Jon Chait notes that opponents of using standardized tests to measure teacher quality are often exaggerating the extent to which anyone is proposing to rely on such tests:
Different states have different ways of measuring teacher performance. But none of them use student test scores as more than 50% of the measure (PDF). Classroom evaluations and other methods account for half or more of the measures everywhere.
This is, obviously, an important aspect of a well-designed evaluation system. In part, that’s because no test can capture everything that matters. In part, it’s because adding alternate evaluation methods into the mix is an important check against “teaching to the test” in too direct a way. And last it’s important because you need different evaluative methods in order to try to check and see if your methods make sense. If the results of your tests and the results of your classroom evaluations are constantly sending diametrically opposed messages, then you’re probably doing something wrong. If you’re interested in this sort of thing, Susan Headden recently published a detailed account of what specifically happens in DCPS’s super-controversial new IMAPCT teacher evaluation system. If you’ve heard that it just consists of Michelle Rhee administering standardized tests and firing people at random, you’re mistaken.