Charter school achievement follies continue

A while back, Schools Now noted that state test averages for Milwaukee charter schools were disappointing. It also suggested two possible reasons: they were mostly new schools and there was no information on how students were doing before entering the charter school.

Now the New York Times has glommed onto a study done by the American Federation of Teachers of 4th grade scores on the NAEP for charter schools nationwide. The Times swallowed whole the claim that this study shows charters underperform public schools, causing widespread outrage among charter school supporters. (For links to the various points and counter thrusts, see this post by Eduwonk.)

Critics of the Times and the AFT pointed out the claims ignored differences in students. In fact, when matched by ethnic groups, apparently performance by charter schools was about the same as that for public schools. In other words, these critics are damning the AFT for not doing a value-added analysis, for not taking account of the students in the schools.

Schools Now sees a glimmer of a breakthrough here. Many of these same critics resisted the idea that in judging public schools adjustments needed to be made for the students, often with the accusation that doing so would signal to those students that not much was expected of them. Having recognized the need for value-added analysis with charter schools, perhaps they will see the need when it comes to judging all schools.

Eventually perhaps, the silliness of studies that attempt to measure all charter schools or all voucher schools will also be recognized. That a school has a charter or accepts vouchers tells nothing about the school's educational programs or its teachers. So we are likely to see the same variability of educational effectiveness among charter and voucher schools as among conventional public schools. Trying to make generalizations about outcomes from charter schools is a little like trying to prove that all schools whose name begins with B outperform their peers. Depending on the sample chosen (and who is paying the researchers' bills), sometimes they will do better and sometimes they will do worse.