Why Do They Say What They Say?
Matt DiCarlo has another important post up today about Value-Added Modeling (VAM), the error-ridden statistics tool promoted by the corporate reformers as the solution to the "problem" of teacher evaluation (a "problem" I still contend no one has shown to be an actual problem).
Matt does what he does best: debunk think-tanky reforminess with actual research. In this case, he raises some important cautions about VAM's inability to account for student effects outside of a teacher's control (it can't), and about its volatility across years compromising its ability to make high-stakes decisions (it does).
To this, I'd add two things: one Matt will probably agree with, and one I'm sure he won't.
First: VAM is predicated on the idea that the standardized tests administered to students are themselves valid
Matt does what he does best: debunk think-tanky reforminess with actual research. In this case, he raises some important cautions about VAM's inability to account for student effects outside of a teacher's control (it can't), and about its volatility across years compromising its ability to make high-stakes decisions (it does).
To this, I'd add two things: one Matt will probably agree with, and one I'm sure he won't.
First: VAM is predicated on the idea that the standardized tests administered to students are themselves valid
