The No Child Left Behind Act may represent the largest threshold-based government accountability system in the country. Schools are evaluated not by how much progress students make, but by their success in pushing students over the proficiency bar. By now, you’re probably familiar with the discontents of this system: states can game the system by setting that proficiency bar low; some schools have triaged their students, essentially reallocating resources to the kids most likely to become proficient in the very short-term; and policymakers can misleadingly make claims about declining racial achievement gaps based on proficiency rates, even as these gaps are unchanged or growing.
Proficiency-based accountability systems leave us in a terrible spot. On the one hand, we want to push kids and raise the bar for proficiency. But on the other hand, we want to make sure that the lowest performing students aren’t kicked to the curb. The higher you raise that bar, the more likely you are to have a significant proportion of students in any given school below proficiency. And those are precisely the conditions under which it makes sense for educators to allocate their time and attention strategically.
All of this, of course, should have been expected in a system focused on proficiency rather than growth. And contrary to popular belief, NCLB’s growth model pilot doesn’t allow true value-added models, but is instead based on a “projection model” which requires all students to reach a fixed proficiency target regardless of their initial achievement levels.
What am I suggesting? The new Department of Education would do well to let states experiment with a few different accountability systems: 1) dump proficiency altogether and identify schools as in need of improvement based on whether they are making less growth than expected. In other words, drop NCLB’s arbitrary targets and evaluate schools based on how they are doing compared to the schools we already have, or 2) keep proficiency around, but focus improvement efforts on schools that are both low-growth and low-proficiency - not relative to an arbitrary standard, but perhaps those in the bottom 15% of both categories. (That number should be set based on the number of schools to which states can provide targeted support.)
Either of those options would require significant new investments in better tests that are designed to measure growth, and careful attention to building a value-added model that is both valid and reliable. New Yorkers know well that a poorly designed value-added model at the center of the Progress Reports wreaks more havoc than no value-added model at all.
My recommendations will surely fail to impress the “no excuses” crowd (or more aptly, the “nuke the system” crowd--my belated entry into Elizabeth’s Green’s name-the-reformer contest) who see anything short of “100% proficiency” as not radical enough. “No excuses” is great rhetoric, but in the end it’s just that. So my wish #2 is that we move past this bravado in the next four years and develop a more reasonable and effective way of identifying and supporting low-performing schools in getting better.
PS: Check out Richard Rothstein’s related op-ed, Getting Accountability Right, which speaks back to Wish #4 (integrating a broad set of goals of public schooling into accountability systems).