Michael J. Feuer is a senior analyst at the Office of Technology Assessment of the United States Congress and was the project director for the Q.T.A.'S recent study, “Testing in American Schools: Asking the Right Questions.’'
The rhetoric of national standards and testing is a rhetoric of crisis. One typically hears references to the assumed free fall of American productivity and competitiveness, a problem that is laid at the doorstep of the public-school system, whose alleged failures are in turn chalked up to our system of standardized testing. For zest, add a pinch of comparison to those allegedly superior European or Japanese testing systems, and what emerges is a very appetizing recipe for reform: Fix the tests, which will make the schools better, which will get the economy humming again.
I simplify, but that’s the basic model, one that has left many economists--not to mention other social scientists--more than a bit puzzled. My point here is not to challenge or defend the economic argument per se. Rather, it is to suggest that the rhetoric of crisis and salvation trivializes the subtle connections between curriculum, instruction, and assessment, and could ultimately betray the cause of school improvement that new standards and tests are supposed to promote.
The second r is for regulation, not the federal kind with a capital R, but rather the kind that means control. One of the most vexing problems in the history of testing in the United States has been to safeguard tests from misuse. History teaches us that it doesn’t much matter what a test looks like--the better it appears to be in accomplishing one function, the more likely it is to be borrowed for other functions. It’s a kind of “Peter principle,’' pushing tests into uses for which they are fundamentally inadequate--or even harmful--and ultimately eroding their credibility and usefulness even in their original settings.
There is no clear solution to this problem. But the first step is to clarify the purposes of testing: providing useful instructional feedback to teachers and kids, monitoring school or school-system effectiveness, and selecting, placing, and certifying students are fundamentally different objectives that impose different constraints on test design and use. It would be a tragedy if the creative assessment methods being researched and tried out now in dozens of districts and states were smothered by premature or inappropriate uses.
Finally, the third r is for rationality. The idea of national educational standards is very appealing and seems to make a lot of sense. So why is it so hard to achieve? In addition to the historical and cultural reasons having to do with our decentralized pluralistic school system, there may be a cognitive barrier.
Educational outcomes are much easier to agree on after the fact than they are to define or articulate in advance. Everyone has his or her favorite example of what it means to be good in math or reading or writing; but it’s another thing entirely to spell out those criteria in advance, or to find criteria that everyone can agree on. Moving from an awareness of examples of good educational outcomes to a precise articulation of what those examples represent may push us to the limits of our human rationality. Using those definitions in a system of high-stakes examinations could raise many fundamental problems.
Proponents of “systemic validity,’' performance assessment, and the “thinking curriculum’’ are acutely aware of the power of good examples. The National Council of Teachers of Mathematics, for instance, defined its math standards principally in terms of examples of what children ought to be able to do. It’s not clear that the best way to breed good examples of learning is to convert them into high-stakes exam questions.
The national standards process has already fomented a healthy debate, a kind of national soul-searching on what and how we teach our children. Keeping the debate alive should remain a high priority.
