As someone supervising postgrad students, I wonder what is the best way to train these students to write high-quality research papers?

The best way seems to be through coauthorship but ...

On the one hand I can be hands-off and rather give them feedback as best as I can but let them do most of the work themselves, sort of like a "sink or swim" strategy.

On the other hand I can be hands-on and rewrite the paper (or large parts of it) and hope that they learn from the example I give.

On the one hand I want them to learn to write papers without me some day and I want them to learn as quickly as possible to take load off me.

On the other hand I want us to publish in the best place possible with as little stress as possible.

On the one hand I want them to see the peer review comments from what they themselves have written.

On the other hand I do not want them to be discouraged by the peer review system early on and do not want to submit papers I know we could have written better.

On the one hand I want them to get confidence by being published and presenting work at good venues.

On the other hand I wonder if they need to get there themselves to get that confidence boost.

At the moment I'm very hands on. There's a few factors for this: I don't like submitting papers that are not up to a certain standard, I'm a native English speaker in a non-English speaking country with non-native speaking students, publications are a high priority for me, etc.

But recently a friend/colleague warned me that I should not be so hands-on with the writing of the papers of my students. He pointed to recently graduated researchers in my area that published good work as students but who struggled after leaving their PhD supervisor saying their supervisor did them no favours in the long term ... that by being so hands on, while the students have success getting published and so forth with me, I'm not really doing the most important thing a good supervisor should do: making myself obsolete.

My question is how to train students to write good research papers and in what ways can I balance being hands-on and hands-off?

I'm also interested in methods not directly involving coauthorship.

  • 30
    In my experience working with (European) graduate students, a big step was teaching them the structure and logic of writing a paper. Early on, for the first few drafts (that they wrote), my response was generally a one-page outline that was how I thought the paper should be structured. Once they got the hang of the logic and flow needed to lead readers through the paper, the rest was just finding the right words (by far the simpler task). So, focus first on the logic and flow to make the point(s) properly.
    – Jon Custer
    Commented Aug 2, 2017 at 13:05
  • 6
    How much time do you spend with them reading good and bad examples of writing? How much time do you spend discussing how badly written published papers could be better? Commented Aug 2, 2017 at 19:03
  • 11
    Your goal of teaching students to write well is fundamentally at odds with your goal of publishing high-quality work quickly. Which goal is more important to you?
    – JeffE
    Commented Aug 3, 2017 at 1:45
  • 2
    @AntonioVargas Of course I do! The problem is with the desire for speed. Learning to write well takes time. If you don't hage time you dan just write the paper yourself, but then how will yur students learn to do it?
    – JeffE
    Commented Aug 4, 2017 at 2:43
  • 10
    You have too many hands. Commented Aug 4, 2017 at 22:14

11 Answers 11


edit: To clear a misunderstanding that might have arisen (see the answer of Kimball): the papers described here were not solo-papers. Neither me nor my advisor would ever have the idea to write a paper together and then put only one name on it...

My instructor did a really good job with that (and is still doing it), so I'll explain here approach as I experienced it:

For our first paper, we were sitting in her office, she wrote it and asked my opinion from time to time. I was still really new to everything and agreed most of the time, but I also already learned about some questions that should be considered and, of course, I learned the style she was using first-hand.

For the second paper, we started in the same way. At some point, she told me that I should write a part (the proof of a theorem, a short paragraph,...) myself. Of course we discussed the content first, but I was responsible for writing it down. I already had the first part of the paper that we wrote together to use as a starting point, to adjust my style to it, but I was still ultimately responsible for it myself. Once I finished it, she looked over it, gave me comments and asked me to write a new version. And a new one. And a new one,... until it looked just as good as the rest of the paper she wrote. Of course it would have been easier for her to just write it down herself, but doing it this way I learned how to write myself and I had the success of having something I wrote myself in a paper. (Yes, I had to adjust it a few times, but ultimately I did write the final version that we published)

For the third paper, she told me to write it completely myself. Once again we went through a long process of her commenting on things and me correcting it, but in the end I had a paper that I wrote myself, and I felt really proud of it.

After that, she got me some review jobs, so that I could learn what to look for in a paper, how to recognize strong parts and flaws and (I assume) so that I don't just copy her style but also see other authors and learn from them.

All in all, she did a really good job and with only three publications yet I feel like I am now able to write a paper on my own.

Now let's end with some general disclaimers: The approach given above assumes a motivated student that wants to produce good quality papers and is willing to adjust it until it looks really nice. A lazy student who just wants to increase his paper count with the least effort possible is of course still a problem. Also, your students should be able to produce papers in English. If not, point them towards English courses offered (most likely) at the university, help them install spell-checkers on their computers, etc.

  • This approach will fail miserably in disciplines where graduate students don’t usually write 3+ papers. Case in point, I authored two papers during my PhD, and I was above average among my peers (with the median being somewhere below 1). Commented Aug 3, 2017 at 9:28
  • 4
    The question asked for a way to train students in writing papers, so I assumed that they will indeed write some. I don't know of a way to teach paper writing without at least a little experience. If you want to teach it in a way such that students will already be able to produce good papers from the first one they ever write, you might consider this approach and then don't publish the papers, or only publish them locally among the research group, etc.
    – Dirk
    Commented Aug 3, 2017 at 9:34
  • 2
    Right, I see how my comment could be misunderstood. Of course you can only learn by doing. My point is that your approach explicitly requires writing multiple papers; yet this is a luxury not generally available in my field, so the approach won’t work. Instead, the student really needs to do the busy work (ideally with support form the PI) from the first paper. Commented Aug 3, 2017 at 9:37
  • I got the context of your answer, and don't disagree with it, but I wasn't sure about the OP's situation, which is why I thought I'd offer an answer for other situations.
    – Kimball
    Commented Aug 3, 2017 at 13:05
  • Accepting this as the most general answer: start by setting an example writing most of the first paper about their work, and then "ween" them onto writing papers by themselves with extensive feedback. (Other answers of course have complementary ideas.)
    – badroit
    Commented Mar 23, 2018 at 3:43

Good writers read a lot. Writing an academic paper in a vacuum can seem overwhelming, but writing an academic paper after studying the presentation and writing style of some well-written papers (preferably in your field; hopefully your field has well-written papers!) gives some idea of what works and what doesn't in an academic piece. In particular for me, my ideas of how to create and present my figures and what makes good discussion and conclusion sections have been entirely influenced by others' work.

ZeroTheHero's suggestion of actively engaging your student in a discussion about why the good papers are good (and the bad papers bad) is a great one.


A nice starting point can be to ask your student to identify 2 good papers and 2 bad papers, and then ask the student to explain (by writing first one-page justification and defending it orally after) why she or he thinks the papers are good or bad, comparing them and explicitly pointing to the differences.


I think the suggestions in other answers are good, but I think the approach depends on the situation, and Dirk's for instance seems quite different from mine. E.g., solo papers in math, I do not think is it normal or appropriate for an advisor to write any of the paper (cf. Can an advisor write a student's PhD thesis?)

I did my PhD in pure math, where a thesis typically amounts to 1-2 solo papers, and they are often not written until after graduation. I think my advisor was very helpful about teaching me how to write, and this was my experience:

  • First, unlike many cases, my advisor had me write up my results as 2 papers before writing my thesis. I was fortunate in that I got 2 stand-alone results sufficiently in advance of graduation for this to make sense. This meant that my advisor could give me tips and feedback throughout the writing process. (This also meant I didn't spend the first year of my postdoc writing up my thesis, which is far too common in math.)

  • Before writing, my advisor suggested how to pitch and organize my paper.

  • After writing a draft, he read through the paper and made many comments. As I recall, sometimes he read the paper while I was in the office with him, and I took notes on his comments, and sometimes he read the paper by himself (at least with my thesis) and made comments by pen, but then would go through them with me when we met. Many of these comments would be about little details, like how to phrase things to make it more clear, whether something needs a citation, but some of these comments were also on broader principles (e.g., state your main result clearly at the beginning). What I think was important, was he would explain why it's better to write this way instead of that.

  • We would go through a few iterations of this until he was happy (I think) with the paper. One time, after going through extensive comments and rewrites twice, he told me to just start over and write it this way instead. I was frustrated at the time, but in retrospect think it was very good advice, and have several times since found that starting over quite helpful.

  • He also gave me advice on where to submit, and fortunately I got referee reports before graduating, and he helped me deal with revisions.

Besides this approach, which I think is great (provided the advisor is a good writer themself), here are a couple of other comments.

  • It takes time to learn how to be a good writer, and getting feedback from different people is instrumental in this process.

  • While I agree with one of the other answers that reading good papers is helpful, what I actually found more helpful was reading bad papers. After spending loads of time of struggling to understand things which are poorly written, you realize certain things to avoid in writing, and how you would prefer it to be explained. Perhaps more importantly, this should impress upon one the importance of writing well.

  • General English writing guides (e.g., Strunk and White) can be helpful for instilling basic writing principles, as well as grammar. There may also be various articles on writing advice in your general field--we have several in math, and there are general science ones as well--that you can recommend to your students.

  • Good writing is of course subjective, and depends on where you're coming from, and one person I thought was a great expositor, another colleague thought was hard to understand. Anyway, I think it's good to encourage students to develop their own style and realize different approaches have different advantages.


First, congratulations for taking care of the students and your proposed goal of making them better writers.

As a self-educated non-native speaker, this is what I would suggest. I know that this might sound simplistic, but you really don't need many principles (which doesn't mean writing well is easy to accomplish):

  1. You can't be a good writer if you're not a good reader. That means not only that your students should read a lot but more importantly they need to pay attention and care about what they read. Asking students to read and explain articles helps a lot. Then, they are able to identify first-hand how writing style and clarity affects understanding in readers, which is the whole point.
  2. Know your audience = learn your genre/style. For better or worse, scientific writing is not creative but absolutely straightforward. The structure is standard, even the writing style is standard, but of course, you need to learn it. This is sometimes difficult, particularly to young/non-native speakers that associate long-grandiose-impressive-but-unintelligible prose to "importance" or "high status". I would suggest you give them a copy of this excellent book: https://www.amazon.com/Science-Research-Writing-Non-Native-Speakers/dp/184816310X
  3. Practice, practice, practice. And get feedback. The more you write, and the more you see that audiences get what you mean (even if they disagree), the more confident you become.

Above all, please teach them that clarity, conciseness, and particularly writing a clear conclusion or proposition (something new the reader can learn from them) that derives from the experiment/work is the main goal of their writing, not to show how creative or eloquent they can be.

  • 2
    Is writing style really standard in science? In math, there are some basic structural aspects that are standard, but there are certainly very distinct styles (e.g., beginning with generalities or beginning with "examples"), with different people preferring different styles. Broadly, I believe these styles grew out of different mathematical cultures, e.g. the French school versus the Russian school, and now within even small communities you can find various styles.
    – Kimball
    Commented Aug 3, 2017 at 13:02
  • 1
    That's a good point - I didn't express myself correctly. There are variations depending on many things (field, theoretical vs. experimental paper, etc...), but what I mean is that it's not "creative writing". There are well-established styles and you should (particularly at the beginning) pick one.
    – Pablo
    Commented Aug 3, 2017 at 13:18

The way I found very effective, is indeed to have the student as a co-author for a paper. The student would do most of the practical work (or it is split between multiple students). Then it is discussed how to organize the paper and everyone would describe their part of the practical work in a section - that way, if there are multiple students involved, they have to discuss and normalize their terminology for it to make sense as a whole. The advisor would take a look and comment on the structure and content mainly (top-level) until a complete basic version is established. In the meantime the advisor would write an introduction and/or related work section etc. Then students and advisor take turns "improve"-reading the whole paper or distinctive sections. Each time the advisor does a pass, he adds commentary where needed or roughly restructures sections to give the students the general direction of how to improve then let them do the details. Initially this means a lot of big changes, but in time they get smaller. Finally it gets more fine-grained and the advisor changes individual sentences to improve the language, shorten paragraphs etc - in essence he goes from top-level to fine-grained changes in a way that he'd always spends about the same time with his pass.

The students would exactly see what he was changing and could object - for instance, if they thought the reformulation didn't match what they initially intended to say with regard to the experiments etc. Through those discussions and being hands-on involved in the changes (after each advisor iteration, the students work with the text again and can also do whatever they feel as improvements), the students should feel as still having control over the text, but they also learn by doing. I think this is a quite motivating approach and also very helpful to really improve our own writing style - in the end, everyone can learn from everyone and by iterating over the text of others, it is made sure everyone understands the whole content to some degree at least. For the advisor it means he has a more or less constant level of effort - sure, towards deadlines the iteration frequency may increase ;)


There are some good suggestions in other answers that mainly are related directly to writing the paper. A further point: in my opinion, to become good at writing papers you need to try reviewing them.

  • Where possible, and with the permission of the editor, work with your student on some real reviews that you've been sent. Even better, if you are reasonably confident in their ability, pass the review onto them entirely (again, via the editor).

  • Make sure that your department has a journal club that your students attend. And attend it yourself so that you can highlight points that may have been missed.

  • What about having a "review exchange" prior to submission? Your student exchanges manuscripts with another student (not necessarily simultaneously), and they participate in a mock review process. This could help with your worry about submitting a paper that is below par. Disclaimer: I've not actually tried this...


To the immediate question, I don't think co-authoring is the best way to reach this goal.

  • Like most sports, arts, games, and crafts, writing isn't one skill. Each is a cluster of skills, rules, templates, techniques, and strategies. FWIW, I think editing is it's own cluster of skills that merely overlap writing.
  • Being "good" at a given sport, art, game, or craft isn't about methodically following a massive script or checklist, even if it's full of observations from the best of the best. The lion's share is about "naturally" synthesizing elements you've repeated often enough that you can can perform them on autopilot.

If you think about it this way, hopefully the following approach will make sense:

  1. Break what it means to "write high-quality research papers" out into a list of priority-ranked skills, rules, formats, techniques, and strategies you'd like your students to know. Be pragmatic about a skill's usefulness. Correct citations might be important, but it's probably a bad idea to spend a week learning to flawlessly cite without referencing the style guide before you can clearly explain methodology.
  2. Know what the baseline looks like, and start with a diagnostic. Get a sense of whether any students need significant foundational work in writing or English to be ready to work on your initial goals.
  3. Chart a progression that balances focused skill acquisition with synthesis. Students should be continually picking up a handful of skills and working towards synthesizing them into increasingly complex forms and longer work.
  4. Approach every skill from multiple angles. What's appropriate will vary, but some ideas are: having them break down pre-selected examples, find their own examples, compare examples they found, get together in groups to debate the best/worst examples anyone in the group found, revise/improve "bad" examples, write their own, compare theirs with classmates, revise or critique previous work in light of new techniques
  5. If it's important enough to mark, require corrections/revisions. It only takes a little more work on your part to make sure students promptly put your feedback into practice.
  6. At the risk of stating the too-obvious: know if your university/college has a (graduate-)student writing center, and if so what they offer/are capable of. It's not like you'll be able to outsource this task on them, but they may already have workshops/groups on some topics you want to cover. If so, attend one and see if it's a suitable standin.

I can't tell if you're supervising all of the students in a specialization/degree/program/etc., or if you're only supervising students who seek you out as an advisor.

This is a lot of extra work for you and a captive audience, but if it's an elective relationship and you can get away with making it clear that this is a requirement for working with you, you'll at least filter for motivated students.

It's probably good for time and resource reasons if this opportunity is built into degree requirements as a single or multi-year course track. If budgets aren't austere and politics don't intervene, the funding request for a pilot program aimed at improving the quality, quantity, and placement of publications would seem to write itself.

Because writing happens to also be a good way to learn, it's even better if certain proficiencies are a core goal of your program reflected in all courses. If there are courses in which students read existing literature or formulate small research projects, these are opportunities to build literature reviews, critiques, research proposals, and other forms of writing into the pedagogy.


One thing that my advisor has always done that has been helpful for many reasons (not just improving writing) is to encourage me start writing up problem statement, partial results, etc. as a project is progressing.

Often other students seem to wait until they think they have enough results to publish a complete paper before even starting writing. I've tried this approach in the past as well, but I think early writing has significant advantages.

  1. It ensures that both myself and my advisor have the same understanding of the problem.
  2. It ensures that I understand the problem (or the proposed solution) well enough to articulate it in writing.
  3. I can get early feedback on the structure and content, and particularly the description of the motivation for the work.
  4. It can be used as a starting point for paper later down the road.


Some of these may be less appropriate in some fields; take with a grain of salt.

  1. Get them to critique/review poor, fair, good, and excellent papers. It is much easier to see problems in someone else's work than your own; but after having done so you start to see the same mistake in your own work.

  2. Compile a list of good resources on good writing, and encourage students to start there and go beyond on their own. Or you can write a "writing manual" or "top 20 things to avoid" yourself. I read my supervisor's advice on writing many times while writing papers and my thesis.

  3. (This one is tough but important.) Do whatever you can to teach students how to answer the question "Is this good?" for themselves. It's relatively easy to spot grammatical errors, confusing wording, and problems with flow; it's much harder to look at a work overall and comment on the quality.

    Here's the advice I would try to drill home: you must go to first principles. It is about the individuals who read your paper and how they will react. Some will be students, some professors, some from other fields, perhaps even some laypeople. The more you respect their time, the less effort it is to read the paper, and the more you can inspire and excite them, the better.

  4. (Also tough but important.) Try to break the "follow-the-rules" mentality. It is a creativity killer. Actually I think breaking this mentality also leads to better research. Of course, we need to follow the rules some of the time, but the education system creates students who believe that following rules is sufficient to create success, because that's how it works at school.

    What I would drive home: Remember that you have a blank page. You can put anything on it if you think it will be effective (but make sure to get opinions on it before submission). Usually this means creative diagrams or poignant examples which get several ideas across effectively. Try to think of various ways to explain things and go with the one which will have the best impact on the readers.


I think it's admirable that you are trying to look after your students; many people have the attitude "they are adults" and I think it's a lazy and narrow-minded approach. So please continue.

That said, it sounds like you are mollycoddling them: "as little stress as possible", "do not want them to be discouraged", etc.

Putting one's hard work up for scrutiny by experts in your field is supposed to be stressful. Students need to know what the reality is like. Some of them will tolerate it very well, and you're holding them back by hand-holding them. Others will struggle with it, and if you hand-hold them then they will still struggle once they graduate. There's a good chance that if you have high (but realistic) expectations, they will rise to the challenge—or they may realise this career is not right for them, and if so, better now than a few years down the track.

By all means help them get emotionally prepared, step in when they are about to hit a crisis, and so on. But give them as much independence as they can handle, have high expectations, and when possible, allow them the satisfaction of publishing their own work, even if you know it could be improved (provided it's fit for publication).

To be clear, co-authoring a paper can be a very valuable experience for students—not because it insulates them from stress and discouragement, but because they learn about the thinking process and the tricks of the trade.


I would set a timeline for revisions. I would pick a format [APA for instance]. I would fail any student who did not follow the format or copied too much. Create a way to make the students select a topic and 'working' title. Next, and since your main issue is sensitivity. I would force the students to turn in a rough draft after a couple of weeks. Then, I would collect all the papers. I would shuffle them and hand them out for the other students to critique.

After my professor did this for me, I have become a way better writer!

  • 2
    I think the main flaw with this approach is that if you let the drafts correct from people who do not know how a paper should look like, they will miss a lot of important points and gain a wrong self-confidence.
    – Mayou36
    Commented Aug 4, 2017 at 11:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .