Followup: consider a "control" experiment (which, due to ethics cannot be conducted) whereby schools are replaced with "classrooms" of hundreds of students and a single custodian. There is no curriculum, there is no teaching, students sit for several hours a day at desks. They are allowed only to read or write, sit doing nothing, or do anything that looks vaguely like doing homework (though it could be doodling), as long as it's done quietly. Every month their parents are asked to fill out a survey of how they think their children are doing education wise, but there are no consequences attached to any of this.

Now, the big question is how would students fair in terms of education in this experiment relative to existing schooling, especially when correlated by socio-economic and cultural background? I highly suspect that we'd get similar results to today's educational system. Kids with educated and/or affluent parents would tend to acquire some modicum of education, other students would do less well.

The work the OP has done tends to lead to this conclusion which is not anything like an endorsement of the current public educational system.