On testing, the buck stops where?

This morning, I was reading John Kuhn's latest post, "The Big Error in School Accountability," in which he says this:

Educational accountability is designed as a low-cost replacement for social responsibility.

 Because that--lack of support is the hole in the national accountability push--is his main point, I'd urge you to read it.

But in catching up on the news this morning, I came back to this passage:

Do districts and campuses sometimes benchmark too much? Yes. Why is that? The same reason they sometimes narrow the curriculum and teach to the test—because the consequences levied for poor scores on the standardized tests that McKenzie and Kress love so much are so overwrought and over-punitive that school leaders and teachers fear for their livelihoods if their students do poorly. As such, they give frequent benchmarks to attempt to determine if students are on track to pass the standardized tests that McKenzie and Kress insist are the ultimate yardstick of successful schooling. They use benchmarks to determine who needs intensive remediation, because they want those kids to pass.

I have heard from Mass state officials mulitple times that most testing is district testing, that overtesting is the district's fault, and that the state has no responsibility in this regard. It's frustrating, because what Kuhn says above is precisely right: that much of the district testing is about making sure that kids scores on similar assessments (which are not history essays and science labs) are going up, so their state scores will go up.

This has never been acknowledged by the state in my hearing, though, until I read the coverage of Holyoke's Monday School Committee meeting with the state, which is threatening takeover.  Rob Curtin, from DESE, offered the critique:

Curtin also critiqued the amount of testing in the district. The district recently switched from the state-recommended testing system, Achievement Network (ANet), to implementing a district-created program.

"Is it a concern that we're building our own and not paying ANet for it?" Sheehan asked.

Curtin said the concern is the amount of testing, that the district could benefit from quarterly rather than biannual testing.

Several school committee members questioned the need for more than the required amount of testing, which they argued keeps students from learning on testing days.

Curtin countered that city teachers could learn more on their student achievement levels from such data.

"When did chapter tests or spelling tests not become enough," Brunelle argued. "I bet if you asked [teachers] they'd know without having to turn in some data."

 Right. So, to sum up: overtesting is a district problem, not a state problem, unless you aren't doing enough of what we think is the right testing, in which case it could help us decide if you get to keep your schools.