Test score watch: Bungling the SATs further!


Who needs summer school now: Now they’ve even started to cheat on the SATs!

So says the New York Times, which reports that some students have hired brighter folk to take the SATs for them. One kid even got arrested!

In its report, the Times describes the lazy security protocols which have governed such affairs until now. This recalled a recent Times editorial, at which we mordantly chuckled.

“Ways to Prevent Cheating,” the headline said. As they began, the editors described the way New York State is toughening up its own testing procedures:
NEW YORK TIMES EDITORIAL (9/21/11): Ways to Prevent Cheating

The Board of Regents took an important step this month when it directed the New York State Education Department to develop a plan for eliminating glaring weaknesses in the state testing system. As a first move, the Regents voted to require that state tests for third through eighth grades be given on the same day all over the state. And, from now on, teachers and administrators will have to certify that they have received and will follow the same administrative protocols.

The state is relying ever more heavily on standardized exams to make schools accountable for student performance. The Regents exams now determine whether high school students can graduate, and the yearly tests in elementary and middle schools are used in decisions on how schools are run. In coming years, teachers will also be judged, in part, on how much their students improve on state tests. Given their growing importance, the tests have to have better security measures than exist now.
Sad. Cheating on standardized tests has been an issue for decades. (We wrote columns on this topic in the Baltimore Sun—in the late 1970s!) In New York itself, questions were raised about various testing practices all through the last decade. But the editors of the New York Times kissed the ass of their billionaire mayor, applauding his brilliance, singing his praise and enjoying free booze at his parties. When Gail Collins ruled this roost, she mocked the rubes who were raising those questions—the rubes who turned out to be right.

The paragraphs we have posted speak to a sad dysfunction. Had it really never occurred to anyone that teachers and principals have to follow standardized “administrative protocols” in administering standardized tests? This represents a failure of the state board—but also of the Times itself. The Times likes to talk a good game on these topics. But its own board has slumbered and snored in the face of those “glaring weaknesses.”

Then too, we thought of the letters the New York Times published about this year’s SAT scores.

Average scores dropped a bit this year, even as student participation grew to record levels. Anyone would understand that average scores may tend to drop if more students take part—anyone but the Times and its far-flung correspondents. The paper basically missed this bone-simple point in its news report on the scores. It then published this unfortunate column by E. D. Hirsch, which bungled the point even further. (For our previous posts, click this.)

The Washington Post explained in detail about the way growth in participation might tend to affect average SAT scores. But this bone-simple factor escaped the Times. And then, they published these letters.

The letters appeared beneath this headline: "How to Stop the Drop in Verbal Scores." That headline extends the sense that the drop in scores means that something is going wrong. In fact, the drop in scores may reflect nothing beyond the fact that many more students were tested.

There’s nothing “wrong” with those four letters, but we were struck by the sample the Times selected. Surely, someone must have written the Times to alert them to life as it’s lived on the planet—to let them know how tricky it is to use average SAT scores as a measure of the student population’s progress.

Everyone has always known that the SAT isn’t made for that use. Everyone has always known that, except the New York Times.

Our public school students are very dumb, our biggest newspapers like to cry. When we read complaints of that type, we try to consider the source.


  1. The idea to fight cheating is to have the cheaters say that they followed certain protocols? There isn't a big jump from cheating to lying...

  2. Following "standardized “administrative protocols” in administering standardized tests" means having documented procedures for verifying the identity of test takers, integrity of their answer sheets, and so on, that is subject to independent, impartial review or even real-time objective observation of the processes (like the UN does with contested national elections, for instance).

    It's not difficult to figure out what Somerby was getting at with that statement, since other testing groups routinely follow such procedures (MCATS, medical board exams, bar exams, etc.). But you probably knew that.

  3. If the Times and others were to acknowledge the simple fact that "the drop in scores may reflect nothing beyond the fact that many more students were tested," they'd have to acknowledge analogous differences between populations taking the international tests that are constantly used to evaluate US students' performance. I'm not just referring to, e.g., the advantages Finland enjoys as a homogeneous society compared to our own, but even more to which groups are tested in these different countries. For instance, do countries that divert a substantial percentage of their 13-year-olds to trade schools have these students take the NAEP test alongside their elite, university-bound students? I really don't know the answer (and I can't imagine the Times thinking to ask it, much less answer it for me), but until I do, how can I even begin to compare one country's students' results' to another country's?

    The sloppiness in our world's use of statistics/anything numerical is so widespread that I'm almost beyond hope. We seem to expect analyses to which numbers are attached to provide quick-fix, "bottom-line" answers, rather than a set of tools (among many) for refining the further questions we need to ask.