As college-admissions committees across the country begin sorting through this year's applicant pool, imagine for a moment what might take place at a specialized postsecondary institution such as a dart-throwing academy. This type of institution would clearly wish to accept only those applicants showing the highest potential as professional dart players. It would therefore be reasonable to expect such schools to use a dart-throwing "exam" as an admissions criterion.
Now suppose that the dart-throwing academy employed the following method of assessing ability: Each applicant arrives at a dart-throwing center early on a Saturday morning, pays a $25 fee and, after a three-hour wait, throws exactly one dart at the dartboard. Those who place their darts closest to the center of the board are considered superior applicants. Of course, because there is a certain amount of chance involved in where the dart lands, this method will imperfectly assess ability. Some very good dart players will have a bad throw; some poor dart players will have luck on their side.
Moreover, suppose that applicants who believe their first throws underestimate their true abilities have the opportunity to try again -- by spending another $25 and enduring another three-hour wait on a Saturday morning. There are several reasons to think that these dart throwers would perform better the second time around: They are more familiar with the test procedure, they may have practiced their throwing technique in the interim or they may have truly just had a bad throw during their first exam.
Finally, suppose that the dart-throwing academy considers only each applicant's "best" dart thrown during the admissions process. This would mean that applicants risk nothing by arranging to throw a second dart: If their second try turns out better than their first, they have improved their chances of admission. If their second try is worse than the first, they have done themselves no harm.
This is clearly an admissions system biased in favor of those who can afford to throw the greatest number of darts. Given a sufficient number of Saturday mornings and sufficient funds to continue paying the testing fee, even the least-skilled dart player could be assured of a favorable admissions ranking -- so long as other applicants lacked similar resources.
Indeed, the dart-throwing academy would be rightly open to criticism that it was providing advantages to the wealthiest dart players rather than the most accurate ones. Surely such a practice would be inconsistent with the meritocratic values of a society such as ours.
Right?
Not exactly: Most institutions of higher education in the United States employ a system alarmingly similar to the fictional dart-throwing academy described above. When evaluating applicants who have submitted multiple standardized-test scores, most selective colleges consider only the highest score submitted. Thus, the typical college applicant faces the same set of incentives as the dart players considered above: There is no chance of harming one's application by retaking the SAT or the ACT, and, given unlimited resources of time and money, an applicant would do well to take the test as many times as possible. Because any individual's scores tend to fluctuate between test administrations, sooner or later an applicant will -- entirely by chance -- achieve a test score that overstates his or her true ability. Many applicants who would prefer not to retake the test undoubtedly do so simply to maintain their relative position in the face of widespread retaking by their peers. High-school seniors are most certainly aware of this incentive system, as many college guides instruct their readers to take the SAT early and often.
In a study to be published in the January issue of The Journal of Human Resources, Charles Clotfelter and I examine SAT-retaking behavior among applicants to three selective colleges. We find considerable evidence of an SAT-taking "arms race" among the applicants in our sample. The typical applicant took the SAT two or three times, only 15 percent stopped after taking the test once and a small number took the test five or more times. We also found that applicants tended to improve their scores upon retaking, for each of the reasons cited in the case of dart throwers: greater familiarity with the test, improved knowledge or ability and the fact that people who decide to retake the test really were more likely to have had a bad outcome the first time around.
Our analysis shows that one particular type of applicant is highly unlikely to retake the test: one who receives a very high score the first time around. This is an intuitive result, as applicants who do well have little to gain by retaking the test. Controlling for scores on the first test, applicants who decide to take a second test tend to have higher family incomes and more educated parents than those who don't retake, and they are significantly more likely to be white or Asian American. In other words, students from disadvantaged backgrounds -- with the greatest difficulties in meeting the financial and time commitments involved in retaking the test -- are the least likely to do so, holding their first test scores constant. Among other things, these results show that SAT fee waivers, which can reduce the financial obstacles to retaking the test, fall short of creating a truly level playing field. Fee waivers do not compensate for the free time or wages that applicants forgo when they retake the test.
The effect of this behavior is to reduce the standing of disadvantaged applicants in the eyes of admissions committees. Consider a scenario where a committee must decide between two applicants who received identical scores on their first SAT administration. Our study shows that the more socioeconomically advantaged applicant is more likely to have retaken the test, and is therefore more likely to gain the upper hand in the admissions process.
For many disadvantaged applicants, affirmative-action policies probably offset a good portion of this handicap. However, after the U.S. Supreme Court rules in the University of Michigan admissions case this spring, the situation could change dramatically -- and not for the better.
Could the current system be changed in order to eliminate the benefit admissions committees confer on advantaged applicants? Our study shows that the answer is an unequivocal "yes." If admissions policies considered the average of all test scores -- or only the most recent of all test scores -- there would be less incentive to retake the test. Under either of these new systems, applicants would actually be risking something by deciding to retake the test -- and therefore might think twice before doing so. Applicants who genuinely felt that they had tested poorly, meanwhile, would still be free to try again. Empirical evidence suggests that incentive structures do influence test-taking behavior: Most law schools, for example, consider the average of all submitted LSAT scores in ranking applicants; the rate of LSAT-retaking is correspondingly much lower than the rate of SAT-retaking nationwide.
Changing test-score-ranking policies would not eliminate the advantage that wealth provides in the college-admissions process. Applicants of high socioeconomic status will still enjoy greater access to test-preparation courses, professional counseling and other costly means of improving their chances of admission. But altering the procedure for evaluating test scores would eliminate one very visible source of unfairness on the admissions playing field, and one of the few such sources that colleges themselves can actually control.
In addition, an alteration of test-score-ranking policies would provide American high-school students with millions of hours of additional free time each year. Absent strong incentives to retake the test -- and therefore freed from their justified fear of being downgraded relative to their peers -- teenagers would find themselves with one or two extra Saturday mornings to work part-time jobs, polish their application essays, volunteer in their communities or just sleep in. The admissions process in which they were competing would not only be more fair, it might also be more sane. You don't need to be an aspiring dart player to understand that such reforms would be right on target.
Jacob L. Vigdor is an assistant professor of public-policy studies and economics at Duke University.
