Education reformers like to point to a handful of charter school networks that deliver excellent educational results for low-income children as evidence that appropriately-managed schools can make a huge difference in kids' lives even in the absence of comprehensive resolution of all problems of social justice. Conversely, teachers unions have loved to point to a 2009 study by the Stanford Center for Research on Education Outcomes (CREDO) which showed that on average kids in charter schools learned slightly less than kids in non-charters. There is actually no contradiction between these messages, but it's been a large point of contention so now that there's a 2013 update (PDF) showing that charter school students do slightly better than non-charter students, we should expect to see a titanic shift in talking points even as nobody actually changes their mind.
That said, as Sara Mead argues the real story here continues to be the enormous variability in charter school outcomes. If you think of the idea of charter schools as being roughly "let's let people try some different ways of running schools" variability is about what you'd expect. But there seems to be systematic state-by-state variation.
In DC, Tennessee, Rhode Island, Louisiana, Massachusetts, New York, and New Jersey the charter students do way better on average than the non-charter students. In Arizona, Arkansas, Nevada, Ohio, Oregon, Pennsylvania, and Texas it's just the reverse. Some of that may be a coincidence, but a fair amount is probably a result of systematic policy issues. Texas' traditional public schools are above average in terms of student learning outcomes and yet it has very loosey goosey chartering practices. DC has both a stronger charter authorizing system and a weaker traditional public school system, so DC charters look great. In general a strategy of letting a thousand flowers bloom only really works if you then cut down the flowers that turn out to be really ugly and let then let the better ones replicate. Some states do that and others don't.
This table is not my favorite piece of graphic design in the world, but it does nicely illustrate how everyone can be right about charters:
As you can see, for the kind of kids that school reformers like to talk about charter schooling seems to have been a big help. Black kids are doing better in charters. So are low-income Latinos. So are English-language learners. So are poor kids. (obviously these are overlapping groups). On the other hand: white kids, middle class Latinos, and Asian kids seem to be doing worse in charters than in regular schools. This is kind of the hinge on which all education politics tends to founder—traditional public schools seem to work quite well for middle class white kids but not well at all for low-income minority kids. When reformers focus on fixing the big problem in public education they often end up with prescriptions that are ill-suited to the needs of the median American family.