A meta-analysis of the impact of COVID-19 on student achievement

Abstract

Despite increasing evidence calculating the extent of COVID learning loss, few researchers have attempted to collect and examine the evidence through meta-analysis. To fill this gap, our meta-analysis seeks to explore the existing research regarding the effects of COVID on learning in reading and mathematics. Our findings illustrate that the learning loss was real and significant compared to previous school years. Applying our rigorous inclusion criteria informed by best evidence synthesis methodology, we identified 30 eligible studies. On average, meta-regression results showed that students lost 0.21 (p = .006) standard deviations of learning during the pandemic school closures. Moderator analysis further showed none of the group comparisons (i.e., subject, grade levels, country and test-type) are statistically significant. In the final section of this article, we provide recommendations for educational policymakers to address this challenge.