I think the confounding variables aren’t even known a lot of the time, because so many decisions are being made on the very local level and the data isn’t even available. I don’t know how many students are attending in person in any school but mine, really: that data is not released. When we went remote, the district talked about “fewer that half” were intending to come to in-person school. About 40% said they’d come to mine. But fewer than that actually showed up, and then that number dropped to almost nothing when they realized that at-home was working pretty well and that with most kids remote, the lessons and class structure were designed around that expectation and they didn’t get much from being here in person, anyway. But I don’t know if that second drop happened in other schools.