Why Teacher Evaluation Is So Difficult
Taxpayers are entitled to know if teachers are doing their job. But trying to devise a fair method is much harder than it appears because so much depends on the students teachers happen to inherit ("Teacher Tenure Helped Me in a Hard Situation," The Wall Street Journal, Jul. 5).
Even sub-standard teachers can look good if they are assigned a class of Talmudic scholars. Conversely, even exemplary teachers can look bad if they get a class of miscreants. If schools randomly assigned students, that would be a different story. Looking at outcomes over, say, three years would be a way of drawing valid inferences because a distinct patterm would be evident. But that's not how most schools operate.
When I was teaching, I saw how conditions beyond a teacher's control largely determined how that teacher would be evaluated. Teachers who fell out of favor with one principal were given five classes of the most difficult students. Not surprisingly, ratings for those teachers were poor. It was a way of punishing those teachers.
The value-added metric is supposed to control for such factors by measuring the progress made by students, even if the final status remained below average. Conversely, five classes of top-notch students would not be an advantage if no progress were made. Yes, those students would be above average, but they would be expected to move beyond that level in order for teachers to receive a high rating.
The trouble is that the value-added metric assumes conditions outside of school remain static during the course of the school year. That's rarely the case. As a result, great caution needs to be taken in evaluating teachers.