Have you considered the possibility of a selection bias in the study? The literacy assessment tool in the article is not mandatory to use which calls into question how representative the 97K student sample is of the student population.

The benefits of hindsight here. In March 2020 it surely looked like we might be headed to the world of "Contagion": www.youtube.com/watch?v=4sYS...
Sadly, once COVID struck, learning losses were inevitable, school shutdowns or not: they just would have played a bit differently.