from Fairness and Accuracy in Reporting (EXTRA! April 2011 "Not Much Value in 'Value-Added' Evaluations," y Daniel Denvir): >Sam Dillon, who regularly reports on education policy for the [New York] >Times, wrote (9/01/10) that the “federal Department of Education’s own >research arm warned in a study that value-added estimates ‘are subject to >considerable degree of random error,’” and a National Academies expert panel >wrote a letter to Education Secretary Arne Duncan expressing “significant >concern” that Race to the Top put “too much emphasis on measures of growth in >student achievement that have not yet been adequately studied for the purposes >of evaluating teachers and principals.”
> Dillon quoted Stanford professor Edward Haertel, a co-author of an August > 2010 Economic Policy Institute study criticizing value-added measures, saying > the system was “unstable.” University of Wisconsin–Madison professor Douglas > Harris described how taking different student characteristics into account > can produce different outcomes. Dillon detailed more problems: students > changing classes mid-year and thus being associated with the wrong teacher; > the impossible-to-discern influence of a given teacher or tutor, since they > teach overlapping skills; a “ceiling effect” limiting the measure’s > sensitivity to gains amongst the highest-performing students. > Sharon Otterman (12/27/10), who covers New York schools for the Times, > reported that “the rankings are based on an algorithm that few other than > statisticians can understand, and on tests that the state has said were too > narrow and predictable.” She also pointed out that “a promising correlation > for groups of teachers on the average may be of little help to the individual > teacher, who faces, at least for the near future, a notable chance of being > misjudged by the ranking system.” > Otterman quoted a Brooklyn elementary school principal, “Some of my best > teachers have the absolute worst scores.” She cited a July 2010 U.S. > Department of Education study that found a teacher would probably be misrated > 35 percent of the time with one year of data taken into account, and 25 > percent of the time with three years. With 10 years of data, the error rate > still stood at a stubborn 12 percent. > Making things all the more complicated, Otterman pointed out, is the fact > that standardized tests are adjusted from time to time, making it difficult > to compare one year’s test scores with the next—and New York’s were just > toughened.< -- Jim Devine / "Segui il tuo corso, e lascia dir le genti." (Go your own way and let people talk.) -- Karl, paraphrasing Dante. _______________________________________________ pen-l mailing list [email protected] https://lists.csuchico.edu/mailman/listinfo/pen-l
