If these were charter-school kids, they'd have have taken off their silly hats and gotten a job. Photographer: Michael Okoniewski/Bloomberg
If these were charter-school kids, they'd have have taken off their silly hats and gotten a job. Photographer: Michael Okoniewski/Bloomberg

The best indicator of whether an education initiative is successful is, obviously, classroom achievement by the students involved. Other data are often just noise: For example, it's interesting that children enrolled in Head Start may be less likely to take to crime as adults, but it's pretty much irrelevant to judging the efficacy of an expensive government program that's failed to show much in terms of student performance.

So temper the enthusiasm for a much-discussed new study from Mathematica Policy Research that found that alumni of charter high schools in Florida and Chicago made nearly 13 percent more per year in their mid-20s than graduates of traditional public schools. Yes, bigger paychecks are probably a better measure of a program's success than keeping out of prison, but is it really the goal here? (Especially given how random the incomes of 20-somethings are; after all, someone in business or engineering graduate school probably has a low income at 25.)

College attendance (and completion) and test scores are usually seen as the top standards for judging success. So charter supporters (and I am one) would do better to push the Mathematica study's findings that charter students were at least 10 percent more likely to go to college, and that students from Florida charters showed a statistically significant advantage in "college persistence," which means attending for two consecutive years. (Interestingly, Mathematica has previously found that the charter school students in these places didn't perform any better on standardized tests.)

The methodology was also interesting: All the subjects attended charter schools in the eighth grade, then the researchers tracked students who went to charter high schools against those who didn't. Still, the students were in high school between 1998 and 2003, which was not only a long time ago but also before the boom in charter schooling. And while Mathematica controlled for race, sex and family income, it didn't have a best type of control group: students who had applied to charter schools but lost out on the lottery to attend one.

That said, the study is yet another reason to increase the experimentation with charters schools at a time of increasing backlash from some politicians and the teachers unions. One direction for supporters to push is in bringing charters to rural places, a need highlighted by a new report from Bellwether Education Partners. The more diverse the schools are, the more authoritative the evidence will be that students and their parents need school choice.

(Tobin Harshaw writes editorials and national security and education for Bloomberg View. Follow him on Twitter @tobinharshaw.)

To contact the writer of this article: Tobin Harshaw at tharshaw@bloomberg.net.

To contact the editor responsible for this article: Zara Kessler at zkessler@bloomberg.net.