America’s big-city public schools have a generally poor reputation, which tends to dissuade middle-class white families from enrolling their kids in them. But we also know that socioeconomically disadvantaged kids tend to do poorly on tests compared with whiter and more affluent ones, regardless of local school effects.
So controlling for the impact of demographics is helpful in telling the difference between a city whose kids test poorly because the schools serve a very deprived population and one whose kids are struggling for some reason we wouldn’t predict based on their backgrounds.
Kristin Blagg from the Urban Institute ran the numbers on this, using NAEP scores compiled by the Trial Urban District Assessment program and letting us see how 21 major cities stack up with and without adjustments:
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6678279/Screen%20Shot%202016-06-20%20at%2012.22.21%20PM.png)
You see a few things here.
One is that school performance in the big Midwestern cities of Chicago, Detroit, Cleveland, and Milwaukee is better than it looks at first glance, though with the exception of Chicago it’s still pretty bleak.
Conversely, the large California districts of Fresno and (especially) Los Angeles are doing worse relative to other cities than the unadjusted numbers would suggest. The raw numbers for LA, for example, put it way ahead of Washington, DC, or Baltimore, but demographic adjustment reverses that.
Third, it’s noteworthy that almost every city does better when you apply demographic adjustments. When you do the demographic exercise on state results, you see that some states go up and others go down. But when you restrict your attention to large urban districts, almost everyone does better with demographic adjustments.
School improvement isn’t just gentrification
Blagg’s report also sheds light on one issue that I know gets discussed a lot in my neighborhood — the link between school performance and gentrification. Several large American coastal cities are experiencing a structural shift in their population base toward a whiter, more affluent population whose children one would expect to do better in school.
Blagg shows that in most cities, the improvement in NAEP scores between 2005 and 2013 has been larger than one would expect based simply on the shift in demographics:
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/6678365/Screen%20Shot%202016-06-20%20at%201.45.05%20PM.png)
In this case only Cleveland is doing worse than would be predicted by demographic change, and a bunch of cities are doing a lot better.
DC also kindly reported separate numbers for the city’s traditional public schools (DCPS) as well as a composite that includes charter schools. You can see here that kids’ performance has gone up in the DCPS system, and it’s gone up even more if you look at the city as a whole. There’s been no starving Peter to pay Paul in terms of student performance, even if the two parallel systems fight over resources and attention.
