Spring 1998, Volume XIII Number 2
Affirmative Action: A Dialogue on Race, Gender, Equality and Law in America
by John Michael Eden
Merit is a notion that is rarely subjected to scrutiny in the context of the affirmative action debate. While some proponents hold that affirmative action is carried out in the service of diversity, others claim that it constitutes an effective way to ameliorate the effects of oppressive historical conditions. Conservatives have criticized both approaches, contending that the first rationale fails because genuine diversity is not necessarily achieved by implementing racial parity, and that the second fails because the central tenets of individualism are violated in the process. Yet, participants on both sides of this debate have attempted to defend or attack affirmative action without reevaluating the prevailing conception of merit, leaving our flawed admissions practices unscathed and unchallenged. It is time to reevaluate what we mean when we invoke "merit" to legitimate our admissions decisions, precisely because by doing so new light can be shed on the affirmative action controversy.
Critics claim that affirmative action allocates places in universities and professional schools to under qualified or less qualified applicants, a contention which they defend by citing the lower test scores of admitted minority applicants. Grade point averages, even conservatives would admit, are a function of the difficulty of one's major, the standards of one's professor, the general quality of the academic institution, and other related factors, which implies that admitting students with lower GPAs is not necessarily a sign for alarm. Low standardized test scores, however, arouse serious suspicion because it is thought that they are objective measures of a student's academic "potential." Although this practice is challenged by studies which show that the fulfillment of potential is not predicted by test scores, the significance of these findings has been ignored or minimized.
Social scientists have challenged the legitimacy of standardized tests, pointing out that statistics fail to corroborate the supposed relationship between high standardized scores and strong academic performance. In a recent study by Wendy Williams and Robert J. Sternberg, of Cornell and Yale Universities, respectively, scores from the Graduate Record Examination (GRE) failed to consistently predict long-term performance in graduate school (see Sternberg & Williams, American Psychologist, June 1997), including dissertation quality, faculty evaluations of graduate student work, teaching ability, and research skills. Although the GRE weakly predicts first-year grades, it does not predict second, third, or fourth year grades, nor does it illuminate who will produce high quality scholarship. The only statistically significant relationship Williams and Sternberg found was one between the GRE "Analytical" subsection and dissertation quality, a relationship which was relatively weak and had validity for men only. Given the GRE's narrow band of applicability, its continued use by so many universities to discriminate between candidates seems suspect.
The problems inherent to the GRE are also symptomatic of standardized tests in general. For instance, when university deans claim that they want the "best students," this often means that they desire to enroll those students who have impressive standardized test scores. Enrolling students with such scores bolsters the prestige of a university because these figures are published in widely read, well-respected national publications. However, those versed in statistics understand the problem with this admissions philosophy: If only students with high predictors (GPAs, standardized test scores) are admitted into "good" universities and colleges, then the statistical claim that such predictors are legitimate is suspect because only a small portion of the pool can have its "predictors" correlated with academic performance. Admitting students in this fashion begs the question of the whether the predictor is significantly correlated with performance because this hypothesis cannot be disproven.
There are simply too few students admitted with poor scores to nullify the sentiment, prevalent in the minds of admissions officers, that test scores are indicative of high academic potential. For those applying to medical school and law school, this reality is especially pernicious because test scores play a pivotal role in the admissions process. Although medical schools have attempted to quell this tendency by placing more emphasis on interpersonal skills in the interview process, the fact remains that a low score on the mcat can rarely be overcome by good, or even superior, grades and portfolios. Moreover, the Law School Admissions Test (LSAT) has had such an unwarranted amount of impact on potential law students, including minority applicants, that the United States Department of Education announced on June 15, 1997 that it will undertake an investigation to determine whether the LSAT's current use violates the Civil Rights Act of 1964.
To the extent that affirmative action's critics base their objections upon the idea that students of color with insufficient merit are being admitted to academic programs, their arguments have much less weight than is commonly supposed. If it is true that standardized test scores are not correlated with relevant aspects of academic performance, then it is an illusion to think that we currently admit the most deserving candidates, including white candidates. Anecdotally speaking, I know of too many lazy or mediocre white students who have compensated for uninspiring transcripts, poor community service records, and less than compelling educational experiences by obtaining high test scores, thus gaining admission into prestigious and well-regarded graduate and professional programs. Statistical data on standardized testing further confirms this, which strongly suggests that any discussion of affirmative action that takes the common notion of "merit" as its departure point is likely to be misguided (see Crouse & Trushheim, 1998, Public Interest, 93 [SAT]; see Clark & Centra, 1982, GRE Board Research Report No. 76-2r; ets Research Report 82-18 [GRE]; and see Gough & Hall, 1975, Research in Higher Education-3 [MCAT]).
Admitting someone with a "substandard" test score cannot be construed, prima facie, as admitting someone with less merit. But if standardized tests should not be considered in admissions decisions, what will replace them? Dr. Williams of Cornell is developing a replacement test for the GRE which is designed to measure the skills necessary for performance in her particular field of study, psychology. Prospective students are given a list of the six attributes of a cogent critique, and then asked to analyze a research article accordingly, a task which more closely approximates the nature of their graduate work. Whatever objections might be leveled at this approach, it does have parallels in current admissions practices. Some top-rated doctoral programs in philosophy, for instance, do not require gre scores and instead rely on student portfolios, consisting of letters of recommendation, measures of previous academic performance (grades/ distinction in one's field), and a substantial writing sample. This mode of evaluation recognizes that the ability to write with a high degree of precision is patently more germane to one's probable performance in a graduate program in philosophy than scores from the GRE. In that connection, one student now at a top-five law school suggested that prospective law students be required to submit writing samples, because a disturbingly high percentage of her classmates cannot write coherent prose.
Her solution may be ultimately beneficial, especially when one considers the aggregate costs of allowing a significant number of poor writers to enter the legal profession. But more importantly, her observation underscores the need to move away from grounding the affirmative action debate upon the results of standardized testing. This does not imply, however, that we should exclusively examine the GPAs, writing skills, and academic proficiency of applicants when making admissions decisions, because this could be misconstrued as a novel variant of "colorblind" admissions, which could result in reinforcing already entrenched patterns of exclusion for minority students. In other words, finding the most defensible admissions calculus will not dispense with the need for enlightened discourse on affirmative action. As Glenn Loury notes in the Dialogue on Affirmative Action, merit is not something that we choose or are born with. It is the product of social forces and economic resources, and as such, we, as a society that is concerned about treating citizens equitably, are compelled to take these realities into account when admitting students into various academic programs. Thus, merit becomes something that cannot be evaluated as if it evolved in a vacuum, but must be considered in the context of the cultural, economic, and social history from which it developed.
Affirmative action remains a desirable public policy, but its implementation needs fine tuning. Those who have faith in standardized tests are comfortable with the current status of the affirmative action debate because they overvalue efficiency, a position which will prove to be problematic, and perhaps even deleterious, in the end. Can we really afford to produce lawyers who cannot write or who lack a commitment to serving people? Or doctors who have poor interpersonal skills? Though this list is not exhaustive of the pitfalls we should avoid, I think the answer to these questions, and many similar ones, is clearly no. Similarly, we cannot afford to exacerbate the profound social and economic conditions which have disparately affected African-Americans, Mexican Americans, Native Americans, as well as others. On the other hand, the prospect of reconceiving affirmative action in a way that takes economic status more seriously seems patently desirable. As Richard Kahlenberg has pointed out (see pg. 14), those of all ethnic and cultural backgrounds, including whites, who are without the financial resources to prepare themselves adequately for an increasingly complex world should be considered disadvantaged for purposes of affirmative action. This goal is not necessarily in tension with increasing diversity and enrolling students from historically underrepresented groups in our undergraduate, graduate, and professional schools. To the contrary, by treating socioeconomic status seriously, we may be able to diversify our student bodies in a way hitherto inconceivable.
John Michael Eden was the Editor of Focus & Program Assistant for the American Bar Association Division for Public Education. In September of 1998, he began the doctoral program in Philosophy at Stanford University.
Spring 1998 Issue Home | Why Race Matters
Affirmative Action as Social and Legal Policy
Affirmative Action, Diversity and College Admissions
Gender, Race, and Affirmative Action
Reconceiving Merit | Affirmative Action in the Workplace
Constitutional Status of Affirmative Action
Book Recommendations | Contributors
Focus on Law Studies Home | Subscribe to Focus | Questions/Ordering Back Issues