Now that I’m supervising students of my own, who are writing their first papers, I think it’s timely that I reflect on some of the things I learned from mine — whether the easy way, or the hard way…
By the time I got up to writing my first paper, I had an Arts degree, a Science degree with Honours (requiring a short thesis), and I’d been researching for a couple of years. I’d presented eight posters, and given two talks, at conferences and workshops — and along the way, written some large number of essays, reports, etc. I thought my communication skills were pretty good; well above par. Boy, was I in for a shock.
You see, my English teachers in high school were right: there are many forms of writing, and each one needs care, attention, and practice. Scientific paper writing is definitely one of these. (So is blog writing, and I’m practising that right now.) All of the exercises through undergraduate chemistry, mathematics, and physics courses had served well in giving me a good base to work from (and the first draft reflected that), but I was not prepared for the level of rigour I was about to experience in the drafting process. How could you be, when no tutor or lecturer is paid enough to take that much time improving your undergraduate work?
As far as the minutiae of spelling and grammar went, there were no problems. My education had taken care of those. The fun began with sanity-checking the mathematics behind the equations we were using in our analysis, particularly the sign conventions — which in this case, meant the difference between one candidate model being considerably more likely than the other.
We were either saying something so true it (as many other things do) looked almost trivial with hindsight, even though it wasn’t clear at all beforehand, or we were saying something so spectacularly wrong that embarrassing would be an insufficient descriptor. So we checked it. Double-checked it. Checked it again. Had one-on-one discussions. Drew pictures on whiteboards. Had group discussions. Drew more pictures on whiteboards. Then we checked the words describing the equation to make sure they were consistent. Someone got confused, so we went around the loop again just to be sure. And we were sure, the best kind of sure, the kind where the good people have checked it out and they all agree, so we moved on…
To a paragraph in the introduction. Were we falling into the trap of assuming too much knowledge on the behalf of the reader? It’s a slippery slope as you accumulate knowledge in a field: you forget the existence of the foundational knowledge you spent years attaining before you understood the concepts and contexts that now seem trivial, and which you now regard as “foundational”.
Whilst each successive draft came with comments throughout, each led to changes that had to be checked themselves. Sometimes, we’d just try writing a paragraph another way, and it would be beautifully succinct and clear. Other times, cutting the number of words down meant that the meaning was difficult to interpret — so we’d go back to the original. Occasionally, rewriting one paragraph had effects rippling through a whole section, as better ways of expressing a key concept emerged.
Sometimes, we’d critically look at the logic behind the paper, and restructure things to hopefully make the links easier to follow. This meant reordering paragraphs, and sometimes swapping sections around, just to see what it looked like. We didn’t always like the result.
Oh, and by this point, after ten or fifteen drafts and six months, we’d forgotten just enough of the reasoning behind those equations that someone asked someone else about them again. Which led to more one-on-one discussions, more diagrams on whiteboards, more group discussions… and another draft attempting to more clearly explain that idea.
Eventually we settled on the “final” version, which was submitted to Physical Review B. Of course, it had to go through the review process — two or three external scientists (referees) got to go over the document and ask their own questions, make their own suggestions, and issue their own demands. So I got to draft the work again. Of course, my collaborators wanted to double-check how I’d changed the manuscript in response, so I got to draft again. Twice. Then there was the letter I had to write to the journal’s editor (and referees), explaining our responses point-by-point and including details of those changes we’d made to the manuscript. That letter got drafted several times as well — both for scientific reasons, as we debated our way through the physics, and diplomatic reasons, as I learned the hard way that what is clearly written to me is not always evident to others.
The end result, of course, was published, and I’m still happy with it today. The science is still interesting (I’m still involved with experiments in that area), and I’m happy that even with several years’ separation, I can easily understand the writing. So can others, who have cited our work in their own.
I have to say that the process, though seemingly agonising and inefficient, was incredibly useful to me (and sensing this at the time kept me sane). Having spent many years after continuing to write and draft papers, there were key lessons from this experience that I think any fresh scientist about to make those opening keystrokes in their first manuscript can take away. I certainly used them to ensure that the next manuscripts enjoyed fewer cycles around the author group.
The writing:
- No matter how good you think the draft is, there’s always a way to improve it. Not all of these are necessary.
- Take the number of drafts you expect to write, and double it. Double it again. Now you have a number that you can be pleasantly surprised about if you don’t reach it.
- Before you write, start with a skeleton of the logic as you see it with the figures you will use to support your arguments. A separate tree diagram often helps here too. You can worry about fleshing it out with formed sentences later.
- If it’s your project, and you’re the main author (usually first in the list for scientific documents), don’t be surprised when your collaborators/supervisors can’t recall little details from that thing you did months or years ago and mentioned once in a meeting. Be prepared for them to insist on including those details in the manuscript so other readers can follow your logic too.
- Don’t be afraid of trying different ideas about writing out. Sometimes, you have to see the results to properly gauge their success.
- Make sure you have another task you can switch to when you get frustrated. You will get frustrated at some point.
- Figure out why you get certain types of comments. Be alert to these causes when you read over your next draft before sending it around.
Working with collaborators:
- Stay in contact enough to know your collaborators’ time pressures — are they likely to have time right now to read your work thoroughly enough to give useful comment?
- Sometimes, you can get two drafts through one person before another has the time to read your work. If this is the case, make sure they have the up-to-date version.
- If you need comments soon to flip another draft before something comes up in your schedule (holiday, other tasks/deadlines, etc.), talk to your collaborators and let them know (nicely). They’re not mind readers.
Having written a few more since then, I’d add these to the lists:
- Keep a running list of things you admire/hate in the way other papers are written, as a reader (do the same for posters and talks). Compare your list periodically with a colleague to gain perspective.
- Your more experienced co-author(s) will make the call on when to submit for the first few papers. Make sure you understand how they make that decision, and what key indicators they look for. If you have several, note how their reasoning agrees and contrasts, and make your own critical evaluation.
- Be prepared for some co-authors to sign off several iterations before others.
- Figure out which type of comment each co-author gives best. If you need something particular checked, ask the appropriate person.
- When you work with new authors, expect the number of drafts required to go up until you are used to the new perspectives on writing and science, and they get used to yours (by writing more papers with them).
And finally, probably the most important point: just because you get good at writing scientific papers doesn’t mean you will be any good at writing in another form, if you don’t take the time to appreciate that form for itself — and practise.