Following up on my post earlier this week on the tentative agreement in Los Angeles around how to incorporate student growth in teacher evaluations, there at least four more reasons why combining school-level student growth scores and raw test scores in a teacher’s evaluation is a bad idea:
- Most importantly, it’s not clear that the district’s solution addresses the problem. Los Angeles has been compelled to act because of a parent’s lawsuit claiming the district was ignoring the provisions of the Stull Act, which requires teacher and principal evaluations to consider “pupil progress” on the state’s academic content standards. A county judge agreed with the parents and ordered the district and union to comply with the law. But by definition, “progress” has to be judged from a beginning and an ending point. If the agreement explicitly forbids the district to measure the student achievement gains between two points in time, it’s hard to argue this meets the “progress” requirement of the law.
- It’s interesting to observe the dissonance that’s happening across the country on this issue. When Tennessee rolled out their evaluation system last year, they were lampooned in The New York Times for using school-wide growth measures to determine teacher evaluation ratings in non-tested grades and subjects. It argued that a music teacher, for example, shouldn’t be held accountable for the math and English scores of the entire school. Economists call this the free rider problem. It’s a fair concern, and it’s one we’re also seeing play out in Florida. Now, here comes Los Angeles saying it’s going to be using school-wide growth measures and explicitly forbidding growth measures based on a teacher’s own students, ignoring the warnings of economists and the on-the-ground experiences of actual teachers.
- By spending all their political capital resisting efforts to include student growth in a teacher’s evaluation, unions are losing the opportunity to engage in really important technical questions: How to design and select a growth model, which variables to control for or not, how to establish norms about mobile students or teachers who co-teach, how to use confidence intervals to express uncertainty in the results, etc. They could also be engaging in things like ensuring there’s a proper roll-out timeline, that the measures are piloted, that observers receive training and support, and that teachers are given ample feedback and support for improvement. This is a huge loss, and it leaves the district to make critical and inherently political decisions alone. I don’t know of many places where unions are engaging on these questions, and it’s a real loss for policymaking.
- In some of their competitive grant programs, the U.S. Department of Education is moving to require classroom-level growth be used in teacher evaluations. The recently completed $285 million Teacher Incentive Fund competition (Disclosure: I worked on this competition while I was at the Department) explicitly required classroom-level growth; Los Angeles’ recent agreement would not meet that standard. The Obama administration has requested $400 million for this program next year, and it’s likely they would use this definition again. Los Angeles would not be eligible for a grant.
As I wrote on Monday, in Los Angeles, the debate over value-added has now gone too far. The United Teachers of Los Angeles is celebrating about the exclusion of student growth in favor of a much more flawed measure. It’s not a good precedent to set.
Photo Credit: American Renaissance