Matt has a set of graphs showing that if you control for demographic characteristics, most urban school districts are perfectly fine, and a fair number even out-compete the national average (DC, alas, is not one of them). The problem is folks look at the averages, rather than the statistics that are meaningful for their situation. And the averages look bad:
But if you compare like populations, the results are substantially different. The following graph tests proficiency for students who aren't eligible for school lunch (meaning poor students are exempted). The findings shift radically:
Now, Boston and New York are as good or better than the national average. (DC still sucks. Sigh.) And Boston and New York, remember, have extraordinarily high private school enrollment (as do most all large cities), so the population they're dealing with is arguably a bit tougher than the national average, as many concerned parents have preemptively ripped their kids from public school. All of which is sort of a shame. The average K-12 private school cost $6,779 per year in 2000 (it's certainly much higher today). Let's say you send your kid to that school for eight consecutive years, and the price stays constant. That's more than $54,000. $54,000 that could have been spent on a broad array of other projects and experiences for your child. You could give them Summers abroad, send them to space camp, or install two bionic knees. You could get a Congressman to come over each night, read them a bedtime story, and discuss the finer points of participatory democracy. You could pay for a lot of their college and let them graduate without debt. But lots of folks don't know they have that choice because they're sure -- sure -- that all urban schools are a cross between the Watts Riots and Srebrenica. It's a misperception that does everyone a disservice. Except, of course, folks who have an ideological commitment to privatizing the public school system.
But if you compare like populations, the results are substantially different. The following graph tests proficiency for students who aren't eligible for school lunch (meaning poor students are exempted). The findings shift radically:
Now, Boston and New York are as good or better than the national average. (DC still sucks. Sigh.) And Boston and New York, remember, have extraordinarily high private school enrollment (as do most all large cities), so the population they're dealing with is arguably a bit tougher than the national average, as many concerned parents have preemptively ripped their kids from public school. All of which is sort of a shame. The average K-12 private school cost $6,779 per year in 2000 (it's certainly much higher today). Let's say you send your kid to that school for eight consecutive years, and the price stays constant. That's more than $54,000. $54,000 that could have been spent on a broad array of other projects and experiences for your child. You could give them Summers abroad, send them to space camp, or install two bionic knees. You could get a Congressman to come over each night, read them a bedtime story, and discuss the finer points of participatory democracy. You could pay for a lot of their college and let them graduate without debt. But lots of folks don't know they have that choice because they're sure -- sure -- that all urban schools are a cross between the Watts Riots and Srebrenica. It's a misperception that does everyone a disservice. Except, of course, folks who have an ideological commitment to privatizing the public school system. 
