State Assembly Bill 5 removes the requirement that increases in student test scores be included in teacher evaluations (“10th graders sail through exit exam,” Aug. 23). Good idea.
A number of studies have shown that rating teachers using student test-score gains does not give consistent results. Different tests produce different ratings, and the same teacher‚Äôs ratings can vary from year to year, sometimes quite a bit.
In addition, using test-score gains for evaluation encourages gaming the system, trying to produce increases in scores by teaching test-taking strategies, not by encouraging real learning. This is like putting a match under the thermometer and claiming you have raised the temperature of the room.
There is another problem. Studies show that children of poverty typically lose reading proficiency during the summer, while more advantaged children improve. This means that we would need pre-tests in the fall to measure the effect of school from fall to spring, without the effect of summer. This would double the already excessive amount of testing now required.
We are all interested in finding the best ways of evaluating teachers, but using student test-score gains is a lousy way to do it.
University of Southern California