A few months ago, Shannon stopped by my office and asked a really simple question for which I did not have a simple answer. She asked, “How do you write a Discussion section?”. I think we’ve all asked ourselves that question, and usually answer it with the same answer we always give in science: “Well, it depends”. So we devoted an entire lab meeting to discussing this issue recently.
When we sit down to write the Methods and Results section, the organization is usually straightforward: describe what you did, then describe what you found. I might even argue that there is exactly one right way to write the Methods and Results for a given paper. The rules and style of the Introduction and Discussion sections of a paper are much more vague though. Because Dynamic Ecology posted earlier this week about how to write a good Introduction, I thought it might be timely to share our collective thoughts on how to write a good Discussion.
First of all, there is no single best way to write a Discussion, and part of why this is such a confusing issue, especially for students writing their first paper, is that each one is different than the last. So in that respect, coming up with a set of rules or a firm outline that applies to all papers isn’t that simple. But we tried anyway. The most important thing is that your Discussion ties the paper back into the ideas you presented in the Introduction. In it’s best form, an Introduction should start with a broad topic (in my opinion, if you’re not doing this, then you’ve already written a manuscript with too narrow of a scope that few will be interested in reading) and narrow down to a specific question that is being tested. If we consider the structure of a paper to be an hourglass, then the Discussion should answer that specific question and eventually broaden out to explain how the results inform the broader topic of the paper.
This is a rough outline for a Discussion with each number meant to indicate a separate paragraph, but obviously some papers will require multiple paragraphs to explain something, while others will not require some of these paragraphs. One might shift the order around too, although I’d argue that this structure best maintains the hourglass shape of a manuscript. Regardless of the order, be sure to have strong topic sentences for each paragraph. And end paragraphs with good transitions to the next paragraph:
(1) Summary of the results. This is particularly important if your results section was complicated or multifaceted. Use this paragraph to give a concrete answer to the specific question(s) you posed at the end of the Introduction
(2) Integration of results. Again, this is most important when results are multifaceted. Link the various results together. The results for photosynthetic rate showed one pattern, but the results for growth rate showed something else? This is the place to explain how one these traits might interact with each other.
(3) Explanation of mechanisms. Take your best shot to explain WHY you found those results. If you can do this with the data you’ve presented, that’s ideal. However, the mechanisms driving the patterns you observed might be beyond the scope of your study, so lay out some hypotheses that you or others might be able to test in the future. Keep your hypotheses reasonable and avoid arm-waving about mechanisms that are too far beyond the scope of the study.
(4) Caveats about results. Are there alternative hypotheses that might explain your results that you didn’t test in your experiment? Discuss them here and explain why they may be valid, or why they are unlikely to be true.
(5) Compare and contrast your results with previous work. Compare your work to other work in the same system, if possible. However, be sure to also compare your work to that in other related systems. For me, I might be interested in other studies on Medicago polymorpha, but more broadly how my results relate to other studies on invasive species.
(6) Future directions. What is left unanswered by your work? What’s the next exciting question to ask? Students often treat this as the “here’s everything I did wrong and would like to fix” section. Avoid that. Discuss how your work is moving the field forward towards even more exciting new questions.
(7) Conclusions about the broader picture. If you’re going to do some hand-waving speculation at all, this is the part where you might have a little more license to do so. If you’re going to speculate wildly, be sure it’s in the spirit of provoking new ideas and questions, rather than relating your paper to something that’s too far unrelated. Be sure to apply your findings back to the big general picture addressed in the first paragraph of your Introduction.
Happy writing, everybody!
