How to write a strong discussion section for your CHI paper

Discussion sections are often considered by reviewers to be the make-it or break-it moment of your paper. Here, you show that you can synthesize your findings with the existing literature. You show your skills in arguing the unique scientific contribution of your work and your consideration of existing literature.

Knowing how to write a strong discussion section has the following benefits:

• If the reviewer had any doubts about your contribution, this is your big chance to convince them
• You can contextualize your findings, imbuing them with more meaning
• You can outline and shape the work that comes after this research
• You get another chance to discuss related work in light of your findings, engaging with the existing literature

Unfortunately, academic writers often miss this opportunity to strengthen their research and engage with it based on their understanding of the existing work. Too often I see people just writing a plain old summary of their work. That is not enough.

Discussion sections are like business cards for your current study. Do them well and you'll make connections.

This is really the best way to view a discussion section. You have a great chance here to convince people that your work has meaning for the scientific community in which you publish. If you write this well, it will lead to more citations and more engagement with your work. And this should be a goal for important research, you want it to be noticed and used by the scientific community.

Therefore, you should focus on the following areas in your discussion section:

1. Importance statement. At the beginning of your discussion, you want to quickly (really only very briefly) summarize your main results and then say right away why they matter. If you have many results, use this structure: "This was result x. It is important to HCI research because ..." and then do this for each result. One or two sentences outlining the summarize result and why it matters.
2. Findings extend previous work. Now, it's time to engage with existing results. Revisit your related work section (in writing mode, this is also a great moment to revise it and add some papers to discuss there). Make a table to tie your main results to existing related work. Did you find the same? Did you find something different? Can you explain this? Compare.
3. Scientific implications of findings. Here you want to reach out and oracle within the space of reason of your results will change things. Either a theoretical understanding of a concept or the way we approach a method. Or implications for life at large. Focus on the impact on the scientific community you are writing for first, then broaden to people at large.
4. Forward look to facilitating future work. What comes next? How are you opening up the research space for more work to come? Think about the future. What future work that either your yourself are capable of doing or others in your field? The more specific you outline the next steps, the better. Researchers often read discussions looking for inspiration on what to with their research questions. Provide this context for them.
5. Limitations of current approach. It is important to be honest about what you were able to achieve without lessening your contribution. Try to be succinct but fair in your evaluation of your own work and what it can explain. It is important to reign any visionary moments you had in point 3 back into context and say what conclusions are reasonable to draw from your approach.
6. How this research is important to the current field? Yes, we already mentioned this at the start, but here we are going full circle by just putting a little oomph at the end of your paper. It is good to end on a high note and emphasize your contributions once again. This will leave reviewers with a positive understanding regarding your impact.

And that's it. Focus on these six steps and you'll give your results section your best chance to make an impression in your current paper. The benefit of a strong results section is that it can really elevate a mediocre paper and tap into a committee's emotions when it comes to making a decision about the paper. Good luck with your writing.

BTW: Daniel Buschek shared his 10 tips for making a discussion better on Twitter with me. I specifically like how he recommends:

• You should already be writing your discussion section right now if you are planning to submit to CHI. It's no good starting too late.
• Discussion should be about 2-3 pages long (in the new format) don't write a discussion that's too short and lacks depth.
• Use subheadings to structure your discussion section. Nobody wants to read a wall of text. Also, if your subheadings are actual findings or takeaways, it helps skimmability of your paper, which is an asset for you.
•  Answer the question: "What does it all mean?" in your discussion. This is crucial for reviewers and readers.
• Connect back to the literature and don't just write a summary (I already mentioned this above).
• If you consider the academic hourglass, be aware to broaden your implications, move from specific to broad in your discussion.
• Explore alternative interpretations. This is best done in discussions with colleagues online or at your institution. Make your work matter beyond your niche.
• Point out contradictions and disagreements with existing work. Be critical but courteous.
• Explain your decision in your research that might limit your findings candidly.
• Here are the concrete writing prompts that Daniel gives at the end of his article:
  • Discuss the impact of each experimental condition or effect in a separate subsection. Relate it to other studies on this condition or effect in detail.
  • Discuss two ways of interpreting a certain result. Do you favour one explanation? Why?
  • Discuss divergent/convergent findings.
  • Discuss your study design choices. Reflect on them with your results. Would you recommend your method to others?
  • Discuss limitations (like sample diversity and size, study duration, novelty effects, development shortcuts, and aspects of internal vs. external validity).
  • Discuss broader impact (regarding, for example, society, privacy, and human biases).
  • Was there a lighthouse paper inspiring you to do this work? Relating your findings to that paper, and others who have cited it, is worth a discussion subsection of its own?
  • Addressing a critical or curious question from a peer would be interesting to many and is worth a subsection.

Buying an online course that can improve your CHI paper writing is like getting an oil change for your car. You can go a long time without it, but everything runs smoother with it. If you feel you could use the extra help for your CHI paper and you are a student, please reach out to me via email for a discount.