Women & ICTs — Closing the Tech Skills Gap

Information and communication technologies remain one of the largest gender gaps in the modern STEM workforce. Twenty years of policy work has narrowed parts of it dramatically and barely touched others — here's what the current evidence shows and what the programs that actually work look like.

The skills gap, not the access gap

The conversation about women and information and communication technologies has shifted fundamentally over the past two decades. In the early 2000s, the framing was access: who has a computer, who has internet, who has a basic email account. In high-income countries and increasingly in middle-income ones, those basic-access gaps have closed. What hasn't closed — and in some cases has widened — is the gap in advanced ICT skills: programming, network engineering, data analysis, system administration, AI and machine learning literacy.

The shift matters because the policy interventions that work for an access gap don't necessarily work for a skills gap. Putting a computer in a household raised girls' technology exposure in 2003; it doesn't, by itself, produce a computer scientist in 2026. The intervention has to be deeper, more sustained, and more deliberately mentored.

What twenty years of data shows

The clearest finding from the longitudinal evidence is that girls' interest in computing peaks in middle school and declines through high school, in almost every country that tracks it. This is not because girls become less capable. It is because the social context of computing — peer groups, classroom dynamics, media representation — becomes less welcoming as girls move through adolescence. By the time they're choosing tertiary programs, the share of girls intending to study computer science is dramatically lower than the share who said in seventh grade that they liked it.

The programs that consistently produce results target this exact transition window. Year-round programs operating across the middle-school-to-high-school span — clubs, mentorship pairings, structured curriculum — outperform single-event outreach by a wide margin. So do programs with visible role models who stay engaged with the same cohort over multiple years.

What works

Across many evaluated programs, a consistent set of features predicts which interventions actually change girls' subsequent enrollment in computing degrees:

• Cohort continuity — same group of girls staying together across years. Builds peer-group support that withstands the high-school transition.
• Project-based learning, not lecture-based — girls build something they value (an app, a website, a robotics project). Tangible artifacts predict continued engagement.
• Mentor proximity — actual working women technologists, not generic role-model talks. Ideally women who completed similar programs themselves.
• Path-to-college clarity — programs that explicitly map onto specific tertiary programs and scholarships outperform unstructured "exposure" programs.
• Family engagement — programs that bring parents into the conversation about computing careers see higher persistence rates.

Most modern girls-in-tech programs — Girls Who Code, AI4ALL, Black Girls CODE, dozens of national equivalents — incorporate some subset of these features. The ones with all five, sustained over time, are where the measurable enrollment gains are concentrated.

Where the gap remains widest

Even with two decades of programs, three specific areas still show large gender gaps:

• Computer-science undergraduate enrollment in high-income countries. Many OECD countries are still below their 1985 peak in women's share of CS degrees. The reasons are contested, but the data is clear.
• Senior engineering roles in semiconductor, AI infrastructure, and systems-level computing. Even where overall tech-workforce gender ratios have improved, these subdisciplines remain dramatically male-dominated.
• Founders of deep-tech ICT companies. Women-founded companies in core ICT (chips, networking, AI platforms) are well under 10% of VC-backed startups in those categories.

Where the gap is closing fastest

Equally important: in several areas, the gender gap is closing or has effectively closed. Bioinformatics, computational biology, UX research, and data science (especially in health and social-science applications) show near-parity participation in many high-income contexts. Cybersecurity has seen the fastest recent improvement, with major increases in women's enrollment at university-level programs over the past five years. Many of the women working in these fields came up through programs that didn't exist a decade ago.