There's also some cost-of-living bias in the opposite direction. The study measures the difference between incoming and outgoing characteristics of students, so schools in low-cost areas that produce students that tend to move to high-cost areas will see a large apparent value-add. I'm pretty sure this is part of what's happening with my undergraduate school, which is pretty high on this list: it's in rural Missouri, which has a ridiculously low cost-of-living (and it takes a lot of local students), but it trains engineers that tend to move to the coasts.