When Science Feels like Math: Quantitative Content in Science Assessments and the Gender Achievement Gap

This paper shows how math-intensive evaluation drives science gender gaps. Using question-level data from a large standardized exam in grades 3 to 10 in India, we find boys outperform girls in mathematics in each grade. In science, girls outperform boys in early grades when few questions are math-intensive, but underperform in later grades as math intensity increases. A substantial portion of the gender gap is explained by question type: girls perform comparably to boys on non-quantitative science questions, but underperform on quantitative ones across all grades. Conditional on previous year performance, girls face a larger penalty from increasing math intensity in science exams. These findings show that gender differences in science are primarily driven by quantitative content in science assessments, helping explain why gender gaps are concentrated in math-intensive science fields. Strengthening foundational mathematics skills and confidence for girls could reduce gender gaps in science, with implications for gender-based sorting in higher education and careers.