You may have your data, but do you have your data analysis plan?

25 Feb

Fast forward from 28th January (my last blog post) to today – supervision this morning with my data analysis plan in hand. The outcome? I can now get back to my data – nine months after I left it. I really can’t tell you how good that feels!

The development of my qualitative data analysis plan was that last little hurdle I needed to get over before I could return to my data, but it wasn’t straightforward. At times it even felt a little uncomfortable. However, forcing myself to ask very focused and specific questions about my research has enabled me to be pretty damn clear about how I am now going to deal with my data. Backtrack well over nine months ago when I had actually started analysing my data – I had jumped right in with little insight as to what I was looking for and why.

So… after writing my last blog post, I spent some time actually looking for examples of a qualitative analysis plan or even some indication of how to develop one. Despite lots of literature telling me that a data analysis plan is extremely important, I couldn’t actually find evidence of one. I found lots about different analytic strategies/frameworks to use – but this is not an analytic plan. From reading various papers and books and talking to people on twitter, I came up with some key questions that I needed to answer – which subsequently formed my analytical plan:

idea, plan, actionWhat are my research questions? (Extremely important as we know, but I needed to put a lot more thought into them)

What is the overall purpose of my analysis? (If I didn’t know the purpose of my analysis, how was I going to plan how to do it? This linked to my overarching research aim, philosophical and methodological approach in addition to the clinical relevance of my research – the so what question)

How does each of my (large) qualitative data sets connect to my research questions? (I needed to ensure that by analysing my data sets, my research questions could be answered – this was easiest explained as a diagram). This part also made me realise and accept that I have a lot of data that will not feature in my PhD. This is quite ok though as I can use them for future potential publications)

How am I going to conduct the analysis? (The trickiest one for me because I have 4 large (different) data sets, all gained by different methods – yet they are all connected in some way. Further discussion with my supervisors was required for this part. My two options were (1) Will I analyse each data set separately then pull them all together in a discussion? Or (2) Will I analyse them all together and present them in one findings chapter. Considering each option carefully, I listed pros and cons and talked through them in depth with my supervisors. Due to a number of reasons we all agreed on option 2. Within this section, I also had to consider my analytic framework. Again for a number of reasons, Miles and Huberman’s was the winner.

What do I need to ask of my data? (This was also a tricky one because I need to ensure that I have both a deductive and inductive approach to my analysis. I must ensure that I am not ‘blinded’ by my theoretical framework and that it is only there as a scaffold for my analysis. The development of some initial broad propositions helped me with this. This also connects closely with how I am going to conduct my analysis – always keeping my research questions at the forefront of all my decisions).

What resources do I need to conduct my analysis? (This was a simple one – time and patience! Also NVivo – lucky to have a licence with my institution and was fortunate to habe been funded to attend NVivo training).

What is my timeline? (Given all the issues that I have had to address over the last few months, I do not wish to be a slave to the calendar. However, having a rough idea of time is not a bad thing. I am looking at analysing and interpreting for the next 10 months – but who knows! As a part time PhD student, I am extremely fortunate to have a generous amount of protected study leave each year (in addition to a whole semester during the end stages))

So here I have presented important questions that I studied carefully in order to develop my analysis plan. I found this exercise to be extremely beneficial and it also helped my vocalise and rationalise my thoughts and ideas with my supervisors. I am now very excited about finally being able to return to my data, reading my transcripts and listening to my audio recordings (even though I cringe hearing myself ask strange questions, ask 3 questions at once, not letting participants finish, laugh inappropriately, not pick up on a cue, sometimes not make any sense, or just sound ridiculous!)

On a final note, at the end of my analysis plan, I have this – which I like very much:

Lyn Richards (2005) and Bazeley (2009) list five key signs which indicate that the analysis is sufficient:

• Simplicity: a small ‘polished gem of a theory’, rather than a mere pebble of truism
• Elegance and balance: it is coherent
• Completeness: it explains all
• Robustness: it doesn’t fall over with new data
• It makes sense to relevant audiences

Thank you for taking the time to read. Did you develop a similar plan or have you approached this differently. I would be very interested in hearing different/similar approaches so please do leave a comment ….


9 Responses to “You may have your data, but do you have your data analysis plan?”

  1. Julie February 26, 2013 at 6:37 pm #

    Hi Emma,

    I had the same issues and developed a similar plan. Here are my comments. I hope these help you in some way:

    What are my research questions? My research questions were developed from my literature review. These did not change during my data collection/analysis phase.

    • What is the overall purpose of my analysis? To help me during my analysis I created nodes (folders) for my three research questions. As I analysed my data, I placed the findings into the relevant folders. I kept in mind the philosophical underpinnings of my research (interpretivism/constructionism) and that the findings were the perceptions of the participants (not my own). I continually had to keep reminding myself of this during the analysis.

    • How does each of my (large) qualitative data sets connect to my research questions? Some of my findings were not used in my thesis but I could use these to develop additional papers.

    • How am I going to conduct the analysis? I analysed the different data sets separately at first. For example, as I started to collect observations in the first month, I analysed the findings each day data collection proceeded. In the second month of data collection I started the interviews. I kept the analysis of the interviews separate from the obs at first. As the analysis proceeded, I combined the findings, but I could easily identify the different sets of data. When I wrote up my findings, I placed them in one chapter. I used different examples to show that I had triangulated the data, explaining how the methods I had used complimented one another, providing different pictures of what was going on. I used Miles and Huberman however I also used bits from framework analysis (Richie & Spencer, 1994), and Lyn Richards (2005) and Bazeley (2009). This is fine too, provided you justify what you did.

    • What do I need to ask of my data? I explained during my discussion chapter that this way of analysing my data was deductive (I was mapping my findings onto my research questions or propositions). It was inductive because I allowed categories/themes to develop and I tried to put my theoretical framework (a psychological framework I used to analyse my data) to one side during this process. Only once my themes had emerged did I try to map these themes/categories onto the psychological framework. This allowed new ideas to emerge which may not have, if I had mapped my findings directly onto my framework. This produced new findings (my original conceptual & empirical contributions). (See also comments provided in: What is the overall purpose of my analysis?).

    • What resources do I need to conduct my analysis? I had a licence to use Nvivo with my institution and had funding to attend NVivo training. I also read lots of books about Nvivo analysis, used online tutorials provided by Lynn Richards about how to analyse your data and learnt as I went along during the data collection/analysis phases.

    • What is my timeline? I started analysis as I collected the data. Doing this full-time I estimate that it took me 8 months full-time to analyse my data. It surprised me how long the process took. Other people told me that it takes longer than anticipated as it is a messy process. But I did enjoy it.

    I cringed on occasions when listening to my taped transcripts because I realised that I had not let participants finish a sentence and I had potentially missed a line of enquiry which could have been interesting. I included this during my discussion.

    I hope this helps!


    • Emma Burnett February 26, 2013 at 8:07 pm #

      Hi Julie – thank you once again for taking the time to provide such insightful comments and feedback. I can certainly relate to a lot of your analytical plan – it’s very reassuring. I liked how you analysed your data sets seperately in the initial stages, but then combined them as the anlaysis progressed. As I also have different data sets to analyse – naturally this will occur sequentially, so I suspect I will also start by analysings seperately before combining (it’s probably a more managable way of approaching it). I loved how you also cringed listening to yourself on audio!! Thanks again Julie – as always, I have enjoyed reading your response 🙂

  2. mzsura February 28, 2013 at 3:30 am #

    I also cringe listening to myself! Not only do i realize little conversation mistakes but its just awkward listening to myself talk lol

    • Emma Burnett February 28, 2013 at 9:29 am #

      I’m so pleased I’m not the only one! It’s definitely awkward listening for me too! On the plus side though, it helps us to improve as researchers 🙂

  3. Dudu Ndlovu (@mandlods) March 1, 2013 at 4:57 am #

    Thank you for an insightful post. I am at the data collection stage, I am going to adopt this idea, I realise already how important it is to map out a strategy otherwise it’s easy to lose direction and focus

    • Emma Burnett March 1, 2013 at 3:23 pm #

      Thank you for your feedback – I’m glad it was helpful. Having already been at the data analysis stage (and not really understanding where I was going with it), this has certainly given me much more focus and direction as I return back to my data. Hope you are enjoying your data collection stage (it was my very favourite part!). Look forward to hearing how you get on 🙂

  4. jterry201 October 22, 2013 at 11:40 am #

    Hi Emma, this is so helpful, many thanks. Have recommended your blog today to the Mental Health Nurse Doctoral Students’ Network today. There is such a lot of useful information here, thanks for sharing it

    • Emma Burnett October 22, 2013 at 11:45 am #

      Thank you for your kind words and for recommending to the Mental Health Nurse Doctoral Students’ Network. I am really pleased it was helpful. If anyone from the Network wishes to ask anything or get in touch, they are more than welcome!


  1. #AcWriMo 2013 | Candid PhD Talk - October 31, 2013

    […] found this post by @EmmaBurnettex on her data analysis plan very useful in helping me think about my data and how […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: