This paper provides evidence on ability peer effects in university education. Identification comes from the random assignment of students to sections. We find that students on average benefit from better-ability peers. Low-ability students, however, are harmed by high-ability peers. We introduce...