 In spite of poor policy design and implementation, NCLB has kids learning more. Photo by Old Shoe Woman |
The anti-testing and accountability drumbeat is constant: A once-rich curriculum has been narrowed to English and math. The arts have been squeezed out. Teachers are teaching to the test. There's no time for recess. And No Child Left Behind is to blame.
These claims are coming not only from the typical anti-test crowd but, increasingly, also from state legislators, governors, and even reformers.
That’s because while some of these claims are probably overblown, many of them are true. Our failure to evolve NCLB and its accountability policies has led to a host of negative unintended consequences, including the aforementioned, the myopic focus on "bubble kids" just below the proficiency cut, and the endless gaming of state tests. But what too few leaders seem willing to admit is that these problems are eminently fixable.
Even more importantly, they are worth fixing. While many would have us believe that there is no value in standards- and accountability-driven reform, the reality is this: In spite of poor policy design and implementation, the vast majority of the high-quality research on standards and accountability policies in general and NCLB in particular finds they've had some positive intended consequences. Chiefly, kids are actually learning more.
With congressional reauthorization impossible for the foreseeable future, Secretary Duncan recognized the need to evolve NCLB and began offering waivers to states, allowing them to scrap some of the worst parts of the law and establish more creative, better-designed accountability systems. My colleagues and I have begun examining the waivers and have found that despite being given fairly broad flexibility, states that have applied for waivers have done very little to correct the flaws of NCLB.
The failures of NCLB and its accountability policies are eminently fixable.
For instance, while waiver states still have to test in mathematics and English language arts, they can test in other subjects and weight the subjects however they want. Yet rather than broadening the focus of their testing and accountability, roughly half of the approved waiver states have actually narrowed it. Specifically, while 13 states added accountability for science tests (which were already administered under NCLB), six for writing, and five for social studies, fully 16 of the 34 waiver-approved states[1] proposed to take away even the tested (but previously unaccountable) subject of science.
What’s more, while states could establish whatever criteria for accountability they wished (e.g., replacing the proficiency-rate measures used under NCLB with measures that actually accounted for schools' contributions to student learning), nearly every state chose targets either fully or largely based on student proficiency. As more than a decade of research has shown, these kinds of status-based measures of achievement disproportionately punish schools serving poor kids, regardless of the extent to which these schools are actually effective at raising achievement.
To be sure, some states did incorporate student-level growth models in their accountability systems, which is a step in the (right) direction of actually holding schools accountable for the portion of student achievement for which they are responsible. However, all but one of these states chose a variant on the student growth proficiency (SGP) models that many researchers argue are not as appropriate as value-added models for actually measuring schools' impacts.
In other words, despite the anti-NCLB rhetoric, states have opted to stay the course. And where they have made changes, they have not chosen methods that are supported by research. The fact is that if states can't—or won’t—tweak these “updated” accountability systems before they're implemented, we'll get more of the same results we've seen for the last decade.
What’s most frustrating, though, is that fixing these problems wouldn’t take much. A few tweaks would go a long way towards improving the effectiveness of state accountability systems:
1) Test in additional subjects beyond math and reading, and include the results in accountability systems.
2) Be creative about the use of non-test-based measures such as graduation, attendance, and even student and parent engagement or satisfaction.
3) Measure both growth and levels of achievement, and create accountability systems that incorporate both types of data (e.g., different interventions for low-achieving, low-growing schools than low-achieving, high-growing), rather than forcing them into an arbitrary index.
4) Measure levels of achievement using points along the distribution (e.g., 3 points for advanced, 2 points for proficient, 1 point for basic) rather than percent proficient.
5) Use appropriate growth models along with multiple years of data to improve stability.
Each of these fixes directly targets one or more of the unintended consequences mentioned above, and each would go a long way toward improving our accountability systems.
So it's time to put up or shut up. If policymakers believe that standards and accountability are important and can help drive instructional improvement (and I think they should), then they need to stop ignoring all that we know about effective policies and start getting serious about fixing testing and accountability in the U.S. Otherwise, those drums will just keep getting louder.
Morgan Polikoff is an assistant professor of education at the University of Southern California's Rossier School of Education; he studies standards, assessment, and accountability policies. He is also an Emerging Education Policy (EEP) scholar with the Thomas B. Fordham Institute.
[1] We did not analyze Washington state because the measures they will use to identify low-performing schools were not identified in their application.