We should avoid flattening the curve in education – Possible scenarios for learning loss during the school lockdowns

|

This page in:

A typical distribution of learning data. The grey area to the left of the red line indicates the share of children who are in ‘learning poverty’

 

Social distancing has been a necessary strategy to reduce the spread of the novel coronavirus (COVID-19), leading most countries to close their school systems. But with 1.5 billion children out of school in 175 countries (as of April 10), there are more and more concerns about the longer-term effects on learning. The world was already in a learning crisis, and the ongoing emergency will put further strain on hard-won gains in learning. In our ongoing work simulating these effects, we propose to think about the effects that school closures will have on the “learning curve.” Doing this will help us focus on the poorest and worst-off students, and to design better mitigation strategies that are in the best interest of children.

Learning curves are typically drawn by the makers of national assessments (such as National Assessment of Educational Progress - NAEP) or international ones (such as Programme for International Student Assessment - PISA, Trends in International Mathematics and Science Study - TIMSS, or Progress in International Reading Literacy Study - PIRLS). The average scores (represented by the top) of these curves are probably most famous, because averages are often used to rank countries. But there is a lot more that we can learn from these curves. The width of the curve (i.e., the standard deviation), for instance, is one of the indicators of inequality within school systems. Another very important feature of these curves is that they can be used to rank students based on proficiency levels.

Over the last few years, a lot of effort has gone into making learning curves more comparable with regards to the bottom rank of performers, or kids reaching only minimum proficiency. (Students falling below minimum proficiency are displayed as the grey area to the left of the red dashed line). Last year, the World Bank committed to focus on the “learning poor” – the students below the minimum proficiency level who cannot read and understand a basic text by age 10. We are concerned about this group, because children who do not learn to read early enough often fail to thrive later in school or when they join the workforce.

Figure 1 – Three possible scenarios of how the learning curve may evolve in the coming months: a lower average, a higher standard deviation, or a sharp increase in low learning at the bottom.  

Figure 2 – Three possible scenarios of how the learning curve may evolve in the coming months: a lower average, a higher standard deviation, or a sharp increase in low learning at the bottom.

In our ongoing work, we are looking at three possible scenarios for the learning curve, which may substantially affect the levels of learning in countries with school closures. Each scenario is caused by a different mechanism that is affecting students right now. The first is the most straightforward transformation, which is caused by a reduction in average learning levels across the distribution (the blue curve). This is an extremely likely scenario despite the best efforts of school systems to offer distance learning. Variation in instructional time is associated with learning loss. Previous crises such as the 2008-09 recession had a substantial negative effect on learning, particularly in districts with higher proportions of disadvantaged and minority children. There is also evidence that shocks like floods substantially affect learning outcomes across grade levels. Children who are not in school learn less, despite the best intentions of distance education and home schoolers.

Secondly, consider how the curve may flatten (or skew) due to highly unequal effects of the crisis (the purple curve). This is a scenario in which children who are at the top will pull ahead, while students at the bottom fall further behind. Even if the virus does not care whether you are rich or poor, the rich are much better placed to mitigate its effects. Wealthier families are in comfortable homes, have good internet connections, can hire a private tutor, and may be better placed for home schooling by well-educated parents. Poor families, especially the extreme poor, live in inferior homes, may not have even a radio let alone internet connection or digital gadgets, don’t have the resources to hire a tutor, and will struggle to keep up with their children’s homework. The bottom of the income distribution may also see a sharp increase in poverty from lack of opportunities to work, or from unemployment. In this scenario, the wealthy will pull ahead, and the poor will fall further behind.

Thirdly, consider how the curve may change due to dropouts (the green population that is now permanently out of school). We have learned from earlier crises, such as the 1997-98 Asian financial crisis and the 1916 polio pandemic, that school enrolment can fall sharply, due to both demand and supply side effects. On the demand side, the income shock leads families to ask their children to work, and they never go back to school. We are particularly concerned about girls, given that they are usually the first to be withdrawn from school. On the supply side, we might see increased numbers of school closures. Governments will be cash strapped, as the global economic system is taking a hit. This may lead ministries of education to furlough teachers and to close or merge schools. Also, many countries have expanded schooling through low-fee private schools. These schools are typically operated on tiny margins, and we don’t know if they will survive this crisis.

It will take time to know how large the effects of the COVID-19 crisis will be. But we cannot wait that long to act, and thus we are simulating the impact on learning now. Building on existing evidence of the effect of crises of learning, and our databases such as the Harmonized Learning Outcomes and the Learning Poverty dataset, we will model how the curve will evolve if we do not take appropriate action. We will look at different scenarios like the ones depicted above, and how different mitigation strategies might help.  

We are not powerless to influence the learning curve.  

This will be a living document. As results and new forecasts become available, we will update this blog, and try to assess how this emergency is unfolding. In the meantime, please supply us with your thoughts and projects, or let us know if there is something specific you would like us to estimate.

Join the Conversation