Broadening participation and success in AP CSA: Predictive modeling from three years of data
The AP Computer Science A course and exam continually exhibit inequity among over- and under-represented populations. This paper explored three years of AP CS A data in the Chicago Public School district (CPS) from 2016-2019 (N = 561). We analyzed the impact of teacher and student-level variables to determine the extent AP CS A course taking and exam passing differences existed between over- and under-represented populations. Our analyses suggest four prominent findings: (1) CPS, in collaboration with their Research-Practice Partnership (Chicago Alliance for Equity in Computer Science; CAFÉCS), is broadening participation for students taking the AP CS A course; (2) Over- and under-represented students took the AP CS A exam at statistically comparable rates, suggesting differential encouragement to take or not take the AP CS A exam was not prevalent among these demographics; (3) After adjusting for teacher and student-level prior experience, there were no significant differences among over- and under-represented racial categorizations in their likelihoods to pass the AP CS A exam, albeit Female students were 3.3 times less likely to pass the exam than Males overall; (4) Taking the Exploring Computer Science course before AP CS A predicted students being 3.5 times more likely to pass the AP CS A exam than students that did not take ECS before AP CS A. Implications are discussed around secondary computer science course sequencing and lines of inquiry to encourage even greater broadening of participation in the AP CS A course and passing of the AP CS A exam.
Exploring Computer Science
Phillip A. Boda and Steven McGee. 2021. Broadening participation and success in AP CS A: Predictive modeling from three years of data. In Proc. ACM Tech. Sym. Comp. Sci. Ed. (SIG CSE’21). ACM, Virtual Event, USA, 6 pages. https://doi.org/10.1145/3408877.3432421