"Revisiting Pediatric COVID-19 Cases in Counties With and Without School Mask Requirements"

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4118566

> We failed to establish a relationship between school masking and pediatric cases using the same methods but a larger, more nationally diverse population over a longer interval. Our study demonstrates that observational studies of interventions with small to moderate effect sizes are prone to bias caused by selection and omitted variables.

If you look at the plot - there were more cases when masks were required!