A couple of colleagues asked about how I write, and what they could do to improve their writing. There’s no catchall solution for everyone, since everyone works differently. So I though that I would start by outlining how I go about organizing a manuscript (or report or thesis chapter). This is the first part of a series of Friday posts on how I write, and some tips or tricks to help folks along. This is by no means THE way to write a paper, but it works for me. The only way to find what works for you is to try different things until one clicks.
First, I start with an outline – the major sections (abstract, key words, introduction, methods, results, discussion, acknowledgements, literature cited, tables, figure legends, figures). I might end up deleting some of these later (if I don’t include any tables, for example), but that’s OK. Then I break each section down into subsections or even paragraphs, usually beginning with the methods.
If we’ve done a study, we all know what methods we used. Subsections here could include study site, sample preparation, and usually one for statistical analysis. As I go through, I’ll make little bullet-point notes to myself like “treatment differences – ANOVA/Tukey’s”, which would remind me that I used an analysis of variance with Tukey’s post-hoc test to look at differences among treatments.
Then, I usually go to the introduction and discussion. “Wait a minute!” I hear you say, “Those are at opposite ends of the manuscript!”. Yep, they sure are. But like bookends on a shelf, they should mirror each other, so I outline them together. I don’t know if my technique has a formal name, so I’ll call it the “inverse pyramid” technique. I start in the introduction and write the main topic of each paragraph. The first paragraph is usually very broad, and outlines the conceptual framework. The next series of paragraphs get progressively more specific until the last one, where I outline my study’s objectives. Here’s a quick example of an outline for a manuscript on writing blog posts about writing:
- Writing is an important communication tool
- Organization and presentation are important
- Scientific writing is specialized
- Academics receive little formal training in scientific writing
- Especially try for grad students – expected to write a large thesis / series of papers
- Most grad student training in writing comes from peers
- Objective: outline my writing method
But recall that I said I outlined the discussion at the same time? It’s paragraphs / subsections are the inverse of those in the introduction. I mean, why introduce something if you’re not going to discuss it? The first paragraph summarizes the main finding(s), and subsequent paragraphs work back up the ladder of topics covered in the intro. The last paragraph or two puts my results in a broader context. In ecology, this is usually the “who cares” part – why would someone not studying your species or your site etc. want to read your paper? How does it fit with the conceptual framework you presented in the intro?
Then I “fill in the blanks”, so to speak, fleshing out each subsection to a full paragraph or two (depending on how detailed my outline was). Not only does this help me keep from wandering off on digressions, but it also makes writing manageable. If I can knock off a small paragraph in the intro, it’s like crossing that outline point off my list.
Sections like literature cited, tables and figures (and their legends) usually settle themselves in the course of the other sections. Then I go back and write the abstract. In my 3rd year of undergrad, I took a course in theoretical and evolutionary ecology. This was my first exposure to writing abstracts, and my prof at the time gave us a paper with the abstract removed. Our job was to supply the summary, and she gave some simple guidelines: 2-3 sentences for the intro, 1-2 for the methods, 2-3 for the results, and 3-4 for the discussion & main conclusions.
Subsequent posts will focus on particular aspects (figures and tables are up next week!). If there’s something you want me to cover, leave it in the comments below.