Knowledge decay between semesters

•We expand the study of summer learning loss into higher education.•We regress the length of time between two related courses on the later course's grade.•Naïve analysis suggests knowledge decay, but it vanishes with student fixed effects.•Knowledge decay remains in languages, for low-SAT students, and for non-STEM majors.

Summer learning loss has been widely studied in K-12 schooling, where the literature finds a range of results. This study provides the first evidence of summer learning loss in higher education. We analyze college students taking sequential courses with some students beginning the sequence in the fall semester and others in the spring. Those beginning in the fall experience a shorter break between the courses. We test whether the length of that gap explains the students’ performance in the subsequent course. Initial results suggest that a longer gap is associated with lower grades. However, including student fixed effects eliminates the observed knowledge decay with a few exceptions: knowledge decay remains for students in language courses, for students with below-median SAT Math scores, and for students with majors outside STEM fields.