Weaving together your next paper: 6 “whats” and a “why”
When I was an undergrad, the professor I most admired was always hammering home the point that when we write a scientific paper, we must not fail to include these “3 Essentials,” in order: What we did; what we found; what it means. These are the barest requirements for telling any good story – and if a single element is missing from your paper, you can be sure that it won’t get more than a mournful shake of the head from the editor of any self-respecting journal.
I think that because we were undergrads, our teacher let us off easy by giving us just three elements to focus on; yet, it’s surprising how many papers come across my desk that are either missing one of these elements (usually the third one) or have gaping holes in two or all three. Without help, these papers would be rejected out of hand, or at best, the editor or reviewers would try to thread the disparate fragments together, filling in the missing pieces with inferences that might or might not be correct.
Trust me when I say that even if reviewers return a detailed summary of all that is missing from your paper, and even if they offer you the opportunity to revise and resubmit, somewhere in the backs of their minds will lurk the sneaking suspicion that you might be too much of a nincompoop to ever get it right.
But we both know that’s not true.
So in the interest of helping you avoid this painful fate with your next paper, I’m taking an expanded look at the “3 Essentials,” transforming them into the “6 Whats and a Why.” If you have a manuscript you’re currently working on, or a project you’ll be writing up soon, take this checklist to hand. I guarantee that it will boost your chances of a favorable review significantly. And even if you don’t get a glowing review, nobody will be able to accuse you of being a nincompoop. I’ll go into more detail about each element in separate blog posts; for now, let’s focus on the basic structure:
WHAT is the subject of this research?
I’ve edited far too many papers in which it took several paragraphs (and occasionally the entire manuscript) before I had any real clue as to what the paper was going to be about. When you set the stage for your paper, don’t start at the beginning of time (unless you’re studying the origins of time); let the reader know within a couple of paragraphs what you’re up to.
WHAT is already known about the subject?
Journals differ in how deeply they expect you to review the existing literature, but suffice it to say that if you spread the lovely picnic blanket of your paper by declaring that you are going to study “Unwanted intrusive thoughts and the growth of facial hair” (Jurac, 1996), your reader is going to expect you to explain why you think these two subjects are related. Tell the reader just enough to convince her that you have a good reason to think this topic is worth exploring.
WHAT do we predict, given what is already known?
If I had to choose one aspect of the Introduction that most often needs significant work, it’s this one. Many (astoundingly many) authors make no prediction at all; make a prediction that is so obvious, one wonders why they are bothering to test it; or make a prediction that doesn’t follow from what has just been discussed. Oops! The reader just tumbled through a large hole in the fabric of your story, and must clamber through your Introduction again, hoping to find his way back to a place where your paper made sense.
WHAT did we do to test this prediction?
This is another area where the logical thread tends to get frayed or broken. Somewhere around the last paragraph of your Introduction, when you’re making a clear prediction that follows from the research you’ve just reviewed, give the reader a hint about how you’re going to test that prediction. The “hint” could be as simple as saying, “We used a repeated measures design with three levels of facial hair (unrestricted, trimmed, and shaven) to test this hypothesis.”
WHAT did we find when we tested it?
The second-most common weakness in this area is to use an inappropriate analysis or to report the results incorrectly or incompletely. My best advice is this, and I can’t emphasize it strongly enough:
If you don’t have the statistical expertise to confidently choose, execute, and defend your methods for analyzing your data, find someone who does.
There are few experiences more daunting than trying to respond to a reviewer who is tearing your paper apart for its poor analytical approach, and not being able to understand the criticism because you lack statistical literacy. We needn’t all be statisticians, but we all should know where to find a statistician when and if we need one.
WHAT do our findings mean?
I can’t recall ever editing a paper in which the authors neglected to provide any results of their investigation, but it’s quite common for the results to be reported without reference to any specific prediction that was made. You’ve told the reader what you expected to find; now tell her whether that prediction was supported. When you report your results, state specifically how they relate to your hypotheses, e.g., “The shaven condition was associated with fewer negative thoughts than either the unrestricted or trimmed condition [insert results of statistical tests]; therefore, Hypothesis 1 was supported.” For every prediction you make, at least one of your analyses should test that prediction explicitly.
WHY should anyone be interested in these findings?
A slightly less genteel version of this question is, “Who cares?” You’ve set up the research topic and your predictions about it; you’ve tested those predictions and reported your results. If your predictions were supported, discuss why it matters; i.e., what should stakeholders in this subject do with your results? Do they point to another study that will reveal new insights? Do they suggest an innovation in some existing treatment or policy? If your predictions were not supported, what can future researchers take away from that? If I had to estimate what percentage of papers fail to close the loop by saying why their results matter, I’d put it at 60%.
As important as these elements are, what’s really magical about them is how they are interwoven to produce a linked story that readers will be able to follow without using a rope swing to scale the chasm you’ve left between the logical points of your paper.
Durac, J. (1997). Unwanted intrusive thoughts and the growth of facial hair: a cognitive analysis. Behaviour research and therapy, 35(4), 371–372.