Does Peer Ability Affect Education Choices?

Abstract

Zero or negative peer effects in academic achievement have been explained in the literature assuming that better peers negatively affect beliefs about own ability (self-concept), motivation or peer interactions. This paper provides clear-cut evidence on such mechanisms. I study peer effects among Swedish compulsory school students as-good-as-randomly assigned to classes with different levels of cognitive ability. My paper highlights the limits of assessing peer effects on academic performance alone: while students assigned to better classes perform better on standardized tests, they systematically underestimate their ability, receive lower grades in subjects lacking anchoring tests, and shy away from challenging courses. Negative effects persist after compulsory school: students exposed to better peers exhibit lower well-being and GPA in high school. Finally, I find that negative effects in self-concept and grade assignment are concentrated among disadvantaged students, who also suffer from lower parental support when assigned to higher ability classes.