In the New York Times, Motoko Rich reported Arne Duncan’s scathing criticism of Arne Duncan’s policy of test-based evaluation for teachers. The story shows that Duncan dreamed up this policy, that he promoted it in Race to the Top, and in the waivers he offered states to avoid the onerous conditions of No Child Left Behind. Rich points out that Duncan borrowed the rhetoric of his most scathing critics in offering states a delay. The story includes an excellent quote from Anthony Cody, recommending that the federal government butt out and leave decisions about teacher evaluation to states and districts.

Kevin Huffman said that Tennessee will continue with Duncan’s policy, even though Duncan has denounced it. “In Vermont, by contrast, the state board of education recently adopted a resolution saying formulas based on test scores would not be included in teacher evaluations.”

It is a good story about the politics of the issue.

The only point missing from the story is that the policy has failed to make a difference wherever it has been tried, that teachers in states like Florida are rated on the performance of students they never taught, and that the American Statistical Association warned that teachers affect only 1-14% of test score variance. In short, the policy doesn’t work. It demoralizes teachers to be judged by a false metric. It has failed. But its advocates can’t bring themselves to admit failure.