Earlier in December, we received more bad news about the achievement of American students: Our 15-year-olds made no significant progress in math and reading on PISA, the largest of the international tests. This followed on the heels of a new report from our National Assessment of Educational Progress (NAEP), which showed no real progress in reading or math for fourth or eighth grade students for the past decade, and longer for reading.
The routine debate is underway about how bad this news is, but such arguments mostly miss a core lesson: America’s school reform movement has plainly failed. It’s time to face up to this failure and think about new approaches for improving education.
The routine debate is underway about how bad this news is, but such arguments mostly miss a core lesson: America’s school reform movement has plainly failed.
There have been numerous reforms over the past two decades, but at the heart of them are efforts to pressure educators to raise test scores. The idea is deceptively simple. Tests measure important things we want students to learn. Hold educators accountable for raising scores, and they will teach kids more. And by focusing accountability on low-scoring groups — most often by setting uniform targets via state or federal laws, such as No Child Left Behind and the Every Student Succeeds Act — we will close achievement gaps.
Unfortunately, this concept has turned out to be more simplistic than simple, and it hasn’t worked. Even though the primary focus has been reading and math tests, reading hasn’t improved. Test-based accountability has contributed to math gains among younger students, but these improvements ended a decade ago, were achieved in part by taking time away from other subjects, and don’t persist until students graduate from school, making them of questionable value. The effort to improve equity has also failed.
As I showed decades ago, the gap between racial and ethnic minorities and non-minority students started to narrow before the rush into test-based accountability, but that progress has ground to a halt in recent years. At the same time, as Sean Reardon at Stanford University has shown, the gap between rich and poor students has widened on a variety of independent tests. The gap between high- and low-scoring students overall has also recently grown larger.
But what about the score gains, some truly dramatic, that districts and states sometimes report? Some are no doubt real, but research that I and others have done has shown that some of these gains are bogus. Teachers have been encouraged to game the system by focusing on the content in their state’s specific tests at the cost of other important material and by teaching their students tricks that take advantage of irrelevant characteristics of the questions on that particular test. Scores go up, but students’ actual learning doesn’t improve.
Then there’s the actual cheating. Far too many educators have been driven to cheat, often at the direction of supervisors — for example, the well-publicized cheating scandal in Atlanta. And the increasingly abundant cheating scandals are just the tip of the iceberg. Most cheating will necessarily go undocumented. More important, many of the supposedly legitimate methods of test prep resemble cheating in that they produce only illusory gains. One that many of us experienced firsthand is “plugging in”: teaching students that rather than trying to solve a math problem, they should simply plug in the answer choices to see which works. Why is this deceptive? Because in the real world — in later education, in the workplace, or in daily life — students won’t be given answer choices to plug in. Arsenals of these techniques have been compiled, and they are peddled openly by test prep firms and given away by states and districts.
The very limited real gains in learning caused by high-stakes testing should be weighed against the immense harm it has produced. Cheating and pervasive bad test prep are only two of these negative effects. In many schools, instruction in important subjects that are not tested for accountability has been gutted. Schools that effectively use bad test prep are mislabeled as successes, while some teachers and schools that are in fact good are rated poor.
In the desperate push to tie everything to test scores, some states and districts have resorted to unreasonable and even absurd policies — for example, “evaluating” teachers who don’t have relevant tests based on the scores of other teachers who do. High-stakes testing has even corrupted the very notion of good teaching. Many young teachers — who are trying to do right by their students — are told explicitly that bad test prep is good teaching. Having never seen any other way, they believe it, and their students suffer. And, of course, the public, parents and students themselves are routinely misled by spurious gains in test scores.
Evidence of this pervasive failure isn’t new; studies documenting it have been accumulating for more than a quarter century. The paltry gains in real achievement have been documented by NAEP every two years. But helped by rapid but bogus score gains on high-stakes tests, many of those sitting at or near the top of our education system have simply ignored these inconvenient facts.
Evidence of this pervasive failure isn’t new; studies documenting it have been accumulating for more than a quarter century.
It’s time to stop pretending that test-based accountability is improving schools and turn to other approaches — including other methods for holding educators accountable — that have more promise and that will create less harm.
What would this entail? We have to measure more of what really matters, including educators’ practices and school climate as well as student achievement. Our current system is premised on the assumption that if we hold people accountable for just a few important things — primarily scores on a few tests — the rest of what matters in schools will follow along, but experience has confirmed that this is nonsense. We need to reward good practice rather than credit gaming the system, as we do now. We should prepare students for tests by teaching them the underlying knowledge and skills the tests are intended to measure, not by drilling them on the details of whatever tests are used. We need to improve teacher training and provide better support to teachers who struggle.
For all of the many problems it entails, we need to rely more on professional judgment in evaluating schools, as some higher-achieving countries do. We need to monitor the system carefully and routinely — among other things, checking to see how educators raise scores. And we need to be prepared to make mid-course corrections when the system doesn’t work as intended. One of the main reasons for the failure of test-based accountability was reformers’ refusal to evaluate their innovations before imposing them wholesale on students and teachers. We can’t repeat that mistake.
This will be much harder than simplistic test-based accountability, and because education is such a complex enterprise, it will take time. Nothing will produce real improvements as rapid as the spurious gains many schools have seen under the current system. However, it can produce the real improvements that students, teachers, and the public deserve.