Anyone who has followed the teacher assessment debate knows the complexity of the issue. Students who live in wealthier areas tend to do better than children from impoverished. Teachers who have special needs children assigned to their rooms have far more challenges to their skills and time than those who don't. Evaluation itself is a very inaccurate science.
But there's hope. An article in Slate tells about value-added approach to teacher evaluation. This looks at the growth children make over the course of a year; it also may look at successive years of a teacher's record. This tends to allow years when teachers have a more difficult class (and we've all had them) to be averaged in with more manageable years. The value-added approach doesn't expect all children to reach one single benchmark (the bane of NCLB) but rather grow in their own abilities.
It's certainly worth looking into. I'm eager to see where this might lead. I do think we need some type of assessment that helps guide teachers and administrators in an honest direction.