This article uses a variety of data sources to examine the extent to which economic disadvantagement explains race bias in admissions and college attendance patterns, then addresses the possibility of reaching racial diversity through economicallybased affirmative action. The first section presents evidence to support reconsidering the validity of admissions measures. Some admissions measures will be shown to exhibit racial bias and race-blind policies do not mean that racially biased measures can be used to determine college admission. The underlying cause of bias will be shown to be economic disadvantage. Second, in spite of phenomenal growth in financial aid programs, statewide attendance patterns will show that financial barriers to attendance exist even among public colleges. Students do not necessarily attend the most selective school that would accept them and their attendance suggests that economic resources play heavily in the decision. The third section reports the results of attempts to produce racial diversity by adjustments to offset students' social and economic circumstances. The technique produced a more racially diverse student body than would be the case using economically neutral race-- blind policies.
The development, evolution, and current status of racial affirmative action programs will not be reviewed for two reasons. First, there are several excellent reviews available (Heffernan and Bazluke 1996; Preer 1981; Friedl 1999). Second, racial affirmative action policies have been fairly ineffective. After more than 20 years of affirmative action, minority graduates have declined as a percentage of the minority population in general (Astone and Nunez-Wormack 1990). From 1976 to 1988 the number of 18- to 24-year-old African-Americans increased by nearly 8 percent; the proportion going to college decreased by about 5 percent. So while the number attending college increased over this period, the number not attending increased even more (Astone and Nunez-Wormack 1990).
Why has affirmative action been ineffective? One reason may be that it became myopically focused on superficial student characteristics and forgot that its original mission was to help people overcome the economic circumstances that were the legacy of racial discrimination and legal segregation. It was the consequence of these circumstances that affirmative action was to offset, but evaluating success became a strictly enumerative determination of racial composition. Maybe affirmative action would be attacked with less zeal if it focused on economic disadvantagement and worked to alleviate barriers to attendance faced by all poor students, whether minority or not (Fuller and McNamara 1978).
Would an economically-based affirmative action policy yield a racially diverse campus? Alexander Astin (1978) struggled with the question over two decades ago. Astin tried to create admissions policies that would yield the proportional racial distributions then required by law without directly considering race. His effort was founded on the principle that public support for special minority programs was largely support for helping people overcome the social, economic, and educational handicaps of discrimination. He noted that public resistance increased as special policies emphasized race rather than social and economic disadvantage. Using a disadvantagement index computed from the sum of standardized scores on parental education and income, Astin compared the minority composition of eight alternative admissions strategies. Those strategies varied by selection-ratio used, measures included, and weights assigned. When applied to the applicant pool, models that incorporated test scores, either singularly or in combination with other measures, produced the least representative freshman class. Class grades were less of a problem and a strategy that equally weighted grades and disadvantagement yielded nearly proportional representation. In sum, Astin was able to produce accepted applicant groups of nearly any racial composition by varying the admissions measures and weights assigned. …