Scott Alexander comments on The Implicit Association Test