THE DOCTORATE SERIES (6 OF 7) – ANALYZING DATA, WRITING YOUR THESIS, AND THE ELUSIVE GOLDEN THREAD

“…it takes ages to get any real coherence out of them. Alas, this is the curse of academia.' Terry Pratchett.

"You write to communicate to the hearts and minds of others what's burning inside you, and we edit to let the fire show through the smoke." Arthur Plotnik.

I did wonder if splitting this blog series into 7 would work. On the one hand, how do you reduce 3 years or so of work (roughly 1800 hours of study plus classroom time) into any number of blogs? It’s lived experience and hard to relate to in writing. And yet…that’s exactly what the thesis is as well….except that’s 80,000 words!

You are instructed early on that a doctorate degree is an academic pursuit that hopefully has some real world relevance. The currency of academia is the written word. Consequently, expect to write a lot, be told you need to edit and revise a lot, and come to terms with the fact that your thesis will never really be finished. There will always be another dimension to explore…another edit it may benefit from…a point of emphasis to draw out.

In this penultimate blog in my informal guide to pursuing a doctorate, I plan to share my experience in: (1) developing my overall thesis story board; (2) analysing the data I gathered in ‘the field,’ and; (3) writing this up into a final thesis. Along the way, I will try to explain about coding interview transcripts, description versus analysis, the development of theory, and what is meant by the ‘golden thread.’

Developing Your Core Narrative

As I mentioned in my previous blog, I don’t recommend waiting to start your thesis until you’ve completed you field work. I found it much easier to work them in parallel. Consequently, by the time I finished my field work, I had a long document (about 35,000 words) that represented the rough first draft of the empirical sections of my thesis. However, when I reached the point of trying to reshape this into a more complete dissertation, I got lost….terribly, terribly lost. I remember being stuck for about a month. What helped me get unstuck was a simple exercise in visualisation (see Figure 1 below).

Figure 1: My PostIt Visualization Exercise

My problem was that as I had been writing and summarising my field research, I had lost track of how it fitted into an overall narrative. I had spent so much time wallowing in my data that I was struggling to find a way to make it all flow. So…..I went to office depot, bought some large multi coloured post its. I then completed the following steps:

(1) Write down the key questions ANY thesis must address:

• What is the problem and why should we care?
• What is currently known about the problem?
• What is not known?
• How do you plan to tackle the gap?
• What data did you find?
• What does it mean?
• How does it compare/contrast to what we already knew?
• How does this make a contribution?
• What are the limitations of the research?

(2) I then placed these as a top row and developed a second row which were my chapter headings? My second row were:

• Abstract
• Intro
• Literature Review
• Methods
• Field work
• Analysis
• Discussion
• Conclusion
• Impact

(3) I then placed post it’s beneath these sections which represented the major themes from my work to date. Essentially, I wrote my section headings on post its and arranged them under these headings.

This may sound silly but this simple exercise resulted in a major breakthrough for me. Once I could see the entire thesis in a single visual I was able to see what was missing and rearrange things to fit better. Perhaps you won’t need this, but for me, it felt like a light bulb moment.

The point here, however, is whatever tactics you employ, you will need a framework. Some form of story board into which you plan to plug your data, your analysis, your literature review and so forth. You need a master plan.

Analyzing Your Data

One of my challenges following my field work was appreciating the difference between description and analysis. You would think this is straight forward…but it wasn’t for me. I had built a case study of a company attempting change. I thought I had written some great stuff explaining why and how this company was attempting change. Even what results it had achieved. The feedback I received, however, was that I was heavy on description and light on analysis. This was not said in a complimentary way!

For some qualitative work, I think there is a blurry line between description and analysis. If for example I say “The company decided to implement the following strategy,” we might all agree it is simply descriptive. If, however, I add, based on my observations and discussions with members of the leadership team, the decision to implement the strategy was based on a belief system which manifested itself in the following ways…..” then I think it becomes more insightful.

My learning here, however, is that subtlety doesn’t work. You are expected to deliver a chapter which shares your data and analyses it. Your analysis needs to be clear and distinct. Consequently, I found it was expedient to do exactly that:

(1)  Describe the data you collected

(2)  Explain how you plan to analyse it (I used Grounded Theory Method to code my interview transcripts)

(3)  Provide your conclusions as ‘findings’

(4)  Provide evidence to substantiate those findings

Coding Your Data

I used Grounded Theory Method to code my interview transcripts. What I found very useful was Urquharts explanation of how to use GTM. As a technique, this allowed me to focus on my data and build it up into an conclusion that I could then use to speculate on theoretical frameworks. In short, it worked for me. You will likely find your own technique. The point though is that this is the core expectation from your data analysis chapter. What did you find and how did you analyse it? You essentially don’t get much credit for the finding data without some analysis.

In my thesis, I covered this as follows:

Chapter 4 – Description of my case study subject (history and motivations)

Chapter 5 – Description and ‘light’ analysis of how their change program compared to other contemporary approaches

Chapter 6 – Participant observation analysis, interview data, and hard-core analysis (using GTM) of my interview transcripts.

Chapter 7 – Development of a theory of strategic change in family owned business

My chapter 6 was the longest chapter in my thesis. I shared 5 different tables on how I coded and reached conclusions. One of my tables was 5 pages long and provided examples of quotes from respondents to substantiate my findings.

One final word on coding data…..it can be a long and painstaking process. When I coded my first transcript it took what seemed like a whole day. Once I had the first one done, the others took less time. I think the value though is in trying to take your own bias out of the process. Code each transcript at a sentence level and challenge yourself to see if people really did say the same thing or something else.

Writing Your Thesis

I have already covered the importance of a master plan and visualisation of your complete thesis. So, when you get the stage of bringing everything together, there are only five things left to do. Sadly, they are rather big things:

(1)  Create the integrated draft in your university’s preferred format (I’ve mentioned this before)

(2)  Adapt your Literature Review, Research Design and Impact Plan’s to fit your thesis

(3)  Develop your conclusion and contribution chapter

(4)  Ensure you have a ‘golden thread’ throughout

(5)  Proof read

If your program is like mine, you will have pieces of your final thesis from earlier work. The literature review in particular. However, the problem is that when you wrote those you did them in ignorance of your field research. They were one-off exercises that were written to some extent in the abstract. Now, once you have your empirical research findings, you need to integrate them into your master narrative.

I found that it was easier to build out from the empirical findings. This is to say, now I know what my findings are, how does the literature I reviewed inform those? How does it link? How does it address the question for the reader of what we already knew about this subject in light of what I now believe I’ve discovered? I think its healthier therefore not to think of these previous documents as finished and ready to be merged into the thesis. Instead, I think you have to see them as useful drafts but content that you should not be married to. You have to rewrite them so that they complement and flow into the empirical.

Roughly for me, the re-writes were as follows:

1 – Intro; 20%...I did a good base with my empirical research. 

2 – Literature Review (SLR); 50%....the key rewrites here were:

• To put in some important term definitions which were missing from the thesis
• Reinforce the literature search by doing a supplementary hand search
• Shape the discussion and conclusions to fit the thesis

3 – Methods; I tried to use the research plan (Del 4) but this was a 75% rewrite.

4 – Findings; As I iterated this I found even this changed about 35% as I tried to get the flow right (and based on lots of feedback).

5 – Conclusions; Essentially a new chapter so 100%

6 – Impact; this was the easiest to meld in. Probably 15% rewrite.

If I boiled this down to one thing though it would be write, rewrite, and rewrite. This at least for me was the only way to make it better. I did 17 versions before my internal Cranfield panel and a further 7 before I submitted for Viva. Each revision was a substantial rewrite of at least one but often multiple chapters.

Let's Talk About This 'Golden Thread'

What I found in my feedback, particularly from Cranfield faculty, was the constant debate about whether I had a strong ‘golden thread’ throughout my thesis. My initial interpretation of this was the same as the ‘Developing Your Narrative’ section above. In the end though I decided it was actually something else.

My new academic friends have a certain expectation about how an argument is going to be set out. Certain words and phrases must be defined and then systematically one must address the research question. What I found was that if I subtly signalled something early on but delayed the reveal and analysis until later….I would get this ‘golden thread’ comment coming back. My proclivity was to try to make the thesis ‘interesting and readable.’ One way I attempted to do this was to write in a more need to know order….here is an issue that you need to know and this is why. Later I will pick up on this and that…because it becomes relevant when I share my findings. 

My learning here is that you are being corralled into an ‘expected format’ for an external examiner. They will not have had the debatable pleasure of working with you and likely, other than the two or three days they spend working on your Viva, will not see you again. They want things in a familiar format and order. So…..

(1)  Layout your arguments clearly in your intro

(2)  Make sure your research question is shown early (within the first 4 pages) and is consistent throughout

(3)  Define ALL major concepts, theories, and terms before the end of your literature review. Anything that is important and may come up in findings or conclusions needs to be clearly defined here.

(4)  Make sure every thread you dangle is concluded…what I mean here is don’t write any throw away lines like “Contrary to what you might expect…” or “Unlike conventional wisdom…” unless you are prepared to close those loops and explain what they are.

(5)  Reasonable repetition is OK. Saying at the beginning and end of chapters how the contents fit into your overall narrative and why they are important is helpful.

(6)  Answer the research question….obvious perhaps…but you’d be surprised that it may not be obvious to someone reading your work for the first time that you have unambiguously answered the question.

Proofreading and Editing 

I think I’m probably average when it comes to writing. I write a lot for work and certainly wrote a lot for my doctorate. What I found was that I made many typographic errors. Some were obvious, others were darn tough to spot. I debated how to tackle these. Some suggested using editing services. One of my friends though scared me off this. Consequently, here is the system I ended up using:

(1)  Get a complete draft finished where I was happy with the core narrative

(2)  Review it twice myself on screen (once the day I wrote it…the second the next day)

(3)  Print a copy, read it and mark it up manually

(4)  Go back and edit the copy based on 3

(5)  Send it to friends/advisors

(6)  Correct all the errors they found (some were mildly embarrassing)

Wait a week

(7)  Print the whole document

(8)  Review the hard copy manually

(9)  Correct the errors

Wait a week

(10)   Have a friend read the thesis aloud to you

(11)   Mark it up based on what sounded wrong

(12)   Correct the errors

I spent endless hours editing and found even then, I would look back later and find more. My conclusion was it probably may never be 100%...there would always be a word or a sentence I could have written better. So, the job is to get it to 99%. The point at which when you read it you don’t cringe when you hear a sentence.

A Few Words About Developing Theory

Not all dissertations will necessarily involve theory. However, it is helpful if they do! In my experience, the development of theory has a mystique around it. I remember distinctly becoming even scared of mentioning theory development for fear that I would be chastised as unworthy and ill-equipped. In the end though, I found theory flowed from my findings as follows:

1. Here is the problem I’m investigating
2. Here is what we currently know and believe about (1)
3. Here is what data I found relative to (1)
4. My data/findings suggests a different conclusion or perspective about (1)
5. Consequently, given my findings, we could adjust (2) what we currently know in the following ways….
6. In making the adjustments for my findings a theoretical contribution is made by either adjusting or replacing a current theory or understanding with a new one.

In short, don’t be too scared of this. I would advise caution about making large scale theoretical claims. These just call out for criticism. However, I found in the end that theory simply flowed naturally from findings when you compared them to existing literature.

 Conclusion, Summary, Tips.

Let me wrap this blog up again with some final thoughts and my list of key reflections. Where conducting the empirical research was fun, writing up the results, while still more fun than the early stages of my doctorate, had some rather frustrating components. In particular, the coding and writing stages can feel repetitive. Moreover, sometimes you get advice or critiques of your work that are designed to be helpful but can result in massive insecurities and (at least with me) bouts of existential angst.

Beyond anything else, when you get to this stage of a DBA you are hitting the point of the marathon when you have little energy left. It is about perseverance and doggedness to some extent.

What I found helpful at this stage was:

1. Visualise your overall story board
2. Come to terms with the reality that your thesis needs to fit a set of parameters and format…..don’t waste too much time fighting that…it’s pointless.
3. Understand the difference between description and analysis
4. Be thorough in your coding and whether you use NVivo or something else…keep your workings clear and easy for you to refer back to.
5. Write, re-write, re-write some more….
6. Don’t get married to earlier versions of your SLR, research design, or impact plans….you are better to start from your empirical and build back out so that everything fits.
7. Find some kind and patient people who are prepared to read your work….at least one of them out loud.
8. Accept your work will probably never be 100%. When you get to 99% be prepared to let it go….

To read more of my blogs please click here to visit my website.