Equity, Diversity, Inclusivity and Accessibility in Research Software Engineering: Guidelines for Making a Difference

The Research Software Engineering (RSE) professional community has continued to grow since the role was first named at a Software Sustainability Institute (SSI) workshop in 2012 (1,2). Unfortunately, RSE has a pronounced diversity issue: in the most recent international RSE survey conducted in 2022, around 75% of UK respondents identified as white and 71% identified as male (3). In fact, RSE is less diverse than software engineering in industry, academia as a whole, and the general UK workforce (4). The UK RSE community has responded to this with various Equity, Diversity, Inclusivity and Accessibility (EDIA) working groups and schemes, and EDIA is often a theme seen in workshops and conferences. However, it is possible that these efforts, despite their good intentions, could replicate often-noted criticisms of EDIA initiatives, such as generating documentation without any change to working practices and putting the onus of changing the profession on those who are most marginalised within it.

In this blog post, we share some background on our research with members of the UK RSE community, outline components and criticisms of EDIA initiatives, and offer the approaches we believe most likely to lead to improved equity in RSE specifically and in Science, Technology, Engineering and Mathematics (STEM) more broadly.

The InclusiveRSE research project

Our understanding of EDIA in RSE comes from working on the InclusiveRSE research project (5). This ongoing project is funded through the SSI and aims to understand the barriers to equity within the RSE profession and consider how these can be mitigated. So far, we have conducted semi-structured interviews to explore the experiences of research software professionals with minoritised identities such as race, gender, social class and disability status (6). We drew on the wealth of social science research of inequalities in STEM to develop the questions and conduct the analysis. The questions and analysis focused on education, career choices and how professional and personal identity interact. We will be expanding this research over the coming years to incorporate the experiences of a wider swath of the RSE community. Importantly, we are also considering how to use our understanding of both disadvantage and privilege in RSE to make practical recommendations for addressing these inequalities.

What we need to do to improve EDIA in RSE

To improve EDIA within RSE work practices and culture, we suggest that a number of related areas all need to be understood and addressed. First, raising awareness of the RSE role across different educational and career stages, as well as across employment sectors and academic domains, is needed so that a broader range of individuals can consider RSE as a potential career (4). However, this work also requires challenging the stereotypes often associated with software engineering (7–9). Developing multiple pathways into RSE roles would support individuals from varied academic backgrounds and institutions, including post-92 universities, which have more diverse student populations (10), to enter the profession.

As with many professions, reviewing recruitment and promotion policies and practices, alongside introducing associated accountability measures for these, is necessary to reduce bias (11). Related to this is the wider issue of crediting RSEs for their work in academia. Attention should also be paid to both formal and informal work allocation (12) to support equity of opportunity when in role, particularly given multiple short-term projects RSEs often manage. Professional development within the role is also supported by transparent access to training and computational facilities. At a different level, promoting inclusive everyday interactions among colleagues is critical for fostering a sense of belonging. Finally, providing accountability and transparency for EDIA commitments through measurable criteria, such as monitoring gender, race and disability pay gaps as well as representation data (13), is required for sustained progress and making visible any diversity issues (14).

Common critiques of EDIA initiatives

Progress towards equitable opportunities in RSE is unlikely to be without its challenges, as EDIA initiatives in academia are frequently, and fairly, criticised for failing to produce meaningful change (15). Some of the strongest critiques describe EDIA initiatives as box-ticking exercises that enhance the CVs of those already in positions of power and that generate income for EDIA consultancies while leaving the day-to-day realities and career prospects of individuals with marginalised identities largely unchanged (16–18). EDIA charter marks in UK higher education—such as Athena Swan and the Race Equality Charter—have been shown to provide accountability and contribute to progress (16,19). However, these frameworks also face criticism for being mainly bureaucratic with limited impact (17,19). For example, in the past Athena Swan did not account for the intersectionality of gender, race and social class (19). Even staff networks, which are dedicated to, and run by those from marginalised groups have not escaped the accusation of being predominantly performative and unlikely to lead to structural change, despite sometimes providing a space of solidarity (20). They allow institutes to demonstrate a commitment to EDIA whilst avoiding making changes to their institutional structures and processes.

Developing the guidelines

Within the InclusiveRSE project, we have been discussing how to move from empirical evidence of EDIA issues, both from our own research and the literature, to progress towards effective impact in RSE. Rather than jumping to detailed solutions we want to articulate guidelines for the underlying components that are likely to lead to improved EDIA rather than performative solutions. As a team, we bring together social science expertise and RSE-related professionals from early career to senior levels. This places us in a strong position to co-develop these guidelines.

We have developed the guidelines from the substantial literature critiquing EDIA initiatives (e.g. (16–20)). We also draw on our own experiences of EDIA initiatives, both positive and negative, including compulsory institutional courses that rely on online material, staff networks, and software coding communities for LGBTQ+ individuals. In particular, we’ve drawn on Anita’s experience of participating in a diversity training programme delivered by* BeWhatYouSee* consultancy (21) where she found that the programme’s in-person, cohort-based, interactive and research-informed approach facilitated deeper understanding and changes in behaviour. So, while we are critical of the priorities of EDIA consultancies generally, we know that when well-designed, they can be an effective route to impact.

Our guidelines for how to make a difference

1. RSE leaders are key to effective change. They have the influence to address structural issues (16,22) and, by investing their time, can reduce the burden that those with marginalised identities often shoulder for EDIA work (23). It is important that they do not see the responsibility for change as lying elsewhere, which can result in inaction regardless of good intentions (22,24). Equally, RSE leaders need to be aware of the tendency to legitimise inequalities through supporting the myth of meritocracy (14).
2. An understanding of EDIA issues in RSE and STEM more broadly is important to motivate change and inform meaningful impact. These understandings should include areas such as the social history of inequalities in general and computing in particular (9); EDIA-related legislation (present and historical); and how disadvantage and privilege currently present in a STEM context (25).
3. Knowledge of the pitfalls of EDIA work is essential to develop evidence-based solutions that produce impact from effort rather than replicating tokenistic action (15,26). We have given a summary of the pitfalls in the section above “Common critiques of EDIA initiatives”.
4. The voice and experiences of marginalised individuals should be centred*** ***without placing an emotional burden on those individuals by requiring them to retell their stories. This provides RSE leads the opportunity to empathise with the lived experience of RSEs from marginalised groups (18). An understanding of their experiences can be gained from already collected empirical evidence and well-planned research activities (27). This guideline links strongly with Guideline 8, being clear about the moral intention.
5. Structural incentives are needed to encourage RSE leaders to act to change a structural problem. RSE and research leaders are likely to have multiple competing priorities and need to have research aligned incentives to support moving from good intentions to action. For example, linking EDIA work to prestige and funding. It is important that these incentives are balanced against Guideline 4, centering marginalised individuals, and Guideline 8, being clear about the moral intention.
6. Training on how to act when witnessing discrimination is required, such as practical sessions on active bystander techniques. This has been identified as a key competency required in research leaders to effect change (25). It can lead to structural change through influencing areas such as recruitment and promotion processes. This training relies on Guideline 4 – an understanding of the lived experience of RSEs with marginalised identities.
7. Learning to navigate emotional discomfort is part of EDIA training. This discomfort can come about through gaining awareness of your own privileges or acknowledging that you may have caused harm previously, even if unintentionally. Avoiding this discomfort has the potential to reduce EDIA training to an ineffective, money generating exercise for consultants (18). Additionally, participants who own this discomfort and start to take small practical steps are more likely to instigate institutional change (22). Even with emotional discomfort, EDIA training, in our own experience, can still be rewarding, constructive and enjoyable.
8. Being clear about the moral intention is important for effective change. This is the intention for all individuals with the potential skill set, regardless of their identity, to be able to effectively pursue a career in RSE. Often the business case (improved productivity and creativity (28), and increased labour pool) is used to justify EDIA initiatives and gain funding. However, there is a risk of colluding with and perpetuating regimes of inequality in HE through focussing solely on efficiencies and outputs (14,18). For the RSE profession to have the highest level of impact, addressing wider societal inequalities within the RSE careers would need to be adopted as a core goal (18).

Leaning on the Social Sciences

Our recommendations, particularly Guidelines 2, 3, and 4, draw strongly on the social science literature. From our own experience, having a cross-disciplinary team that includes social science expertise has been essential to the Inclusive RSE project. It has helped us understand the patterns of inequalities that exist within STEM and gain an in-depth understanding of how broader societal, political and historical trends are likely to influence EDIA work in RSE. While all the references cited are important to our understanding – indeed you could learn a great deal just by reading the titles - we would like to recommend four articles that we hope will allow the reader to gain an insight into published literature relevant to RSE in UK HE. We have chosen these articles because they are accessible to those without a social science background.

First, Ensmenger (8) gives an amusing and recognisable description of the masculinised culture in computing and how this developed. Second, Śliwa et al. (25) share a theoretical approach to understanding accumulating privileges and disadvantages in UK HE. We particularly recommend reading the juxtaposed quotes if the full article feels a little too theoretical. Third, Feront et al. (22) explore how holding discomfort and acknowledgement of privilege can lead to transformational institutional change. This is a key reading for RSE leaders. And finally, Dagan et al. (12) offer a less theoretical approach to looking at good and not-so-good team management of diverse software engineers in Google.

Finally, we would like to make our own commitment to putting these guidelines into practice as we continue the Inclusive RSE project. Whilst they are, as far as we can see, the best approach to institutional change, we also acknowledge that any attempt at transformation may well meet unexpected barriers (29), especially in the current changing socio-political environment. These barriers can then be shared to form the next level of knowledge generation from which to build EDIA work that makes a difference.

Authorship and Contribution

The authors are listed above in alphabetical order, and our contributions are as follows. Anita Banerji leads the Inclusive RSE project. She was involved in the discussions that led to this blog, wrote the initial drafts, and produced the final version. Jasmine Folz brings social science expertise to the Inclusive RSE project. She provided the peer-reviewed publications on which the discussions that led to this blog were based and led those discussions. She reviewed and edited both the drafts and final version of this blog. Pauline Karega and Ella Kaye are coproduction research partners for the InclusiveRSE project. They were involved in the discussions that led to this blog and reviewed the final version. Caroline Jay is the academic sponsor for the Inclusive RSE project. She reviewed the final version of this blog.

References

1. Hettrick S. A not-so-brief history of Research Software Engineers.

2. Williams M. Society AGM 2024 Report.

3. Hettrick S, Bast R, Crouch S, Wyatt C, Philippe O, Botzki A, et al. International RSE Survey 2022

4. Chue Hong NP, Cohen J, Jay C. Understanding Equity, Diversity and Inclusion Challenges Within the Research Software Community. Lect Notes Comput Sci. 2021;390–403.

5. Banerji A, Folz J, Jay, Caroline. Software Sustainability Institute. Research strategy for exploring and improving diversity and inclusion within RSE and computational research domains.

6. Banerji A, Folz J, Kaye E, Karega P, Jay, C. A co-production approach to understanding diversity issues in digital research.

7. Alegria S. Escalator or Step Stool? Gendered Labor and Token Processes in Tech Work. Gend Soc. 2019 Oct 1;33(5):722–45.

8. Ensmenger N. “Beards, Sandals, and Other Signs of Rugged Individualism”: Masculine Culture within the Computing Professions. Osiris. 2015 Jan;30(1):38–65.

9. Hicks M. Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing. Cambridge, Massachusetts: The MIT Press; 2017.

10. Karamalla-Gaiballa S. Social Inclusion Dashboard

11. Dobbin F, Kalev A. Getting to diversity: what works and what doesn’t. Cambridge (Mass.): the Belknap press of Harvard University press; 2022.

12. Dagan E, Sarma A, Chang A, D’Angelo S, Dicker J, Murphy-Hill E. Building and Sustaining Ethnically, Racially, and Gender Diverse Software Engineering Teams: A Study at Google. In Proc. of the 31st ACM ESEC/FSE. San Francisco CA USA: ACM; 2023. p. 631–43.

13. Karamalla-Gaiballa S. Why Our Values Need EDI Data to Come Alive.

14. Dobbin F, Kalev A. Why Doesn’t Diversity Training Work? The Challenge for Industry and Academia. Anthropol Now. 2018 May 4;10(2):48–55.

15. Bhopal K. ‘We can talk the talk, but we’re not allowed to walk the walk’: the role of equality and diversity staff in higher education institutions in England. High Educ. 2023 Feb;85(2):325–39.

16. Ahmed S. ‘You end up doing the document rather than doing the doing’: Diversity, race equality and the politics of documentation. Ethn Racial Stud. 2007 July 1;30(4):590–609.

17. Carrillo Arciniega L. Selling diversity to white men: How disentangling economics from morality is a racial and gendered performance. Organization. 2021 Mar;28(2):228–46.

18. Bhopal K, Pitkin C. ‘Same old story, just a different policy’: race and policy making in higher education in the UK. Race Ethn Educ. 2020 July 3;23(4):530–47.

19. Madriaga M, Blaisdell B. Breathing space amidst the permanence of racism in the academy. Int Stud Sociol Educ. 2025 Apr 3;34(2):217–34.

20. Be What You See Consultancy.

21. Feront C, Bertels S, Hamann R. With Privilege Comes Responsibility: Why some privileged insiders transform institutions for societal benefit. Organ Stud. 2025 Feb 1;46(2):187–214.

22. Porter KB, Posselt JR, Reyes K, Slay KE, Kamimura A. Burdens and benefits of diversity work: emotion management in STEM doctoral students. Stud Grad Postdr Educ. 2018 Nov 19;9(2):127–43.

23. Dancy M, Hodari A. How well-intentioned white male physicists maintain ignorance of inequity and justify inaction. Int J STEM Educ. 2023 June 23;10(1):45.

24. Acker J. Inequality Regimes: Gender, Class, and Race in Organizations. Gend Soc. 2006 Aug 1;20(4):441–64.

25. Śliwa M, Gordon L, Mason K, Beech N. “That’s bang out of order, mate!”: Gendered and racialized micro-practices of disadvantage and privilege in UK business schools. Gend Work Organ. 2024;31(5):1852–72.

26. Wang ML, Gomes A, Rosa M, Copeland P, Santana VJ. A systematic review of diversity, equity, and inclusion and antiracism training studies: Findings and future directions. Transl Behav Med. 2024 Mar 1;14(3):156–71.

27. Tenquist M, Azman A, Meaden R, Onikan A, Jay C, Banerji A. Recommendations for developing effective inclusivity initiatives in Research Software Engineering. Comput Sci Eng. 2025;1–10.

28. Damian D, Blincoe K, Ford D, Serebrenik A, Masood Z, editors. Equity, Diversity, and Inclusion in Software Engineering: Best Practices and Insights. Berkeley, CA: Apress; 2024.

29. Ahmed, Sara. Conclusion: A Phenomenological Practice. In: On Being Included. Duke University Press; 2012. p. 173–90.