want to find, for example, all the documents that you had tagged as being ‘beliefs about convicts’, then you can bring them all up onto the one page.
If you click on the underlined hyperlink, it takes you to the source document where you coded it in NVivo. It’s better not to code great slabs of material; just enough for you to get the gist and then go back to the source document for the surrounding material. The real advantage of this is that it means that you don’t forget about material that you read years earlier, especially once it mounts up. It also brings together the primary and secondary material again when the tendency is to develop the tunnel vision of “now I’m working on letters”, forgetting about the insights you’d discovered in secondary sources. You can write ‘memos’, which are your own reflections on a particular point, which can also be coded and thrown into the pot as well.
You can also go back to a particular document (or ‘source’) and at a glance see the themes that you’d identified in it. For example, here’s my screen for Paul de Serville’s ‘Port Phillip Gentlemen’. At p.128 I’d found information about ‘the nature of Port Phillip society’, ‘party split’ ‘Kerr’ and ‘Fawkner’ and coded that paragraph accordingly.
It is intended that you develop the nodes as you go along which means that at some stage they become big, baggy unwieldy monsters. At this stage you need to think- do I need to split this concept into smaller nodes? Or alternatively, you find that you’ve made several nodes that are really talking about the same thing- are they really separate concepts? would I lose some particular quality of the concept if I combined two similar nodes?
It does also mean that sometimes documents you read earlier in your research have concepts that were not apparent to you at the time, but that’s true of research generally. At least with a digitized program like this, you’re likely to come across the documents again under a different code and can go back and add extra codes as you go. It’s intended that it keeps growing and changing. For me, it’s certainly more dynamic than having notes filed in hard copy in folders where you forget you’d ever even seen the document at all. It also means that my own fleeting reflections in ‘memos’ are brought back to life again.
As I said, I’m using an old version and I note that NVivo 9 has fixed up one of the real bug-bears – being able to see the codes while you’re actually coding- which for some reason you could no longer do, even though very early versions of the program did have this feature. It would be worth getting NVivo 9 for this feature alone- in fact, it may even prompt me into splashing out for a new computer. It also claims that you can use PDFs but I’m not sure- earlier versions claimed this too but it only worked for OCR’d or text-based PDFs – not image based PDFs which it inevitably seemed mine were. I usually just save the PDF on the computer and make a proxy document (i.e. a dot point summary)- it’s too time consuming mucking around with it.
The drawbacks? The major one is the fear that the whole system is going to crash and that you’ll lose everything. Also, there is the limitation that you can only get out of it what you put into it- it takes discipline and routine. I type up my notes, save them, print it off (yes, I do keep hard copy- 2 sheets to a page), put it into Endnote, put it into NVivo, code it, tick on the top of the hard copy that it has been endnoted and N-Vivo’d and then file it in a ring folder alphabetically by author. The folders on my bookshelves are multiplying alarmingly. It also has to be an ongoing process- I have to be prepared to go back and fix up the deficiencies in the coding when I happen upon a document that I’d read early on, and sometimes this is a bit distracting. But this is the price of keeping it current. I do have parallel systems: I also tag in Endnote and, to a lesser extent Zotero, which I use for internet-based material and somewhat less methodically.
And the advantages? Particularly with my family history of Alzheimer’s, I’m frightened of losing track of all this! I think that any researcher has this fear, Alzheimers or not. I’m relatively confident that I can put my hands on the main documents fairly easily. When I’m working at a conceptual level, it’s easy to grab together all the examples of a phenomenon e.g. ‘loyalty’, and tease it out further because it’s all in one place. Because primary and secondary sources are intermingled, then I can find concrete examples relatively easily. I’m very well aware that I only use it in a rudimentary fashion and that I could probably do other things with it, but I haven’t got time to learn them, and it works just fine for me.
Of course, it doesn’t always work, as this sad experience shows. I still haven’t found the damned document that I was looking for, and I still don’t know whether it ever existed or whether I read into the document something that really wasn’t there. In the end, I wrote around it and found other evidence that was good enough- but I still live in hope that one day I’ll stumble across it again.