Gary Rubenstein has two excellent posts where he analyzes what happened with the New York Public School System’s value-added measurements for teachers, which were just released.
He discovered several very important things:
(1) There is almost no correlation between a teacher’s score in 2009 to that for the following year.
(2) There is almost no correlation between a teacher’s score when teaching math and when teaching reading – to the same kids, the same year, and in the same elementary class.
(3) There is almost no correlation between a teacher’s score when teaching different grade levels of the same subject (i.e., Math 6 versus Math 7, and so on).
In other words, the Value Added Methodology is very close to being a true random number generator — which would be great if we were playing some sort of fantasy role-playing game or a board game like Monopoly or Yahtzee. But it’s an utterly ridiculous way to run a school system and to evaluate teachers.
After you read them, you need to pass the word (email, word of mouth, twitter, Like, facebook, whatever).
We need to kill this value-added mysticism and drive a special wooden stake through its evil, twisted heart.