Monday, June 17, 2013

General Advice on Writing!

Hello! It's been a long time. I am preparing for an epic trip to Barcelona for the International Congress of Vertebrate Morphology ( I am hurriedly re-analyzing data and putting together my talk. I'm also attempting to submit at least one manuscript before leaving in three weeks (we'll see!), which reminded me of a topic I wanted to post about, writing! Disclaimer: I am not necessarily the world's expert on writing scientific manuscripts but I do have considerable experience and hopefully some useful advice. This is not intended to be an instructional on "how to write an introduction/discussion section." You can consult numerous books and courses on scientific writing.

So, you're part way through your graduate school experience and it's time to start writing your first manuscript for submission to a scientific journal. What's the first step when staring at the awful blinking cursor?

1) Outlining

Yes, since high school we have all been told we should create an outline before starting the writing process. But we're not necessarily taught how to make a good outline. Here are what I consider to be bad and then good outlines. I have used the introduction section as an example but the general advice applies to all sections of the manuscript.

Paper title: Cuteness index for prairie unicorns


Write something about unicorns
Write something about cuteness
Unicorns are awesome and breathe rainbow fire


Unicorns are related to horses

  1. Closest relative is extinct Dinohippus (cite Fraser et al. 2013)
  2. Possess horns
  3. Much cuter than other living equids (cite Equus and Fraser, 2012)

Unicorn cuteness can be measured as the index of horn length to tail length

  1. Horn length is a measure of breeding vigor (cite James and Cutie, 1988)
  2. Longer tails make unicorns more attractive mates and are proportional to strength of rainbow fire breathing (cite D'ior and Gabbana, 2001)
I won't go on any further. The point I am trying to make here is that a good outline should be detailed and guide your writing on the topic (even if it isn't unicorn cuteness). A good outline will bring together the relevant information and literature.

2) Start "barf writing" (also called stream of consciousness writing, if you prefer something less grotesque)

I don't mean projectile vomiting at your computer screen (although I have had the urge once or twice!). Once I have made an outline, I start filling in the gaps. Importantly, I don't spend much, if any, time editing my writing at this stage. Why? It's better to get your ideas and thoughts on paper and then cut them down or add to them later as needed. Additionally, if you "edit while you write" it will take exponentially longer to write the paper. 

Your "barf writing" will likely be terrible. But that's okay! No one has to see the paper at this stage. I have written some absolutely awful sentences during barf writing but that's what editing is for!

3) Edit like mad (re-write entire sections, if need be)

I am using the word edit very broadly here. I say I am editing even when I am re-structuring paragraphs and moving sentences from one paragraph to another. This is the stage where I asses the flow and structure of the paper. The introduction, for example, should introduce the problem of the study early on and then explain the details of the problem and how it will be solved. I carefully assess the flow of each sentence and it's placement in the paragraph/entire section. Anything that doesn't fit, is cut. The introduction/discussion (at least) should tell a story that flows logically from one element to the next. Resist the urge to tell the reader everything you know about the topic. We don't want to bore them! We want to lead them logically through the problem, convince them it is important/interesting, and tell them how our study is a great way to address it.

I admit this is the most challenging part of writing a scientific paper and it takes the longest. But I find it the most rewarding. As a coherent paper starts to form, I enjoy the sense of accomplishment. 

I wouldn't call this post exhaustive advice on writing scientific papers but it is a start and I consider the points above to be among the most important. If you have any tricks or advice, leave them in the comments!


  1. Ha! I call 'barf writing' the 'word vomit' stage of writing.

    This probably makes me weird, but I very rarely write much of the introduction first - usually I just write a little bit of background to get me started. Methods and results are the easiest to write first, so I usually do those, then spend a long time on the discussion, then go back and write the introduction to make sure it actually introduces what I'm talking about.

  2. I totally agree. I often start writing the materials and methods before I am fully done the project. I will generally also write the results second. Sometimes, if I am waiting for analyses to come in, I may start the introduction early but I do sometimes have to edit it as my thinking about the paper evolves.

    I only used the introduction as an example. I did intend to add something about the order of writing but I forgot!

  3. I disagree. Unicorns are superficially equid-like pecorans.

  4. You'll just have to write a strongly worded response to my unicorn paper!