How to write a paper (and get it published)

The following post is based on a document I made a couple of years ago to circulate to MSc and PhD students I was supervising to encourage them to turn thesis work into papers and to show the process that takes papers from submission to publication. It seemed to me a good idea to recycle this here for a few reasons 1) it might be of use to other students out there who are trying to write papers for the first time, 2) I still use this document, so any feedback from other professionals that I can add in and update would be welcome, and 3) it might serve as an interesting essay for students or others as to the mechanics of writing and publishing a manuscript. I have covered bits of this before in other posts but this is far more extensive and practical rather than a simple review.

I don’t pretend to be much of an expert on this subject, but I do hope to be able to offer some significant points and reminders that will make the whole experience much easier for you since I have been an author on various kinds of manuscripts, acted as a referee and an editor so all the bases are pretty much covered.

The first key point is that there are a vast number of styles for putting the paper together. I know people who have the whole thing ready mentally and simply type it out in one go, whereas I am the polar opposite simply putting down thoughts and notes as the occur to me and then building up from there with much editing and rewrites. The discussion is frequently mostly finished before I start on the methods (writing, not doing!). Provided the whole thing is coherent and well written you should not have too many problems however you chose to write it.

If possible when you start have a target journal in mind and an idea of what you want to say and how. Obviously this will depend on the size and scope of your paper, how good the results were etc. but it does make things easier if you can already put the style and formatting down while you write. Do try and read a few recent papers of the journal or on similar topics to help to develop your paper in an appropriate style‚ there is a big difference between a review paper and a formal description for example.

Some journals have a very strict layout (like you cannot include a ‘conclusions’ section) whereas other are quite free about how you structure the paper (with the framework of abstract-introduction-discussion). If it is purely a theoretical or review paper you might of course have no methods or results section for example. You might want to combine the discussion with the results in some circumstances, or reduce the introduction to a single paragraph. Don’t be afraid to structure the paper so that it reads in the best way possible, but do be mindful of straying too far from the standard format.

You are writing not just for experts in your field but others who might be interested, or are using your work for an unexpected reason. Make sure you keep things simple, explain complex ideas with figures or make sure they are clear, explain unusual terms of abbreviations, don’t assume the reader (or reviewer for that matter) knows everything that you do. You want to keep things short, but if it takes 2 pages to explain a point, then it takes 2 pages.

Title:
Often this is the last thing to be written, as like anything else in science as you progress you will find your work drawn to other areas and render the title misleading or inaccurate. However it is worth putting one down to try and keep yourself focused on what you want to say. People will not read the paper if it does not appeal to them, so try to make it as broad as possible.

Authors:
This can be a prickly area when dealing with more than 2 other people. Lines of communication are not always clear and people may not realise what contributions other authors have made. You will get people trying to force their way up the authorship chain without the effort and those who are just happy to be on there at the end of the list despite having done most of the work.

Personally I have found myself low on the list despite having done (as far as I could tell!) most of the writing and analysis and I know of people attach themselves (or attach others) without the knowledge or consent of the primary author. In short, it really should be the first author’s decisions who goes where based of what they did, but this can get tricky with office politics. Just be careful.

Abstract:
This should be short and succinct. Except in rare cases it does not contain any references. Do make sure you elucidate what you are testing, what methods you used, the results you got, and what you think this means. Do try and convey why your research is interesting, novel or important‚ even if it is only for a very narrow field. Again this is easiest to write at the end. Most people only ever read the abstract and conclusions of a paper so make it good!

Introduction:
This can vary enormously in length depending on the journal and the individual manuscript. Typically it is less than 2 pages (published manuscript pages are typically about 2 pages of a normal Word document in 12pt, single spaced text), and the purpose is simply to introduce the subject in hand. With palaeontological studies this often means both the group in question (e.g. pterosaurs) and the idea being tested (e.g. flight mechanics).

Introductions can often be pared down to a very short section. This is not an essay or thesis to be marked and you do not need to show the reviewer just how much you know about the subject. You are trying to convey to another palaeontologist who may read the paper enough of the background to the topic that he can follow what you have done and why it is important. A few paragraphs and choice references can be enough if it is a popular area (e.g. dinosaur feathers etc.). It can be worthwhile finishing with a sentence or two stating the purpose of the paper and what you intend to show.

Don’t forget a single good figure can save you pages of complex explanations. You would also be surprised how often editors like figures showing the animals in question. The editors and readers are human too, so a small figure of a ‘typical pterosaur’ or ‘famous feathered dinosaur’ or ‘variations in ceratopsian neck frills’ can help to sell the paper. Not everyone knows what a tanystropheid looks like! Be careful not to get drawn into the discussion here, it is easy to go beyond what is necessary to introduce the paper. You might want to talk about all the available interpretations of some morphological feature, but save it for the discussion, just mention that there are lots and they are controversial and contradictory and save the details for later.

Institutional abbreviations:
These go here. Do make sure you note them correctly as they are not always intuitive – the Natural History Museum in London has the ID code of BMNH as its old name was the British Museum of Natural History.

Materials & Methods:
This should cover just what you did and how you did it. If necessary this can be kept short (e.g. these methods follow those employed by Smith 2005 with the following revisions…..).

Do include what data you collected and how, and make sure that the methods were appropriate. If you had to make a choice over methods (e.g. to use wingspan over mass) then state why a given one was chosen.

Systematic palaeontology / locality:
For descriptions of material, this section comes here.

Results:
Try to limit this section to just the results. This can actually be hard to do without giving any elucidation which you want for the discussion. Make any necessary graphs, figures etc. simple and clear. Try to avoid colours and do not confuse the reader by including too many subsections or cross-references.

Discussion:
Obviously this will normally form the bulk of the paper. You should start by describing the results (the graph goes up) and then interpreting them (this suggests a progressive increase in factor X). From this you can expand to what this may mean for the group / ecology etc (an increase in factor X may promote speciation and thus recovery from extinction etc.).

Try to keep speculation to a minimum and justify your assertions, but do be inventive and if you think you have good evidence for a trend then say so. Do also say if you think this is stretching the point and acknowledge any flaws in you data collection / analysis etc. Remember people (and especially editors) want to read something interesting so try to be positive (‘this means that’, not ‘this *may* mean that’) and exciting. This is annoying as I would prefer to be less forceful in my statements if I am not convinced of my results but that is not always an option.

Conclusions:
Keep them short and to the point. Try to make the summary clear and suggest what other research is required in the future etc. Again many people will only read this is detail so it is worth spending extra time on this.

Acknowledgements:
Thank reviewers of the paper, anyone who sent you data, curators who gave you access to collections and the source of your funding.

References:
Do check they are all listed correctly and formatted properly. I have had more problems with these than anything else!

Appendices:
These are getting increasingly large with online publication. They remain a place to put all of your source data (with references where appropriate) and extra analyses, graphs etc. that could not be fitted into the main manuscript. On many letters to Nature and Science the appendices exceed the length of the publication with text let alone the actual tables etc.

Submission:
When you have written your paper (and prepared illustrations, tables, graphs etc. and cleared it with your co-authors [very important this]) you can prepare for submission. By now you should have a target journal, so ensure that all of your formatting is correct and appropriate.

I like to have two different people read my papers before submission, one who is within the field and one outside. This should ensure that what you are saying is correct and well reasoned, and can be understood by someone who does not normally work in your area. This should catch any obvious flaws in the reasoning or areas that are fudged or badly explained.

When you submit (and online submission can be really fun), you should also include a covering letter to the editor. This need not be longer than a few lines and should state the title of the paper, what is discusses and why it is appropriate for the journal. Obviously everyone is doing the same thing, but try to give it a bit of spin‚ why the results may be of interest to researchers in other fields, does the specimen represent something new (oldest known taxon, first from a continent etc.). Again, this is something I dislike, your work should be valuable of itself, OK a new Archaeopteryx will be more interesting than a partial theropod tooth, but I fail to see why a paper has to be ‘sold’. Still, that is normal, so play the game and give it a bit of a shine if you can. I have seen terrible papers do superbly well though selling the concept rather than the content and brilliant papers languish because they went to a mediocre journal with no fanfare.

Sometimes you are asked to recommend reviewers or even to select those you do not want and these should be chosen carefully. Do pick an expert in the field who is not closely tied to you or your work, but beware of overly aggressive reviewers, or those whose work you are criticising. Everyone tries to be fair, but well, we are only human and you are unlikely to get a review from ‘Smith’ if you have just submitted 50 pages explaining why his last 20 years of work are wrong.

The review:
Assuming you have chosen the right journal and written something good you paper will go for formal review (if not it will come back very fast and you will have to choose again, of course this is far more likely with some journals that have a huge submission rate like Nature). This can take time so do be patient, but if you have not heard anything for 3 months it is probably worth writing a short e-mail to the editor or handling editor to see what the problem is. One of my papers was once in review for 15 months! So this can take a while. Typically papers are reviewed by 2 different people with perhaps some extra comments from the editor though rarely this will be 1 or 3 reviewers (I once managed 4 referees plus the editor!).

There are 4 possible outcomes to a review:
1. Straight to publication. Basically no problems with the manuscript bar a few spelling errors etc. The Holy Grail, this will probably never happen, and certainly don’t expect it.
2. Minor corrections required. There are issues with some of the data, methods or discussion. This will need to be resolved to the satisfaction of the editor, but should not be too hard to complete.
3. Major corrections required. There are (potentially) serious problems with the paper and it will have to go to review again. This is not necessarily a disaster. The manuscript can still be published in the journal, and the required corrections are not always that bad. Some top journals like Palaeontology often send things for a second review that have only very minor errors as a matter of policy. Read the letter very carefully to see exactly what they are saying and what it means.
4. Rejection. The manuscript will not be published and cannot be resubmitted to the journal (sometimes this is possible if the whole thing is completely redone, but it is rarely worth the effort). Still this is often better than an immediate outright rejection without a review as at least you should have some extensive feedback on the text and can use that to improve the paper. Hard work awaits, but it is not lost.

Rebuttals & Resubmission:
Once you have your manuscript back, you can make the necessary alterations to satisfy the editor and / or reviewers. Typically this will involve picking up all the minor errors (missed references, spelling and grammar errors etc.) and then revising those sections of the paper that the reviewers were unhappy with. This can be an exhausting process and requires a lot of work. Pay careful attention to this as failure to do so is likely to get your paper rejected. As you do these corrections you should keep track of what you have done. When you return the paper you should include another letter to the editor to cover what you have done (in the case of a second review, the reviewers will see this letter). Major areas of revision should be noted carefully (e.g. ‘In line with the views of referee 1 we have carefully revised the results section. Four new graphs have been created as suggested and the existing ones revised to include the suggested statistical measures….’).

Minor corrections can be ignored in this letter (provided you have done them) but where you disagree with a referee you should state why and justify your omission of his suggestion. You should of course thank the editor and reviewers in the acknowledgements too.

With this done, you can now resubmit the paper. Now you will receive a straight rejection or acceptance. If there are a couple of minor flaws it will probably be accepted and you will have to quickly revise a paragraph or a figure to satisfy the editor, but more than this and it is likely to come back.

If you have a major problem with the comments of the editor or more likely one of the referees, you are entitled to ‘appeal’. Write back to the editor and state clearly why you feel the review was unfair, incorrect or badly skewed or whatever. Editors are willing to list to a well worded case, but don’t do it on impulse just because you got a bad review. I had problems once where the referee had clearly mixed things up somewhere and was asking me to correct figures that did not exist and edit sentences that had already been removed. When I pointed this out I got a letter from the editor to tell me the referee had complained that my comments had been ‘offensive’, so be careful about how you approach this.

Proofs & Publication:
A few weeks after the manuscript is accepted you should receive the proof copies. These should be checked carefully for errors and returned as soon as possible with amendments. Your paper is now officially in press, but sadly with the speed (or lack thereof) of most journals you will be lucky for it to be published online within three months and actually in print within nine.

That is everything I can think of for now, so it only really remains for me to add: good luck! Things can be slow and tortuous, and it can take years to get something through to publication but it is (generally) worth it. Don’t be put off by huge swathes of comments form co-authors or referees and big gaps in time. It takes time, practice and experience to write a good paper (not that I’d know) and you will get better. The first can be a real slog, but hopefully this will give you a decent headstart.

Now, get back to work.