I was always somewhat passionate about reading and writing. Throughout high school and my BSc/MSc studies, I also considered myself to be a good writer (for the most part). When starting my PhD, it was therefore a harsh awakening for me to realize that I knew little about good scientific writing. Some things I had been taught in German high school were even completely the opposite of good scientific writing in English, e.g., "never use first person". It took many years and several courses on academic / scientific writing until I regained confidence in being a competent writer who can effectively convey their thoughts to others.
My friend Roberto Verdecchia is similarly passionate about high-quality writing. When we discovered this shared passion and that we had independently curated a collection of writing advice for our colleagues and students, we thought: why not combine and publish this? So, that's what we did. The paper titled "Notes On Writing Effective Empirical Software Engineering Papers: An Opinionated Primer" was published by ACM SIGSOFT Software Engineering Notes (SEN) and there's also a publicly available preprint. This document represents the paper we wish we had as PhD students. It's pragmatic and admittedly opinionated: not a formal writing textbook, but a distillation of the advice we find ourselves repeating over and over to our own BSc, MSc, and PhD students.
It was fun writing this paper. Plus, it was interesting to discover that we also had some different opinions on things, e.g., whether to write the study design section in simple present or simple past. In the end, I think we managed to create a somewhat nuanced collection of writing advice for ESE papers and hope that the result is useful for at least some people who feel overwhelmed by scientific writing. Especially in the age of GenAI, I believe it's important to empower junior researchers to find their own voice regarding scientific writing. Hopefully, our paper can support some of them on their writing journey.
What's Inside
Good scientific writing in Empirical Software Engineering (ESE) is one of those skills that everyone agrees matters. It's an explicit or implicit review criterion at basically every SE venue, but one almost nobody actually teaches. Most of us pick it up the hard way: by getting our first few submissions rejected, or desk-rejected, or quietly dinged in reviews for "unclear presentation" without much actionable feedback.
The primer walks through the typical anatomy of an ESE paper: Title, Abstract, Introduction, Study Design, Results, and so on. For each section, it explains what it is for, what tends to go wrong, and concrete tips for getting it right. It also covers more general principles of scientific writing, such as:
- Concreteness Over Abstraction
- Avoid Wordiness and Be Concise
- Use an Active and Simple Writing Style
- Be Consistent and Conscientious
- Know and Guide Your Reader
Lastly, there's also guidance on language, formatting, and consistency, since those small things add up to a lot of the "reviewer appeal" a paper either has or doesn't.
Who It's For
As primary beneficiaries, we had our own students in mind, so this is definitely something you can point a BSc/MSc student or first-year PhD student to, or come back to yourself when you're stuck on how to phrase a tricky Related Work or Discussion section. But even experienced researchers might find the opinionated bits interesting to agree or disagree with.
If you know someone who could benefit from such a paper, feel free to share it with them. We are also interested to hear your thoughts on the paper, e.g., points you strongly disagree with or important things we didn't cover. Lastly, we are considering converting the paper into a slide set (especially if many people find that helpful).
Read It Here
- ACM SIGSOFT SEN version: https://dl.acm.org/doi/10.1145/3743095.3743100
- arXiv version: https://arxiv.org/abs/2506.11002