The Great Debate Over Grading Teachers With ‘Value-Added’ Models
THE SCIENCE OF VERGARA: Researchers played a key role in Vergara v. California, the controversial lawsuit filed by Students Matter in 2012 and closely watched around the nation. (In a preliminary ruling, the judge deemed the state’s teacher tenure and seniority statutes to be unconstitutional, but it is currently appealed by governor Jerry Brown.) Writing in data blog FiveThirtyEight, Andrew Flowers dissects the research presented by both parties on whether “value-add” (VA) measures can reliably predict a teacher’s impact on students’ academic performance, as determined by their test scores.
At the heart of debate is whether the model used to create teachers’ value-added scores (used by the Vergara side used to support its position) is a randomized experiment. For example, does the model isolate variables like students’ socioeconomic well-being—which are outside teachers’ control? A Berkeley economist argued “the association between VA and long-run outcomes is not robust and sensitive to controls” ( PDF), prompting his counterparts (one of whom testified on Vergara’s side) to issue a rebuttal.
Flowers appears convinced by the research cited by the plaintiff, saying that the value-added models “are being replicated in many different settings.”
While Vergara v. California is pending appeal, Students Matter has filed a second lawsuit, claiming that 13 California districts did not include students’ test scores in their teacher evaluations, as mandated by state law.