In May, three notorious letters of the alphabet -- SAT -- once again
set off a maelstrom of bickering in education and government
circles. When the Chronicle of Higher Education reported that the
Department of Education was drafting a pamphlet of guidelines for
universities and school districts on how to avoid being sued for placing
too much weight on a standardized test, the fulminations began.
Although the pamphlet hardly mentions the offending letters, the
very notion that the government might be commenting on the appropriateness
of the controversial college entrance exam was enough to reignite a
decades-long battle about the efficacy of standardized tests and
their role as gatekeepers to higher education.
Like so many debates over education in our country, the battlefield is
sharply divided along partisan lines. On one side are the liberal to
left-leaning activists who believe the SAT and other number standards (like
grade-point averages) should be junked or de-emphasized in order to allow
more minorities into college. The Department of Education pamphlet --
called "Nondiscrimination in High-Stakes Testing: A Resource
Guide" -- has been accused of abetting
their argument. On the other side are the conservative defenders of the SAT as an embodiment of our meritocratic educational standards.
The columnist John Leo, in U.S. News and World Report, called the
education department's paper "an attempt to decapitate traditional
assessments of merit at a single stroke and push the colleges to accept
large numbers of applicants who are well below their standards." The editors at the Wall Street Journal ridiculed the pamphlet and
offered a stronger endorsement of the exam: "The truth is, [SATs] are
highly accurate predictors of college performance, of who is and isn't
likely to graduate."
But is this true? Or is the SAT debate like so many other culture-war
battles -- fueled by misinformation?
"They weren't quoting us," says Kevin Gonzalez about the Wall Street
Journal's claims. Gonzalez is a spokesman for Educational Testing
Service, which writes and administers the SAT. "We wouldn't say that."
What the ETS does claim is that SAT scores help predict
first-year college grades. Yet that's not the only thing the test seems to
reveal. Year after year, College Board statistics show that well-off students score
better than less-well-off students, white and Asian students better than
black and Latino students and boys better than girls. Critics of the
SAT claim that a low test score is a better indicator of
economic background, race or gender than a student's chances of first-year academic success.
According to the College Board (which owns the exams, and created ETS to administer them), SAT accuracy in predicting first-year grades hovers around 50 percent. According to most independent studies, the correlation is lower, around 30 percent -- odds Ralph Nader once described as "a little better than throwing dice." But the correlation between family income and test scores has been as high as 80 percent. So choosing between, say, scores of 1,200 and 1,100 on the SAT is likely to amount to a decision based on class, more than on academic potential.
"When you're using the SAT as a big old sloppy paintbrush, it's fine," said Paul Kanarek, who runs a West Coast franchise of the Princeton Review test-prep course. "I'm gonna argue that a kid who gets a 1,400 combined on the SAT is probably gonna do better than a kid who gets 1,000. I've got no objection there. But that's not how they use the test (at the University of California)."
The University of California was sued last February for "discriminating" against a group of minorities who didn't get into the Berkeley campus under new admissions guidelines. Part of the suit complains that Berkeley made overly fine distinctions based on test scores. This, incidentally, is just the sort of legal vulnerability that the education department's pamphlet aims at showing educators how to avoid.
But Kanarek goes further than the lawsuit's implicit criticisms: "UC should make the SAT optional because it tests babkes," he says. "That's a Jewish word meaning shit. Using the SAT to predict how well kids'll do in college is about as asinine a form [of prediction] as I can possibly think of. That is the reason I think the SAT should be dropped, because it doesn't help them figure out which kids are gonna do better in college."
Of course, Kanarek is professionally biased. In addition to teaching logic and vocabulary, his classes train students to outsmart the test without necessarily knowing the material. Integral to Princeton Review's philosophy is the idea that SAT is a manipulative game to be undermined. For Kanarek it's a cause. Nothing, he says, would make him happier than not having to give test-prep courses. The SAT is "the mismeasurement of man; it is unfair to give it any significant weight whatsoever in determining whether a kid will be successful in college," he says.
"The mismeasure of man" is a reference to Stephen Jay Gould's book on the besotted history of intelligence testing. It shows how claims for IQ tests changed with the political weather during the first part of the 20th century. First, in France, they were just tools for "identifying mildly retarded and learning-disabled children who needed special help," but not "for ranking normal children." Then American psychologists used them as intelligence thermometers in the Army to sort soldiers into different jobs. This idea of sorting carried into the SAT, which the College Board inaugurated in 1926. Although the test had not originally been conceived as a liberal tool, it became one when the GI Bill allowed millions of ex-soldiers to apply to college. Suddenly the SAT was helping undermine the upper-class hold on higher learning by replacing the laziest aristocrats at Harvard and other top colleges -- who counted on getting in just because they'd gone to the right boarding schools -- with members of an ambitious, burgeoning middle class.
But even then, the number of applications a university had to cull was small compared to the number that now floods offices at the University of California every year, and no one was accepted or rejected exclusively based on their scores. In the meantime the basic theory behind measuring "aptitude" -- or inborn brain potential -- was debunked, and in 1994 ETS changed the name of SAT from "Scholastic Aptitude" to "Scholastic Assessment" Test.
Now not even ETS claims the scores measure anything quantifiable about the human brain. "Do they measure your creativity? No. Do they measure your success in life? No. Do they measure how smart you are?" asks ETS spokesman Gonzalez, "Indirectly." And the idea that the tests might be good college-performance predictors is far enough in question for ETS to prod educators to read the scores "properly" (meaning, for the most part, heavily weighted with grade-point averages), as the Department of Education has done with its pamphlet.
"SAT is meant to be used in combination with other factors, especially grade-point average, to predict success in the freshman year," Gonzalez explains. "That's all."
What upset a lot of college officials about the Department of Education paper was the idea that over-weighting a test would not only expose them to lawsuits but might also get them into financial and legal trouble with the government. Despite all the SAT's apparent shortcomings, colleges have no better standard for immediately winnowing out thousands of student applications; and yet misusing the SAT, according to some of the pamphlet's language, might have amounted to a federal civil rights violation, which struck more than a few people as absurd.
Members of the House of Representatives were so concerned they called Norma Cantz, assistant secretary for civil rights, (whose staff wrote the pamphlet), in front of a subcommittee at the end of June. She told the committee that since the pamphlet was still in draft form, its questionable language would be changed. "We are not trying to create new law," she said, and reassured the congressmen that the Clinton administration "supports the use of standardized tests."
But Kanarek thinks the whole debate is off-base, and wants to see an investigation into serious alternatives to the SAT. "There's nothing wrong with a national standard," he explains, but so far a real conversation over the shape of a new one has been lacking, and in the meantime current tests are becoming more and more deeply entrenched. The state of California will spend $50 million over the next five years to coach students for the SAT, which is surely not what John Leo had in mind when he wrote that "better schools in minority neighborhoods" were needed to fix the test-score gap.
"I certainly don't have the answer," Kanarek said. "But what a fun and engaging debate it could be, with some really bright academic minds -- we could have the opportunity to reinvent the way we admit people into college. Let's have that argument."