This brief was developed using Microsoft Copilot and edited by Charlotte Sutcliffe, Duke undergraduate research assistant; for full text and references see
Kim, H. (2025). Who takes computer science in high school? Intersectional and longitudinal evidence. Retrieved from https://doi-org.proxy.lib.duke.edu/10.1177/14782103251341396
Background:
This study investigates disparities in high school computer science (CS) course enrollment using 14 years of student-level data (2005–2019) from North Carolina. As CS becomes increasingly important in the global economy, ensuring equitable access and participation is a policy priority. However, prior research has either failed to use intersectionality, focused only on gender or race, or used cross-sectional data that doesn’t track changes over time. The study applies the CAPE (capacity, access, participation, and experience) framework - specifically focusing on "Participation" - and examines disparities across gender, race, disability, economic disadvantage, and Limited English Proficiency (LEP) status. It is particularly timely given North Carolina's mandate for CS coursework for graduation by 2026–2027.
Findings:
CS enrollment has steadily increased from 3% in 2005 to 7.1% in 2015. However, this growth has been uneven, disproportionately driven by male, Asian, non-disabled, economically advantaged, and non-LEP students. Asian students showed the largest gains, while American Indian students showed the least. Additionally, gender disparities persist with males enrolling in CS at nearly twice the rate of females across all cohorts. There are also significant gaps in CS participation by gender, race, and other identities, with overlapping inequities. For instance, while Asian students have high CS enrollment overall, Asian females lag significantly behind Asian males - demonstrating how high group averages can hide inequalities within a group. Conversely, American Indian females face minimal gender gaps due to universally low enrollment.
School-level factors partly explain the gaps for American Indian and Asian students, suggesting differences in school CS offerings. In contrast, within-school factors (such as sense of belonging or teacher bias) likely explain the lower enrollment of Black students. Intersectional analysis also reveals compounding disadvantages: female students with disabilities or LEP status consistently enroll at lower rates than their male peers with the same characteristics.
Takeaways:
Equity in CS participation cannot be addressed by focusing on single identities or grouped data. This study offers robust evidence that overlapping student identities (e.g., gender and race, or gender and disability) significantly shape access to CS opportunities. Effective policy must acknowledge and target these complex, intersecting inequities - particularly for marginalized students. Simply expanding CS access is insufficient; participation and inclusion must be central to policy design.