California has announced that it intends to take a breather for a year from the annual testing required by federal law until it can put much better tests—aligned with the Common Core—in place to drive its accountability program.
The Administration says the state can’t do that. It violates the law. Never mind that, in our Alice in Wonderland world, the law has expired, the Congress can not agree on a replacement and the Administration has taken the position that it can legally waive any provision of the law it does not like, but will do so only for state behavior of which it approves, so the law is what the Administration says it is.
But I digress. The much-reviled law is about accountability. It used to hold schools accountable for student success and failure, as measured by student performance on standardized tests. As modified by the Race to the Top program and the waiver rules, it now holds teachers accountable for student performance on standardized tests.
Pick your poison. Either way—test-based school accountability or test-based teacher accountability—it doesn’t work. Russ Whitehurst and Katharine Lindquist, in a recent edition of the Brookings Brown Center Chalkboard, talk about the "...undeniable advances in student achievement...during the era of high stakes accountability....” You could have fooled me. NAEP has been monitoring the national reading performance of 17-year-olds since 1971 and mathematics performance since 1973. The overall scores for that whole period have been stagnant. There was improvement for blacks and Hispanics early on, but it leveled off. Scores for 9- and 13-year-olds have risen over that period, in both reading and mathematics, but the rate of improvement before NCLB was passed was greater than the rate of improvement afterwards, except for Hispanic 13-year-olds.
Test-based accountability has been tried and it has failed. Indeed, the data I just shared strongly suggest that test-based accountability has resulted in a lower rate of improvement for minority and low-income students than they would have experienced if test-based accountability had never become the law of the land.
So why is the Secretary of Education going to the mat to prevent California from declaring a testing holiday and why are Whitehurst and Lindquist so alarmed at the prospect of less test-based accountability? The answer can be found in their essay. In it, they talk about an Education Week commentary I co-authored with Linda Darling-Hammond and John Jackson in which we argued the merits of having fewer, but much higher quality, accountability tests. They rightly point out that, if the nation abandoned its requirement for annual accountability testing in order to afford much better tests, it would be impossible to implement the form of test-based accountability for teachers most ardently embraced by the U.S. Department of Education and Brookings: annual assessments of the value added by teachers to the achievement of their students.
But Whitehurst and Lindquist offer no evidence that calculating the value added by each teacher to the educational achievement of our students has been used anywhere to drive an accountability system that actually improves student achievement.
Whitehurst and Lindquist acknowledge that we only have data for reading, mathematics, and science, so there is no data that can be used for the value-added assessments of a teacher’s contribution if that teacher happens to teach another subject or teaches at a grade for which such assessments are not required. They do not acknowledge the many very serious problems with value-added assessments that have been described by the National Academy of Sciences. Nor do they even address the main point we made in our Education Week commentary, that the requirement for so many tests makes it virtually impossible for schools and states to afford assessments capable of assessing many of the most important outcomes specified in the Common Core State Standards, thus forcing our teachers to teach to a much more limited range of outcomes, and deeply alienating them from the accountability system that is supposed to be determining their merit as teachers. What they do is lament the fact that less frequent testing will make it harder to hold teachers accountable for the performance of disadvantaged students.
I take it that the main objection of the U.S. Department of Education to a departure from the regime of year-by-year testing is their fear that such a departure would make it impossible, as Whitehurst and Lindquist point out, to continue to require the states to implement systems of test-based accountability for teachers. Evidently, the reason they care so much is that they believe that test-based accountability is the key to the improvement of student performance. But they give us no evidence for that premise.
Indeed, the evidence seems to suggest that teachers of students from low-income minority families leave the regular curriculum to drill their students endlessly on low-level basic skills tests, while teachers of students from majority and upper-income families do much less of this and spend more time teaching a more demanding curriculum. In this way, an accountability system designed to support the education of minority and low-income students may actually result in lowered expectations and less useful instruction for those students.
If test-based accountability held the key to high performance across the board and to equity for poor and minority students, then no country could have gotten to the top of the international league tables without test-based systems of accountability for teachers and value-added methodologies for assessing teacher competence. But that is not true at all. In fact, no high-performing country has gotten to the top of the league tables that way.
The argument that California should not be allowed to skip a year because that jeopardizes a system of accountability that has been proven to work nowhere does not make sense to me. In my next blog, I’ll talk about the pernicious effect of test-based accountability on teachers and students. In the following blog, I’ll describe accountability systems that work.