How to write the discussion section of an academic paper

This is one of the most challenging questions people have ever asked me, because after looking through dozens of journal articles in my Mendeley database, I could not find a lot of them who used Discussion sections. I believe this idea of the Discussion component of an academic journal article (or book chapter, in some cases) comes from the IMRAD model of publishing, that is, papers that have at least the following five sections: Introduction, Methods, Results, Analysis and Discussion (hence the acronym).

Personally, I neither like, nor do I often write this type of journal article. Even when I was a chemical engineer, I can’t recall that I read many papers in the IMRAD model, as they all had a variation (merging Discussion with Results, or Results with Conclusion, or Discussion with Conclusion). As I said on Twitter, I read engineering, natural science and social science literatures. Thusly, the Discussion sections that I read vary QUITE A LOT.

HOWEVER

All Discussion sections I’ve read are

1. analytical, not descriptive,
2. specific in their interpretation of research results,
3. robust in their linkage of research findings with theories, other empirical reports and various literatures,
4. good at explaining how a paper’s results may contradict earlier work, extend it, advance our understanding of X or Y phenomenon and, most definitely:
5. NOT the conclusion of the paper.

What I think is important to remember when writing the Discussion section of a paper, is to really ANALYZE, not just describe. Link theories, methods, data, other work.

My post on the difference between Description and Analysis should help you write Discussion sections. https://t.co/oxz8uIY3Pd you should all read Graf and Birkenstein's They Say/I Say https://t.co/yDXHawbez1 as preparation to write Discussions – for the rhetorical moves. </fin>

— Dr Raul Pacheco-Vega (@raulpacheco) March 14, 2019

As usual in my blog posts, I here link to a few resources that may be of help (written by other authors).

• Dr. Pat Thomson, as usual providing great advice on Results/Discussion sections of journal articles.
• A handy handout on what goes in each of the IMRAD sections.
• Note how this article by Sollaci and Pereira on 50 years of IMRAD articles does NOT have a Conclusion section (oh, the irony!). However, their Discussion section is quite nice, albeit brief.
• This article by Höfler et al offers good advice on integrating substantive knowledge with results to create a solid Discussion section.
• In this article, Öner Şanlı and coauthors provide great suggestions on how to write a Discussion section of a journal article.

In my Twitter thread, I suggested ways to discern (and learned from) how authors have written their discussion sections.

If you now read the Discussion section, you'll see that in my yellow highlights, I've noted how this particular article contributes to the literature. This is part of what should go in the Discussion section. More than explaining results, how your results link to broader debates. pic.twitter.com/a19hE5FB9d

— Dr Raul Pacheco-Vega (@raulpacheco) April 9, 2019

Discussion sections are particularly used in articles that follow the IMRAD model https://t.co/FzunG4tnce I like this Power Point on what should go in each of the IMRAD sections https://t.co/SQLVLsD6JB – what I've found is that often times, Discussion sections are blended/morphed

— Dr Raul Pacheco-Vega (@raulpacheco) April 9, 2019

There are times when scholars blend Discussion and Conclusions, or Results and Discussions sections. This is not even discipline-dependent, it’s author-dependent.

For example, in this #Free2DownloadAndRead World Development article, the discussion section is blended with the results. https://t.co/cgB82kYXla This is common, and I personally have no objection to doing this. As for PhD dissertation and discussion chapters: this is challenging

— Dr Raul Pacheco-Vega (@raulpacheco) April 9, 2019

Another example, now from the criminal justice field.

If you notice how these authors start their Discussion section, you'll see that they bring back their empirical results to the broader debates. That's what I have seen in most Discussion sections of journal articles (in engineering, public health and some pysch). pic.twitter.com/wpH9jGghjk

— Dr Raul Pacheco-Vega (@raulpacheco) April 10, 2019

Note how the authors link what they did in their paper with "other research that should have been done or that *somebody* should be doing". Overall, that's how I distinguish Discussion from Results (linkage to broader works/debates). Piece extracted from: https://t.co/mO4bN6VrBI pic.twitter.com/IQUzHNr3Pk

— Dr Raul Pacheco-Vega (@raulpacheco) April 10, 2019

Hopefully these notes will be helpful to fellow scholars writing these sections!

You can share this blog post on the following social networks by clicking on their icon.