Better testing could mean brighter future for schools
Photo by Shannan Muskopf
Perhaps the most frequent complaint of teachers and parents today is the degree to which high-stakes testing infects the learning environment. An "opt out" movement with growing coordination and publicity, including in Texas, speaks to parents' fears: Drilling on practice test questions is displacing the kind of rich exploration of engaging materials that can evoke a love of learning in their children. Rather than preparing young people for real-world tasks of work and citizenship, principals — whose careers rest on showing short-term score gains — may be turning schools into joyless test-preparation factories.
In the education world, researchers and reformers are pursuing two different approaches for escaping this pitfall. The first approach is to make the tests better. This "test improvement" reform track embraces the notion that tests will affect education but tries to make the tests good enough that they shape teaching and school leadership positively. The tasks they contain, it is hoped, will orient teachers' and principals' work toward desired student skills. The second approach, which is a little further on the horizon, is to make the tests invisible. This "comfortable tests" reform track aims to use passive or "stealth" data collection over the course of a year to accurately measure learning growth without distorting instruction and school culture.
A recent study by the Thomas B. Fordham Institute gives a progress report on test improvement, and it's an encouraging one. The "next generation" assessments developed by PARCC and Smarter Balanced, for use across multiple states, do an excellent job of assessing not just what's easiest to measure — a frequent knock against testing — but what it's essential to teach, such as conceptual understandings and a proficiency at synthesizing evidence.
Where skepticism creeps in for me is the second step of the "test improvement" track: the assumption that using better tests can elevate instruction at schools where teachers are not currently fostering those higher cognitive skills. Supporters of improved testing sometimes point to Advanced Placement (AP) classes. Because AP exams require students to demonstrate strong command of content and to think critically, the argument goes, we are happy for teachers to "teach to" those tests. Doing so necessarily involves broad and rich instruction, because the tests cannot be gamed through rote test-prep.
They're partly right, but AP classes work not just because they have good tests. They also work because, by and large, the incentives and responsibility are where they should be. Teachers are the ones figuring out how to help their students get ready. Students (and parents) realize that the tests are, in a real sense, "for" them, because they are important to a life goal they care about: college admission. And administrators, whose careers and compensation typically are not tied to AP results, have an incentive to find and train a quality teacher, but to rely on the teacher's judgment of how and when test practice serves their students' learning goals. Capable and enthusiastic teachers are drawn to teach AP classes because those ingredients are present.
By contrast, standards-based reform linked with high-stakes testing has tended to centralize control rather than inspire great teaching. State boards exert control over curriculum, district leaders try to direct more of what principals are doing, and principals become more prescriptive of what's happening in the classroom. Single-shot, end-of-year testing may continue to "allocate responsibility in ways that can conflict with traditional educational governance," even as the tests themselves improve.
The "comfortable tests" reform track, therefore, has some appeal. The July 2016 draft report of the Texas Commission on Next Generation Assessments and Accountability tries to nudge Texas in this direction. It recommends a system of computer-adaptive assessments throughout the school year to measure individual learning growth and suggests further study of whether random sampling might even be able to replace whole-population testing, thus reducing testing's intrusive effects. This year's faulty STAAR administration should only add willingness and urgency to consider the Commission's ideas.
This concept may never be fully realized; at best, it's well down the road. The current development efforts around stealth assessment seem mainly directed to gathering learning data unobtrusively within digital games. The gamification of education is something many educators justly have reservations about, and it's hard to see how these efforts ever translate to assessing creative or analytical writing, for example.
Nevertheless, it's pleasant to imagine a day when learning gains are constantly monitored by teachers and school leaders using a process that does not displace so much of the adults' anxiety onto the students. Experiments in the classroom such as these make it possible to foresee changes that re-empower teachers and boost student and teacher morale.
