To Begin at the Beginning

It seems obvious: any piece of writing should have a beginning, a middle and an end. But how often have I heard the lack of structure moaned about by those folk who have to read some prose, in particular supervisors reading their student’s work. Not infrequently I have been given a draft manuscript which has seemed to meander all over the place without obvious structure. The discipline of structured writing is often missing from our (English) education system and science students early on in their careers are given little opportunity to write extended prose on which to hone their skills. Getting to the stage of PhD thesis-writing without having had opportunities to practice their literary skills on more modest pieces of text can be a formidable challenge for an anxious student and it’s not surprising some of them struggle.

Clearly, any thesis should have an introduction and literature review to begin with, a middle meaty bit where the description of methods and all those fascinating results (even null results) are described; finally there is the ‘end’, some conclusions to wrap the whole thing up and some imaginative (possibly even fantastical) ideas and speculations about what could be done next by any succeeding student, building on what has been teased out in the thesis.

If writing a novel – say a whodunnit – it seems pretty obvious you need to start writing at the beginning, set the scene and introduce the cast of characters, work through all the complexities of the plot designed to obfuscate just Who Did Do It, and finally have the denouement which Reveals All. Although I have never attempted to write a novel, I find it hard to imagine not writing the chapters in that order. To do so would be immensely challenging to keep consistency of the plot: little clues need to be introduced at the correct moment so they can be developed later; characters have to mature and grow in the text as they would in real life.  Maybe it is possible to write brief vignettes out of order, descriptive passages without key parts of the plot lurking within, but could one write large chunks of the end before the beginning? As I say, I find that hard to imagine.

However, when it comes to thesis writing, I find students often start in the middle. For them, at least some of them, it seems to work. Perhaps they want to start with their results so they know if there are holes that need to be plugged with further experiments; perhaps they want to start with something simple like the experimental methods chapter while they try out their composition skills. If this works for them, well and good. As with so many things in life, there is no single right way to tackle a thesis and far be it from me as a supervisor to impose some rigid structure which merely paralyses rather than assists them. But, and of course there is a but or I wouldn’t be writing this, it does make it hard for me the supervisor trying to critique the writing as it lands on my desk non-sequentially.

If the results are written before the methods, then how can I be sure the methods will actually contain all those vital little bits of detail that certainly haven’t appeared in the results chapter?  A confident student can assure me that of course they know they need to do this but, with deadlines approaching and the student getting ever more stressed, those small points may ultimately escape their attention.  If the literature review, because it can seem straightforward verging on the boring, is left to the end, then all those cross references during the discussion may get hopelessly entangled not to say mangled. It is embarrassing if it is one of the examiners who picks up, either that crucial references are completely absent or, more commonly, the papers listed in the bibliography have got garbled. Modern software ought to make this easy to keep a check on; but attention to detail can get mislaid in the final, frantic push to submission.

So, I still try to nudge students to the linear model. To begin at the beginning so that all cross references are easy to make, back to their first appearance – by reference or section number as appropriate; to maintain a steady progression through experiments of increasing complexity until, as in the whodunnit, the ultimate denouement reveals the coherent story which brings the disparate chunks of work to a logical(ish) conclusion.

I should end by pointing out that these remarks do not relate to any current student I am supervising, or indeed to any particular student from the past. They are general remarks based on reading many theses over the years and were provoked by talking to Elizabeth Simpson, a retired scientist-cum-writer from Imperial College, whom I met last week at an evening celebrating science and mathematics at my old undergraduate college, Girton College, of which she was also an alumna. Being back in the college certainly took me back to my own Cambridge beginnings.


I haven’t been back to Girton more than a handful of times since I’ve been on the academic staff in the University, and it was definitely a case of Memory Lane, both in terms of the sheer atmosphere and familiar architecture, and the faces I saw there. The tutor who had fined me (the vast sum of £10 I believe) after my 21st birthday party – shared with about 10 others –got out of hand in the very same Stanley Library in which I was now called upon to talk about the joy of physics; she was there! The long echoing corridors felt very much the same, down which the sound of girls sobbing (only girls, it was at the time a single sex college) during exam season would echo sending shivers up the spines of those who had not yet succumbed. And, unbidden, the memory flooded back of my own first timid entry into the college back in 1970 when power cuts, I think due to striking power workers, meant my admission interviews were conducted by candle light and we dined off cold spam. That was my beginning in Cambridge. Perhaps it didn’t look auspicious at the time, but it has led to some excellent times in the more than four decades since!


This entry was posted in Communicating Science, Research and tagged , , , . Bookmark the permalink.

10 Responses to To Begin at the Beginning

  1. Sen says:

    Sorry but I completely disagree. Writing a thesis is a scary prospect. Writing one of the chapters with the results and analysis that they’re most confidant about sets them up to write the best thesis they can. My introduction, theory and conclusion chapters were the last things I wrote. I edited them all in order to make sure they linked together but first drafts were results chapters first and supplementary bits after.

    • Sometimes that’s right. Sometimes the lit review can seem easier to a nervous student who’s still testing the waters of writing and trying to make head or tail of their experiments. As I said, i don’t think there is a single right way and each student has to do what works for them (and in my experience that is exactly what they do do). I just think not writing in a linear fashion can lead to problems in continuity that may ultimately be damaging to the overall thesis.

  2. The best thesis-writing advice I got was to do all the figures first, so that I then knew what to write about them. Maybe not a useful approach in all fields, admittedly, but it did seem to help me organize the “plot line” to go around them.

    I found the worst part to be the “summary and future directions”, which I did indeed write last. Writers’ fatigue had set in by that point, and it also ended up being the shortest chapter.

    Finally – would it be cheeky to request a blog post with a little more detail of this “out of hand” birthday party in a library (of all places!)? 😉

  3. Anna Gronert Álvarez says:

    I also have gotten the advice of preparing all the figures first. I hadn’t really considered to do so until now, but I see how organizing the figures in a way that you have a storyline might make the writing of the results part easier.
    From what I have observed, usually most PhD students leave the introduction part to the end, when everything is done (frantically written before submission date). My supervisor strongly advised me to start with the introduction and I have found it very helpful. (It’s by far not done, but then again, I haven’t really started writing writing 🙂 ). It has helped me keep the big picture in mind (helpful when you are in the midst of your research not to succumb to frustration) and to get interesting input that I can still integrate in my research.

  4. Grant says:

    Advice I find myself giving is to try write the research papers that would come from the thesis first. (Too many things I could say here, so I’ll say none of it for now!)

    Personally, I feel the literature review ought to be started well before writing the thesis as something akin to a review paper to aid the student in their work; they really ought to be doing that concurrently with, or prior to, the research work – just my thoughts. Ideally this might be done in the first year and updated informally as they thesis progresses. (In practice, of course, people mostly maintain a mental tally of new papers!) With that in hand, the literature review in the thesis would be an updating and brush-up of the earlier review.

  5. Grant I entirely agree the literature review ought to have been started much earlier. I expect my students to do much of that in their first year. But, projects evolve, emphases may change, and an early draft is bound to need substantial reworking with new references added. If it is done after, say, writing the results, then the issues about cross-referencing I refer to apply.

    Anna and Richard Interesting idea about doing the figures first. I’ve certainly used that route to help students structure papers, never thought of doing it for the more substantial task of writing a thesis but it’s certainly an interesting idea. However, it doesn’t resolve the issues about making sure you have all your ducks in a line for cross referencing that remain a concern to me. It simply may help break down what can otherwise seem a monolithic and impossible task.

  6. BB says:

    gosh I had no idea that was your view….I have definitely subscribed to the deal with the results first as they are the bit you are actually an expert in. Deal with the ‘hard’ things like introductions and lit review later when you have worked up some momentum….

    Abstract and conclusions to be written last – when you actually can stand back and see the overarching themes of your thesis….also they are to be written at 3am the morning before your hand in deadline thus ensuring your supervisor has no chance to check them….

    • I expect you’re right and your supervisor did have no time to read them! As I say, I don’t think there’s a single right way, but I obviously have a personally linear approach to this. It’s probably good you didn’t know that was my view if it isn’t your’s…. your way worked, anyhow…..

  7. Emma says:

    I think a lot of the confusion comes from conflating three different aspects of the writing process:

    * the external structure or conventional divisions (lit review, methods, results, etc.)
    * the argumentative flow
    * the act of writing itself – setting pen to paper, or fingers to keyboard.

    Balancing these three aspects is what makes the creation of a thesis so difficult, and not distinguishing between them can lead to a lot of anxiety.

    I entirely agree that from the outset you need to have an idea of where you’re going , and also be familiar with the conventions that dictate the final shape of the thesis (e.g. the word count and the structural divisions). But these aren’t the same thing! Knowing where you’re going – the ‘big picture’ of what you’re actually trying to do in a thesis – really comes down to three questions:
    WHAT is your argument/claim? (research question)
    WHY does it matter? (research context)
    HOW will you prove it? (method)
    Setting down the answers to these three questions and sticking them on the fridge is the best way I know of to ‘begin at the beginning’ – and recognising that the answers may evolve in the course of the research is an important part of moving on from there.

    However, the act of writing itself – the actual setting pen to paper – is a different aspect of the process. Many students feel hamstrung by the perceived need to occupy an authoritative position, and/or express themselves in “proper academic” language from the start, while they are still at the exploratory and questioning stage of the thesis. At this point it is VITAL to feel free to engage – in writing – with whatever aspect of your topic you feel most confident in addressing, and furthermore to feel you are “allowed” to write in a non-academic, possibly quite messy way, simply in order to get those ideas out of your head and onto paper, so that they can be reworked and developed in the redrafting process. This exploratory form of writing and re-writing (redrafting) is a fundamental part of the thinking process.

    Managing these different aspects of the process is an ongoing negotiation, and even towards the end of the thesis process it doesn’t feel linear; even at the point where you’re creating the strutural and argumentative “pathway” for the reader to follow, it feels less like following a line and more like stitching together a garment, or putting together flatpack furniture : ) The result as seen by the reader should be a compelling linear chain of argument; but the process that gets you there sure isn’t.

    (There’s more background to this, and links to some excellent academic writing blogs, at if anyone’s interested.)

  8. Iolanthe says:

    Aside from the discussion on effective sequences for producing thesis chapters, the much, much more critical point which resonated with me was the need for structure in the finished item! I run a research group in industry, and my heart sinks in sympathy on this. It’s not just research reports, but business cases and powerpoint presentations: a declining number of people seem to be able to see the multiple ‘lengthscales’ required in good communication and that a sound over-arching structure is essential to producing something coherent without avoidable and even dangerous loose ends all over the place. Meandering magpie-constructed acres of waffle drive me mad! And for my own appreciation of this aspect of technical communication, I am eternally grateful to my PhD supervisor twenty years ago.

Comments are closed.