This post, with an anonymous author, reviews the research on value-added measurement, with frequent references to those who claim that the rise or fall of test scores is the best way to judge teacher quality.


The basic question he or she addresses is whether the actions of your kindergarten teacher or your third-grade teacher can affect your lifetime earnings, as Raj Chetty and his team asserted in a study a few years back.


The author goes into a lengthy back-and-forth about whether such claims make sense.


But the one essential fact that his post is missing is that 70% of teachers do not teach tested subjects. A district or school can evaluate teachers with VAM only when there are enough years of test scores to document the effects of the teacher over several years. Teachers of subjects other than reading and mathematics in grades 3-8 will never get VAM ratings.


But many states have solved this problem by assigning VAM ratings to the 70%. Their ratings are based on the scores of students they never met in subjects they never taught. This is called an “attributed rating.”


That makes sense, said no one ever.


That may be why Hawaii and Oklahoma have dropped VAM. It is expensive and gives false positives and false negatives. Expect more states to join these two states.