What if poor children tend to have distracting home lives, stressful situations, parents who aren't educated enough to help tutor, poorer diets, higher rates of disease and fewer positive peer and authority figure examples?

If these are what makes poor children perform poor at school, then the strategy of making good schools to boost house prices is not viable -- we simply have no idea how to make kids facing the problems described above achieve good scores, because if we did, we'd already be doing this. Moreover, this wouldn't attract wealthy buyers, because their children don't have these problems in the first place, so they won't get any improvement.

The point here is that we have no idea how to make a school good, other than by kicking out bad students (whatever the reasons for them being bad might be) and getting good ones. If you knew a way to take a random bad performing school in the country, and bring its students educational outcomes to above national average, you'd make lots of money contracting out consulting to local government.

> The point here is that we have no idea how to make a school good, other than by kicking out bad students

I would disagree with any metric claiming to measure school quality that is not constant when the demographics of the school change. If our metrics simply gauge the socioeconomic status of a school's population, we need better metrics. We also need to develop interventions that overcome the educational costs of poverty, perhaps, dare I say, by eliminating--or at least attenuating--poverty.

Who are you arguing with? You don't like poverty? Neither does anyone else.

And if the metrics don't appeal to your preconceived notion of how the metrics should work, you want to change the metrics? You want metrics to not notice changes in demographics?

> If you knew a way to take a random bad performing school in the country, and bring its students educational outcomes to above national average, you'd make lots of money contracting out consulting to local government.

Converting them to boarding schools and essentially cut off contact with the rest of the world may work. It would also be both really expensive and creepy, reminiscent of the times when native American children were forced to attend English-only schools away from their families.

I still don't know whether quality children usually have rich parents. Whatever quality and rich are taken to mean, exactly.

It's perhaps worth to consider what does "quality children" mean in the context of a "good school" - and it could certainly be argued that one of the major criteria for parents moving in and choosing a school is to get their kids into a school where the other kids are of a comparable or higher socioeconomic status. The other things about the school are fixable by heavily involved parents, but you can't (easily) alter the peers which your kid has in school, only pick a school with "good" peers. Other things being equal, having your kid in a school with (other) poor kids puts him/her at a disadvantage, all kinds of poverty-correlated problems (e.g. parental neglect, parental drug abuse, nutrition problems, crime and imprisonment) drag down not only affected kids but also their classmates.