The uneven distribution of school closures in the US since September 2020 threatens to exacerbate regional, racial and class-based divides in educational performance, according to research by Zachary Parolin, of Bocconi University's Department of Social and Political Science, recently published in Nature Human Behavior. For example, in October, only 35% of White students were on distance learning, compared with 52% of Black students, 60% of Hispanic students and 65% of Asian students. And schools recording the lowest math scores were 15% more likely to be closed.
Professor Parolin and Emma Lee (Columbia University) found in fact that exposure to distance learning from September through December 2020 was more common among schools with lower academic performance (measured with third-grade math scores), a higher share of students experiencing homelessness, more students eligible for free/reduced-price lunches and from racial/ethnic minorities.
"If the schools and students with the greatest pre-COVID disadvantages are also those most exposed to school closures and distance learning, inequalities in learning outcomes may worsen," Professor Parolin says. Recent studies have demonstrated, in fact, that distance learning is less effective than traditional schooling, and that reductions in test scores appear to be particularly steep for students with less educated parents.
The authors measured in-person attendance using an anonymized mobile-phone database able to track traffic around locations, which covers 94% of US school districts and 98% of counties. A school was identified as "closed" or "mostly closed" if it experienced a 50% year-over-year decline in in-person visits during a given month.
"The race/ethnicity and math score gaps are particularly striking," Parolin and Lee write. "In October, 35% of White students were exposed to distance learning, compared with 52% of Black students, 60% of Hispanic students and 65% of Asian students. Moreover, schools recording the lowest third-grade math scores prior to the pandemic were, on average, around 15 percentage points more likely to be closed during September to December 2020 relative to schools with average test scores."
Disparities seem to be mainly driven by geography: larger and denser cities are both at greater risk of COVID spread and more diverse in population, while rural and less populated areas are more frequently white. Politics can also play a role, as Democrats (in power in denser, race-mixed areas, such as California or Washington DC) tended to be more rigorous in COVID restrictions than Republicans.
"The results of the study don't imply we are taking position on whether schools should close", Prof. Parolin warns. "School closures may save lives if they prevent the spread of COVID and the decision to turn to distance learning is undoubtedly difficult. Our study only observes that closures may be widening the educational gap."
Nature Human Behaviour