Value-add measures for teachers are complicated. Two takes fresh out today. Shorter versions:
From the teachers’ union-funded EPI (pdf): We don’t want to say don’t use value-add, but use it only a very wee little bit! We’re more bullish on peer review, but ignore the evidence there please!
From U of W’s Dan Goldhaber: Use it responsibility and beware of the limitations. Why on earth is the LAT doing what it’s doing?
Goldhaber’s take is sensible. EPI is right that the fetishising of 51 percent of evaluation from value-add isn’t wise (and it’s also not practical as a comprehensive tool). And they sensibly call for a federal push to innovate with various evaluation models. But isn’t that what’s happening under Race to the Top and related initiatives?* And since we really don’t know what works here yet there is nothing wrong with states innovating with heavy value-add models (meaning weighted at 50 percent or more), too, is there? Besides, it’s worth nothing that models that use value-add for much less than 50 get attacked, too.
In fact, I’d argue the underlying issue is less the specifications of any value-add model, or any evaluation system that uses value-add, and more the underlying issue of outcome-based evaluation. Most of the debate today is camouflage for that.
*Take for instance the DC IMPACT model, which is a pretty good tool.