In this chapter, we situate how the insights in equity and inclusion we have gained through the FemTech research can inform how organizations such as computer science departments or tech companies can step up and improve inclusivity. We cannot offer a complete set of guidelines, but we can propose agendas, questions, and considerations which hopefully can assist organizations in creating their own strategies for intervention.

First, it is important to state that increasing diversity and reducing homogeneity related to gender in computer science organizations will not benefit from an essentialist focus on cis- and binary gender differences. No relevant insights will come from using specific stereotypical characteristics of, for example, women and men related to computer science as a lens through which to make a change. Women and men are not two exclusive categories possessing certain characteristics related to technology, computer science, or programming. It is not more difficult for women to learn how to program, and plenty of excellent women are very technically inclined and extraordinary programmers. Similarly, some men have no prior experience in programming when entering a bachelor’s-level program in computer science and struggle to get through. Just to be clear: women do not have specific characteristics that make it more difficult for them than men to work with computer science topics, methods, or domains.

We propose that organizations, instead of considering the unbalanced gender statistics as the central problem to solve, will benefit from considering the statistics as a symptom of how systemic structures, traditions, and culture within the organizations privilege and award certain kind of behaviors and interactions while constraining others.

Empowering People Considering Multiple Diversity Dimensions

The main problem is not the innate inability of specific social groups to engage with tech design and development (as some prejudiced views still hold) but rather the existence of inequitable conditions for true access to the tech playing field that derive from a combination of factors that are more or less relevant in different social contexts: exclusionary cultures in computing, ableist infrastructures, digital divides, preparatory privilege (indicating the many extra-curricular courses and activities mostly engaging boys, for instance), and social norms linking choices of tech careers and even the existence of tech skills to specific genders, cultures, and ethnicities. It is vital that we expend effort and resources to allow equitable access to and conditions of computer science, which sometimes includes programming workshops designed for certain populations.

Improving and finding new ways to teach programming for all is critical. Jane Margolis and Allan Fisher document that before interventions that led to 50/50 gender diversity, the computer science curriculum and teaching structures at Carnegie Mellon University were hurting all students, and that these structures particularly served as a barrier for women and under-presented groups who felt vulnerable in an unfamiliar territory when they began their education. For example, large courses trying to teach too broad a range of students made “students who are less experienced feel that the professors assume students know more than they do” (Margolis and Fisher 2003, p. 83), since they experienced their peers describing the curriculum as ‘easy’, ‘boring’, and ‘repetitive’, while they themselves were ‘drowning’ (ibid.). Similarly, reporting from the experiences from Harvey Mudd College, Alvarado and Libeskind-Hadas (2012) found that large classes with students spanning a wide range of experience in programming upon entering a program was identified as a barrier. This insight led the institution to create two parallel initial computer science courses (Gold: ‘no prior experience in programming’; and Black, ‘prior experience in programming’), which both led to the same second computer science course in the next semester (Alvarado et al. 2012). The ‘outsider-ness’ experienced by under-represented groups in computer science made the students much more vulnerable to problematic teaching environments, which led them to leave the field even in an otherwise welcoming environment (Margolis and Fisher 2003). It is not our agenda here to discuss how to best teach computer science; instead, we aim here to encourage organizations to remember that time spent improving the conditions for under-represented groups is fundamentally about improving the conditions for all.

Improving conditions for under-represented groups in computer science organizations includes considering all the diversity dimensions and their intersections (see Chap. 7). Meredith Ringel Morris, Andrew Begel, and Ben Wiedermann studied the challenges of neurodiverse software engineering employees at Microsoft and found that a major challenge was working in noisy environments (Morris et al. 2015). Software engineers working in open offices is not uncommon; however, this office layout might compromise the efficiency of neurodiverse software developers. From an intersectionality perspective (Crenshaw 1989), combining gender and cognitive abilities (gender and disability diversity dimensions), women software engineers diagnosed with neurodiversity risk being disproportionally harmed by noisy work environments as well as educational structures of large classes. Research on open office spaces shows that such designs are negatively related to employee satisfaction and productivity (Brennan et al. 2002). Thus, by accommodating software engineers with neurodiversity in terms of office layout, we also improve teaching environments for neurotypical software engineers.

We propose that organizations take a multi-dimensional perspective on diversity including, but not limited to, gender, ethnicity, disability, age, socioeconomic background, and so forth, and that they consider the ways in which infrastructures, including physical layouts, technological platforms, and work processes, enable or constrain a diverse group and take action to accommodate an inclusive environment.

Diversifying Computer Science Stereotypes

Working explicitly with diversifying tech organizations (education and IT industry) benefits from breaking down existing narrowly defined stereotypes within and outside formal and informal spaces. Historically, computer science began as female (Hicks 2017); yet, over time, the prevalent computer science stereotype became masculine, celebrating the male subculture of computer hacking (Ensmenger 2010). Nathan Ensmenger (2010) documents how this change was accidental but continues to be reinforced and institutionalized today. Ensmenger points out that the stereotypical notion of “the antisocial programmer, wearing sandals and a beard” was a deliberate self-construction rather than emerging from the initial field of computing (Ensmenger 2010, p. 240), a stereotype repeated in public culture such as in popular TV series like Silicon Valley, The IT Crowd, and The Big Bang Theory.

A crucial part of becoming an inclusive organization is breaking with the singular stereotype of the computer scientist as a male geek and instead opening up to alternative definitions of computer scientist (Frieze and Quesenberry 2013, 2015). It is important to challenge stereotypes and to extend and multiply narratives about who can belong to and succeed in the field. The emergence of new co-existing and parallel alternative narratives about who belongs in computer science is profoundly important for the field’s long-term transformation. It is about developing the organization professionally. The organization must be developed in tandem with institutional support for and a professional organization of inclusive initiatives driving the change for diversity while improving the organization for all (Frieze and Quesenberry 2015, p. 77). Diversity initiatives should not only be about outreach and communication or about organizing yearly recruiting events or celebrating Ada Lovelace Day and International Women’s Day. Diversity initiatives are not about changing members of other under-represented groups to make them fit into existing structures. Instead, the diversity, inclusion, and equity agenda is about re-thinking institutional structures, including language and symbolic representations, events, norms, and artefacts embedded in certain cultural perceptions and assumptions, and opening the field of computer science in new ways and for new groups of people. Acknowledging our own privileged and subjective perspectives on computer science (we are insiders to the field), we do enjoy the nostalgia of the 1980s geek culture. Taking an honest and dedicated interest in computer science retro and nostalgia as a playful and creative way to scaffold new types of engaged interactions can assist equity initiatives in changing from within. By rewriting the history of computing to include the invisible women through intertextual design (Bjørn and Rosner 2021), the first author worked closely with Daniela Rosner in creating AtariWomen artefacts that manifest the important contributions of women in the early days of the computing gaming industry. We use these AtariWomen artefacts as a vehicle to bring in past stories about the women in gaming to the present with the aim of impacting the future of computer game development (ibid.).

There is a need for multiple parallel narratives about computer science – and computer scientists – that can coexist and benefit each other. Our agenda is not – and has never been – about making things ‘womanly’ or painting technology ‘pink’. Instead, in each activity, in each design artefact, in each intervention, we always consider how our intervention designed with inclusivity in mind will be perceived by everyone. We want everyone to find the new learnings, abilities, technological artefacts, designs, and interventions made for and together with commonly under-represented groups interesting, and thus to appreciate the inherent qualities of the technological artefacts.

We propose that organizations actively work towards identifying and challenging stereotypes, not just as part of branding and communication but, more importantly, in the organization culture, considering all the stereotypical markers present in artefacts, technologies, and organizational layout, and that explicit interventions allowing for multiple parallel narratives can coexist.

Equity Mainstreaming

Gender mainstreaming has in recent years been a political approach adopted by the EU and the UN to measure and consider gender equity (Daly 2005). Inspired by this term, we propose equity mainstreaming to extend the focus from gender to the multiple diversity dimensions relevant to technology design. Equity mainstreaming in tech organizations and computer science departments is related to people, work processes, and the actual work of designing IT systems.

Equity mainstreaming includes diversity data collection. Collecting data about the state of diversity is not a simple task, and often organizations lack access to and insights into current situations, which also makes it difficult to evaluate whether new initiatives have the expected impact (Bjørn and Borsotti 2021). Often diversity data are limited to binary stats about women and men. In many situations, there are no data available across diversity dimensions, and the lack of data makes it difficult for decision-makers to consider intersectionality. As we argued in Chap. 7, classifications and categories have politics (Suchman 1994; Bowker and Star 2002; Bjørn and Balka 2007), which means that the categories and classifications we have for data collection and analysis considering diversity also have politics. By choosing to collect certain data while omitting other types of data – by making it possible to combine certain data and omit others – organizations make choices (maybe unintentionally) not only about what is visible and what remains invisible in the organization but also about what is excluded and what is included. For example, when job-posting software systems require applicants to state their gender as one of just two categories, applicants who identify as non-binary are forced to fit into this classification. Besides forcing applicants into categories they do not identify with, the system also renders invisible important insights about gender diversity in hiring. Further, if the classification of sexual harassment cases in an organization does not specify the situation and location of an event, it is difficult to act to prevent future harassment. Differently, if we knew from the data that harassment cases most often took place during social rather than professional events, and whether there are special locations and areas of the organization that are more prone to harassment situations, this would provide important insights for the organization to act on. Collecting detailed, yet anonymous, data about harassment might show how the introduction of technology blurring the barriers between work and private life produces new risks of workplace harassment (Tenorio and Bjørn 2019).

Available diversity statistics are essential for organizations to make strategies for inclusive environments. We need access to diversity statistics in order to make diversity data visibly available within the organization, emphasizing that diversity is important to its agenda. While diversity data are a multiplicity, not easily collected in a template, we have identified three main types of relevant diversity data, both quantitative and qualitative: retention and career development data to help organizations help everyone succeed within the organization; harassment and discrimination data to help organizations understand the contextual nature of events and to work towards reducing the risks; and organizational citizenship data to help organizations balance important service and care work required to function while reducing the risk of cultural taxation for under-represented groups (Bjørn and Borsotti 2021). Combining these data sources, while considering various diversity dimensions as well as intersectionality (Crenshaw 1989), will allow organizations to make interventions that support diversity and to consider how to use the data to inform them about important qualities of individuals that are often overlooked in promotion cases. Further, collecting data that can help the organization improve productivity for specific groups, such as neurodivergent software engineers, would also be beneficial.

Equity mainstreaming is about improving the work environment for all and utilizing opportunities for innovative thinking that accompany increased diversity. Technology design is fundamentally built on classification and categorization schemes and on processes by which users are often configured in resemblance to designers (I-methodology). These biased processes can affect the interface design and testing, database structures, and procedural design of, for example, workflow systems. Organizations should both increase diversity in software engineering teams and provide them the space, power, and resources to promote actual change – to identify and extend the edge cases and complex combinations of categories required for the IT systems being implemented. For example, in designing IT systems for interaction between schools and parents, software developers need to consider the wide variety of family constructions in a society such as “rainbow” families, single parents, or divorced parents (see Chap. 7). Having a diverse software engineering team increases the chances that different participants can provide different insights into the complex nuances of family structures, improving the quality of the system’s data structure. Further, having a diverse team will also reduce the risk of overlooking important aspects of technology design such as how light reflects differently on different skin colors, constraining people with darker skin to interact with certain digital systems (Benjamin 2019a, b). Values and ethics are hugely important for the design of IT systems (Møller et al. 2020). In this way, introducing equity agendas to tech organizations and computer science education programs also means improving the curriculum in computing and the ethical aspects of technology design. Such initiatives will contribute to reducing the risk of discrimination by design (Sachs 2015; Rubin 2017).

Finally, equity mainstreaming in tech organizations can be used as a vehicle for tech innovation. Using diversity as a vehicle for tech innovation has particularly been emphasized in research on accessibility in technology design, where amazing researchers such as Katta Spiel, Kathrin Gerling, Cynthia Bennett, Emeline Brule, Rua Williams, Jennifer Rode, and Jennifer Mankoff (Spiel et al. 2020) have for years demonstrated how designing technology for people with a disability is an innovative way to develop technologies for all. For example, Cynthia Bennett demonstrates how the experiences of blindness and interaction with technology can be used to re-imagine technology design (Bennett et al. 2019). Over the years, accessibility research has argued for how tech experiments and user studies should include people with disabilities, since assessing, for example, a website using screen readers will allow software engineers to make their technologies more accessible and available – improving technology use for all.

We propose to implement equity mainstreaming both as a strategy for improving the work environment in tech organizations and educational settings, increasing the diversity of employees and students, and as a strategy for technology design. Using equity mainstreaming in technology design requires considerations of categories and classification schemes, ethics, and how to account for diverse user groups when designing and implementing IT systems in society, improving technology use for all.