June 25, 2008 2:34PM

So Much, Yet So Little

Depending on where you live, you might have seen a story in your local paper on a new report finding that test scores have improved under the No Child Left Behind Act, implying — but not outright saying — that NCLB is working.

So why would getting this news depend on where you live? Because the report looked primarily at state tests, and only about 28 states had sufficiently consistent data to do meaningful score analyses. The other 22 had changed their testing in so substantial a way since NCLB’s passage that not even three years of consistent results could be strung together. Which is the biggest non‐​finding finding of the report: NCLB has instigated so much test engineering — often to make assessments easier — that nearly half of all states have no useful long‐​term data. And don’t automatically assume that the other 28 haven’t changed things: the report itself notes that they could have made questions easier or harder, changed relative weights of questions, and made other, more subtle changes to their tests.

There are other problems with the study that make it impossible to credit score increases to NCLB, most of which the report is upfront about but news stories rarely feature: There’s no “control group” unaffected by NCLB against whom to compare students “treated” by the law, no ability to tease out the effects of NCLB versus other reforms, etc. The most obvious problem, though, is that in large part because of NCLB, lots of states have testing systems incapable of providing the consistent, long‐​term results that federal policymakers promised the law itself would deliver.