But take the results with a grain of salt: schools were closed for ~1/3 of the time, and we only studied PCR test results (i.e., no asymptomatic or mild cases).
We also found that most variation stemmed from individual and school-level factors, not geographic location.