After a year of researching for my Masters project, I am writing a manuscript for publication. I have spent nearly 3 months writing it now. My advisor has read my drafts, but after nearly 4 drafts that I thought were successively better, he says that I am in still the pre-draft stage and that I don't know how to write a manuscript properly. Of course, he gives me adequate guidance and is very helpful, but I am demoralized by my own shortcomings.

I have a hard deadline of 3 weeks to submit my edits to my advisor, after which he has said he will finish editing for me.

I understand this sort of thing happens to all new writers, especially in technical writing, and I could hear tons of personal stories about how difficult his/her experience was - but in the end, pain is very personal.

Anyways, the purpose of this post is not just to get motivated or feel better - it is to ask for concrete advice on where and how to begin, especially given that I don't have that much time?


5 Answers 5


Concrete advice on where and how to begin, especially given that I don't have that much time

I work in biomedical science so my advices may not fully apply to you, feel free to take whatever that is useful.

Don't write too much: Most students I work with who are trying to turn their paper into a journal article tend to write too much. It's usually described as a painful process because it would feel like chopping off a lot of work from their thesis (which was done with much blood and sweat.) But the trimming and refinement need to happen because thesis format is not suitable for journal.

Concretely, these newbie articles tend to

  1. Have an introduction that is overly long
  2. Contain abundance of details in the methods section, documenting the conception of the idea to every little details, with meticulous follow up of citations
  3. Justify every single decision, some of which trivial
  4. Cover too many hypotheses or fail to boil down the structure to bite-size, thinking everything in the chapter has to go into the paper
  5. Elaborate lengthily in Discussion, failing to highlight the most important aspect

But also don't write too little: And yet, most students tend to overlook these components:

  1. The paper goes into technical way too abruptly, doesn't account for the fact that some non-specialist may be reading this. And this is really an art, you'll need to see your target journal and get a sense of who are the readers, and then decide how broad to start with

  2. Fail to explain the "so-whats" in the Discussion

  3. Did not address the strengths and the weaknesses of the study. Or quoting weaknesses that actually can be alleviated but were not. For instance, I have a lot of students stating weaknesses like "this is a cross-sectional study so causal inference is not feasible," which I tend to agree; but some may say "in this analysis we did not adjust for income levels" while the income level variable is actually in the data set. In those case I'd request a re-analysis. In short, if you can have done something about it, don't list it as weakness.

General style mismatches: Sometimes the articles have nothing really wrong about it but the style just make it clunky as a journal article. For instance:

  1. Overly sensationalizing or flooded with emotional adjectives. E.g. "the obesity epidemic is indeed a human tragedy," "our results are surprisingly encouraging."

  2. Mixing up facts and arguments, ending up with a lot of statements that are not backed up or statements that are accompanied with a tsunami of facts without a trace of synthesis.

  3. The objectives and the conclusion do not match. And this is very common in theses because we set off with a plan and things may fail to work in the middle and we have changed direction. In journal article, the objective statement and the conclusion need to resonate so that they read nearly like Q&A.

  4. Trying too hard to make the whole paper self-contained and complete. Common signs are including too many details that can be included in the form of citations.

  5. Calling the same concepts or things with many different names. In English composition we are often encourage to diversify our word selection but this is a big no-no in scientific writing. If the variable is "occupational stress," call it so throughout. Do not use variants such as work-induced stress or even worse: work-related mental problems or job-site related depression, etc.

  6. Drafting, writing, and editing at the same time. Most of the students lost steam because they want to churn out perfect sentence from the get go. This causes the process extremely slow and painful. It's easier to do a brain dump and then come back to edit the work. On a good day you may save 50-60% of the dump, on a bad day maybe 10%, but I found it worth it because I can be more focused in either laying out ideas or perfecting the wording, separately.

So, how to start? Before writing, read a bit. Download the "guideline to the authors" from the target journal and use a highlighter to indicate all stylistic requirements. Follow them to the t.

Then, download about 3-5 articles related to your fields from that journal, and read for word-uses and syntax. Generally, when reading each paragraph, ask yourself "Why did the authors put this information here?"

Now, when it comes to writing/drafting. I find it easiest to do in this sequence:

  1. Identify 1-3 main analysis outcomes (table, graph, regression model, etc.) Paste them on a wall or a board and they will be the staple of this article.

  2. For each of these output, write about 2-4 main points or interpretations people should take away with. Imagine you're at a conference manning a poster. What would you tell the audience to focus on in each table or each graph? If one of them has overwhelming amount of talking points, you are trying to pack too much. Consider breaking the analysis output down, and trim, trim, trim. In this stage, do not throw away unused outputs, put them aside in a folder or a box in case you have a second thought.

  3. For each of these output, write out the methods. Remember to only contain enough information to allow readers to replicate, not too much so that the readers have to endure the whole story. For instance, reason for picking a certain concentration or threshold can be explained with a citation rather than writing it out.

  4. Now combine your selected output into the Results section. Write it in the driest style you can think of. No interpretation, grievance, celebration, defense whatsoever, just plain, hard, cold data.

  5. Break Discussion into segments A, B, C, D, and E. In A, use one paragraph to summarize what you found. In B, incorporate the talking points you generated for each of the output, so that the first paragraph makes sense. In C, compare and contrast your work with other literature and discuss the potential differences. In D, delineate the strengths and weaknesses of your study. For each weakness, discuss what you had tried to make it less a matter, or what else could have been done in future work. Finally, in part E, go slightly wider and wilder: elaborate what these results mean to your target audience and their field. You can quickly see that this section can get overwhelming if you have too many objectives. For starters, one main objective with a sub objective will be a good challenge.

  6. Write the Introduction. Focus on some "what's known" to bring your audience to the same page so that they can understand your work. Then, discuss what is not known or what challenges the field is facing. After that, explain how your work may help, and lastly, lay out your objectives clearly.

  7. Cut and paste the objective at the end of the paper (or use split screen so that you can see the objective section), and start writing the Conclusion. Introduce no more new information there, simply answer the objectives point by point. Make sure the objective part and the conclusion part resonate. When you feel it's good enough, you can then take out the reference objective you just pasted there (or close the split screen function.)

  8. Go back to write the abstract.

Other nitty-gritty

Don't feel bad that people are not happy with what you write. A lot of the time they do not even concretely know what's wrong with the piece. Ask for actionable suggestions and then thanks for their comments; take whatever applies to you.

Your supervisor may not appreciate your writing quality, but will definitely appreciate your attention to details. So, make his/her life easier by following the journal's guideline tightly when it comes to formatting.

Check if all the figures and tables are correctly indexed.

Check if all the pronouns connect well and leave no room for misunderstanding.

Use subtitle to introduce another layer of structure if you feel too overwhelmed.

It's true that some papers can take years to write. But you have been working on it for more than 3 months and it is more than probable to write a decent article in 3 months. Some other users may say a paper can take years, but remember the duration is also a factor of i) how much time the person is putting in every day, ii) how much the work is being held up by co-authors, and iii) personal writing style. I myself have papers that were done in the span of 2 months to 6 years; there is pretty much no rule when it comes to time.

  • I think this is very good advice. Another tactic that I find useful is to work out how I would convey the main narrative using approximately one figure per page. I then make those figures and build the rest of the paper around them.
    – jakebeal
    Nov 22, 2014 at 21:44
  • "Most students I work with who are trying to turn their paper into a journal article tend to write too much.": Yep! :-) This fact is frequently based on the false implication "I've done a lot of work -> I should write a lot, describing all the work done". Nov 22, 2014 at 21:57

A great online course called "Writing in the Sciences" is now available free of charge, and it provides really great advices. When I was writing an article recently, I just had the time to watch the first week's lectures, and it helped me a lot to finish my manuscript.

To summary some quick tips to do a good writing:

1- Read papers and copy (everything but the content) from the ones you like. This will give you hints on how to structure a paper, style you can have and many other things. Mimicking is a fundamental learning mechanism of humans that helped us achieve the knowledge we have, don't overlook it.

2- Just get to write, even if you don't feel like it. A good paper is reworked several times before it's ready for publication, so just get going on your first pass, you'll enhance later.

3- Don't try to make overly long or complicated sentences. Just try to make short sentences, and try to use simple formulations (present tense instead of past tense, standard sentence form: subject-verb-object). Also avoid jargon and initials as much as possible, this will be one less burden for your reader (even when they're specialized in the domain you're writing in).

4- Be logical in your progression. Try to be logical and gradual from chapters to chapters (for this, generally there are common templates, like the OHERIC methodology: Observation/Problem introduction, Hypothesis/Your Model, Experiment/Simulations, Results/Interpretation, Conclusion/Opening) but also inside each chapter, so that the end of a chapter naturally leads to the next.

Usually, the introduction is the hardest part and usually too much overlooked so you should really focus on this part ; the conclusion is the easiest part (just summary what problem you tackled, what you did and what you've found and potentially future avenues that could extend the findings on this problem) ; and the abstract and title are done at the very end when you already wrote the whole paper.

5- Add lots of pertinent references. A good example is any well-developped article in Wikipedia (since this encyclopedia follows some common editorial guidelines that are just as well used in scientific publishing). For any claim you make that isn't yours, try to reference, and most importantly, your reference must be pertinent (avoid referencing an article you barely read or that isn't focusing at all on the claim you are making, ie: an article about brain's memory making an hypothesis about consciousness at the conclusion as an opening isn't fit to be referenced for any claim about consciousness since this is just speculation).

6- Be yourself. This can be difficult if you are not comfortable in the language you're writing your article in, but if you are comfortable enough, try to keep (or create) your own writing style.

7- I think this one is less important at first, but later if you want to be a pro: uniformize your editorial line. For example, if you use American English words in your abstract, don't use British English words suddenly in later parts. A very good listing of editorial tips can be found here: http://www.iaria.org/editorialrules.html

8- Use a spellchecker, always.

9- Ask other people to read your paper, and be open to feedbacks. This is crucial and the final step to make a good paper, as it is very difficult to see the big picture and the small glitches yourself. If possible, try also to get your paper read by non-specialized persons, like your relatives, they will tell you if your work is pleasant to read even if they can't grasp every technical subtletlies.

Good luck for your paper. And remember: writing is always painful and feels unnatural for most people (including professional writers), so don't feel out of place, just try to do your job and try to be proud of the result.

  • Your spellchecker from item 8 did not complain about "peoples" in item 9, did it? ;) Interestingly, this is precisely one of those small glitches that are only seen by other people ... Nov 23, 2014 at 18:55

Three months? Try two years!

I'm trying to write a manuscript since two years. This is because my advisor is a perfectionist, which makes me an extremely lucky guy (in this case, you as well). I have a friend whose advisor's expectations are below the average and he ended up submitting a thesis even lower quality than below the average.

Believe me, what you go through is a perfectly normal stage, for who wants to be in academia. At least, this is what I've seen so far.

I suggest you to do the following:

First, write the core of your work. Not the intro, not the conclusion, not the experimental results or methodology. Just tell what your work is about and why is it precious. Tell it like you're telling it to a five-year-old. Try to tell it using the least number of sentences as you can. If possible, tell it in only one sentence.

Afterwards, read. Read until you memorize it. You will realize that it sucks (unfortunately). Then, change it. Try to change your point of view.

I have written at least thirty manuscripts if not fifty. And I have started over at least ten times. Now, I got "OK" from my advisor for the last version of, wait for it, introduction chapter.

"The most important thing is organization of the manuscript", my advisor told me. "It should both teach and entartain the reader." By entertaining, of course, he means that the reader should tell to himself "this is good job!" Like reading a really good novel.

I suggest you to keep your expectations low at this stage, since your last work is still a pre-draft. In a period like three weeks, it is extremely hard to complete a work. Besides, if you do not rush, you will complete it in a shorter time.

I hope I'm wrong and you publish your work.


It is surely painful, as you're being assessed on something as intimate and self-defining as cognitive potential, and think you're not measuring up. And academia is a rather terrible environment when it comes to taking into account how this sort of thing affects productivity and happiness. But.

The thing is, the goal is not to make the manuscript good for your supervisor, it's to make it good for a research journal. In this process, the supervisor will help. The help will come in the form of a very red, very re-written manuscript, but it will be useful. For me, the single most helpful event in academic writing (and I have taken courses on it previously and I generally like writing) was when I received my first manuscript back all in red. I looked at it sentence by sentence, and paragraph by paragraph, and man have I learned a lot! So don't look at that deadline in three weeks as the end of the road, but as the beginning. After all, it's not as if the supervisor doesn't know your output so far. So take it exactly for what it is: feedback that's more detailed than what you have received so far.

Other than that, it's very helpful to find similar papers, papers you thought were particularly clear, and look for hints there. Personally, I like short papers that try to drive a single point home, where the intro is short and to the point and where the discussion doesn't go too much beyond the data obtained. Such papers give up on telling you everything the authors know about a question, or even discussing all possible implications of the data, in order to keep concise. But that's just me.


To echo cagirici. THREE MONTHS IS NOTHING!!!!! I, took, over a year to finish my first manuscript.

My biggest advice is to put the paper away for a week and revisit it after you've done some reading (and not necessarily in the area that you are writing in, but in areas surrounding that topic). I cannot emphasize enough how putting away your work for a bit of time can help clarify the process and help things become clear within your writing.

You must log in to answer this question.

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