How schools rebound from pandemic learning loss: Longitudinal findings from mandatory large-scale assessments

Background: COVID-19-related restrictions on schooling resulted in learning losses, which were larger for socially and economically disadvantaged students. While recent empirical results point towards a recovery of learning loss in some subjects, it is unclear which school-related context factors may have impacted the recovery from the pandemic. Aims: We investigate proxy indicators for three school factors, (a) socio-cultural composition, (b) proportion of students with migration background, and (c) the duration of school closures, and how they relate to differential learning gains in repeated measures from grade 5 to grade 8 for a pre-pandemic cohort compared to measures in pandemic-affected cohorts. Samples: We analyzed educational large-scale assessment data, covering all public schools in one German state (total n > 200,000 students, k > 1000 schools). Competence test data in reading and mathematics were available for a pre-pandemic cohort (2015-2019: assessed in 5th grade in 2015 and in 8th grade in 2019) and for two cohorts affected by the pandemic (2018-2022 and 2019-2023). Results: Difference-in-difference-in-difference analyses showed no significant moderator effects for the three risk factors at the school-level (i.e., socio-cultural capital, proportion of non-German speaking students, lost school days) regarding the recovery from learning losses (i.e., similar learning gains from grade 5 to 8 across cohorts). Conclusions: The null effects tentatively suggest that the social divide between schools did not widen in the aftermath of school closures. On average, schools seemed to recover from restrictions on schooling regardless of the schools' social composition