By Susan Carter

At lunch the other day an academic friend declared that it was absurd for students to ask how many references were needed in an essay or thesis, as though you tallied these up quantitatively.  ‘How long is a piece of string?’ was her standard response.

I didn’t agree; over time I’ve taken a different tack on this issue. Measuring out that string quantitatively gives another way to think quite deeply about thesis writing.

This insight was triggered when I’d organised a panel of academics for a workshop on thesis writing and a doctoral student asked them how many references were needed in the works cited list. He had over a thousand—was this too many? And a professor from Engineering almost instantly replied ‘Yes, it is too many. You need about 200.’

I was startled that an academic made a quantitative suggestion so surely—tallying numbers seemed at odds with the requirement for critical evaluation and analysis.

Yet I could see that it was in that instance helpful. A thousand references would signal a lack of the critical evaluation needed to choose wisely— and instead of saying this and embarrassing the student, the academic answered in the terms of the question. The student was also in Engineering, so they were talking the same language.

Getting a sense of ‘how many’ helps judgment of how the research is positioned relative to what is expected in the thesis. If there is a general convention that a thesis in this discipline has round 200 items in its works cited, knowing this begins to scope what can be in there, and what might not be. What isn’t really central enough for citation can be seen more immediately.

It’s a bit like packing a small suit case: you choose what you will be wearing in advance because the spatial dimension dictates. You MUST decide what your criteria are for choice and then adhere to them.

So after thinking about this principle I suggest thesis writers try taking a bean counter approach to the task. How many words in total are desirable? How many chapters are needed? About how long will the introduction and conclusion be? How detailed will the description of methodology need to be, and then if methods are to be contextualised within a methodology, how long is that likely to take? If the introduction is to be no more than 4000 words, what tasks must it do within this limit?

In New Zealand and Australia usually the total may not exceed 100,000 words, with an unspoken assumption that around 80,000 words suffices for an average thesis written in prose. Thinking through the length factor bit by bit may reveal that the scope of the thesis needs to be cut back to ensure delivery of the right length of string, or at least a reasonably suitable piece of string.

Honing questions about quantity increasingly finely can show more clearly approximately how long there is for each little section or chore. surety about the limitation on how much detail there is room for gives a more direct route to writing each section. It breaks it into doable pieces.

It may become apparent that there are only 1,000 words to describe the international context of a local problem, for example. References to literature might signal that literature was reviewed appropriately, but the sentences will possibly need to paint a broad brush picture in order to fit.

But in practice I find a sense of limitation can do more than render things doable because in small chunks: it can also give traction with the whole cognitive process of deciding what will be in and what won’t be. It compels you towards the importance of each section in relation to the other sections, so towards a structural overview and understanding of the work as a whole.

This is my second blog prompted by a good rigorous debate with colleagues (see ‘Thesis length: what’s the limit?’ below) and I wonder if you have more to say on the topic of bean-counting or not. Do you work on writing exercises or support by thinking with numbers? Or oppose the idea strongly?

For more on thesis structure see too

Carter, S., Kelly, F. and Brailsford, I. (2012). Structuring your research thesis. Houndsmills UK: Palgrave MacMillan.

And as a postscript for those who were interested in the earlier poll on restricting thesis length, the results this morning stand at 81% students (n25) agreeing restriction is helpful and 80 academics (n43) agreeing that it is.