Test Results You Can't Rely On

Email a Friend

Almost everybody who is interested in education can agree that accountability is a good thing. But like many parents, I am increasingly convinced our No Child Left Behind testing is flawed at the core. When we get accountability wrong we don’t just waste money, we waste children’s lives.

DESCRIPTIONLorraine Levey Matthew Levey

At a community meeting in the Cobble Hill neighborhood of Brooklyn last week, Shael Polakow-Suransky, the Department of Education’s chief accountability officer, did his best to defend state tests that he acknowledged "are not as good as we would like."

But it's a hard position to take these days, and the crowd, whom he described as containing “many friends,” grew angrier as the night wore on.

Before the meeting, I had a chance to speak to Mr. Polakow-Suransky, and he told me that for low and high performers, the New York State tests can’t do more than provide a "best guess" as to the child's actual achievement level. Graded on a 4-point scale, they can vary as much as 0.5 in either direction, he said.

A score of 3.4 is more accurately described as somewhere between 2.9 and 3.9. Or between "just barely proficient" and "off to M.I.T. on a scholarship." Except we can't really tell you which one.

Last month, when the Department of Education released teacher data reports that were based on these same test scores, Mr. Polakow-Suransky said that "the tests are meant to measure students in the middle, and not subtle changes among the highest and lowest performers.” He cautioned principals that test scores higher than 3.4 or below 1.68 are therefore unreliable.

Yet it is exactly these imprecise tests that drive parents from Cobble Hill to the Upper West Side insane trying to ensure their kids earn the highest possible score. For many parents in my neighborhood, a middle or high school that screens kids based on these state tests is better. Schools that take any students in the neighborhood, the logic goes, are not as good.

Even though their children are already getting hours of in-school test prep, some parents spend thousands of dollars trying to make sure their children ace the tests.

But it turns out all this effort may not matter. Differences in student scores higher than 3.4 are more "noise" than "signal." As Mr. Polakow-Suransky told The Times last month, the tests are “too blunt a measure to track small amounts of progress made by students at the very ends of the spectrum.”

Some might be pressed to weep for the progeny of brownstone Brooklyn who are denied their first choice of middle school because of a sloppy test. But the hallmark of this administration has been its concern for the voiceless and too-often ignored children for whom school is so critical, but whose parents struggle to go the extra mile.

Shouldn’t Mayor Michael R. Bloomberg and John B. King, the state education commissioner, be outraged that the test they rely on to judge the progress of our struggling learners doesn’t work? Shouldn't we all?