A new Stanford study says so!
Ok, first off? You can't measure "learning" by the state test. And then? You can't compare some hypothetically average student (made by comparing all state scores into some weird amalgam "average") to a kid in the charter. You'd need to compare that student to the students in the public school which he'd attend ordinarily.
Also? I could have a freaking brilliant teen who hates school and is ready to drop out. Let's pretend hypothetically for a moment that he "learns less" in the charter. The charter is still in my opinion a wild success if it keeps the child from becoming a hobo without a degree. My little hobo? Will have a degree, darnit!
Lies, damn lies and statistics. Responses on the Joanne Jacobs blog were pretty skeptical as well. Interestingly, Jay P. Greene has come out with an article showing that virtual schools in Arizona far outperform traditional brick and mortar schools in that particular state. That's at least closer to comparing apples to apples.
My scientific conclusion: just do what works for your kid. Caveat emptor and all that.