Being a (diverse) Research Software Engineer

The views presented in my writing are my own, and, alas, sometimes don’t reflect the values or positions of my employer.

I recently joined the Society of Research Software Engineering. Having been employed as an RSE for a couple of years, I am keen to make new connections to the wider community, and participate in our growing movement. In the spirit of getting involved, and given my ongoing enthusiasm for diversity & inclusion, I volunteered to participate in the Society’s EDIA working group. This post is not related to any conversation that we’ve had so far, but collects some of my own thoughts on EDIA in the RSE profession.

The software engineering industry is rightfully criticised for being very white, and very male. TechNation reckons that about 26% of software engineers are women, climbing to 28.4% in London. The proportion of people from minority ethnicities isn’t great either: 20% of people in the UK; only 15.2% in tech (but the top jobs, CTO, CEO, etc are way whiter). TechNation focuses almost exclusively on these two indicators of diversity in the linked report. There is mention of class inequality, but zero consideration is given to LGBT+ inclusion or disability, and the understanding of gender employed is binary.

I was surprised to find that Research Software Engineering, by all accounts, is doing worse by all of these indicators. Neil Chue Hong gave a talk to the Software Sustainability Institute in 2020 at the International RSE Leaders Workshop quoting the following stats for UK RSEs, against the BCS statistics for the wider software industry (in brackets). The BCS stats are different to the TechNation ones, but tell a similar story of exclusion:

  • 14% of RSEs are female (to 14% of SEs)
  • 5% are of non-white ethnicity (21%)
  • 6% have a disability (11%)

Let’s call these categories the ‘big three‘. I have skin in the game (please excuse the use of that particular idiom) for two out of three of them. I am female. I am neurodiverse, specifically autistic, which is disabling for me in a neurotypical society. I am white, and I want to be a better ally for my racially diverse colleagues. And, unlike the majority of RSEs in our community, I am not straight, cisgender, or allosexual.

We don’t have good diversity statistics for LGBT+ representation in the RSE community, perhaps because of a lack of engagement, or perhaps a lack of visible LGBT+ representation. The RSE Society’s Landscape Review of 2022 struggled to attract enough survey responses to draw statistically rigorous conclusions about the structure of the community, but suggested ~75% of the 34 respondents were straight, 15% gay or bisexual, with the remaining 10% preferring not to say. This is roughly in line with 2021 Census statistics for the UK more broadly, where almost 90% of people identified as heterosexual. It’s not really possible to draw any conclusions comparing the RSE community to the UK as a whole, as the sample size is just too small.

I was surprised that the RSE community’s diversity statistics are as bad as they are, because I’ve worked in two RSE teams in the North East of England which seemed to be doing better, in a region which I had generally assumed to be below-average. We have high-profile diversity initiatives (the N8CIR Women in HPC chapter), and I have been surrounded by incredible women in the RSE teams I have worked in. That said, the most senior roles in these departments, just as the top jobs in the wider tech sector, have all been held by white men. This isn’t a comment on the quality of those men, or of their commitment to improving diversity, but a reflection of the status quo in our sector.

Executive leaders in Universities are often men too, who in my personal experience have not been as attentive to Equality, Diversity, Inclusion and Accessibility (EDIA) as they should be. My employer, Durham University, had a recent high-profile EDIA disaster which resulted in the resignation of a trans staff member instead of (as far as we know, any) material consequences for the perpetrator. The UCU Union branch representative commented at the time that “this sends a troubling message about the sincerity of management’s commitment to promoting equality, diversity and inclusion.” It is a pattern repeated across the UK.

A Way Forward?

So, what can we do about it? One approach has been to focus on recruitment, widening participation from underrepresented groups. To do this without quota-filling positive discrimination (illegal in the UK) requires the notoriously difficult work of challenging systemic and societal biases. The societal barriers to participation for women and girls in computing have been recognised for well over a decade, but attempts to address this through national curriculum reform may even have had a negative effect on participation. While perhaps the most obvious target, it’s clear there are few gains to be made by focusing on recruitment alone.

Another approach is retention. To keep existing RSEs in the job by improving conditions for those who are already here. It’s well known that universities struggle to attract software engineering talent because of rigid pay structures, and a decade of real-terms pay cuts in the sector can’t help. Staff pay is now 25% behind inflation compared to 2010, so join a union! It stands to reason that RSEs work in universities for other reasons, and for me, this includes the perception of a generally more accepting and inclusive culture. If I was motivated solely by financial remuneration I wouldn’t be working in a university at all, but it’s clear they’ll struggle to keep many like me without a good deal on pay this year.

To me, retention is about organisational culture. Foregrounding EDIA in strategic planning is necessary to create a culture where people are treated fairly, and microaggressions or incidents of sexism, racism, ableism, and (homo|trans|bi)phobia are treated with appropriate gravitas. Active bystander training is also fantastic for helping staff to be better allies. I would love to see it provided with greater frequency, not just in response to incidents, and not just through online training modules (although even that would be a start). Visible displays of organisational support are valuable, too, but ring hollow and tokenistic without commitment on other fronts. Flying the Progress Pride flag on university flagpoles in June is deeply appreciated, but a slap in the face if you are a queer member of staff experiencing organisational homophobia.

Community is also incredibly important. One challenge is in building these community links. I was at Newcastle University as a student or a member of staff for a decade and was closely involved with the LGBT+ network there. When things went wrong, I knew exactly who to talk to, and they knew me. So often, getting help is about who you know, not what you know. I’ve moved institution a couple of times since then and it’s been difficult to build up those social links again (especially working from home during the pandemic). Social capital is important for all of us, but our autistic colleagues may have particular difficulty building it, which can lead to isolation.

Accessible, cross-institutional communities are a potential way to cope with this. My PhD colleagues and I set up as a network for feminist researchers in and around the area of human-computer interaction, growing out of a feminist reading group. For the past few years, we’ve been running a Digital Teabreak via Zoom: we maintain an actively updated timetable on the events page of our website, and if you’ve been enjoying this post then it’s probably for you: sign up to the mailing list and come along! I wonder if there is any interest in doing something similar at the RSE Society: a regular networking event with a low barrier to engagement and a focus on diversity, or perhaps even part of a wider LGBT+ or wider EDIA-focused RSE network.

Finally, I want to give a nod to Dave Horsfall’s work on mental health, which shows that 85% of RSEs have experienced some kind of mental health difficulty. It’s incredibly important that we can have open conversations about it, as knowing you’re not alone makes a world of difference. Beyond this, we also need to highlight the negative impacts on mental health from systemic trans/homo/biphobia, ableism, and racism. We understand through intersectionality that the pressures experienced by people with more than one protected characteristic are multiplied! This means that a neurodiverse, racially diverse, LGBT+ colleague is statistically more likely to experience mental health difficulties, and may be more likely to need to take time off work for disability-related illness.

My aim within the RSE Society EDIA working group will be to advocate for intersectionality in our work, and to engage in community-building for minority RSEs. I believe those are key to improving the RSE experience, and therefore retention. For the broader issues around pay, I would encourage you to (if possible) join a Union, or (if not) to respect and not cross picket lines when we are on strike this semester. The only way to get a better deal is through collective action.