How do you write a PhD thesis?

This question sounds like the start of the old joke: “How do you eat an elephant? One piece at a time”.

A PhD thesis is a document which must explain, summarise and defend three or more years of research in your chosen field, often running to 40,000 words or more. Despite the daunting nature inherent to writing such a long piece of work — where internal consistency and coherence are paramount — the same answer also applies to writing a thesis: break the task down into small, bitesize chunks.

After writing my own thesis, I came up with some guiding principles that I realised helped me finish writing with plenty of time to spare before the deadline, and in this post, I’ll explain those principles in easy-to-digest pieces.

The eight principles are:

Get your house in order

This step lays the preliminary groundwork to the thesis, and can be done long before you start writing in earnest, perhaps even months in advance. First, think about the physical tools you will need to write: a laptop or computer, perhaps an additional monitor and keyboard if your laptop keyboard is uncomfortable to type on, textbooks, your own handwritten notes. Know where you will get these things from if you will need them.

Second, think carefully about the software you will use to write the thesis. Cosmology is a mathematical field, so a text editor that can handle equations was imperative for me: LaTeX is the industry standard. It’s so powerful, and produces such nice results, that I would recommend its use even if you’re not writing any equations. You can use LaTeX via Overleaf, which is essentially the Google Docs of the LaTeX world, but it can be slow for very long documents. I prefer to use a local editor, my favourite being TeXstudio.

One key component of the thesis is the bibliography. You will hopefully have been collecting references to papers, conferences and books as you have conducted your research, but if not, now is the time to start. Since I used LaTeX to typeset my thesis, I used BibLaTeX for my bibliography. I collected all my references (nearly 350) in one .bib file and ensured they were all consistently formatted, for example checking that no dates or journal titles were missing, and that all diacritics were present in the names of authors. This is a time-consuming task, but very worthwhile, as it refreshes the literature in your mind and makes your final bibliography look polished and consistent.

Finally, think carefully about the content of your thesis. In my field it’s common to base the majority of the research or results chapters on papers that you have published during the PhD, but in other fields it’s more common to do it the other way round, turning thesis chapters into papers. Either way, these chapters should be the easiest to write, as it’s your own work.

What I personally found much harder was planning the introductory material, which for me made up the first two chapters. I wanted to be as discursive and expository as possible, but without going down too many rabbit holes of unnecessary detail. To strike the balance, I collected as many theses in my field as I could find online (in the end, about fifteen) and read (or at least, skimmed) their introductions to better prepare for writing my own, by getting a feel for what topics people explained in depth and which were commonly left out. Of course, you have to tailor your introduction to what comes later in the thesis, too.

Decide things once, and never again

Of all the steps, this is probably the the most fundamental: make decisions, and don’t change your mind. If you’re like me and prefer to work close to a deadline (more on this below), you don’t have the luxury of time. Trust yourself to make the right decision the first time round and stick with it.

What sort of decisions am I talking about? Pretty much everything: when to start writing, when to stop writing (and I’ll expand on these points in a moment too), what content to include, whether your figures are camera-ready, whether you’ve included enough background material, whether you’ve explained your methodology clearly enough… the list goes on.

The scariest and most exhilarating part of thesis writing is the realisation that it’s all on you. It’s ultimately your decision what goes in and what doesn’t and it’s imperative that you trust your own decision making.

Know when to start

My self-imposed, unofficial deadline was in December last year (the official one was at the beginning of January, but I wanted a proper Christmas break) and I had a moment of panic in October when I sat down to properly start writing and realised just how much I needed to get done in a relatively short space of time. This stress was stopped in its tracks when I decided to put everything else aside (I was still working on two research projects at this point) and do nothing but write for two solid weeks.

I had already set up my LaTeX document, sorted out most of my bibliography (the bulk of which comprised the references from the three published papers that made up three of the four research chapters of my thesis) and pasted in some of my own personal notes that I’d made on various topics over the years. In the end, I either discarded or heavily edited most of this initial content, but it was a great morale boost to not be starting with a completely blank page. You can do the same, even if you only start with bullet points of the topics you want to cover.

I had actually finished getting everything in order by mid-summer and I was tempted to start writing in earnest at that point, giving myself about six months of writing time. But this would have almost certainly been the wrong approach for me. The deadline was not close enough to allow me to remain motivated. I needed that little frisson of October panic to focus my mind and to sit down and write. I think Parkinson’s law very much applies to my approach to work: the duration of a task expands to fill the time allotted to it. I knew deep down that it was not going to take me six months to actually write the thesis. In the end, I wrote the vast majority of my two introductory chapters, the preface and the conclusions in the space of those two weeks in October.

Know when to stop

This is equally important as knowing when to start. Once you have a complete draft of the whole thesis, you obviously need to proofread and edit it. The questions that should be at the front of your mind at this stage are “does this make sense?” and “is my method/result/conclusion clear?”. The thesis is a narrative, and it’s helpful to signpost the reader with linking sentences as much as possible. Refer to previous sections wherever you can, and repeat your main questions and answers as many times as you dare.

One of my personal goals for my thesis was to make it actually enjoyable to read and for the chapters to flow naturally so that it could be read from start to finish. This meant having a strong opening — I wrote a three page preface in which I laid out the historical achievements in the field and presented three questions that I would attempt to answer in the thesis — and a narratively satisfying conclusion. I deliberately went a little further than concluding chapters I’d seen in other theses, discussing some methodological flaws common to the whole field, and what can be done to mitigate them. I wrapped up the thesis with a reiteration of the three questions I’d posed in the preface and how we still don’t have the full answers — which is a good thing from a scientific perspective. I was expecting my supervisor to tell me that the opinions I’d espoused in the conclusions were superfluous or detracted from my scientific results, and to remove them, but fortunately he liked them, and so they got to stay. I’m glad, as it’s probably my favourite part of the whole thesis.

I could have gone further and written more in almost every single section of every chapter in the thesis, but there comes a point when you have to be happy with what you have written and make peace with the fact that you cannot say everything you want to (or even need to). Sometimes a footnote or a reference or two to direct the reader to more information on a topic is all you can provide, as you have neither an infinite amount of time in which to write nor an infinite word limit.

I recommend setting deadlines for yourself throughout the writing process for finishing sections, chapters, figures and other content, and especially a very hard deadline about two to three weeks before you intend to hand in. After this deadline you are no longer allowed to write new content: you can only proofread and fix typos.

This is also the time for your supervisor to read through the whole thing and give you comments (though hopefully you have been sending them chapters as you finish them and they will therefore have very few new suggestions). Your supervisor could (and should) give you plenty of feedback on both the content and style of what you’ve written, but you are allowed to disagree — though try to have the wisdom to realise when they are right, and don’t just argue for argument’s sake.

Write what you know

This old cliché really does apply. Don’t pretend to be something you’re not. Write in your own words, your own voice and your own style. Write what you know about. In fact, I took this one step further:

Write what you wanted to know when you started your PhD

Near the end of my write-up I was struck by just how much I’d written and how confidently I’d done so. I couldn’t help but think back to the start of my PhD three years previously. My knowledge of the field at that time was limited to a sketchy understanding of general relativity, the standard model of cosmology and some quintessence dark energy models, content that I covered in about five pages of my 220 page long thesis. I felt horrendously out of my depth and wondered for a long time (perhaps close to two years) if doing a PhD in cosmology was the right thing for me.

When it came to writing, then, it was as though all the stuff that I’d been learning almost by osmosis over the preceding three years suddenly came pouring out. The best way I can describe it is it was as if all the papers I’d read and all the talks I’d attended got mulched down into compost in my brain, and then finally, in the last few months of my PhD, things started to grow.

So, I wrote my thesis primarily for myself. I wrote down everything I thought I should have known back then, explained everything that I’d wanted to understand, in my own words and in a language that I could follow. I included asides and footnotes and extra titbits of information that I’d picked up along the way. I knew that I could not make it as pedagogical as a textbook (and indeed, that would have been the wrong tone to aim for), but I could teach myself in the process: the present me, but also the me from three years ago, fresh out of undergrad and wondering what on earth I’d got myself into.

Be consistent and coherent

Hand in hand with decision-making are what I think are the two most important things to consider when writing a PhD thesis: consistency and coherency. The whole thing must hang together and make sense when read from cover to cover.

To help myself with consistency, I created my own style guide, in which I listed all the decisions I’d made to maintain it: how I was going to hyphenate certain words, how I would spell certain words (parameterisation was my particular foe), how I would typeset tables and long equations, whether I would capitalise words like “section” or “chapter” when referring to other parts of the thesis, which acronyms I would use and which I would write out, and many other small details. They don’t seem like much on their own, but taken together they make the difference between a thesis that’s technically well-written and a thesis that’s actually easy to read. It doesn’t matter so much what you decide, as long as you are consistent.

In terms of coherency, I read and re-read my thesis a lot. I read individual chapters and sections in close detail, to catch typos and small mistakes, but once I had the complete draft I printed it out (thank you free university printing) and read it front to back in one sitting, to hear how it sounded in my head and whether it flowed together nicely. I did this a couple more times before submitting, paying particular attention to the sentences at the beginnings and ends of chapters.

Remember, it’s yours

And finally: take ownership of your thesis. It’s your thoughts, ideas, results and opinions. Frankly, it’s your chance to show off what you’ve learned, not only about the background of your field, but what your research has contributed to human understanding of the world. Don’t squander that chance. Aim to produce a document that you will be proud of in years to come, that your own future PhD students could read and understand, and that your peers would be happy to cite in their own theses or papers — and don’t settle for “good enough”!

I turn coffee Lapsang Souchong tea into code; if you found this post useful you can contribute to my habit here: https://ko-fi.com/nataliebhogg.