Gender Differences in High-Stakes Performance Among Early Adolescents: Evidence from a Middle-School Exit Exam Reform

We study whether high-stakes testing affects the performance of female and male students differently in early adolescence. Existing evidence from high-school, university, and admission settings suggests that female students often perform relatively worse when stakes are high, but evidence from earlier stages of the educational career remains limited. We exploit a 2017 reform of Italy's lower-secondary exit exam that removed the score on a nationwide standardized test from the calculation of the final grade while preserving mandatory participation. Using administrative data covering the full population of students and a gender-by-reform difference-in-differences specification, we compare the female -male performance gap across the high- and low-stakes regimes. Girls outperform boys in both regimes, and their relative advantage is modestly larger, by about 0.03-0.05 standard deviations, when the test carries direct grade consequences.