In a recent post, I wrote that Randi Weingarten, the current president of the American Federation of Teachers (AFT), has (finally) expressed her full opposition against the use of value-added models (VAMs) to evaluate and measure teacher effectiveness. She has elaborated on her reasons why in a recent article, also about the Common Core and its relationship with VAMs, in the Huffington Post.
She writes: “Just look at what’s happened with the over-reliance on tests and value-added methodology (VAM). VAM is an incomprehensible formula, at least to those who don’t have a Ph.D. in advanced statistics, which attempts to predict how a teacher’s students will score in the future by using past test scores and other various assumptions — and then compares that prediction to actual results. Like predicting the weather, VAM is subject to many factors that influence the final result. That VAM score is then used to sort, rank and evaluate teachers.
The AFT has always been leery about VAM — and we’ve said since day one that VAM should never be the singular measure of student learning used to evaluate teachers. In fact, I questioned the fairness, accuracy and reliability of value-added metrics in a 2007 New York Times column. We have enough evidence today to make it clear that not only has VAM not worked, it’s been really destructive and it’s emboldened those seeking to turn public education into a numbers game.
Pittsburgh teachers acted in good faith to partner with the district on an evaluation system that included VAM with multiple measures of student learning. But while the system was being designed, anti-public education legislation was passed in Pennsylvania that hijacked a promising professional growth system by making it a numbers game fixated on ranking, sorting and firing teachers.
In Florida, the system went completely haywire, giving teachers value-added scores for students they had never taught or who weren’t even in the same building. One example is Mrs. Cook, an elementary school teacher who was named teacher of the year by her colleagues but was labeled unsatisfactory based on a VAM score calculated the performance of students she hadn’t taught.
In 2011, the average margin of error for VAM scores in New York City was plus or minus 28 points.
We have heard similar stories in Los Angeles, New Mexico, Houston and elsewhere. But what happened in Washington, D.C., was really the last straw. Last month, right before the holiday break, the district announced that some VAM scores were incorrect due to a technical glitch — a technical glitch that affected the lives and livelihoods of the educators who received these scores. As of today, 44 teachers have been told their scores from last year were wrong (including one teacher who was fired). And the district’s response was simply to say it was a minor issue. Would the district have the same reaction if it involved 44 students? When you use a system for such high stakes–a system that lacks transparency, accuracy and reliability on so many levels–how can you ever expect the teachers to trust the system?
I may have labeled VAM a sham, but many others built the evidence base for it.
The RAND Corp. and the Board on Testing and Assessment of the National Research Council of the National Academy of Sciences both conclude that VAM results shouldn’t be used to evaluate individual teachers.
It doesn’t have to be this way.”