Thread by @arijitchakrav

Okay, let’s talk about “learning loss”. (H/t @McKinsey)

US high school test scores (ACT) are out - they’ve fallen steeply for the 4th year in a row. Now at their lowest level in 30yrs.

Let’s look at the data & the media narrative and weigh potential explanations, shall we? (1/)

So, first, the data. (H/t @wabbit011235813 )

The little red mark on the graph is 2020. As you can see, for 25 yrs (1995-2029), scores were stable, bouncing around the 21 mark.

There was a slight dip in 2019, and then Covid hit. Since then, scores have been in free fall. (2/)

@wabbit011235813 These are historic declines. Multiple consecutive years of falling scores, with big drops every year.

Important point- the pandemic *didn’t* accelerate a downward trend. Here’s the graph again, to make this clear.

Scores were stable from ‘95-‘20, then plunged after that. (3/)

@wabbit011235813 After last year’s big drop in test scores, the media coverage was very clear- this was because of the “lockdowns” in 2020.

The idea was that “learning disruption” experienced during the first year of the pandemic was still “playing out” in falling test scores. (4/)

@wabbit011235813 It’s helpful to have the facts straight on the actual magnitude of disruption.

First, for 2wks after 3/13/20, schools were mostly closed nationwide.

Next, from 04-06/20, schools were figuring out remote learning. During this period, many students were effectively truant. (5/)

@wabbit011235813 Next, starting in 09/20, a minority (~40%) of students attended remote school, with a similar number attending in-person every day. By 09/21, remote learning options had largely vanished across the country. (6/)

So, just to be clear. We’re talking two months of (at worst) no school in the ‘19-‘20 academic year for the average student, followed by one year of remote learning for 40% of students in the ‘20-‘21 academic year. (7/)

With two kids in elementary & middle school at the time, I know how much this sucked. And it’s clear that the closures hit inequitably. For example, our school in Lexington MA was doing remote learning in 2 weeks, while other schools didn’t implement it until September ‘20. (8/)

Note that the inequity argument was already weaponized as early as June ‘20 by the good folks at McKinsey (all of whom, presumably, have good health insurance) to push the children of people without good health insurance back into high-exposure in-person learning (9/)

Many schools reopened for in-person learning in poorly ventilated buildings,albeit often with mask mandates in place. (The Urgency of Brunch crowd would deal with that problem later, of course, once again weaponizing inequity in service of inequity) (10/): nejm.org/doi/full/10.10…

It’s worth noting that the “learning loss” concept invented by McKinsey had a specific meaning. It counted days that schools were closed as lost learning (no debate there) and days that students were learning remotely as a fraction of days of in-person learning (debatable) (11/)

Regardless, the important point is this- *there is nothing in the concept of learning loss that can explain continued learning loss once in-person learning resumes*.

Learning loss is measured against what students would have learned if they were in person. (12/)

So, once in-person learning resumed -by that definition- learning loss should have ended.

Only it didn’t. Many different datasets from middle and high schoolers worldwide have shown that the rate at which children are falling behind is the same. (13/)

Not only are kids not catching up on their “learning loss”, *they’re moving in the wrong direction*. As the ACT test scores show, it’s not that it’s taking time for things to get better with children’s learning.

Things are getting worse. (14/)

Just for laughs, take a peek at the handwaving & charlatanry accompanying yet another year of falling ACT scores. Expect to hear a lot in the coming days about the “pandemic accelerating trends of inequity”.

Inequity is real & it sucks. The explanation is bullsh*t though. (15/)

Note the “ six year trend” framing. Very clever ;>

Of course, some news organizations (aimed at perhaps a more gullible audience) are still trying to pin falling scores on “lockdowns”. (16/)

The idea that two months of extended summer vacation in 2020 is the explanation for this is just silly.

The idea that ACT scores continue to fall in ‘23 because a minority of kids were remote in ‘20/‘21 is equally silly. (17/)

And finally, the “pandemic accelerated a trend of declining scores as a result of structural inequities” is a complicated explanation that would make sense if the data supported it. Only it doesn’t. (18/)

So what could possibly explain this (mysterious) ongoing learning loss?

We knew in the fall of ‘20 that reopening schools without robust mitigations would lead to widespread Covid. Sure enough, most kids have had Covid by now, many more than 1x. (19/): journals.plos.org/plosone/articl…

Most kids have been infected.

Most infections originate in the school setting- in fact, as we predicted, schools are a major driver of Covid spread (🧵 coming soon).

Each infection brings with it an almost constant risk of long term health effects.

These are the facts. (20/)

So what are these long term effects, and how could they possibly lead to falling test scores, you may ask?

The thing is, Covid is not just a cold. The virus can reach the brain even after a mild infection, and is capable of causing brain damage (21/): time.com/6294762/how-co…

There’s literally dozens of papers documenting this. (Google “Covid” and “cognitive damage”).

Cognitive damage after Covid can happen to anyone, and is equivalent to about ten points off an IQ score for severe cases (22): .health.harvard.edu/mind-and-mood/…

Now of course, severe Covid is rare in kids. But Covid in kids has been documented to impact cognitive ability, memory and concentration, sleep and anxiety. Here’s an example, many more can be found on this topic (23/)

And hey- who woulda thunk it?- ACT performance correlates quite a bit with general cognitive ability. It’s not factual recall- based.

Maybe a virus that causes brain damage is a more reasonable explanation for falling ACT scores than “lockdowns” or Zoom? (24/)

Beyond falling ACT scores, the impact of repeated Covid infections on cognitive ability in kids needs more study. The folks who were handwringing about learning loss seem pretty chill about cognitive damage in kids, though. (25/25)

@dgurdasani1 @lfwhite14 @EpiEllie @DrZoeHyde