So California is about to abandon the use of the SAT test. This seems likely to produce cries of joy from millions of parents. The Scholastic Aptitude Test (SAT), as anyone who has teenage children, or has had them in the past thirty years, surely knows, is a test designed to predict how well children are likely to perform in college. The California education system discovered that the trouble with it, in brief, is that it doesn't.

This is actually an oversimplification. The test is available in two parts, the SAT and the SAT II. The latter is designed to measure actual knowledge of certain subjects, and the California educators found that the SAT II served as the best of several possible predictors of college performance. The only other measure that came close was how well a child had done in high school. Saying a child who has done well will do well does not strike most people as profound. It is rather like saying that a child who is above average height for her age is likely to grow into a taller-than-average woman. People have also dismissed high school performance as a biased criterion. The arguments have gone both ways: high school performance is not reliable because some high schools are inferior and do not prepare students well for college life; or, high school performance is not reliable because some high schools are inferior and it is much easier to emerge from them with high grades. Take your pick.

So the SAT II appears to be the measure of choice. The basic SAT, on the other hand, proves to be far less useful as a predictor of college performance. This is unfortunate, since although almost everyone takes the basic SAT, far fewer take the SAT II.

It seems reasonable to ask what is going on here. How did a test which does not do what it is supposed to do become such an overwhelming influence in American education, to the point where ambitious parents routinely enroll their kids in expensive tutoring for the SAT exam? And, a second question which actually interests me rather more, if the SAT is not measuring what it was designed to measure, what is it doing?

I want to tackle the second question first. One objective of the SAT was to remove any special advantage accruing to a taker because of family wealth, position, or quality of prior education. This sounds like a worthy goal, since we would like to ensure equality of opport unity in college education to everyone. Unfortunately, it ties the hands of test designers so much that what is left becomes, no matter how much our education system may hate and avoid the term, a form of intelligence test. As I pointed out in an earlier column, tests like the Stanford-Binet, designed to measure general intelligence, are now so "out" that it is in bad taste even to think about them. And our education system, no matter how it lags behind other nations in multinational testing, can be relied on to lead in political correctness.

It was not always so. In the early part of the twentieth century, intelligence tests were all the rage. People boasted (some unenlightened few may do so still) of their high IQ. Such tests declined in popularity as it became clear that the tests were being used too often to confirm an opinion already held by the testers. "Intelligent" had the implicit meaning, "like me."

The designers of the SAT, to give them credit, have done their best to avoid such a trap. No longer should educated white males always come out on top just because the tests are made up by educated white males. Everyone, from other races to uneducated people to women to immigrants, ought to have an equal chance of doing well. There should also be no way that a student, by practicing, should be able artificially to inflate a score, since access to such practice is itself not available to all and another form of discrimination.

Regrettably, after purifying the SAT until no form of ethnic or cultural bias can be detected in the questions, what is left often appears as little more than a set of tricks. The number of those tricks is limited, and they can be learned. Scores can be dramatically improved by special training for the SATs, even though it is hard to think of anything less likely to be of value in college than techniques needed to improve SAT scores. Rather than take this as intuitively obvious, the California educators went ahead and did a large study that confirmed it experimentally.

Being told a trick and remembering it does not to my mind qualify as intelligence. If I ask for the next number of the sequence, 1, 4, 9, 16, 25, or for the next number of the sequence 2, 3, 5, 7, 11, 13, 17, you have a good chance of being able to deduce the answer for yourself. However, if I ask for the next letter of the sequence, O T T F F S S E, or of the sequence A, E, F, H, I, and you provide it to me, you are probably not working out the answer. Far more likely, you have seen these particular sequences before. (By the way, to point out one other defect of this type of test, I know two equally valid answers to the second letter sequence.)

Discovering a trick for yourself, yes. That displays a form of intelligence. But certainly not the only one. Where in our measuring system do we fit creativity of other types - the ability to invent melody, to improvise on an instrument, to draw, to sculpt, to make up stories, to dance, to invent new mathematics and physics? The answer is, of course, that it is not there. We can't test in groups what we only know how to evaluate one by one.

And here we have the answer to my first question: How did the SAT come to loom so large as a determining factor in who goes to which college? It came because our colleges lack the people and resources to spend lots of time with every applicant. If they could do so, any form of standardized test would be unnecessary.

An improbable dream? I know, I know. But we ought to work on it. Because until something like one-on-one evaluation is realized, a SAT, maybe under another name but with the same weaknesses, lurks in your child's or grandchild's future.

"Borderlands of Science" is syndicated by: