Computer Science is currently experiencing booming enrollments at the university level and increased interest throughout the pipeline. Several academic institutions have noted that CS enrollment has tripled or quadrupled in recent years. This type of computer-science boom has happened twice before--each time followed by a drop in enrollment.
These enrollment cycles aren't necessarily bad. Total enrollment in CS continues to grow, and these cycles aren't unique to CS. However, one feature of these booms and busts is alarming: their impact on gender diversity. During each drop of total enrollment, the proportion of female students choosing to pursue CS degrees also drops. In other words, women leave the field at a higher rate than men. During the late 1980s, for example, male enrollment dropped 35% whereas female enrollment dropped 53%. Similarly, in the early 2000s, male enrollment dropped 30%, compared to 56% for females.
To compound the inequity, enrollment booms do not seem to increase gender diversity. In fact, the proportion of females enrolled in CS remained steady throughout the last two booms, and data suggests that we may face the same scenario for this boom as well. So even when CS is popular, we are not attracting enough women to make up for previous reductions in diversity.
Why is this happening?
People interested in gender diversity want to understand these trends and prevent disparities in female enrollment during the current boom. Although the issue is complex, we can point to some key issues.
First, booming enrollments put significant stress on CS departments. Universities often don't have the faculty and staff to support all the new students who wish to enroll. Departments typically have a few choices. They can increase faculty and staff, which is often too expensive and time-consuming. They can increase the workloads of existing staff, which may lower faculty morale and educational quality. Or they can restrict access to CS programs. Often, universities end up doing a combination of all three, but it's the last that seems most harmful for diversity.
Some programs institute formal restrictions such as enrollment caps and more rigorous acceptance criteria. These practices can disproportionately harm female students since research has frequently shown that women with equal credentials are often viewed as less qualified for STEM positions than their male peers, especially in male-dominated fields. Enrollment caps can also increase perceptions of CS as too hard, competitive, and cutthroat, which may further discourage those female students who already doubt their abilities.
Departments may also place informal restrictions on access to CS by making introductory courses more difficult in the hope of "weeding out" students. Again, such restrictions more often harm female students, who tend to underrate their abilities even when their work equals or surpasses that of their male peers. Furthermore, women who preform well in STEM areas are also likely to perform well in the humanities, while the same does not necessarily hold true for men. Thus, women may have more options if they chose to leave CS compared to men.
Finally, how we brand CS may be an issue--both when recruiting students and restricting them. Historically, when a field needs more workers (a category that CS has fallen into recently) messaging tends to suggest that the work is easy, that "everyone can do it." The implicit or even explicit implication here is that "everyone, even women, can do it." When a field has reached its capacity and supply exceeds demand, the messaging shifts: the field is now branded as "elite," open to only for the most qualified workers. In addition, the field takes on a stereotypically masculine persona. We have seen this trend in CS: the media often depicts computer scientists as hyper-intelligent, though socially awkward, males. Current efforts to attract more interest for CS, especially from women, seem to be falling back into the trap of using language such as "anyone can do it." While this approach raises interest, it may ultimately harm efforts at diversity when CS students inevitably recognize that the field is both rewarding and challenging (i.e., more difficult than "it's so easy" implies).
What can we do?
First, we need to change our messaging. More specifically, we need to stop portraying CS as an easy field. Let's tell students, "computer science can be challenging, but it's a challenge you can meet. And if you do, there's a world of fun and opportunities on the other side."
We also need to look at the formal and informal practices universities use to handle booming enrollments. This is a complex issue: schools face capacity problems they can't ignore. But we can start collecting and analyzing data about policy changes and work to avoid those that reduce diversity. A quick course correction may keep more women engaged in CS as we ride out this next cycle.
Finally, we need to spread awareness about this phenomena. The more university leaders and other stakeholders know about the facts, the better poised they will be to make decisions that promote gender diversity.