Essays on gender gaps in student achievement: Evidence from OECD countries

The gender gap in math and science skills has received much attention as a potential determinant of the under-representation of women in science and engineering fields and the gender pay gap among college graduates. If girls accumulate fewer math and science skills than boys in primary and secondary schools, they might make different higher education choices and earn lower wages. Each chapter of this dissertation examines a different aspect of the interaction between gender, education policy, and educational outcomes during students' childhood and adolescence when many basic math and science skills are developed. The first paper asks whether teacher-student gender matching improves student performance. If teacher-student gender matching does have a positive impact on student achievement, it could be one of the most efficient educational policies to enhance student performance. Using the matched-paired feature of TIMSS, this paper explores within-student variation in the effect of teacher's gender on student outcomes. The results provide little support for the conjecture that students benefit from teacher-student gender matching. This paper further examines whether the teacher gender composition in the education system could explain the observed cross-country variation in the teacher-student gender matching effect. The second paper documents the variation in gender gaps in math and science achievement across developed countries and explores the institutional factors that contribute to the observed international variation. This paper finds that countries with highly selective academic streams have larger gender gaps, even before streaming occurs. It also finds that the pro-female biased class/program assignment policies used in many countries substantially reduce the math gender gap. The third paper examines the impact of single-sex schooling on student academic achievement. Although single-sex schools have been suggested to benefit boys and girls by providing a more favorable learning environment, estimated gains to academic achievement from single-sex schooling are generally confounded by selection bias. The results suggest that single-sex schooling does not always have a positive influence on student achievement and that the selection process often explains a significant proportion of the observed achievement gap between single- and mixed-sex schools.