School closures were very bad policy
Published September 8, 2022
By John Hood
Roughly everyone in the United States — with the possible exception of teacher-union leaders and their pet politicians — knew that learning losses from COVID-era school shutdowns were going to be big. But the latest results from the National Assessment of Educational Progress (NAEP) still retained their power to shock.
From 2020 to 2022, average scale scores for American nine-year-olds dropped by five points in reading and seven points in math. That’s the largest decline in reading scores since 1990. It’s the first-ever decline in math scores. As might be expected, learning losses were largest among students who were already low-performing and among disadvantaged students with less access to parental support and resources while trying to learn at home.
Here in North Carolina, 51% of our public-school students scored at “grade-level proficiency” on state exams in 2022, up from a disastrous 45% in 2021 but still well below the 59% levels of 2017, 2018, and 2019. North Carolina also sets a higher bar, called “college and career ready,” for which the latest averages are even more sobering: 34% in 2022, compared with 30% in 2021 and 45% in 2019.
When the pandemic struck in early 2020, I wrote many times about the difficult tradeoffs our policymakers faced. Although I didn’t always agree with the choices they made, I understood the reasoning behind the initial shutdowns and subsequent restrictions on commerce and travel. Presented with limited information, few therapeutic options, and no vaccines for this deadly disease, policymakers’ options were constrained and inherently costly.
Early in the pandemic, however, it became clear that the disease’s risk profile was greatly skewed by age and preexisting conditions such as obesity. Young children faced (and still face) a tiny risk of serious illness from COVID-19. Other countries started reopening their public schools in the summer of 2020, as did some states. North Carolina didn’t. As some of us argued at the time, and as most in retrospect now concede, this was a very bad call.
Defenders of Gov. Roy Cooper and his administration would hasten to point out that North Carolina has experienced a relatively low rate of COVID-19 deaths. Doesn’t that prove that the state’s approach to school reopening, and to pandemic restrictions more generally, was the right one?
Not so fast. While North Carolina’s COVID mortality compares favorably to that of neighboring states, the story is more complicated than that. For one thing, because the risk of serious illness is so strongly related to demographics, simply eyeballing raw totals is unwise. You have to adjust the data.
The most recent age-adjusted death rates I’ve seen were produced in late August by the Bioinformatics CRO, an international research team. Its figure for North Carolina is 295 per 100,000 residents, ranking the state 29th in the nation. That’s clearly better than the age-adjusted death rates of Tennessee (403), Georgia (359), and South Carolina (356). But North Carolina’s rate is actually a little higher than Florida’s rate of 288 per 100,000.
Remember the furious criticism hurled at Florida Gov. Ron DeSantis for restricting his state’s businesses too little and reopening his schools too early? As it turned out, the Sunshine State’s risk-adjusted COVID deaths aren’t much different from those of, say, Illinois and Connecticut — and are significantly better than those of tightly controlled Pennsylvania and New Jersey.
Indeed, according to a separate statistical analysis by the Bioinformatics CRO, the stringency of state lockdown measures — including school closures, workplace closures, and restrictions on public gatherings — shows no correlation with COVID deaths after adjusting for each state’s age and obesity rates.
We can’t yet know for certain whether states that opened their schools early, such as Florida, experienced significantly less learning loss than North Carolina did. The NAEP reading and math trends I referenced earlier are not yet available at the state level, and it’s best to use a common yardstick for such measurements.
But I’d say it’s a reasonable guess. Lengthy school closures were, in fact, unreasonable.