Gender bias in standardized tests: evidence from a centralized college admissions system

Abstract

This paper aims to analyze the gender gap in educational outcomes from different student assessment methods. I exploit a college application setting in which the centralized admission system allocates students based on a composite score, which is a weighted average of high school grade point average and a standardized test score. Using administrative data, I find that females significantly outperform males in high school grade point average in every subject, and not only on average but also at all quantiles. Yet the situation is reversed when it comes to standardized test scores: males outperform females in all subjects and almost all quantiles, with the largest magnitude of the difference in quantitative subjects and highest quantiles. Based on these findings, I argue that the gender gap is affected by the student assessment method used in a centralized system of college admissions.