POSTS

My Ten Commandments of scientific writing

- 11 minutes read - 2252 words

The writing of a paper (or the dissertation itself) is often a long process, along which many decisions are made: should I send my ideas for feedback now, or generate more polished text? should I think of the target journal now or decide once I have the finished draft? et cetera. To finish this mini-series of posts on writing (why writing papers is hard, how I write papers, and the second part of that writing process), I review here the main principles and lessons that I have learned after more than 10 years of writing scientific papers. I hope they help you navigate these decisions if you are in doubt, or if you have to step out of the usual writing process due to unexpected events.

Let me start with a personal story. A story of failure, and eventual success, after a lot of hard work and harder decisions. The story of a paper I co-wrote, which took more than three years to be published. For a few months, four of us had worked at a literature review, made a conceptual framework out of it, and run some case studies to see if the whole idea made sense to others. We thought there was something there worth communicating to others. We wrote the paper with lots of outlining and feedback, we sent it to a good conference in our field, and back it came… a rejection. Our paper was not convincing or clear enough, maybe tried to be too many things at the same time. One year into this side project, we seemed to be back to square one. Should we run more studies? should we rewrite the whole thing differently? should we just take it as it was and send it somewhere else? or just forget about it altogether?

Decisions, decisions, decisions… we have already established that writing is hard and, even if we have a process to guide us in this endeavor, sometimes we don’t have full control: a colleague comes with a full draft that is messy, or we have to re-do the whole paper after a peer review. In those cases, it is good to have general principles, values, heuristics to help us decide what to do. Below, I offer the most valuable lessons I’ve learnt after more than 10 years of doing this with moderate success. Half-jokingly, I have put them in the form of Ten Commandments to follow when writing scientific papers1:

  1. Always think about the audience first. We often think of papers as our accomplishment, something we possess, which we “hoard” in our CV. However, it really is an act of communication, and the real success is to get somebody to read, understand and be interested in (and eventually cite and build upon) what we found out in our research. Thus, we need to find the right receivers, people like us, who may care about our findings (i.e., choose the target journal/conference carefully). And, of course, we should try to make the paper easy for the reader to get (vs. easy for us to write): provide clear structures, examples, illustrative graphs and photos, cut down unnecessary jargon, etc.
  2. Scientific writing is about getting ideas across. Along with the audience, the most important thing in any communication is the message. We should stop worrying what people will think of us by reading the paper (‘do I write well enough? do I sound smart enough?’). What people will really care about is the new knowledge that we are sharing. Clarity is probably the main value in scientific writing, ambiguity is the enemy, and obscuring or making our text more vague on purpose is the biggest sin. We should be concise and simple, if things can be communicated effectively in that way: people are busy and they will appreciate getting the main point of our research in three pages rather than 30 (see the 1st Commandment).
  3. Go from core ideas to actual text (top-down approach). This may be just a personal pet peeve of mine, but I think this principle is crucial to have a sensible structure, logical coherence… in short, to achieve the clarity of message required by the 2nd Commandment. We should always go through outlines and bullet-point ideas before writing a single line of draft text. If you personally prefer to start writing text directly, that’s fine… but then reverse-outline your text and work on the outline: find logical gaps, repetitions of the same arguments, etc. (and be ready to throw away lots of that draft, in compliance with the 1st and 2nd Commandments).
  4. You never get it right the first time. I have never, ever, written (or seen) a first draft of a scientific paper that is published as-is in a good peer-reviewed journal. We should forget about that notion altogether. We will make mistakes, we will fail to see ambiguities, or people will misinterpret our words. And that’s OK. Breakdowns and repairs are an essential part of any communication. We need to “normalize failure” when writing, fail early and fail often, learning from our mistakes until we find the right way of communicating our ideas. That’s why I see writing as an iterative process, very much like designing any other object. We need to think of each abstract, each outline, each draft, as a prototype of the communication act we’re attempting. And, eventually, build the paper that synthesizes all that we learned about communicating those ideas. But, how can we see the faults in our own prototypes, the creations we have crafted as best as we could? Read on…
  5. Feedback (from co-authors, from the scientific community) is essential to good quality writing. Every time I write a paper alone, I find that the quality of the ideas and the end result are just passable; however, once I mix other people in the process, they start seeing logical gaps, unnoticed connections to other ideas or extant research, and in the end the result is so much better. The same goes for the peer review process: even if I don’t always agree with them, reviewers’ different perspectives (or misinterpretations, even) make me realize that the text was not as unambiguous as I originally thought. Of course, this is not magic: we need to choose co-authors wisely (and choose journals/conferences where we know the review process is tough but high-quality), learn from our mistakes and our good choices. And if we think of writing as something we are always learning to do better (and we should, even after we become the respected Professor), feedback is a well-known, necessary component2. There is no mastery without feedback3!
  6. Release early, release often (or “showing incomplete work is OK”). This is just an attempt to hammer down the 3rd, 4th and 5th Commandments again, by taking a well-known motto from the software development industry. If we want lots of feedback, and we wait until we have a complete paper before showing our ideas to others, it’s going to take forever to do even 2-3 rounds of feedback (if we’re lucky enough to get such patient co-authors). The only alternative if we want four, five or more rounds of feedback, is to show incomplete work. There is no way around it. We should forget our perfectionism and what our advisors/co-authors will think. Sending them an outline today is better than sending a draft next month. Besides, co-authors tend to be busy people – it may be easier for them to find half an hour to read our outline this week, rather than a whole morning to read a complete draft anytime next month.
  7. Communicate effectively. As with any other collaboration, communicating clearly, politely, and respectfully goes a long way. We shouldn’t take people for granted, everyone in the pipeline (from our co-authors to the editors, or the reviewers) is doing this work of helping improve and publish our paper on a voluntary basis, for free (or for quite cheap). Saying “thank you”, “please”, laying out expectations clearly, meeting deadlines, being open to negotiate, etc. are more important than most people realize. Also, let’s remember that most actors in this business are more experienced than we are, and that they may have a point (even if we sometimes disagree).
  8. Learn to use a few tools to make your life easier. If writing is a craft, we need to learn the craft’s tools. Not all the tools out there, but at least the ones that we will need very often. Reference managers (lately I use Zotero), collaborative editors like Google Docs (or Overleaf if submitting to journals that use Latex), the sometimes-hated-but-ubiquitous Microsoft Word, submission systems, diagram drawing apps (e.g., Google Draw), Excel, R or whatever software we use for generating graphs… all are part of the writing process at some point. We should learn to use each of those, and how they interact (can my graphing app export graphs that the submission system accepts?). There’s no need to spend too much time fiddling with them or trying new ones. But, if we find that our current toolbox gets in the way, let’s go out and find another tool that makes things easier in this and future occasions. Ask around; probably others have faced the same problem we now have.
  9. You only learn about writing by writing. Despite the emphasis on outlines, feedback and other tricks to make the writing process efficient, in the end we have to write, and write well, write clearly. And practicing a lot seems to be necessary (but not sufficient3) to be good at anything. So we should write, write, write any time we can. Write for ourselves. Love writing. Have a “craftsman mindset” about our writing4. If we find some good writing courses (scientific or otherwise), by all means we should try that too. Being skillful at clear writing is the single biggest academic and professional booster I’ve found in my life (and I’m not alone in thinking that).
  10. (insert your own) In the end, writing is also a creative activity, and in that sense it is very personal. There is no single best way to do it that works for everyone. Once we comply with the nine Commandments above, what else is important? Find your own principles, your values, your voice, your process. And put the main idea here (or make an 11th, 12th Commandments – who says they have to be 10?).

Just for the fun of it, I also tried to summarize the Ten Commandments above into two Great Principles of writing. I came up with this:

  • Writing is not about you, it is about how you can help others with the new knowledge you discovered, and

  • Papers are designed communication artifacts: prototype, iterate and collaborate around them!

I recommend you have these Commandments, these principles present every time you start writing something scientific. You could even print them out and put them somewhere visible, in the place where you do your writing:

Ten guiding principles when writing scientific papers, which can be summarized into two

My Ten Commandments of writing

Back to the initial story of our failed paper, these principles helped us navigate the decisions during the following two years. It was obvious that our proposal was not clear enough (2nd Commandment) and a change of approach was needed. We re-did the paper entirely (4th), going from a clearer main idea for the contribution into a totally different outline (3rd). However, we still were not convinced that the new orientation was strong and clear enough. Hence, we enlisted the help of a new, fifth co-author, a very well-respected researcher in the scientific community we had as the target (1st). We sought feedback from him (5th, 7th), even before we wrote a single word of actual draft text (6th). He saw the potential of the idea, but pointed out a few flaws in our approach. We addressed those, and started writing and writing (9th). Several rounds of feedback ensued among ourselves (5th again) until we all were satisfied with the text. Of course, throughout this complex process we had to share documents, keep our references under control (especially since it was a literature review), keep track and be able to reproduce our diagrams, tables, and other artifacts (8th). We found a new target venue for our paper by politely contacting the editors (7th) of a journal which was quite interested in the field we were targeting (1st), and sent it there. Of course, revisions came from the journal (5th), and more writing had to be done (9th) to clarify, and polish, and make sure that the audience (1st) would get the message (2nd). But eventually, the paper was accepted!

So, get off the couch and start writing! Heed the 9th Commandment. May you have lots of success with your future papers, and let these principles guide you to a place where writing papers is actually fun. I know that place exists – I’m having a drink there right now.

Do you agree with my nine Commandments of writing? What would you pose as your “10th commandment”? Let me know in the comments below!


  1. Of course, you will notice many similarities between these principles and the 10-step paper writing process I outlined in previous posts. It is these principles that shaped that process! [return]
  2. Hattie, J. (2008). Visible learning: A synthesis of over 800 meta-analyses relating to achievement. Routledge. [return]
  3. Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406. [return]
  4. Alley, M. (1996). The craft of scientific writing. Springer. [return]
comments powered by Disqus