Day: 30 August 2021

We know that the switch to remote learning caused large increases in school dropout in middle-income countries like Brazil. But what about rich countries like the United States? There you’d expect the switch to be more or less seamless.

Not so, it appears. A new paper documents a dramatic decline in U.S. school enrolment last year, of which around 25% is explained by the switch to remote learning. Thomas Dee and colleagues analysed data on a large sample of K–12 public schools, comprising 17 million students across 34 states. Their main finding is shown in the chart below:

Although school districts that retained in-person learning saw a decline in enrolment, the decline was far larger in those that switched to remote learning. The authors confirmed that this finding held up in more sophisticated multivariate analyses.

Their estimates suggest that “offering remote-only instruction increased disenrollment by 42% (i.e., a change from 2.6 to 3.7%) relative to in-person instruction”. The decline in enrolment was concentrated in kindergarten.

How many students are we talking about? Roughly 57% faced remote-only instruction at the start of the 2020 school year. And about 49 million were enrolled in 2019. This means the switch to remote learning caused about 300,000 additional students to disenroll (49M x 56% x 1.1 ppts).

Interestingly, the disenrollment effect of remote learning was larger in rural areas, which may be due to a lack of high-speed internet access (though the authors were not able to explicitly test this hypothesis).

So where did all the disenrolled pupils go? As the authors note, there are several possibilities. They may have switched to private schools or home-schooling, decided to skip or delay kindergarten, become truant, or dropped out of school altogether.

The study’s findings indicate that a large number of parents did not want their children to participate in remote learning. This ‘revealed preference’, the authors note, is consistent with a report published by the National Academies of Science, Engineering and Medicine last year. It concluded that schools should “prioritize reopening with an emphasis on providing full-time, in-person instruction”.  

Last year’s decline in enrolment, which was exacerbated by the switch to remote learning, may have profound consequence in the years to come. If many students do not return to the school system, they may never gain the qualifications necessary to attend college and advance in the labour market.

And if most students do return, having simply delayed kindergarten (a practice known as ‘redshirting’), they will enter an unusually large, mixed-age cohort. The members of this cohort could face large class-sizes and other challenges throughout their educational careers.

Last year, school districts across the U.S. took the unprecedented step of closing schools and instead providing remote-only instruction. Previous studies have found that the switch to remote learning caused sizeable learning losses – at least in the short run. Thomas Dee and colleagues have shown that it exacerbated disenrollment in the U.S.

Will school closures be judged as a wise policy in hindsight? The answer is almost certainly ‘no’.