Backdoor Affirmative Action

Abandoning college-admissions tests is a misdirected effort.

How can we counteract the plunge in the number of black and Latino students on our college campuses? This is the question faced by California educators since the passage in 1996 of Proposition 209, which banned consideration of race or ethnicity in admissions to public colleges and universities. Now, the state of Washington is beginning to confront the effects of Initiative 200, a clone of Proposition 209 that passed in November, and Texas continues to grapple with the effects of the Hopwood decision.

Last year, the University of California landed upon a simple solution to the decline in minority enrollment: Eliminate the SAT as an admissions criterion. “We ... have evidence that the SAT loses us 2,000 Latino students this year alone,” Eugene Garcia, the dean of the school of education at Berkeley, said in a 1997 interview. Although enthusiasm for total elimination of the standardized-test requirement has waned, some California educators and lawmakers cling to the view that admissions testing is the primary barrier to campus diversity.

Would eradication of standardized admissions tests produce a more ethnically diverse freshman class? Recent evidence from several sources indicates the answer is no. Consider:

Let’s take a closer look at each of these points.

• Would eliminating the SAT foster diversity on California campuses?

In December 1997, the University of California office of the president reported on a study designed to determine the effect of various admissions policies on the rates of “UC eligibility,” which is based on the completion of college-preparatory courses, GPA for those courses, and, if the GPA is below 3.3, scores on the SAT or ACT. The data for the study, which were collected by the California Postsecondary Education Commission, came from a random sample of 1996 graduates of California public high schools.

The study’s conclusion was surprising to some: Eliminating the admissions-test requirement (while maintaining the 12.5 percent overall eligibility rate mandated by University of California admissions policy) would produce very small changes in the eligibility rates for Latinos (from 3.8 percent to 4 percent), African-Americans (from 2.8 percent to 2.3 percent), and Asian-Americans (from 30 percent to 29 percent). The largest change would be an increase in the eligibility rate for whites (from 12.7 percent to 14.8 percent).

The indisputable fact is that both high school achievement and admissions-test scores are reflections of the same educational system.

The minimal change in eligibility rates for African-American and Latino students is less remarkable in light of the finding that about 76 percent of graduates were ineligible because of inadequate course background. About 63 percent of graduates had “major course omissions” or grade deficiencies, or attended “schools that did not have a college-preparatory curriculum approved by the university.” The percentage of students with major course background deficiencies was higher for African-Americans (77 percent) and Latinos (74 percent) than for whites (59 percent) and Asian-Americans (39 percent). The analyses by the postsecondary education commission further showed that 90 percent of the UC-eligible public high school graduates achieved their eligibility status solely on the basis of their GPAs in college-preparatory courses. Only 2.5 percent of graduates were ineligible solely because of inadequate admissions-test scores.

• Are admissions tests biased against people of color?

A 1994 College Board study by Leonard Ramist, Charles Lewis, and Laura McCamley-Jenkins, based on 1985 data from 45 colleges, provides a useful context for considering ethnic-group differences in admissions-test results. This unusually thorough analysis of the utility of the SAT as a predictor of college grades found that, as in most previous studies, average SAT scores were higher for Asian-American and white students than for black and Latino students, with more dramatic differences in the math than in the verbal component. (The most recent SAT results revealed similar patterns.)

Among the general public, the persistent pattern of racial-group differences in average test scores is often considered to be sufficient evidence for the existence of test bias. For example, the original language of a California bill vetoed by Gov. Pete Wilson in September stated that “a [standardized] test discriminates if there is a statistically significant difference in the outcome on test performance when test subjects are compared on the basis of gender, ethnicity, race, or economic status.”

When academic researchers study the SAT, however, they don’t focus exclusively on the average scores of each ethnic group. Because group performance differences could arise for many reasons that aren’t a function of the test itself--unequal educational opportunity being the most obvious--the absence of such differences isn’t considered a criterion for fairness.

A typical SAT validity study evaluates the predictive effectiveness and accuracy of the test: How well does the SAT predict first-year college grade point average? Consistent with recent University of California studies and with most earlier SAT research, the College Board study showed that high school grade point average and SAT scores were important predictors in all ethnic groups, and that including the SAT led to better prediction than could be achieved using only high school grades. In the African-American group, the College Board study showed that SAT alone provided slightly more effective prediction than high school grade point average alone.

A particularly important question is whether the SAT tends to produce grade predictions that are too high or low for certain groups of test-takers. The College Board study found that the SAT tended to predict higher first-year college GPAs than were actually attained for African-American, Latino, and American Indian students, while predicting lower grades than were attained for Asian-American and white students. Past research on the SAT performance of black and Latino test-takers has typically produced similar results. (Among many reasons that have been advanced to explain this phenomenon is the possibility that black and Latino college students are more likely to experience life difficulties that interfere with academic performance.)

The honest and direct way to achieve diversity is by considering an applicant’s membership in an underrepresented group to be a ‘plus’ in the admissions process.

"[F]ar from being biased against minority students, standardized admissions tests consistently predict higher levels of academic performance than most blacks actually achieve,” say William G. Bowen and Derek Bok in their recent book, The Shape of the River. (“Book Cites Benefits of Race-Based College Admissions,” Sept. 16, 1999.) Educational researcher Robert L. Linn made a similar statement in a 1983 paper. This well-established finding undermines the claim that removing SAT requirements would be effective in boosting the enrollment of well-prepared black and Latino students.

• Is high school achievement a better admissions criterion than test scores?

While there is little basis for concluding that standardized admissions tests are biased in any formal sense, it’s clear that overrelying on tests or other traditional achievement measures in admissions can perpetuate the underrepresentation of certain groups by--as author Ellis Cose puts it--rewarding “those who have already been well-schooled.” The studies by the University of California president’s office, the California Postsecondary Education Commission, and the College Board all showed that high school grades generally follow the same patterns of ethnic group differences as do test scores. And, as noted earlier, the California studies showed black and Latino test-takers were less likely than Asian-American and white test-takers to complete required college-prep courses.

Because grades and test scores tend to parallel each other, it’s not surprising that the University of California analyses predict that a current proposal to offer admission to the top 4 percent of graduates from every California high school, regardless of test scores, would “have little impact on racial proportions at UC.” Messrs. Bowen and Bok weigh in on this issue as well, asserting that a policy of admitting a fixed percentage of each high school class (as the University of Texas is already doing) “is unlikely to be an effective substitute for race-sensitive admissions policies.” This approach, they say, is likely to lower minority students’ college graduation rates and diminish “the pool of students who can compete effectively for positions of leadership in business, government, and the professions.” (Some educators have raised the additional concern that excluding standardized tests from admissions requirements could lead to rampant grade inflation, rendering high school GPA useless as an admissions criterion.)

The indisputable fact is that both high school achievement and admissions-test scores are reflections of the same educational system, with all its flaws and inequities. By using grades rather than SATs as an admissions criterion, said sociologist Christopher Jencks in a 1989 essay, “you are simply substituting tests designed by high school teachers for tests designed by the Educational Testing Service. ... “

Individuals of every political stripe agree that, ultimately, we must fix the K-12 “pipeline” (or “river,” to adopt the Bowen-Bok metaphor). But this will take years. What can we do in the meantime? The honest and direct way to achieve diversity is by considering an applicant’s membership in an underrepresented group to be a “plus” in the admissions process. Eliminating admissions-test requirements as a form of covert affirmative action is not sound policy. Instead, we should focus our efforts on eliminating the legal obstacles to affirmative action programs.

Rebecca Zwick is a professor of education at the University of California, Santa Barbara, and an associate editor of the Journal of Educational and Behavioral Statistics. Before joining the University of California in 1996, she spent 12 years in the division of statistics and psychometric research at the Educational Testing Service in Princeton, N.J.