The LA Times ("Who's teaching LA kids?" August 15) is practicing educational research without a license, and is practicing it poorly. The value-added analysis of teacher quality used by the Times is based on test scores. It assumes that the test scores are valid, that teachers are randomly assigned to classes, and that the results for individual teachers are stable. None of these assumptions have been validated.
The analysis also ignores the potent influence of other factors on student achievement, such as poverty, which is typically a stronger predictor of student achievement than teaching quality.
The Times also neglected to follow the usual procedure of sharing work with the academic community before making it public to millions of readers. They also put reporters in the role of expert observers of teachers. What's next, a value-added analysis of brain surgery, with newspaper reporters critiquing surgical procedures?
— Stephen Krashen
http://susanohanian.org/show_letter.php?id=1250