June 15: Meet Your Virtual Twin

A new report, just out from CREDO at Stanford University, looks at performance of charter school kids in 16 states as compared to peers in traditional public schools. Mostly it repeats what we already know:  that performance ranges from great to poor, and is strongly affected by the quality of authorizing, among other factors. The report also offers support for Obama’s bust-the-caps position, saying that kids in non-cap states do better. We’d like to stand and cheer about the positive findings (poor kids do well, and specific locales like Denver and Chicago do too) but our enthusiasm is tempered by what we don’t know, and can’t explain, in the overall study, which says that a significantly larger number of charter schools perform less well than those that exceed the performance of district schools. We’ve talked to several researchers who have questions about the study’s methodology; it doesn’t actually follow the kids from district to charter, but instead creates a “virtual” twin in a district school for each charter kid and then compares performance.  A lot of the charter data seems to come from kids in charter schools less than 3 years old – which means that the “mobility effect” may be afoot – and the matching process also seems to introduce issues of motivation and information that are screened out in “gold-standard” studies. 

Expect this to roil the waters for awhile, but put it in the context of a lot of other research to be found here. 

Our comments here and the study’s executive summary here.