Why We Should Replace Aptitude Tests With Achievement Tests
Students perceive a disconnection between the subjects they study in high school and the criteria used for evaluating them in the next stage of their lives. Employers tend not to ask for high school records, a fact which sends a clear message: They do not value the information in them. Although colleges and universities ask for school grades, and some weigh them highly, they also rely on tests which are not keyed to the content of high school curricula. In so doing, they send a signal to students that their aptitude, as measured on certain tests, is more relevant for admissions than what students know.
A disconnection between effort in school, on the one hand, and work and college on the other lowers the incentive to study. It doubtless contributes to the well-documented ignorance of American students relative to those of other nations. Without strengthening the link between work in school and later rewards it is hard to imagine how we can achieve two of our national goals for education for the year 2000:
- Increasing the high school graduation rate to 90 percent.
- Having students leave grades 4, 8, and 12 with demonstrated competencies in challenging subject matter including English, mathematics, science, history, and geography--and prepared for responsible citizenship, further learning, and productive employment.
The most widely used of these tests has been called the Scholastic Aptitude Test, S.A.T. (This test, with questions on verbal, math, and writing skills, is now S.A.T. I; College Board achievement tests are now called S.A.T. II.) The old S.A.T., now S.A.T. I, is not aligned with the high school curriculum, unlike subject-matter achievement tests, now S.A.T. II, which are. Replacing aptitude tests with achievement ones would alter the incentives of students and teachers. Students who now aspire to college build the skills needed to score well on such tests. The math ones excepted, these tests are not aligned with what high school teachers cover in class: in history, the sciences, much of literature, and, of course, foreign languages. Only better students who are aiming for selective institutions take subject-matter tests, and even for them the dominant concern is the aptitude-test score. (Testimony to its power is the popularity of cram schools such as Stanley Kaplan.) Although many institutions urge students to take several achievement tests and most leading ones require them, students seem to believe that little weight is given to subject-matter test scores for admissions. The motivation of teachers should also be improved by a closer alignment between what they teach and the assessment of their students.
The movement toward national curricula standards, a movement which has been endorsed by President Clinton and many state governors, is highly compatible with our proposal. College-entrance criteria should reinforce national content goals and the curricula of the schools. However, there is opposition to national curricula standards in the elementary and secondary school establishment, and their future is uncertain. This makes college-entrance criteria all the more important.
The College Board, which established the S.A.T., is an important player in this process. It has long offered achievement tests in many subjects and is now developing PACESETTER, a secondary school framework and related assessments, in five subject areas. This is a positive development. Yet the College Board does not propose that the S.A.T. I be cut back, and made only incremental modifications to it this year.
In contrast, James Crouse and Dale Trusheim in The Case Against the S.A.T. (University of Chicago Press, 1988) argue that high school grades are better than the (aptitude) S.A.T.'s in predicting college performance. They also cite several studies showing that achievement tests predict college performance about as well as do S.A.T.'s, conclude that there is little basis to believe that the S.A.T.'s are freer from differences in curricula offerings among high schools, and dispute the claim that the S.A.T.'s help black and poor candidates.
A recent study by Jonathan Baron and M. Frank Norman, "S.A.T.'s, Achievement Tests, and High School Class Rank, as Predictors of College Performance'' (Educational and Psychological Measurement, 1992), compares the predictive value of various combinations of the S.A.T.'s, high school class rank, and achievement tests. They found that once high school rank and achievement-test scores are known, the S.A.T.'s add little to the prediction of academic performance. (The main analysis did not incorporate underrepresented minorities, but a supplementary analysis showed that this finding applies also to them.) The authors conclude that "use of achievement tests and high school class rank for college admissions could provide high school students with incentives to achieve academically, but the S.A.T.'s are not designed to provide such incentive.''
John Bishop in "Incentives for Learning: Why American High School Students Compare So Poorly to Their Counterparts Overseas'' (Research in Labor Economics, Vol. II) writes that "improving the teaching of these subjects [science, history, social science, trigonometry, statistics, and calculus, or the ability to write an essay] will have only a minor effect on how a child does on the S.A.T., so why worry about standards?'' Also, he says, "the tendency not to reward effort and learning in high school tends to be a peculiarly American phenomenon. School grades are the major determinant of who gets the most preferred apprenticeships in West Germany, in Canada, Australia, Japan, and West Europe, [where] educational systems administer achievement exams that are closely tied to the curriculum. ... Performance on these exams is the primary determinant of admissions to a university.''
This change in America probably needs to be led by private higher education, although for legal reasons others should participate in the movement for change. Although these institutions bear many burdens, they are less encumbered politically than public ones. Leading private institutions--and some public ones also--are selective; for them, the combination of high school records plus achievement-test scores would produce close to the mix of students that they now get through a combination of aptitude tests, high school grades, and, sometimes, achievement tests. There would be some differences: At the margin, less-bright but hardworking students would be favored and bright ones who are lazy or from schools with poor subject-matter offerings would be disfavored. But these institutions today use various criteria other than test scores to enable them to admit people that they want--for various reasons--and they would continue to do so. For instance, student portfolios could be used if some high schools do not offer all the courses covered in achievement tests.
These institutions have nothing to lose regarding the mix of their students, while all of higher education--and society as a whole--has much to gain from the example they would set. Moreover, even for the elite institutions, over time the better alignment of high school curricula and admissions criteria would produce better-prepared entering classes.
Why hasn't this happened? Among selective private institutions, only two small colleges, Bowdoin and Bates, have dropped the S.A.T.'s.
- One possibility is that admissions staffs in most selective institutions see themselves in something of a zero-sum game with their peers and worry that they will lose out if they switch and others don't. We doubt that any such worries are warranted.
- Another is that they fear being charged with hurting minority access--even though the S.A.T.'s have been criticized on just this score. It is true that a smaller proportion of minority students take achievement tests now, but the gap has been narrowing. More to the point: If such tests were required, the gap would disappear. According to Mr. Bishop, "Substitution of academic-achievement tests for aptitude tests in college admissions improves minority access because minority-majority differentials tend to be smaller (in standard-deviation units) on achievement tests (e.g., the NAEP [National Assessment of Educational Progress] reading and math tests) than on aptitude tests (e.g., the S.A.T.).''
- Low-income students would be disadvantaged by the cost of taking several achievement tests rather than the S.A.T.'s. However, the College Board now subsidizes the price paid by low-income students and it could establish a sliding fee as a permanent policy.
- Cooperation limited to elite institutions to change the admissions system might be interpreted as violating the antitrust laws, a concern since the U.S. Justice Department's action against the Massachusetts Institute of Technology for joining with other universities in allocating scholarship funds. To help avoid this, institutions that come together to promote this change should not be only elite ones. And clearance should be sought from the Justice Department based on the social purpose to be served.
- Perhaps the most legitimate concern is that poor-quality curricula in some schools would hurt the chances of bright but poorly prepared students. Again, changing the criteria for admissions--along with the movement for national standards--would increase incentives for inadequate high schools to improve their curricula. To give these schools time to get ready and to give an adequate "heads up'' to students, there should be adequate notice of the impending change, say four years.
So what should be done? A set of colleges and universities should agree to replace the S.A.T. I with S.A.T. II College Board achievement tests that are revised to be compatible with national curricula standards. If each is reluctant to do so on its own, perhaps for fear that it would be disadvantaged, some outside person or group--say the U.S. Secretary of Education or the chief state school officers--should urge this switch for all.
Michael Kirst is a professor of education at Stanford University.
Henry Rowen is a professor of business at Stanford and senior fellow of
the Hoover Institution.
Vol. 13, Issue 01