How to Derive Meaning (Not Just Codes) From our Data: Ten Tips

I have recently been discussing the issue of making meaning from your data not just listing lots of themes with a couple of my students...trawling through your transcripts and coding similar phenomenon might be time consuming and important but in the end this type of coding is only a tool to help you think about the depth and complexity of what is really going on in the data.....the aim is to move from CATEGORISING to developing CONEPTS, to let higher order meaning emerge from your analysis.

Here are 10 tips I would recommend to help:

1. Go for a walk, untangle yourself from your data and just see what comes to you
2. Try telling someone what it all means in five minutes....you may already know how it hangs together
3. Draw lots of diagrams and pictures until one comes out that seems to fit
4. Talk to colleagues to get it out of your head and bounce it off someone else
5. Cross reference categories with demographics.
6. Have a look at Negative Cases or Outliers...they will tell you alot about the rest
7. Go to theories in the literature..do they kick start your thinking?
8. Summarise each of your participants findings into one paragraph, lay all of them out on the floor and then compare and contrast these, not the codes.
9. Struggle
10.and keep struggling!

REMEMBER..build an argument or a narrative...not a list!

Here is an incredibly good paper by the master, Pat Bazeley Research Support P/L and Australian Catholic University

Sometimes Coding Process is as Important as Content

We had a very interesting collaborative coding session today in our regular coding party for QRIP, reading a transcript about how men respond to penile rehab after prostate cancer surgery.

For the first time it really dawned on me fully why line-by-line analysis can be so important...even though we weren't there at the interview a careful analysis revealed a great deal..so careful, in fact that we we spent 1 hr 30mins looking at about 15 lines of text...

What we saw was very interesting.....we were not only able to see the content (that the participant had a good sex life before surgery) but the process...that his short distancing statements about his sex life suggested he was finding the questions difficult and that they could have been causing him distress because of his losses.......only a very careful line-by-line analysis would have allowed this..we noticed, for example, that the word "it" was used instead of sex, that he referred to "males" in a general sense rather than himself when talking about the effect of the surgery on his sex life....it raised many hypotheses that would not have been possible by rushed analysis, or maybe even slow analysis done by only one person...

Like therapy it seems important to pay attention to the way words are used not just their content...the process can be as important to understand and code as that content.....it would also seem that focusing on these dynamics are critical when actually interviewing.....the way language is used can guide you about the sensitive decision-making required..do I probe more here? Is it too much for the interviewee? Shall I ask for permission to continue down this track? Is it me that's more scared of emotions than the interviewee? Are they responding to my anxiety rather than their own?

Coding: Don't Forget the Forest

It’s a horrible but inescapable fact that it takes more

time to organise, write and present material well than

it does to gather it . . . The sensible researcher will

allow as much free time to write his [sic] report as he

spent in the field. If he is really astute and can get

away with it, he will allow more time. (Wax 1983,

193–4; in Cook and Crang 2007, 131)

I'm writing up a study at the moment and reflecting on the coding process. After wrestling with the data for a few years Im flabbergasted that its only really when Im writing it up that I have come to fully understand it....Ive looked carefully at every bloody leaf and branch, at every tiny free code as it becomes part of a larger set of thematic relationship, Ive even developed a lovely theory, cross checked it with participants but only now as I have nearly finished writing it did i finally get it.

Coding is a combination of paying close attention to every detail and simultaneously keeping your eye on the big picture...hopefully something bigger emerges than the data..call it 'higher order' themes....some that transcends the data and leads eventually to a creative product rather than just a research report.

Here's an interesting paper relating to this 

Coding then, if you are informed by Grounded Theory can go through the following steps:

1. Immerse yourself in the data by reading thru the transcript many times
2. Free code loosely, not every word but proper chunks of meaning
3. As interviews progress you'll start to commit to codes
4. The hierarchy of codes will start to emerge, ie, which are leaves, which are branches
5. You can then look for differences between participants
6. Complex relationships may then emerge between codes and between codes and attributes of participants
7. Your model emerges and you can fill gaps in through selective sampling of new participants
8. You could then augment your analysis through theoretical triangulation: ie..try and use a variety of theories
9. You can also use investigator triangulation and have a group of people help you code, preferably from other disciplines or orientations.
10. Then go for a walk, a run, have a bath, go clubbing, get drunk, go out for dinner, ride your bike and if your lucky all the higher order themes might emerge