The Right Response to the Atlanta Cheating Scandal
For those of us who support academic standards, testing and accountability as strategies to improve public education, the Atlanta cheating indictments are sobering. Here was a system where dozens of employees, over the course of almost a decade, racketeered to rig results (or so it is alleged).
And while one can hope that Atlanta was an outlier in terms of the scope and longevity of its cheating conspiracy, it’s hardly an isolated case, as examples from El Paso, Philadelphia, Washington, D.C., and other locales demonstrate.
As expected, test critics are having a field day, using Atlanta as evidence of why all this must go. They yearn to throw the accountability baby out with the testing bathwater. But they’re wrong. The better approach is to “mend it, not end it.”
Try this thought experiment: What would happen if U.S. schools ceased all standardized testing—and related consequences? No more annual assessments, no more grading schools based on the results, no more interventions in low-performing schools, no more teacher evaluations tied to test scores, no more “merit pay” for high performing teachers or job jeopardy for low performers.
The result: In our most affluent communities, little would change. Schools would continue to drive toward the real-world standard of college acceptance at elite universities, via Advanced Placement exams and high SAT scores.
At schools serving both rich and poor kids, we would probably see a return to the 1990s, when achievement gaps were overlooked, wealthy students were guided toward rigorous coursework and “college readiness,” while poorer pupils were shepherded into easier classes with less challenge and weaker teachers.
And in high-poverty schools—the main target of twenty years of reform and the primary drivers of America’s improved student achievement since the 1990s—a few might keep pushing students toward college and good jobs, but many would return to the “soft bigotry of low expectations” and be satisfied with getting their students to graduation day, whether or not they learned much along the way.
I can’t prove that my forecast would come true, but the burden rests on those who want to eliminate testing and accountability to provide assurance that the system won’t revert back to its bad old ways.
If ending testing and accountability carries huge risks, what might mending it look like?
First, we should embrace testing as a diagnostic tool, not just an accountability weapon. We should get feedback into teachers’ hands much faster and make sure the tests themselves are of higher quality. All of this is the aim of the Common Core assessments currently under development, to be ready for prime time in 2015.
Second—and this is obvious—we need to invest in better test security. The Common Core assessments will be online, closing off current cheating strategies (like erasing and replacing answers on bubble sheets), but surely opening some new avenues. States need to spend the money to make sure test results can be trusted and cheaters can’t succeed.
Third, targets (for schools, students, and teachers) should be challenging but attainable. One source of the cheating scandals was educators feeling that fraud was literally the only way to produce the scores demanded by the system. They might have been right. The focus should transition to achievement growth over time, rather than hitting a particular “cut score.” (Many state accountability systems have moved in this direction recently, thanks to Education Secretary Arne Duncan’s waiver process.)
And fourth, official rankings or grades (of schools or teachers) should be informed by test scores but also leavened by human judgment. School grades might be conferred by British-style inspectors who look at pupil achievement along with much else.
And teacher evaluations should be the province of school principals, who should consider test scores as one set of data among many. The use of human judgment is particularly important if consequences are attached—closing schools, for example, or giving extra pay to great teachers, or terminating poor ones. (In this case, Duncan’s mandate for states to develop formula-driven teacher evaluations is a step in the wrong direction.)
Testing and accountability, properly conceived and implemented, can still be important tools in improving achievement and opportunity in America. Let’s keep the good, throw out the bad, and extinguish the anti-testing fire that started in Atlanta.
UPDATE: Joanne Weiss, Arne Duncan’s Chief of Staff, wrote to make this clarification: “Federal policy doesn’t require ‘formula-driven teacher evaluations,’ only that student growth be a significant consideration in the evaluation, and that evaluations should consist of multiple measures. We don’t stipulate any weights or formulas, nor do we require their use. Further, human judgment is critical to any good evaluation system. While it’s true that many states have implemented formula-driven evaluation systems (often including human judgment factors, like teacher observations and school/community contribution as part of the ‘formula’), we have no such requirements in our policies or regulations.”
This article first appeared in the New York Daily News.