Elements of Science: What are the elements of a great discussion?

Do you know the phrase “the data speak for themselves?” According to Dr. Stephen Heard, professor at the University of New Brunswick, past Subject Editor for FACETS, and author of The Scientist’s Guide to Writing, “The data don’t speak for themselves—or at least, they don’t easily or clearly. That’s why we write papers rather than just publish datasets.”

It’s also why papers have a discussion section. “It offers the writer a chance to explain how they think a study’s results should be interpreted and why they matter,” says Heard. The discussion section, he says, is an opportunity to consider and extend your findings to claim the most robust interpretation and the broadest importance that you can legitimately argue. “The aim is not to exaggerate either interpretation or importance…but to show the reader why they should care and how your work should change the way they think about the field.”

Sounds complicated? Fear not. Here are some tips to help you make your discussion section great.

Start with a framework

“I always found the discussion and introduction [sections] to be the hardest to write because, unlike the methods and results [sections], they don’t have an obvious structure,” says Heard. Fortunately, “they do have pretty standard content.” Heard says authors can take advantage of this to create a discussion section framework. He recommends including the following elements in order:

(1) interpretation of results to answer the research question
(2) consideration of possible limitations
(3) relationship of results to previous literature
(4) broader implications of answering the research question
(5) prospects for future progress

Similarly, using subheadings to highlight key points can help keep the discussion section focused and serve as a guide to the reader. To see an example of an article in which subheadings are used well, Dr. Wendy Ward, Co-Editor-in-Chief of Applied Physiology, Nutrition, and Metabolism (APNM) suggests reading Chan et al.’s 2023 paper, “Breastfeeding in Canada: predictors of initiation, exclusivity, and continuation from the 2017–2018 Canadian Community Health Survey.” Even when subheadings are not accepted or required by a journal, “it can be helpful to include subheadings in a draft version of the discussion section to help keep your focus [when writing] and to be sure that all the aspects that need to be discussed are, in fact, interpreted,” Ward says.

Discussion section from “Breastfeeding in Canada: predictors of initiation, exclusivity, and continuation from the 2017–2018 Canadian Community Health Survey,” by Kathleen Chan et al.

Focus is key

Even with a framework, researchers face the challenge of writing a discussion that meets journal word or page limits. Going into too much detail about each finding or summarizing at length the results of other studies are mistakes Dr. Sylvie Gauthier, Co-Editor-in-Chief of the Canadian Journal of Forest Research, often sees. You need not discuss every result; highlight your main findings. When referencing other studies, provide a brief 1–2 sentence summary of only the most relevant details or results.

Gauthier also cautions against concentrating exclusively on the specifics of the study (its particular region, species, methodology, etc.) and ignoring any implications that can be generalized and might be of interest to a broader audience. She encourages researchers to “place their work contribution into the larger context of their field. It is important to show how a wide audience can benefit from the work.”

As an example, Gauthier points to Reuling et al.’s paper, “Initial tree regeneration response to natural-disturbance-based silviculture in second-growth northern hardwood forests,” published in the Canadian Journal of Forest Research in 2019. Two elements particularly captured Gauthier’s attention—the concise nature of the text and the inclusion of a sub-section on management implications.

Discussion section from “Initial tree regeneration response to natural-disturbance-based silviculture in second-growth northern hardwood forests,” by Laura F. Reuling et al.

Note limitations – but don’t be too critical

All research has limitations. Rather than shy away from them, researchers should address them head-on. “I think it is important to not simply list what the limitations are but explain how these limitations may have impacted the findings and/or modified the interpretation of the findings,” says Ward.

However, Heard reminds writers not to tear their work to shreds. “It’s important to express the limitations of your work, and thus the limitations on your inference, but you can do that without suggesting to a reader that they wasted their time reading your paper.”

Rather than create a long list of potential mistakes or caveats, it’s much more useful to clearly explain possible sources of error and how they may have impacted the results. For example, “We might have underestimated quantity A because X might have failed for reason Y. If so, our conclusion B is too strong. Experiment C could resolve this doubt’.” Regardless of what sources of error may have crept into the study, clarity and transparency around those sources of error and what they mean builds trust in your work.

“This doesn’t mean exaggerating the strength of the work—it means reporting, but not exaggerating its limitations,” Heard adds.

Be aware of journal requirements

While discussion sections largely follow the framework outlined by Heard, it is always worth spending some time to familiarize yourself with the journal’s guidelines as there may be specific points to consider.

For example, APNM encourages authors to follow the SAGER guidelines (Sex and Gender Equity in Research). These guidelines provide helpful, comprehensive guidance on reporting sex and/or gender within a manuscript. Within the discussion section, “it is ideal for authors to discuss the potential implications of sex (or gender) on the interpretation of the study results. “If sex (or gender) has not been incorporated into the study, it is best practice to explain why this is the case and how this may have influenced the findings,” says Ward. “Not considering sex (or gender) can be addressed as a potential limitation, and the reason it is not considered should be adequately explained.”

Another thing you may come across in some journals is a conclusion section. This section should reiterate the answer to the paper’s main research question, explain why that answer is important, and identify a direction for future research to build on the findings. “Perhaps a conclusion section is best thought of as a mechanism for making sure a discussion [section] ends with the most important things,” says Heard.

Finally, embrace excitement!

The discussion section is likely the last thing someone reads of your work, so making the section as engaging as possible is important. One way to do this, Gauthier says, is to imagine you are giving a talk or an interview. “You would be more engaging if you were smiling and looking excited about your work.” The same principle applies when writing papers. Enthusiasm about your study and excitement about its outcomes will engage your reader. The more you engage your reader, the more likely the work will leave a lasting impression.

Tips, techniques, and expert advice on crafting great scholarly communications can be found in our Elements of Science blog series. Read about the elements of a great map, discover how to create a great briefing note or table, and learn how to write an impactful title and abstract.