Aptitude vs. Achievement: Should We Replace the S.A.T.?
The idea that colleges should choose among applicants on the basis of their "academic ability" appeals to both educators and the public. But "ability" has two distinct meanings, which imply different admissions policies. In one usage, academic ability means an existing capacity to do academic work. In the other usage, academic ability means a potential capacity to do such work. To say that an applicant "has the ability to do differential calculus," for example, can mean either that the applicant can already do differential calculus or the applicant could learn differential calculus given opportunity and motivation. To avoid this ambiguity, psychometricians usually call the ability to learn something an "aptitude" while calling current mastery of a skill or body of knowledge "achievement."
When colleges say they want to select the "ablest" applicants, they almost always mean the applicants with the greatest academic "aptitude," not those with the highest current levels of academic "achievement." Yet so far as we know, the United States is the only major industrial nation in which colleges and universities use so-called "aptitude" tests to help them make admissions decisions. This does not mean that British, French, German, Japanese, or Soviet universities are indifferent to "aptitude" as we have defined it. On the contrary, educators in these countries appear to be just as interested in their students' ability to learn as are educators in the United States. Nonetheless, universities in other countries rely on what psychometricians call "achievement" rather than "aptitude" tests to help them make admissions decisions. They do this because they assume that the best single predictor of how much a student will learn in a university is how much he or she learned in secondary school.
Despite the widespread use of tests like the Scholastic Aptitude Test (sat) in this country, the evidence that we have reviewed indicates that they do not measure what we normally mean by "aptitude" any better than conventional achievement tests do. sat scores appear to be just as dependent on home environment and school quality as are scores on conventional achievement tests. And conventional achievement tests predict success in college and adult life just as well as the sat or other "aptitude" tests do. Many at the Educational Testing Service (ETS) conclude from this that, while the sat is no better than conventional achievement tests, it is no worse either. We disagree. The reason is that emphasizing tests like the sat in college admissions undermines efforts to improve secondary education.
Both college teachers and the general public have become increasingly concerned in recent years with the fact that freshmen know less than they used to. The decline in average sat scores is the most publicized symptom of this problem, but other testing programs have found the same trend. The problem is not that primary-school children are not mastering the "three R's" as well as they used to; if anything, they are doing better. But teenagers are not moving as far beyond the "three R's" as they used to.
Talking to high school students, teachers, and parents, one gets the strong impression that high-school students do less academic work today than they did 20 years ago. Yet their nominal grades are higher than they were 20 years ago. This means that a clever student can earn A's without learning much. Such students know that if they earn A's in high school and do well on the sat they will be able to get into a top college. And they assume--erroneously, in many cases--that because they got A's without doing much work, they must be "smart" and therefore likely to do well on the sat If they then do badly, they often blame the test rather than their school.
While we have no quantitative data on the number of hours secondary-school students in different countries spend studying, most observers agree that American high-school students do less academic work than their counterparts in Europe or Japan. Most observers also agree that one crucial reason--perhaps the crucial reason--European and Japanese teenagers work hard is that at the end of secondary school they must take achievement tests that largely determine whether and where they will be allowed to obtain higher education. We would not want to see American teenagers become as preoccupied with academic success and exams as Japanese teenagers are, but this is not a real danger. We would like to see American teenagers do more academic work than they now do. If selective colleges were to base admission on high-school grades and tests that measured mastery of the secondary-school curriculum, instead of basing it on grades and tests that measure vocabulary, reading comprehension, and basic mathematics, high-school students who wanted to attend selective colleges might take their academic work somewhat more seriously.
Emphasizing mastery of secondary-school curriculum would have other benefits as well. First, it would help reinforce the traditional American notion that effort will be rewarded. Emphasizing the sat, in contrast, reinforces the notion that working harder will not help much and that success depends on factors like "smarts" over which you have no control. As we have seen, this notion is probably mistaken, since the sat is also an achievement test. But neither students nor teachers fully realize this. If we want high-school students to believe work pays off, colleges must not only reward applicants who have worked hard but must do so in an obvious way.
Emphasizing tests that measure mastery of secondary-school subjects would also encourage such schools to offer more demanding academic courses. One should not, of course, exaggerate the potential impact of college admission tests on high-school curricula. Only about half the nation's high-school graduates attend college, and only a handful even apply to selective colleges where test scores play a major role in admission decisions. This minority does, however, influence the overall character of secondary education in a way disproportionate to its numbers. Those who attend selective colleges often come from influential families, and many educators judge their school's overall success by such students' accomplishments. Furthermore, while such students are not always social leaders in their schools, they typically exert more overall influence on their peers than their peers exert on them. Changes in college admissions policy that affect these students therefore tend to have ripple effects on schools as a whole.
The effect of substituting conventional achievement tests for the sat would depend, of course, on how the new system worked. If all colleges were to require conventional achievement tests, more secondary schools would feel obliged to offer courses explicitly tailored to such tests. Those who oppose such exams usually point with horror to their consequences in France; those who favor them usually point to England, where national exams have produced more curricular uniformity than in America but far less than in France. Given the small proportion of American students applying to selective colleges, national exams would probably have even less impact here than in England. Furthermore, if our concern is with maintaining diversity, it would be easy to establish a set of achievement tests more flexible than those used in either France or England.
The College Board now offers achievement tests in 14 subjects. If more colleges were concerned with measuring achievement, ETS could easily double or triple this number, even giving several types of exams in more popular subjects. If the College Board allowed students to take as many exams as they wanted, while only reporting their highest scores, students could take a lot of unorthodox courses without jeopardizing their admissions prospects. Such a system would still reward "smarts," as the sat does, but it would also require and reward sustained application in at least a few areas, which the present system does not.
Most discussions of college admissions policy focus on distributive justice. They ask whether the existing system admits the students somebody thinks selective colleges "ought" to admit. If we apply criteria of this kind, the choice between the sat and conventional achievement tests appears, on present evidence, to be a toss-up. The evidence we have examined suggests that both are equally helpful in identifying applicants who will earn above-average grades, who will finish college rather than drop out, and who will be more successful than average after they graduate. Both pose comparable obstacles to blacks, working-class whites, and applicants from second-rate high schools.
But we must also ask how college admissions policies affect the behavior of secondary-school students, teachers, and administrators. Because this issue is harder to quantify, experts tend to ignore it. But in the long run it may be even more important than the questions of distributive justice about which experts worry most. A good college admissions system should encourage diligence rather than sloth, and seriousness rather than frivolity among secondary-school students. The present system does not seem to do this. Selective colleges today base their admissions decisions largely on grades and sat scores. Some high schools distribute grades in such a way as to promote diligence and seriousness, but many do not. Because of its content and its name, the sat does not appear to reward diligence. Furthermore, by emphasizing skills that secondary schools do not explicitly teach, at least after 10th grade, the sat implicitly tells secondary schools that most of what they teach does not really matter.
It is always hard to defend a curriculum concerned with what Matthew Arnold called "the best that has been thought and said." Lacking external support, American secondary schools have largely stopped trying. Changing our college admissions system so as to reward those who learn the most in secondary school would not solve this problem, but it would probably help.
Vol. 01, Issue 35, Page 24, 21