How many times have you looked around your meeting room and realised there are no (or very few) women there? Around 20% of Computer Science undergraduate students are female, but how many research software engineers (RSEs) are? A decade of survey data suggests that we are starting to close that gap.
Our new study "Women in Research Software Engineering: Insights from 10 years of a global survey” uses ten years of data from the International RSE Survey to explore how women's participation and experiences in research software have changed since 2016. The survey is just one example of the Institute for Research Software’s ongoing commitment to conducting research for policy, and the resulting dataset is the international benchmark for understanding the international research software workforce.
While women remain underrepresented in the RSE workforce, the balance is improving. In 2016, women represented only 11% of survey respondents, but that has risen to 18% in the data from 2026. And despite ongoing challenges around career progression, recognition, or access to training, women RSEs report high job satisfaction.
The survey shows that in many ways, the experience of women RSEs is very similar to their male counterparts. They report high confidence in their programming skills, particularly in Python and using Git, and software engineering techniques. When it comes to improving their skills, we’ve seen a shift in the skills women RSEs want to develop. Whereas men have consistently rated management as the top skill they want to improve, since 2022 women have been increasingly identifying management skills as one of their key areas for improvement as well. We believe this represents a cultural shift, with women RSEs recognising that explicit training focused on management and leadership can be a useful tool in overcoming some of the structural barriers to career advancement.
The 2026 survey was the first to ask specifically about generative AI use and revealed that women and men RSEs use these tools at almost identical rates. Women primarily use generative AI for writing or debugging code, or generating technical documentation, just as their male colleagues do. We observed an interesting difference in tool preference, however. While both genders most commonly use ChatGPT, women trust GitHub Copilot more than their male counterparts, which prefer Claude.
Despite the good progress made in the last ten years and high overall satisfaction of women RSEs with their work, there is still work to be done. The survey has identified persistent barriers to career progression, including limited access to advanced training and underrepresentation in senior roles, that will require focused attention from the sector in order to see a truly equitable research software workforce.
The study was originally presented at 13th ACM Celebration of Women in Computing: womENcourage 2026, and provides the first longitudinal, gender-focused analysis of the global RSE community.
The Institute is committed to producing evidence-led guidance for institutions and policymakers in this area. You can read more in our recent report on concrete steps for improving equity in the research software workforce.