Emergency school closures aimed at minimizing the impact of the coronavirus pandemic disrupted the education trajectory of over one billion children worldwide starting in March 2020. However, the length and manner of closures varied greatly from country to country and education system to education system. A new report from the University of Arkansas’ Department of Education Reform looks at twenty years of international trend data, culminating in 2023, that shows clearly that the longer a student’s school was closed, the more academic damage is in evidence.
An international team of researchers uses data from six administrations of the Trends in International Mathematics and Science Study (TIMSS) exam. TIMSS, given every four years, assesses fourth- and eighth-grade achievement in science and math across seventy-one education systems worldwide. It also provides contextual information and characteristics of students through a background questionnaire. This study incorporates both the 2019 results and the newly-released 2023 results, giving us a detailed look at the impacts of the Covid era on student achievement. School closure data, including the types (full, partial, hybrid) and durations (measured in weeks), come from the UNESCO Institute for Statistics. The researchers first created a long-term trends model from pre-pandemic testing data, then used a mixed-effect model to estimate deviations from those trends as revealed by 2023 results, adjusting for various demographic factors.
Overall, performance in mathematics and science for both fourth and eighth graders fell significantly below expectations based on the pre-pandemic linear trends. The global average decline was 0.11 standard deviations (SD) in student achievement, but it’s the specific mechanisms at work that deserve the most attention. In all models, the longer schools were closed, the greater that country’s student achievement deviated from pre-pandemic trends. This negative relationship ranged from about 0.16 points lost for each week of closure (grade four science) to 0.52 points lost per week (grade eight mathematics). This is equivalent to declines in average achievement of between 0.07 to 0.21 SDs. Negative impacts of school closures were strongest among eighth graders, low-performing students, girls, and what are termed “linguistic minorities” (students who reported speaking a language at home other than the one in which the test was given).
The academic damage from Covid can been seen across the globe, and the data are clear that school closure is the prime mechanism responsible for those findings. While it is not directly accounted for in the data or the modeling, most students likely experienced some form of remediation after disruptions ended and before taking the 2023 TIMSS. One can only assume the grim results would have been even grimmer without it. The report’s recommendations focus on the importance of such remediation efforts tailored to not only restoring lost content, but also on providing additional support to the most vulnerable students to get them back to where they should have been absent the disruption, especially for the most-impacted populations. These include increased—and precisely targeted—funding, direct interventions like high-dosage tutoring, and motivational nudges to keep students moving forward. They also believe that the unprecedented learning loss now fully evident in TIMSS and other international tests, like PISA and PIRLS, underlines the need for international planning and cooperation to make sure that such catastrophic educational disruptions do not happen again.
SOURCE: Tomasz Gajderowicz, Maciej Jakubowski, Alec Kennedy, Christian Christrup Kjeldsen, Harry Anthony Patrinos, and Rolf Strietholt, “The Learning Crisis: Three Years after Covid-19,” University of Arkansas, Department of Education Reform working paper series (December 2024).
