How to……1: structure a paper

I don’t intend to write lengthy, exhaustive tutorials the way the SVPOW guys do. They are much better at it than I am, and I’d be covering a lot of double ground. Also, I bet I’d bore most of my readers to death.

What I will do, however, is type up a short “How to” note whenever something important or helpful comes up among the penguins that fill roughly 90% of my brain. Today, let’s have a look at how to structure a paper. And I mean not the content, but the actual document, the sorry heap of words that are supposed to transmit the genius idea from my brain to those of others.

Say, you have done a nice study of something or other, or you have a great idea, a piercing question. Maybe all research work has been done, or maybe you just came up, by pure accident, with the perfect way of describing something or of posing a question. Now, you want to put pen to paper (fingers to keyboard, more likely), and jot down all the great sentences that jump to you mind. Now what?

As explained by Mike here, there are several approaches to this, and all are valid, and none works smoothly. However, for the task of giving what you write a proper structure, there is an approach that has never failed me. That Golden Rule of Structure is simple: Structure your paper first! Then write it.

OK, sounds weak, but I means this very literally. Before you write anything of the text, the abstract, or whatever, structure your paper! The hard way is taking a piece of paper and writing down the headings for the sections, then making a list under each of what needs to be mentioned in each. The easy way is using a word processor’s “navigation view” or “document structure” window (talking MS Word, different version here – most word processors have something like this). Here’s what it looks like when I write a paper on tail strikes in ankylosaurs (first draft).

Obviously, at this point all the words I put into the title and headings are placeholders. They may end up in the final paper, or not – for now they are there to tell me what the section is about. No more, no less. If I can find a zingy replacement later, I may use it.

As you can see I used the pre-defined headings of my word processor, with minor alterations to formatting that make the difference between heading levels more clear. At this stage, and for 95% of writing time, I do not give a flying piece of elephant dung about the journal’s style I will have to comply with later. All I make sure is that

1. there are at most three levels of headings (¹),
2. there are an Intro, a Materials and Methods sections, a Results section, a Discussion section, and, if necessary/helpful, a Summary (²)
3. there are not more than four sub-parts to each heading (³).

1: nearly all journals allow three; 2: check, some journals require a summary even if not needed or have rules on how to call what; 3: if you over-split into too many sub-sections readers will get too confused. This is a “soft” rule, follow as long as you feel good about it.

This framework is not fixed in concrete, and I diverge here by splitting Material and Methods – because of the the third rule: while there would strictly not be more than four parts, the materials part of Materials and Methods would be totally dwarfed by the methods. Split it up, KISS!

Once you have a simple, clearly understandable structure like this, you’re all set – theoretically! There’s still a billion ways you can end up submitting a paper the reviewers will call “chaotic” or “disorganized”, and that’s the first step on the road to an “Unacceptable” rating. Your research can be stellar, but if you manage to piss off the reviewers by annoying them with an unstructured manuscript, you’re very close to doom and rejection.

As you can see, I followed the time honored traditional IMMReDS structure. Introduction, Material & Methods, Results, Discussion, Summary. Is this structure the holy grail? Certainly not, and you should feel free to vary it, adapt it to your paper. Sometimes, stuff comes in two versions, both give you results, both need to be discussed, but they do not mix. In this case, alter the structure like this:

• Introduction
• Materials
• Methods 1
• Methods 2
• Results 1
• Results 2
• Discussion 1
• Discussion 2
• Summary

Or maybe better:

• Introduction
• Materials
• Methods 1
• Results 1
• Discussion 1
• Methods 2
• Results 2
• Discussion 2
• Summary

Depends on what serves a clear, understandable telling better. in this context remember that many people do not read papers in their entirety, but search for key terms or results. The more you deviate from the classic IMMReDS structure, and the longer your paper, the more you need to have a Summary that is worthy of the name, because people will “get lost” within your paper.

BTW: Some journals want these sections in a different order – that’s what [CTRL]+[x]-[Ctrl]+[v] is for. 😉

So why did I say there was an easy and a hard way? Because of this:

This is really neat: using a word processor and its predefined heading formatting, wherever I am in the text, a glance to the left will tell me what sections I am in, what sub-heading, and what else is there in my paper. Trust me, for a 40+ page manuscript this a huge aid. You can immediately see if what you want to write will go into the proper part, and you can see what other parts need a sentence or paragraph or two on the subject. Imagine, for example, that a new study on anky tail clubs comes out while I busy writing this paper. Obviously, I want to compare my results to it, and will add a sub-section to “Discussion” –> “Comparison to previous studies”. I may even be tempted to use (for now) third-order headings to structure that into parts for each previous paper. On the main text window it will be hard to find where else to add stuff on the new paper, but the navigation view tells me, at one glance, that something needs to be added to the Intro, and it tells me (and that’s more important), where exactly I should add that (right under Carpenter et al. 2011, with a third-order heading).

If you slavishly follow this rule of creating the structure first, filling it in later, you’ll never get hit by a reviewer’s disdain for being “unstructured”, and it will save you a lot of time when you make revisions, too. I only wish I had known about the Document Structure View (now Navigation View in Word ’11) before I wrote my monstrously long Plateosaurus and Kentrosaurus papers. Kudos to Martin Sander for teaching me!

By the way: I have found is best to approach this simultaneously in two ways: have a text module document, and the proper paper itself. That way, whenever you come up with a phrase that is “just right”, or remember something that must go into the paper, but you can’t work it into the proper paper, you can just jot it down for later consideration. Even if – and that is an important point, even if it does not fit into the intended structure at that time.

And as a bonus for all those who have read through the entire post, here’s a very nice Plateosaurus hand from the Frick/CH locality.

MSF 23 (Sauriermuseum Frick) Plateosaurus engelhardti
left antebrachium and manus in dorsal view.