Learning Losses

Researchers, policymakers, and school administrators alike are worried about what COVID-19, and its accompanying disruptions to schooling, will mean for children’s learning and development. We know from high-quality studies on hurricane-caused school displacements and even from school closures induced by the polio pandemic of 1916 that these schooling interruptions can lead to short-term achievement losses and lower educational attainment. Even brief disruptions due to inclement weather affect student learning, as evidenced in papers studying the effects of snow days. The research on summer learning loss, also known as summer slide or setback, indicates that the effects of time out-of-school vary considerably across children, with some students actually gaining ground and evidence that younger students experience more reversion in the summer months. The literature on the effects of instructional time may lead to worries that negative effects of COVID-induced school closures will be disproportionately felt by children from lower-resourced families, indicating that already-large gaps between disadvantaged children and their more advantaged peers are likely to widen with more time out of school.

Four recent research projects indicate that these concerns about COVID-19 disruptions are warranted.

• Researchers at Harvard-based Opportunity Insights used data from the online platform Zearn, which is a companion to the Eureka Math curricular program, to look at engagement from before and through the months of widespread school building closures. As one can explore with the tracker at Opportunity Insights, total student participation in online math coursework was down 30% in early May as compared to January 2020. Students from high-income zip codes actually increased their engagement slightly (up almost 5%) while participation among students from low-income zip codes dropped off dramatically (down 52%) in the same timeframe. It is also notable that the patterns vary a lot by state, probably due in part to data limitations. In Indiana, participation decreased substantially across the board, dropping off 62% by early May 2020 as compared to January 2020, but those declines in engagement with the platform were present for students regardless of average socioeconomic status in their zip code. Similarly, the researchers also looked at progress in the online math platform. The overall decline of just over 3% masks substantial differences with students from the high-income zip codes making considerable progress (+45%), those from middle-income zip codes holding relatively stable (-1%), and those from low-income zip codes losing ground (-36%).
• Corroborating these findings, a study released this summer by Bacher-Hicks, Goodman, and Mulhern explored nationwide data on internet searches to gauge search intensity for online learning resources. While search intensity for relevant terms doubled once COVID-19 hit, there was again variation by geography. The researchers observed larger increases in areas of the U.S. characterized by higher income, better internet access, and fewer rural schools. The authors point out that this evidence again suggests the potential for widening gaps in educational progress and performance along these dimensions.
• In another recent study by researchers at the University of Virginia, Brown, and the Northwest Evaluation Association (NWEA), the team relied on NWEA Measures of Academic Progress (MAP) assessment data to make projections of COVID-induced learning losses. They couple both prior estimates of learning losses and gains from instructional time with analyses of typical summer learning patterns in test data covering five million students. While this paper focuses on learning loss due to school building closures in the spring months, the estimates suggest that additional closures and disruptions will exacerbate the patterns. They project that students are returning to school in fall 2020 with 63-68% of the reading progress they would normally accumulate in a school year, and only 37-50% of the math gains. Importantly, they also estimate that the highest performing students (the top third) may actually make more progress in reading than in a typical school year, further widening achievement gaps.
• A team from consulting group McKinsey & Company also used NWEA assessment data in a June 2020 paper on COVID-related learning loss. The analysts model different scenarios for when and how students return to school: returning to school in fall 2020, January 2021, or fall 2021, with average-quality remote instruction, low-quality remote instruction, or no instruction in the meantime relative to typical in-person instruction learning rates. They estimate significant learning deterioration regardless of the scenario with projected negative effects concentrated most pronouncedly among low-income, black, and Hispanic students. The team also projects increased high school dropout rates resulting from COVID-induced school closures with an additional 2-9% of high school students leaving high school before finishing.

The educational effects of COVID-19 are likely to be far-reaching, long-lasting, and particularly severe for the neediest learners. Stalled educational progress and learning losses, widening achievement gaps, and lower educational attainment all have profound consequences for individual students, their families, their futures, and our economic and societal well-being.