Statement on Affirmative Action

Statement from
Bruce Alberts, President, National Academy of Sciences,
Wm. A. Wulf, President, National Academy of Engineering,
and Harvey Fineberg, President, Institute of Medicine

April 1, 2003

This week the U.S. Supreme Court is hearing arguments in two cases that challenge the right of colleges and universities to account explicitly for race in student admissions. At stake is the legality of what has become the broad consensus among the leaders of America's colleges and universities, as well as many leaders of business, the military, and other sectors. For at least 25 years, since the Supreme Court's ruling in the landmark Bakke case, U.S. colleges and universities have initiated and experimented with a wide variety of admissions systems that rest on three basic premises:

This argument is no less applicable to the specific goals and responsibilities of the nation's scientific, engineering, and health care enterprise. We hope that the Supreme Court will permit colleges and universities to apply appropriate programs and policies to achieve the diversity that is necessary to the future of this enterprise, to the mission of higher education, and to fulfilling America's great promise of opportunity for all.

Scientific and technological progress has been largely responsible for the enormous advantages now enjoyed by more Americans than at any time in history. National security, the health of our nation, and the strength of our economy depend heavily on the advancement of science, technology, and medicine. Although there have been those throughout history who have tried to pit high standards for scientific achievement against increased opportunity, the reality is that enlarging the franchise has enriched everyone.

Diversity in science, engineering, and the health professions is both a means and an end. It is a means toward even greater technological and economic progress, because creativity, productivity, and success in these fields depend greatly on openness to diverse ideas and experiences. And it is an end, because science, technology, and medicine – just like the humanities, law, art, theater, sports, or the military -- should not be accessible to only the privileged few. Medicine offers an especially compelling case. Diversity of the physician work force will improve access to health care for underserved populations. And diversity among managers of health care organizations makes good business sense.

There is overwhelming evidence that diversity enhances educational experience and benefits scientific, technological, and medical advancement. Is it attainable through programs that do not consciously and explicitly account for those attributes that combine to create the sought-after diversity? The answer appears to be no. Thus far, attempts to achieve racial diversity by including factors that may appear to correlate with race -- rather than race itself -- have proved to be ineffective.

There is a fundamental difference between denial of due process or of equal treatment and the conferring of preference based on race. It is critical to understand that increasing the probability of admission for some students does not automatically deny admission to other students who may erroneously believe that they are otherwise assured a place. A multitude of factors go into the admissions decisions of selective colleges and universities, and the assumption that any single factor is sufficient to deny admission to an applicant without that factor is wrong.

The idea that the inclusion of race as a factor in admissions policies compromises the principle of merit is a mistaken one that rests on misperceptions about the precision with which existing technologies measure current and future academic performance. It is still a widely held but erroneous perception that standardized test scores, in this case often the SAT, provide an objective and accurate gauge of future student success. A solid body of research leads to a more subtle conclusion that is directly relevant to the case before the Supreme Court.

Standardized test scores do exactly what they were designed to do: When used with other measures, such as high school grades, they predict freshman grade point averages reasonably well. Nevertheless, the strength of that correlation is such that there is substantial room for error. Some low-scoring students will excel in college, and some high-scoring students will fail. A university's complex goals and mission cannot be achieved by relying solely on test scores.

The common misconception that test scores are reliable but imperfect predictors of future performance has far-reaching implications, as noted in a 1999 study from the National Research Council of the National Academies, Myths and Tradeoffs: The Role of Tests in Undergraduate Admissions. Indeed, the claim that a student with a higher test score should be guaranteed admission over another student with a lower score is not scientifically defensible. Standardized tests are not designed to make fine distinctions at any point on their scale, or to provide information about all the factors that influence success in college and beyond. Tests are useful primarily when administrators need to sort applicants into broad categories -- those who are quite likely to succeed academically at a particular institution and those who are unlikely to succeed. Still, flaws remain. Because of the gap between majority and minority scores, a statistical outcome is that a greater proportion of minority candidates are misclassified as unlikely to succeed.

On the whole, scientific, medical, and technical progress goes hand in hand with the goal of diversity. Taking that goal seriously strengthens the nation's economic, academic, and social well-being.

Bruce Alberts
President, National Academy of Sciences

Wm A. Wulf
President, National Academy of Engineering

Harvey V. Fineberg
President, Institute of Medicine