Blog: The Educated Reporter

Maybe we should call it an “unconfidence interval”

Sharon Otterman wrote over the weekend in the New York Times about the complications of value-added scores for teachers, including teacher-of-record issues and a confidence interval that may give observers more pause than confidence. A commenter to my previous post on teacher-of-record noted the confidence interval issue, which is clearly demarcated in the Times story: A city teacher ranked in the 63rd percentile could actually stand anywhere from 46th to 80th.The metric may do okay identifying the worst and best teachers, but in the middle it is somewhat of a muddle, and it is important if you are writing about such systems to be clear about what they do and do not show.

“As a general rule, you should be worried when the people who are producing something are the ones who are most worried about using it.” This was from Douglas Harris, whose University of Wisconsin colleagues produced the New York measure.