Blavatnik School of Government, University of Oxford
RISE Launches Interactive Data Visualisations Estimating Long-Term Learning Losses from COVID-19 School Closures
COVID-19 will exacerbate the learning crisis. Causing schools to close around the world, the pandemic disrupted education as we know it. But the COVID-19 shock to education systems will likely cause severe and long-term learning losses that are far bigger than the ‘mere’ time schools were closed. Learning losses can continue to accumulate after children return to school.
It is important to estimate long-term learning losses
Research following the 2005 earthquake in Pakistan shows how the long-run effects of school closures can compound over time. When schools closed for 14 weeks following the disaster in Pakistan, learning losses were far more severe. Four years after the earthquake, children were 1.5 years behind their unaffected peers, far more than the 14 weeks of school they originally missed. Learning losses can far exceed the actual school closure time if no mitigating action is taken, multiplying an initial short-term learning loss into significant long-term losses.
The same will be true for COVID-19. Modelling the impact of school closures on children in Grade 3, Michelle Kaffenberger shows how an initial three-month school closure could build up to more than a year’s worth of learning by Grade 10 if no mitigating action is taken. This is because many students had already fallen behind before the pandemic caused schools to close, as curriculum and instruction were too ambitious to have all children keep up. And even school closures themselves will likely exacerbate inequalities—not only because solutions for remote learning vary a lot in quality and effectiveness, but also because students from lower socio-economic backgrounds do not have the same household support and access to remote learning options (especially those that require electricity or internet). Hence, differences in learning based on household characteristics are likely to widen during school closure times.
But the outcomes do not have to be so grim. Combining short-term remediation with long-term reorientation of instruction and curriculum to better align with children’s learning levels not only has the potential to fully mitigate learning losses, but to improve learning outcomes beyond what was to be expected under the ‘business as usual’, counterfactual scenario where the world never experienced COVID-19.
Interactive visualisations bring learning dynamics to life
Complementing Kaffenberger’s model to illustrate these potential long-term learning losses, RISE has developed a set of interactive data visualisations that allow the user to explore the dynamics themselves, adapt the model for a specific grade-level cohort between Grades 1 and 10, and to simulate the potential effects of short- and long-term mitigation efforts on different learning measures. The visualisations help illustrate answers to the following questions:
- After schools reopen, how big might long-term learning losses become if no mitigating action is taken?
- What is the potential long-term impact of short-term and long-term mitigation scenarios to remedy learning losses?
- How might different mitigation scenarios contribute to children reaching the SDG 4.1 learning target of minimum proficiency?
- Under different mitigation scenarios, how many children are estimated to be in school but not learning anything because they fell too far behind to benefit from instruction?
The interactive visualisations are a tool to make the dynamics of the learning process more tangible, inviting readers and users to manipulate parameters, adapt the simulations and see shifts in potential outcomes on learning depending on which action is taken. Ultimately, the simulations underscore the importance of aligning systems for learning, by making sure children master foundational skills and classroom instruction is adapted to children’s learning levels, and can help education leaders plan for reopening schools in a way that prioritises learning.
RISE blog posts reflect the views of the authors and do not necessarily represent the views of the organisation or our funders.