Good storytelling is essential to write a successful research paper. And I don’t mean making things up. Science stories need to rely on data. But we need to write up the data as a story so people get it.
Key to good writing is building a coherent story.
It is important to emphasis that your research paper doesn’t have to be comprehensive - you can’t address every caveat or cite every relevant piece of work - but it must be coherent.
But what is a ‘coherent’ story?
One conceptual model (I love models!) for thinking about what a coherent story is to think about a web of cause and effect.
Your paper should build a web, or in the simplest case a chain, of cause-effect actions. If everything in your paper is related by cause and effect, then it will be coherent.
A model of a coherent research story as a cause-effect web.
There are a few more rules to consider, but it will be easiest if we bring up an example.
Consider Dale Bryan-Brown’s recent review of studies of population connectivity in marine ecosystems.
He opened the study with an observation that there are many studies of population connectivity in marine systems. That is the effect. Then he moves on to explain they have never been synthesised - a new ‘cause’. He then builds an argument around why a synthesis is needed to fill this knowledge gap (an effect again). And so the paper goes.
Now, we want to make sure we don’t have any extraneous causes or effects or cause-effect pairs floating around in the text. Such loose cause-effect pairs will contribute to building an incoherent story.
An example of this would be a recent paper we published about a new statistical method, which we applied to study the effects of pollution on coral reefs. In the Discussion we bring up important caveats to the new method and discuss ways future studies could improve on what we did.
We don’t spend time discussing uncertainty around the physiology for how pollution kills corals, because the main ‘story’ is about the application of the statistical method. Discussion of physiology would be extraneous and therefore incoherent.
The Discussion and Conclusion of the paper should tie up most of the cause-effect pairs, ending on effects (i.e. the observation and implications of your study). However, you may want to specify some future research directions too. These are ‘causes’ that you want future studies to go out and answer.
So I hope you find this conceptual model useful. Let me know what you think on Twitter.