Teaching students to write scientific papers

By Grant Jacobs 21/03/2012 27

(Mid-week ramble!)

How did you learn to write a scientific paper? (Also: how did your first paper go?)

This brief thought is inspired by a discussion over at Occam’s Typewriter, at Cath Ennis’ place (aka Cath@VWXYNot). There she was pointing (grumping?) at some long-winded phrases in papers she was helping others with.

Somewhere in the discussion there’s this exchange between us (links and emphases as in original):

Cath@VWXYNot? says:
It boggles the mind how few people seem to have read any scientific papers before attempting to write their own!

(I’m not referring to this specific manuscript, but rather to a general pattern of people just not seeming to understand the conventions of the format)

Grant says:
’It boggles the mind how few people seem to have read any scientific papers before attempting to write their own!’

I find that very hard to imagine. They ought to have been reading papers since senior undergraduate classes, surely. Mind you, I can imagine plenty not thinking about how to write a paper, though, bad as that is.

Cath@VWXYNot? says:
That’s just what I’ve inferred; I don’t see how anyone could read more than a handful of papers and then still commit some of the formatting and other atrocities I’ve seen over the last few years. But like you said, people may just have a problem generalising what they’ve read and then applying the standard format to their own papers

Grant says:
I’m with you now :-)

You’d hope it’d occur to students one day they’ll be writing one of those things. Hmm. I might blog about an aspect of this 😉  (Not tonight, soon-ish.)

Seems I wasn’t right about that last one as I’m writing it tonight.* Ah, well…

How were you taught to write papers?

My memory is way too vague now. I recall there being little bits of it a different stages, but I can’t remember being formally taught to write research papers as such.

I have this fuzzy recollection that we were made to write up some undergraduate laboratory ‘stuff’ in a basic paper format. (I’m quite sure of it, really, but like most people my undergrad years were a blur of schedules, lectures, labs. and everything else in between.)

Thinking back I don’t think I connected these exercises with the research literature in any real way at the time, that it was just how we were ’supposed’ to do it. I think this was in part because at that point we had no real exposure to the research literature. We were still getting formal lectures and reading those enormous textbooks that seem designed to try give skinny students weight training on the side.

It makes me think that those ‘write-up’ exercises would have meant more if we’d been shown a few papers first, let us see that research papers were the lifeblood of the research industry.

Perhaps part of the problem was not having been taught the history the scientific method thing. It’s got a bit to do with why research papers are laid out as they are. Y’know, learn what’s established thus far, develop your hypotheses, how you’re going to test them (and why those tests are appropriate), then results, etc.

It’s more than saying ’it’s done like this’, it’s showing why. A sort of taught-course equivalent of that writing advice of ’don’t tell; show’.

Either way when I came to write my first ‘proper’ research paper I’d long forgotten those exercises. I’m willing to bet that’d be the same for many others. There’s a decent gap between undergraduate studies and your first paper. (In my case longer than most through working for a couple of years in between undergraduate and postgraduate studies.) A Ph.D. student might not write a paper until their second or even final year.

On the face of it it seems simple enough when you’ve read lots of papers, as most Ph.D. students have. Roughly speaking there’s a title, a summary that you try cram the main messages into a set space, an introduction that sets the scene and lays out what you aim to investigate, then what you got up to (Materials & Methods), rounded out by a discussion that examined the results and pulled them all together, followed by a list of the research papers you’d cited in the text.

Nothing to it! Ha. I don’t think there is an easy way to appreciate much work it is to write a paper until you do your first one.

That leaves me with another question that I might ask readers who have written research papers to share to those who have not – how long it would typically take you to write a research paper?**

Perhaps also: advice you’d give to students writing their first paper.


* I was in the midst of a longish article, Structure in Our Genomes, it’s the sort of thing that needs time. This I can write without having to find illustrations and think so much about how to lead non-biologists to what I’m saying.

** I once similarly asked readers how long it took them to write a science blog post.

Other articles on Code for life:

What aspects of biology need to be explained better?

On alternatives to academic careers and ’letting go’

Seeking science-y reading?

Teaching kids critical thinking

27 Responses to “Teaching students to write scientific papers”

  • I teach science writing to university students, and most undergraduates have never read a scientific paper, much less had the structure explained and been told to copy it in their lab reports. I remember I did an entire NZ BSc and was only exposed to the literature in one third-year class. So perhaps we could start there.

    Postgraduates in my experience read a lot of papers but have not really examined the conventions and functions of each section, so they’re floundering when they try to write one. Even exposure to a few different journals’ Instructions For Authors, which are often quite thorough, would help. I’ve read several books on writing in the sciences, but I haven’t found a really good one yet.

  • Actually, what you describe is EXACTLY what happened with me. I had read scientific papers before, a few, just a few. But no instructor ever explained them to me – the format, the peer-review process. I was taking a senior level class which involved an independent project and a write up. I asked him, “how do I write this up?” I had NO idea. He said, follow the format of the papers I was reading. So, it wasn’t until I had to write my first paper, that I really ‘read’ a scientific paper for content and format.

    So, how did I learn to write my first scientific paper? With lots of trial and error. LOTS and lots of error. The next semester I had to write up my senior thesis project. I was so thankful for that class (which was an elective) because at least I knew what to do this time. Again, NO one sat me down – formally in a class or informally in an office – and explained to me how to read a scientific paper, reference one, avoid plagarism (I’m sure I did untintentionally) or how to write a paper.

    That was the mid 1990s. Fingers crossed that it has changed for students today.

  • How long it would typically take you to write a research paper?n
    That is very hard to answer. I’ve never kept track of number of hours spent on the keyboard on a manuscript, because papers can go for long periods when I don’t work on them. For me, writing isn’t separated from things like data analysis, creating graphics, etc. For that matter, many chunks of writing go into making posters and presentations that are the preliminary versions of the paper.

    And are you including the writing that occurs after the reviewers and editors have made their requests?

  • Mike,

    I read a fair number of papers in my final year of my undergrad, but then as it was a BSc(Hons) that year was really a MSc in effect. (That’s cheating, huh?!)

    I’m sure most first-timers lean heavily on the Instructions For Authors; you’d like to think the reason journals make them substantial is to try head off the worse offences, esp. from wayward first-timers!

    Personally, I think for post-graduates it’s one of those things where Ph.D. supervisors just have to make themselves the initial peer-review and give their student constructive pointers & feedback, etc. That was my experience. I’m not sure if it’d be the same in institutions with very large research groups. Perhaps there the post-docs get saddled with this?

    It’d be nice to see a better link between writing up experimental work and writing papers somewhere in undergraduate, even if just to expose the thing once. Or maybe that’s idealism and impractical?

  • Zen,

    Indeed, it’s possible to put a true number to – I was just thinking indicative ballparks so that people have some idea. As you say it’s tied up with other things and the work doubled up, e.g. preparing for talks. (I suspect most academic researchers aren’t into monitoring their time the way lawyers do, either! :-) )

    I was thinking of the whole thing, start through to published.

  • DNLee,

    Welcome and thanks for sharing your experience. (Now your first comment has been approved your comments won’t be held up again.)

    “That was the mid 1990s. Fingers crossed that it has changed for students today.”

    I wonder too – I’d love to hear comments on that.

  • My first-years have to write an essay as part of their assessment for the paper – and they’re expected to do their research in the primary literature. That was a deliberate decision on my part as I recognise that it’s a rather useful skill for them to have. Same for the essay, really, as so many of them have very little experience in producing a reasonably lengthy piece of formal written communication. But – we don’t leave them all alone to do this!

    As well as providing marking rubrics at the same time that they get the essay topics, plus various ‘how-to’ guides (including ‘how to read a scientific paper’) we also spend a lot of time in tut classes, labs, & with individuals going through the various how-to’s & what-not-to-do’s, giving them practice with things like paraphrasing, writing introductions – & reading a paper for meaning. (If we** didn’t do the latter, many would be lost from the outset as they really don’t know how to navigate through a paper & can get totally swamped by the minutiae.)

    ** not the royal ‘we’ – the senior tutor & I. (I sometimes think I would be lost without Brydget!)

  • I had gone to the literature a fair amount by the time I finished my BSc (in 2008) but I think that was spurred by the fact that I was doing a BA in History at the same time, which required me to produce essays referencing the literature (in a broad sense) from 100-level papers on. By the time I got to the 300-level biology and geology papers that required reading and referencing of scientific papers, the idea of using online databases and so forth wasn’t new. The amount of reading ramped up precipitously for Honours, though – we had to read a great deal, and several classes required us to summarise papers as posters or presentations, which meant grappling with structure and deciding what was most important. At least one lab-based class at 300-level also required a report with the traditional introduction/methods/results/discussion structure, so the idea was there. But I never had a class that actually went through “how to write a scientific paper”, as I recall – my first paper was at the end of an summer studentship the year after my BSc, so I got hand-held through the writing I did by my supervisor. (Also, it was a three-page technical note, which reduced the challenge somewhat.)

    My current department at the (American) university I’m doing my PhD at offers a very well-respected postgrad course called “Reading and Writing in the Sciences”, which takes students through all the steps of how papers and grants are structured, how reviews are done and written, and how the whole academic writing system works. It’s an excellent idea and I’m grateful we have it – it’s the sort of thing every PhD student needs but not everyone gets.

  • I was never formally taught how to write a scientific paper. However, I was formally taught how to read them – we covered this in the second year of undergrad, when our profs announced that we should now be using the primary literature almost exclusively and not looking at textbooks any more. So by the time I came to write my first paper, I’d been reading them in detail for a few years (and had also written various reports for internal assessment as part of my PhD programme). I just fit my own results and methods into the standard format that I’d learned by osmosis. It took a few papers before I was truly comfortable with writing Discussion sections, but to be honest I found the rest of it very straightforward.

  • I would say that “start to finish” is usually months – and then it’s not finished because there is still work to do after its reviewed by institutional and journal referees.

    It’s more than just the writing time – there is hunting down of references, maybe of extra data, checking details of what oen thinks has already been found, etc. And then something I strongly believe in – there is the time, perhaps weeks, when the paper is put aside but your subconscious mind still works on it, bgefore you return with a fresh eye to amke sensible changes.

    Of course – this can depend on discipline. My feeling is that is some areas papers can be very brief, full of jargon, covering smaller aspects of work, etc. In oithers jargon is less acceptable and papers usually report more substantial studies.

    Regarding training, I was pleased to see in my last few years at Agresearch before retiring that quite a lot of internal education was promoted – including one course on writing scientific papers. I participated in tutoring this one year – I think its actually a good thing that older scientists were encouraged to communicate their experiences and expertise to yougner scientists. Especially on this issue – there was certainly plenty of interest from younger scientists.

    Mind you – I had a basic disagreement with one other tutor – the use of passive sense. He thought this was compulsory for scientitific writing – I opposed it. (He used to “correct” my writing when he reviewed my papers).

    But I guess that is an issue for each individual to work out.

  • Ken,

    I did a quick ‘survey’ of use of the active voice in research papers a while back, prompted by a science writer (I forget who now) saying he rarely ever saw the active voice in research papers. My own results suggested (tentatively) that while it’s use in the minority, it’s used more than I had expected (although it’s possible review papers and summaries might bias that):


  • Don’t most people write their first paper with their supervisor? I thought this is one of most important things a postgraduate supervisor can do for their student.
    The instructions for authors are an important source of info as well, as is the practice drawn from writing a thesis.

  • I don’t remember having any formal training in how to write. My PhD supervisor just left me to it. My first paper took me ages! I remember being quite disappointed that he didn’t change much before sending it off for peer review. The worst bit was how long it sat on his desk before he got around to looking at it. The experience did leave me wondering what had happened to that great exchange of wisdom and knowledge I had been expecting.. that’s a supervisors job right?! Saying that, I did just make my masters student write a first draft of her paper. I just gave her the instructions and left her to it. But in my defense, she has been working as a freelance medical writer and writes beautifully. On the otherhand, I have a postdoc that I’ve been trying to get to write his first papers and it has been a bit of a disaster. He seems to be suffering writers block but won’t accept any help. Doesn’t help we are not on the same continent.

  • And thanks for linking to your previous post on how long it takes to write a blog post. I missed this one first time around. I was astonished when I first started how long a post took to write. Looking at the comments, I’m no slower than anyone else :)

  • I am encouraging a PhD student of mine to start thinking about his first paper whilst still doing the experiments (he’s in his first year). This helps to get an idea of the story he’s going to tell (i.e. write up) as well as to identify the gaps in the data before he moves on to the next stage of the project. I have suggested that he starts with drawing the graphs & tables in PowerPoint because these are the data he’s going to show and write about. There are a few tricky points in this process, one of which is to get the ‘bigger picture’, and starting with the graphical presentation of the data can be very illustrative (pun intended). The main issue is that I will not do it for him; he has to learn by doing it himself. Only then he’ll find out that reading papers and writing one yourself are like listening to music and mastering the piano (for example); the latter takes a lot of practice. I have not really mastered it (I procrastinate ad libitum) and I am supposed to teach my students! Luckily, nowadays doctoral students have access to so many resources and they’re well looked after so they don’t (have to) rely on only one supervisor. Fortunately for the PhD student he has got a co-supervisor (and 2 advisors) whose writing skills are superb.

  • Siouxsie,

    I guess you have to judge each student on their strengths/weaknesses. I imagine the student who has been writing will have also been mentally thinking how she would have written things (papers) she has read, as writers do, which might help.

    “And thanks for linking to your previous post on how long it takes to write a blog post. I missed this one first time around. I was astonished when I first started how long a post took to write. Looking at the comments, I’m no slower than anyone else :)”

    Good news, huh :-)

  • Frederik,

    Writing the paper as you go is an approach that I suspect isn’t used enough. After all, you’d hope you have some idea of what you’re trying to achieve as you start out!

  • In hindsight I was extremely dissapointed/annoyed that I wasn’t exposed to primary literature at any stage prior to my 4th year at Uni!!. Thank goodness I stayed for my MSc(tech). Absolutely agree that it is up to supervisors to really teach the ‘how to get a manuscript published’ by making it happen. The first one is cruicial to understanding!

  • Rationally Speaking had an interesting podcast relevent to this post and some of Grant’s other posts regtard science blogging and scommunication. Its at RS56 – Howard Schneider on Science News Literacy.

    Here Massimo Pigliucci makes the point that not only are students not taught about scinece writing, lecturers are not taught to teach.

    It’s about 40 minutes long but well worth listening to.

  • Thanks for the link Ken. It’s a good point. For those looking for more on this, Alison’s blog may be a better point of call than mine – science teaching and the studies suggesting how it might be done better is her specialist topic. I’m merely an amateur on that subject!

    Her latest post is that very topic – ‘ tertiary teachers & accreditation’:


  • Grant,

    You wrote “After all, you’d hope you have some idea of what you’re trying to achieve as you start out!”, which made me think of the following quote by Albert Einstein:

    If we knew what it was we were doing, it would not be called research, would it?

  • Hi Frederik, I was writing about what you set out to achieve, not what eventually happens :-) I was just saying that you could start writing based on your aims, etc., as your earlier comment suggested.

  • Over the last month or so my blog has received repeated spam from companies offering to supply dissertations for students.
    This is certainly something that would inhibit students ability to write (if they were stupid enough to risk being caught – I wouldn’t treat the use of such services lightly)
    e.g. http://www.topthesis.com/

  • Michael – been getting those for months too :-/

    Looking back at the responses, it’s interesting how few people have had formal training in writing research papers given how central these are to academic careers. Thanks for all your contributions – and don’t let me saying that stop anyone from adding more!

  • I’d comment but I’ve yet to do a paper. I’ve written a Masters thesis but that feels entirely different. I’m trying to compile things at the moment to go into a paper and thinking about how to go about writing it up. I have no idea how it’s going to work atm. though. Somewhat daunting.

  • Ben, it depends of course on the individual but the best bit of advice I ever got on writing in general is to ignore your attmepts to edit as you write.

    Give your “right brain” full reign, just write down what comes to you and completely ignore any urge to edit or correct. (Some people suggest turning off your monitor and just typing). It helps to do a bit of mind mapping before hand, or a brain storm.

    Only when you have got it all down (probably a major section for a paper) do you then go back to edit. That’s when you “left brain” comes into itself. At other times it just gets in the way.

    I found after doing this a few times that the required editing was not really extensive (perhaps my writing improvfed), I was much faster and procrastination was less of a problem.

    I like to see writing as a right brain/left brain activity – with sepoeration of the two parts.

    It’s amazing also what work your unconscious brain will do on the paper if you just put it away for a while (even weeks) and come back to it fresh.

Site Meter