Diversity 101: The Barriers to Diversity Across the Pipeline

The following is a post from our Diversity 101 series at Atlassian, where we present scientific evidence of why, as a company and as teams we care about diversity. We invite employees to share their stories, reactions, and questions to the post, helping us create a greater shared understanding of what diversity means and why it’s indispensable to the way we work.

Looking at the numbers of people from underrepresented backgrounds in tech, it’s pretty obvious that our industry has done a poor job of recruiting and retaining people who don’t come from a very small set of backgrounds. The level of representation of women (especially in technical roles), Black and Hispanic people, people with disabilities, veterans, and more is much lower than the general population. Here, we’ll break down the actual causes of the lack of representation: a dearth of opportunity. What do we mean by lack of opportunity? It’s a host of things. From social and economic factors that influence who is given access to computers to being constantly–overtly or implicitly–qluestioned about your suitability for a career in tech, members of groups underrepresented in tech face significant additional barriers to entering and staying in tech than their well-represented counterparts.

Meritocracy and the Talent Pipeline

How often have we heard that the lack of women, racial minorities, and many other groups in tech is a “pipeline problem”? While it’s true that there are less people from these groups in the tech industry, this idea both fails to account for the industry’s effect on that number and the fact that there are significantly more minorities in technical training programs than are currently employed in tech. For example, in the United States 11% of technical degrees are obtained by Black and Hispanic students, yet most tech workforces only comprise 2–3% people of color. Simply put, before we can meaningfully talk about a “pipeline problem”, we have to effectively employ the pipeline that already exists.

Another common explanation for the lack of representation in the tech industry more broadly is meritocracy. That is, that we as an industry are very good are recognizing talent, and that tech organizations promote and retain only those that are worthy. But there’s a problem with this ethos: it assumes that the lack of representation is driven by the lack of ability, potential, or interest in working in tech for members of underrepresented groups. Numerous studies have shown that this simply isn’t the case. Rather, while members of well-represented groups are evaluated fairly, studies show that those from underrepresented groups are both given less encouragement and held to higher standards throughout their careers. The cumulative impact of a lack of educational access and encouragement, stereotypes about who works in tech and the unconscious biases associated with those stereotypes, and tech culture itself provide a much more reasonable–and empirically valid–explanation for the lack of representation.

Access

Access to education in technical fields are crucial factors that motivate individuals to choose a career in technology. Students must be presented with technology as a viable, interesting career option. Lack of access can begin even earlier: boys are more likely than girls to be given computers earlier in their childhoods and low-income individuals (who are more likely to be members of socially marginalized groups) simply have more limited access to technology. These effects are often compounded for individuals with multiple minority identities (e.g., women of color). In many schools, there is simply no access to technical education, and this is more likely for schools that serve primarily minority students.

Even where good STEM education exists, teachers show a subtle bias for male students in math and science, which translates into internalized attitudes about who has potential to succeed in technology. This is especially surprising as scientific evidence shows that boys and girls have similar mathematical ability. Studies also show that children are often given subtle cues that reinforce stereotypes of computers as something for white men who prefer to avoid social interaction, which has been shown to be a significant choice-limiting factor. And even if these barriers are successfully overcome, the way much of core computer science is taught can itself be limiting the potential technical talent pool. Traditional computer science courses often focus on pure programming, without emphasizing a strong connection to the real-world application of core computer science concepts. Evidence suggests that when computer science is linked with solving real-world problems, a much wider variety of students choose to study the subject.

Bias

The application of stereotypes doesn’t only affect who chooses technology: they also strongly affect who is chosen to join the tech industry. An overwhelming body of literature has established that much of human decision-making is driven by a set of simple models in our unconscious minds. These models–called heuristics–are prone to error (bias), especially when dealing with the evaluation of individuals. These models rely on culturally relevant stereotypes, and often cause us to make mistakes in our evaluation of people at all stages of the talent pipeline. Unconscious bias has been empirically shown to lead evaluators to:

• Underestimate the level of qualification of minority job candidates
• Find more mistakes in work products created by members of underrepresented groups
• Mistakenly attribute a greater share of “credit” for success to majority group members

The effects of these cognitive errors–of which we’ve covered just a few–lead companies to under-hire from existing talent pools, and is a strong driver of high rates of attrition for members of underrepresented groups.

Culture

Tech companies constantly talk about their unique company cultures. But certain aspects of tech culture actually push out the diverse individuals who overcome all of the other hurdles to joining our industry. More than 50% of technical women leave the industry by mid-career, and they overwhelmingly cite the culture as the key culprit. From a focus on hiring “rockstars” and “ninjas” and increasingly long hours expected out of all employees to company events that focus on alcohol or violent competition, many minorities simply feel that they “don’t belong” in the industry. Intellectually, our industry often says that the “best idea wins.” But research shows that the “best idea” is often judged to be the one presented by the individual who is more confident, rather than more competent. And it turns out that there’s a well-documented confidence gap for women and other underrepresented minorities. Many environments that share this ethos are also more competitive, which women tend to opt out of (studies suggest this may be due to increased punishment women receive for self-advocacy in professional environments).

Like this article, and want to read more? Think about hitting that little ❤ button below, sharing, or following my writings here. Check me out on Twitter at @adblanche.