The new (and improved) theology of the SAT.

There was always a Calvinist streak to the old SAT. A descendant of IQ tests from the 1920s, the test now known as SAT I is supposed to measure something innate and immutable—aptitude—and be impervious to coaching or studying. Good works in school avail you nothing. Brains are bestowed at birth, and results are predestined. On judgment day, the class clown or goof-off may be raised up, his chosen status revealed by a perfect score of 1600, and the class nerd cast down, his stellar transcript unmasked as the product of grade inflation. (Only the “saved” have a prayer of being admitted to Harvard, or the kingdom of heaven.) If you do badly on the test, it’s hard to know what to do about it—except make your contribution to the $100 million test-prep industry, the educational equivalent of buying indulgences.

In truth, of the 1.7 million students who take the test each year, scores will only be a determining factor of admissions for a tiny fraction—50,000 maybe—who attend one of a handful of selective colleges. But as a result, the SAT I is at the center of most debates about affirmative action and educational opportunity. It’s less a test than an entire belief system, governing American notions of merit, opportunity, and fairness.

That theology, however, is about to change. On June 27, the College Board is expected to approve a series of revisions to the test—among other things, introducing a separate writing exam, eliminating the analogies in the verbal section, and adding more reading and advanced math questions tied directly to high-school curriculums. The move is an implicit response to the criticisms of Richard C. Atkinson, the president of the University of California. A former professor of cognitive psychology at Stanford University and testing expert (who in his 70s was considered on the verge of retirement), he stunned the educational establishment a year ago by announcing that the state university system—the College Board’s biggest client—was considering replacing the SAT I with subject-based, achievement-style tests such as the ACT or the SAT II. Conservatives promptly denounced Atkinson’s plan as a cynical ploy to get around Proposition 209, the state’s ban on affirmative action. In their view, the College Board’s subsequent decision to revise the test is a capitulation to the forces of political correctness. Progressive educators were only slightly less unhappy. Most would like to see high-stakes, standardized tests eliminated altogether. They thought Atkinson didn’t go far enough.

Defenders of the SAT I argue that the test is necessary because it provides a common yardstick against which to gauge the scholastic ability of students who have attended widely varying schools. The test has what statisticians call a higher “predictive value”—a measure of how well a student will do during his or her first year of college—than grades or teacher recommendations. But according to a University of California study released last fall, when combined with the high-school grade point average, the SAT II has a higher predictive validity than the SAT I. And not only is the SAT II a better yardstick, it has all sorts of other ancillary benefits as well.

Atkinson’s real beef with the old-style SAT is that it’s an aptitude test, designed to measure some vague notion of innate intelligence. But by divorcing aptitude from actual achievement, he argues, the test sends precisely the wrong message to high schools and students: that what kids learn in class doesn’t matter, because it all comes down to what they’re born with anyway. Uncertain of how to prepare kids for a test that is not about anything in particular, teachers and parents either throw up their hands or push students into test-prep courses that have no real educational value. This sense of fatalism is the most pernicious aspect of the test. It’s one thing to tell a poor black parent that his or her child hasn’t achieved. It’s another thing altogether to say that a child’s aptitude, or capacity to learn, is fixed in the firmament and unalterable. Yet that is essentially what the SAT I says.

The College Board is not revamping the test entirely, but Atkinson has succeeded in pushing it in the direction of achievement tests—no small accomplishment. By adding a writing test, for example, the College Board will almost certainly encourage more writing in the nation’s high schools. Similarly, replacing the analogies section—which has spawned some of the more mindless test-prep—with reading questions tied to high-school curricula might actually inspire the reading of real books. That’s not to say that such reforms will get rid of test-prep altogether or eliminate the class advantage of kids who can afford private tutors. But the College Board is at least nudging test-prep toward the study of real subject matter, such as math, English, and science. This will help all students, whether they are applying to a selective college or not.

Opposition to Atkinson’s ideas comes from both left and right. But conservatives’ criticisms are harder to fathom. Aren’t they the ones who are always saying schools should concentrate more on content? Their objection is that aptitude tests do a better job than achievement tests of uncovering what affirmative-action critics Abigail and Stephan Thernstrom call “diamonds in the rough—students who come from lousy schools and haven’t learned much but who are nonetheless extremely promising academically.” The Thernstroms never quite explain how kids who haven’t learned anything could still be “extremely promising academically.” But they seem to fear that poor, bright kids would suffer if asked to perform on an achievement test that actually requires them to know something specific. (Imagine the scorn they’d heap on a liberal who made the same argument.) But according to the University of California study, there is actually less correlation between students’ scores and their socioeconomic background on the SAT II than on the SAT I. In other words, slightly more disadvantaged kids—the kind who attend those “lousy schools” the Thernstroms are so worried about—are admitted to the University of California using achievement tests.

What conservatives are really worried about is Atkinson’s calls for a more “holistic” admissions process, by which he means simply looking at more than grades and test scores when assessing applicants—a practice, it’s worth pointing out, that fairly describes the way every Ivy League school has long approached admissions. But “holistic” is the kind of word that sets conservatives’ teeth on edge. They worry it’s a cover for lowering standards to ensure that more minority students get in than would otherwise. And in their pursuit of this obsession, the virtues of achievement tests often get lost. Whatever one thinks of holistic assessment, the real issue is whether achievement tests are better than aptitude tests. And by any measure—from predictive validity to the incentives they create for high schools—they are.

Progressive educators have their own hobbyhorse to ride, arguing that any form of standardized test stifles creativity and encourages teachers to concentrate on rote exercises. (This is also the complaint of wealthier suburban parents at places like Scarsdale High School.) But a child who has already blossomed into a minor novelist shouldn’t have any trouble writing a coherent essay with a thesis and a few supporting paragraphs. If kids have already graduated to “higher thinking skills,” the nirvana of progressive educators, they can handle a few lower order tasks, too.

As originally conceived by the Educational Testing Service, the Scholastic Aptitude Test was supposed to create a “natural meritocracy” by helping poor and middle-class students win admission to elite colleges that had previously been reserved for the sons of the East Coast establishment. But as the test has evolved, it has ended up replacing one accident of birth (class) with another, in its way equally dubious one (innate intelligence). The idea of merit, in the sense of actual achievement, has been lost. The College Board’s recent changes to the test are a small step toward resuscitating that idea. It’s the right move.