The Impact of Dropping the SAT

A new research study -- based on simulations using actual student applications at competitive colleges that require the SAT or ACT for admission -- has found that ending the requirement would lead to demonstrable gains in the percentages of black and Latino students, and working class or economically disadvantaged students, who are admitted.

The finding is consistent with what admissions officers have reported at many colleges that have gone SAT-optional. But the basis of this new research goes well beyond the anecdotal information reported by colleges pleased with their shifts. Scholars at Princeton University's Office of Population Research obtained actual admissions data from seven selective colleges that require the SAT or ACT. Using the actual admissions patterns for these colleges, the scholars then ran statistical models showing the impact of either going SAT-optional or adopting what they called the "don't ask, don't tell" approach in which a college says that it won't look at standardized test scores.

These models suggest that any move away from the SAT or ACT in competitive colleges results in significant gains in ethnic and economic diversity. But the gains are greater for colleges that drop testing entirely, as opposed to just making it optional. (To date, only one institution -- Sarah Lawrence College -- has taken that step.)

In terms of other measures of academic competitiveness, the study found that going SAT optional would result in classes of students with higher grade point averages. Dropping testing entirely, on the other hand, would result in higher levels of academic achievement in the entering classes at the public institutions studied, but not the privates. The research will be formally presented next month at a conference at Wake Forest University about college admissions, but the Princeton researchers released the findings Wednesday.

Parts of the findings may be controversial with both SAT critics and fans. The study found that, as the College Board has long argued, the SAT is a good way to predict the first-year academic success of students. But the study's findings on the impact of dropping the SAT as a requirement provide an independent analysis to show that dropping the SAT as a requirement does lead to increased diversity -- and that is something many colleges want to promote.

The study was conducted by Thomas J. Espenshade, a professor of sociology at Princeton, and Chang Young Chung, a statistical programmer there. Espenshade said in an interview that he started the work without strong feelings about whether the SAT should be required, and that the work received no financial support from the College Board or entities engaged in either encouraging or discouraging use of standardized tests in admissions.

One conclusion of the new study that is sure to be closely watched by admissions offices is this one: "The results show unambiguously that increased racial and socioeconomic diversity can be achieved by switching to test-optional policies." This may be particularly important because the colleges studied by the Princeton researchers were all selective. Some defenders of the SAT, noting that the colleges going SAT-optional include a wide range of institutions, not all of them selective, have suggested that selective institutions may not see the desired changes by dropping their testing requirements.

The study was based on a variety of models, all based on actual admissions data for the institutions studied (which were promised anonymity) and patterns in applications once institutions go SAT-optional. (There is much less of a track record, of course, for the "don't ask, don't tell" model, since Sarah Lawrence is alone in that approach and is known for having a highly individualized admissions process.)

In the modeling, the researchers assumed that at SAT-optional colleges, those with high SAT scores would continue to submit them, and be helped by them. Only at "don't ask, don't tell" institutions, the researchers found, would there be no advantage for having high SAT scores. For both groups, the researchers created models of two (partly overlapping) shifts in the applicant pool -- both based on the track records of other colleges, but adapted for the specific colleges studied. One shift would involve a 30 percent increase, in the applicant pool, of various groups that tend to be discouraged by SAT requirements: black and Latino students and low-income students. The other shift would be an increase of 30 percent in applications from students whose SAT scores are lower than the institutional profile.

The findings appear to confirm what SAT critics have said for years: that reliance on the SAT in college admissions favors applicants who are white and/or wealthier than other applicants. At the private colleges studied (mean SAT score 1405), the percentage of admitted applicants who are black would increase from 8.3 percent to 11.3 percent in the scenario in which more minority students apply and the SAT becomes optional.

In this scenario, if the SAT is not considered at all, the percentage of admitted applicants who are black would go up to 13.8 percent. For Latino students, the percentage would go from 7.9 percent to 10.6 percent in an SAT-optional system and to 12.0 percent when the SAT isn't considered at all. Most of the corresponding drops would come in the white applicant pool, although Asian numbers would also go down modestly.

In terms of the class shifts of such a move, gains would be seen (in all the scenarios) for lower, working and middle class students. Upper and upper-middle class students would decline, while still representing a majority of those admitted.

At the public institutions studied (mean SAT 1206), similar shifts would take place with regard to race and ethnicity. The percentage of admitted applicants who are black would go from 7.8 percent to 9.9 percent if (as projected) minority and low-income applicants increased by 30 percent after a college shifted to SAT-optional admissions. In terms of class shifts, publics differed from privates, in that the gains would come from both the top and bottom of the wealth scales, while the middle would lose a bit.

Another key finding of the study is that the "cost" of going SAT optional or even abandoning the SAT completely is relatively small in terms of non-testing academic measures. For instance, the percentage of admitted applicants at private colleges who have high school grade-point averages of A+ would go up (very modestly) under the SAT-optional models, while dropping slightly under the "don't ask, don't tell" approach.

Going SAT optional at private colleges would result in gains (of less than 1 percentage point) in the 81.1 percent of admitted applicants who graduated in the top 10 percent of their high school class. If colleges were to drop the SAT entirely from consideration, the competitive private colleges studied would end up with declines of about 4 percentage points in the share of their new classes in the top 10 percent. For competitive publics, however, either going SAT optional or dropping all consideration of the SAT results in gains in the share of the admitted class from the top 10 percent of high school classes.

Espenshade did have some cautions for those considering changes in SAT policies. One was that there are questions about whether the application gains anticipated by going SAT-optional would continue as more colleges enact such policies. If the application pools don't continue to increase, there is some danger, he said, that colleges might not have as many academically talented (but test-averse) applicants.

Another caution is that Espanshade said that for the colleges studied, the SAT is in fact an excellent way to predict the first-year academic success of applicants. So removing the SAT would take away that tool. (Espenshade acknowledged that the study did not examine whether other measures, such as class rank or grades in college preparatory courses, might have the same predictive value.)

Finally, as with all projections in higher ed these days, the study notes ways in which the current economic woes facing the United States could affect admissions trends. "It is unclear how the current economic downturn will affect application rates to schools that have recently adopted test-optional admission policies," the study says. "The diversity-inducing effects of these newly instituted policies could be blunted if lower-income students are discouraged from applying.

"Admission policies, too, might need to be adjusted to reflect new economic realities. In particular, constraints on financial aid budgets could mean that schools can no longer afford to admit as many students from lower social class categories, even if these students make it into the applicant pool. In short, the current economic climate could produce greater racial diversity at colleges with test-optional admission policies but little more (or even less) socioeconomic diversity."

So where does that leave the authors of the report? Espenshade said that for him, the findings reinforce the views of a special committee of the National Association for College Admission Counseling, which in September called for all colleges to reconsider whether they need standardized testing as a requirement in admissions. The panel did not call for the end to standardized testing in admissions, but suggested that a real review would lead many colleges to abandon requirements. The "one size fits all" approach, in which many colleges have considered testing the norm and thus appropriate, needs to be challenged, the admissions officers' report said, and Espenshade said he has the same feeling.

A spokeswoman for the College Board said Wednesday that the organization's researchers were not familiar with the new study.

Robert Schaeffer, public education director of the National Center for Fair & Open Testing, a leading critic of the College Board, said he wasn't bothered by the findings on SAT validity because his organization has never said that the SAT has no predictive value, only that there are better ways to predict college success without the negative impacts of testing.

Schaeffer said he saw the report as significant. "My initial reaction is that the authors confirm the core arguments of advocates for test-optional admissions," which are that "test-optional admissions unquestionably increase racial diversity at the institutions where it is adopted" and "when students with strong high school records are admitted in place of those with top ACT or SAT scores, there is no negative impact on undergraduate academic performance."