Abstract

Do high-stakes exams widen gender gaps in academic performance? We study this question in early adolescence using Italy’s lower-secondary exit exam. A 2017 reform removed the national standardized test from the final grade while preserving the requirement to take the same test, reducing stakes without changing content. Using administrative data on the full population of students and a difference-in-differences design, we compare gender gaps before and after the reform. We find no evidence that higher stakes disadvantage girls. If anything, girls perform slightly better relative to boys when the test has exam stakes, suggesting gender differences under pressure emerge later in education.