Intersectional Factors that Influence K-2 Students’ Computer Science Learning

Abstract

Much computer science education research tailors curricula to specific demographic groups, yet often overlooks students with intersecting backgrounds. This paper explores the implementation of the Coding as Another Language curriculum for predominantly Latine, multilingual, and low-socioeconomic students. To evaluate student performance, we used a pre-and-post-test design on a validated coding assessment using a hierarchical linear model with fixed effects to control for teacher, grade-level, and parents’ educational attainment. Findings indicated that students began the curriculum with wide disparities in initial coding abilities. Looking at the intersection of language, gender, and socioeconomic characteristics, we found pre-test average score differences between intersectionally identified groups of up to 1.81 with a large effect size of 1.13. Many group differences in the average pre-test scores were significant with medium to large effect sizes. The post-minus-pre-test difference demonstrated significant improvement in all students' coding scores after exposure to the curriculum, with an effect size of 2.63. We found significant heterogeneity in these gains, with greater increases for students who entered the curriculum with lower initial pre-test scores. The largest post-test average score difference of 1.15 with a medium effect size of 0.72 was smaller than pre-test average score differences, mostly representing statistically insignificant differences with trivial effect sizes. This convergence in post-test average scores demonstrates that differential improvements mitigated pre-existing disparities in initial coding abilities. Our results prompt a compelling discussion on the curricular foundations that effectively mitigate disparities among students with diverse and intersecting backgrounds.