Editor’s Note: Education Week Commentary editors asked seven education practitioners and leaders to respond to the White House’s “Testing Action Plan” and a coinciding Council of the Great City Schools study on mandatory testing. Read what each contributor had to say.
The Obama administration’s proposal fails to address the most serious problem associated with testing: the collateral damage to instruction that results from test-based-accountability pressure. If current practice led to large and positive impacts on student achievement, the shift from instruction to test preparation that has occurred in many schools would arguably be justified. But accountability systems like No Child Left Behind have negligible impacts on test scores, as a review by the National Research Council concluded.
Jennifer Jennings
Current accountability practice harms students in three amply documented ways:
- It limits what they learn. Accountability pressure incentivizes schools to narrow what they teach to the most easily identifiable and predictable standards on reading and math tests. Gains on state tests are troublingly absent on other assessments (as with recent NAEP results), even when comparing student progress on the same skills. This suggests that students have been taught to recognize questions rather than master the underlying concepts.
David Cantor
- It determines who gets taught. Focusing on students who are close to passing—so-called “bubble kids"—to the exclusion of others well above or below the passing threshold has been widespread. This can help the lowest scoring students when proficiency standards are low, but multiple studies show that it hurts them when proficiency standards are high, as they are now in states implementing common-core standards.
- It influences how students prepare. Higher accountability pressure is associated with increased use of stimulant medications like Adderall.
Some policymakers argue that these impacts may be acceptable because test data prove useful for informing instruction. Imagine a doctor ordering blood tests to make a diagnosis but not getting results for five months. The critical period for intervention may have passed. Given the lag in returning state test results, teacher-designed assessments are more effective ways to immediately inform instruction.
How might we decrease both the amount of time spent on formal standardized testing and its impact on teaching and learning while maintaining accountability for the learning and well-being of disadvantaged children? The administration’s 2 percent proposal represents virtually no change from the average of 2.3 percent spent on tests by 8th graders.
Setting a cap of .5 percent—about six hours a year—is within reach, if we reconsider test design. To keep schools accountable, states could replicate a matrix sampling approach used in NAEP. This method essentially subdivides a test among several students, with each having to solve a smaller number of items; their performance is aggregated to determine whether the school is serving all of its kids. Alternatively, states could also test a random sample of students in each school.
We hope that the president’s proposal leads to a discussion about the effect of incentives on educational practice. Changing incentives that lead to documented distortions would be a real commitment to improving the day-to-day experiences of children in American classrooms.