This survey was part of the larger study funded by Bill Gates. The claim of "prediction" is not supported by the evidence.. the correlations with test scores were low.. and in any case, this is another example of reifying test scores and stupid concepts such as "a months worth of learning" as if all learning is the same, regardless of the subject, the grade level, the prior experience of the kids and so on. The very low correlations with test scores are not surprising and a by-product of the survey questions, designed to check whether the "good" teacher gets kids to comply with rules, defer to the teacher's authority, think of learning as not making mistakes (and for the sake of Gates, stay on task all the time). Perfect conditioning for students being taught that education is a matter of doing well on fill-in-the bubble tests that Gates and this researcher seem to value as the single best measure of great education. See this and the links within it. http://voices.washingtonpost.com/answer-sheet/guest-bloggers...