Student Evaluations of Teaching (SET) are subjective measures of student satisfaction that are often used to assess teaching quality. In this paper, we show that heterogeneity in students' reporting styles challenges SET validity. Using administrative data that allow us to track all evaluations produced by each student, we are able to isolate student-specific reporting scales. We show that reporting heterogeneity explains at least one third of the within-course variation in SET. We also document that students sort across elective courses according to their reporting style. As a result, the average evaluation of two otherwise identical electives can differ only because of heterogeneity in the reporting style of students attending them. Using a simulation exercise, we show that this type of sorting coupled with large sampling variability severely alter the ranking of courses within a major, calling into question the use of SET to incentivise teachers.
Das Dokument ist öffentlich zugänglich im Rahmen des deutschen Urheberrechts.