Leveling the Playing Field
Litigation involving the statistics of higher education is once again before the Supreme Court in the current term, as Fisher v. University of Texas is set to be heard by the court. Whether the court still considers diversity of the student body to be a compelling interest that permits the use of race in any way—or whether there is any such compelling interest—is at stake, with statistics poised to play a role.
The History of the Matter
Brown v. Board of Education (1954), which relied on social science evidence to find that separate can never be equal in public-school education, was seminal in opening up access by minorities to higher education as well, but 24 years later, Board of Regents of the University of California v. Bakke (1978) marked the beginning of the understanding that achieving equality cannot be as simple as basing decisions solely on race.
In Bakke, the court said that diversity in higher education was an objective that met the constitutional criterion of “compelling” government interest that would permit the use of race in the accomplishment of the objective, but subject to scrutiny showing that no other less-discriminatory tactic might be employed to achieve the same result.
In particular, quotas—such as the practice examined in Bakke of reserving 16 places in the admitted law-school class for minorities—were impermissible, while the use of race as one factor in an admission process might survive constitutional scrutiny.
Affirmative Action Over Time
In Wygant v. Jackson Board of Education (1986), the Supreme Court expanded upon guidelines for affirmative action in the context of the employment of teachers in public schools:
Not only must there be a compelling government interest but the means to achieve its purpose must be not just reasonable but narrowly tailored under the strict scrutiny standard.
Neither the amorphous existence of societal discrimination nor an interest in providing minority teachers as role models to attempt to alleviate its effects is sufficient justification for race discrimination absent convincing evidence of prior discrimination.
The burden on non-minority actors must be as minimal as possible; that is, there can be no less intrusive means to accomplish the purpose.
With the focus on “narrowly tailored,” higher-education institutions adopted a variety of admission procedures through the years, but the first successful attack on affirmative action in higher education came in the Fifth Circuit Court of Appeal’s decision in Hopwood v. Texas (1996). The University of Texas law school had used a composite index of undergraduate GPA and LSAT scores to place candidates for admission in three categories of admission (“presumptive admit,” “presumptive deny,” and “discretionary”). Although the law school also considered what qualities each applicant could bring and the strength of the student’s undergraduate education and major, background, life experiences, and outlook, the cutoff scores for placement in the three categories were racially based.
That the stated purpose of the admission plan was said to be the “aspiration” that the admitted class reflect roughly the percentage of minorities among graduates of Texas colleges raised further suggestions of unacceptable quotas; moreover, the applications were color-coded by race. This was determined to be permissible by the district court, but the appellate court concluded that the law school “may not use race as a factor in deciding which applicants to admit to achieve a diverse student body, to combat the perceived effects of a hostile environment at the law school, to alleviate the law school’s poor reputation in the minority community, or to eliminate any present effects of past discrimination by actors other than the law school.”
When the Supreme Court denied certiorari to Hopwood in part on the grounds that, since the specific procedures in question were no longer in effect, there was no case or controversy before the count, the explicit use of race in admissions became untenable throughout the states in the Fifth Circuit. Through testing the outcome of various techniques, the use of statistics became key to judicial decision-making by determining how diversity in terms of increasing the participation of underrepresented minorities can be achieved in a constitutionally acceptable way.
The effective overturn of the Hopwood mandate came in Grutter v. Bollinger (20003), which held that the constitution’s equal protection clause did not prohibit the University of Michigan law school’s narrowly tailored use of race in admissions decisions to further the compelling interest in obtaining the educational benefits that flow from a diverse student body. Reasoning that, in the school’s review of each applicant where, unlike in Hopwood, no admission decision was based automatically on the variable of race, Justice Sandra Day O’Connor wrote, “In the context of its individualized inquiry into the possible diversity contributions of all applicants, the Law School’s race-conscious admissions program does not unduly harm nonminority applicants.”
In particular, the university had justified its admission plan on the grounds of the “critical mass” needed to ensure that minority students did not feel isolated or like spokespersons for their race and to provide adequate opportunities for the type of interaction upon which the educational benefits of diversity depend.
Peace in the affirmative-action war did not last. While under the mandate of the Hopwood ruling, the University of Texas at Austin had established a procedure conferring admission to the state’s public universities to the top 10% of graduates of the state’s high schools. After Grutter, finding that this did not appear to result in the desired diversity, the university augmented it with a discretionary program designed to achieve a broad diversity of backgrounds and talents, using race as one of the factors to be considered.
This was challenged by Abigail Fisher, who was not in the top 10% of her high-school graduates and was not admitted to the university through the discretionary process. After decisions in the district and circuit courts supported the university, the case was granted certiorari by the Supreme Court in fall 2012. Petitioner, respondent, and hosts of amici filed briefs replete with the results of competing studies of the effects of affirmative action. This is the case that is bringing the issue back to the court in the current session.
Voluminous studies supported the university’s contention that a “critical mass” is required for effective participation by minorities to achieve the benefits of diversity for all students; statistics produced by the university showed that the numbers of minority students enrolled as a result of the 10% rule were insufficient to meet the goal. Moreover, to counter the charges leveled against the use of race, it was shown that other factors—such as outstanding talent in a variety of fields—were considered in aid of acquiring a diverse student body; in fact, a plan permitting the university president to intervene in admissions was officially approved by the regents after several years of controversy over the practice.
Not only challenged was whether the 10% race-neutral procedure produced sufficient diversity, but whether diversity was a desirable educational objective.
Fisher supporters, although unable to construct a race-free plan that would accomplish the diversity goal, also alleged that the university’s plan admitted underqualified minorities for whom the diversity experience would have an adverse effect, based largely on a single study at one law school. The university and some amici countered with a collection of studies and meta-studies showing the contrary.
Although Fisher did not specifically urge that Grutter be overturned, some conservative justices, in oral argument at the Supreme Court, appeared skeptical of the use of race in any way in admission decisions. Instead, a broad majority of the court ruled that affirmative action in aid of diversity must be strictly reviewed, but it did not outlaw such programs (Fisher v. University of Texas, 2013). Reiterated was that the use of race must be narrowly tailored to meet the “strict scrutiny” test and, moreover, that courts must confirm that the use of race is necessary in the sense that there is no other realistic alternative that does not use race to create a diverse student body.
Because the lower court had not done so, the Supreme Court sent the case back to determine whether the university could make this showing. In her dissent to the remand, Justice Ruth Ginsburg pointed out that even the alternative suggested by Fisher—the 10% rule accompanied by an individualized review—would result in race being taken into consideration, if only covertly. As she declared, “[O]nly an ostrich could regard the supposedly race neutral alternatives as race unconscious.”
On remand, the Fifth Circuit concluded in July 2014 that the university’s plan was sufficiently narrowly tailored to withstand the required strict scrutiny. It is this decision that the Supreme Court has granted certiorari, with the case to be heard in the current term. Presumably, once again there will be arguments, statistical and otherwise, for and against the plan currently in effect, and alternatives offered to demonstrate that it is not sufficiently narrowly tailored to withstand constitutional strict scrutiny on one hand and that no less-discriminatory plan can accomplish the stated objectives on the other. But also at stake this time could be whether the court still considers diversity of the student body to be a compelling interest that permits the use of race in any way—or, indeed, whether there is any such compelling interest. In this determination, statistics will no doubt play a role.
Affirmative action is often characterized as “leveling the playing field”; that is, providing equality of opportunity while giving special consideration to those disadvantaged in some way. The assumption is that more is needed once overt discrimination has been overcome as a result of the heritage of the Brown decision of 60 years ago and a series of national and state laws forbidding discrimination on the basis of race, sex, color, ethnic origin, religion, age, and disability. However, not all discrimination is subject to the same standard of review.
The Supreme Court has not applied the same standard to gender equality as to racial equality—thus, leading recently to an incipient push to revitalize the Equal Rights Amendment that passed both houses of Congress in 1972, having been introduced for the first time in 1923, but fell short of ratification by two-thirds of the states before the 1979 deadline.
An equal rights amendment to the constitution would, presumably, elevate the level of scrutiny accorded gender discrimination. There has been some progress under the existing U.S. Constitution, justification for discrimination on the basis of gender requiring more than a rational purpose but less than the compelling interest that leads to strict scrutiny. Men gained admission to the previously all-female University of Mississippi school of nursing (University of Mississippi v. Hogan, 1982) and the Virginia Military Institute was forced to admit women when the alternative “leadership” plan drawn up for women was determined to fall short (United States v. Virginia, 1996).
What has been the genesis of equal educational opportunity for women is Title IX of the Education Act of 1972, forbidding gender discrimination in any program receiving federal funds (and, as amended, in all programs in an institution that received any federal funding). Although historically single-gender institutions were exempt, the number of single-sex schools has reached near the vanishing point and medical schools and law schools are by now close to gender equality in enrollments, science, technology, engineering, and math (STEM) fields have been more intractable at all levels, including faculty. Indeed, the most significant area of litigation has involved sports, with this year’s World Cup success by the U.S. women often cited as a triumph of Title IX. On the other hand, in the recently published The Only Woman in the Room (Pollack 2015), the author laments that all too little has changed since she was a physics major at Yale 40 years ago.
The statistics that have been dispositive in gender discrimination in higher education are frequently very simple. In the landmark First Circuit decision (Cohen v. Brown University, 1996), a simple comparison of the percentage of women among Brown students (51%) to that of women among athletes (39%) was telling, even without resorting to noting a p-value of less than .001. Offered unsuccessfully by Brown as a mitigating factor was a claimed lack of interest by women in sports. Also unsuccessful as an explanation of disparate spending on men’s and women’s basketball at Temple University was that, on road trips, the men were placed two to a room and the women three to a room since they were smaller (Haffer v. Temple University, 1982).
A rare example of a successful challenge involving lack of STEM diversity occurred in a Philadelphia case, under the gender-equity provision of the Pennsylvania state constitution, where, among other factors, lack of equity was shown by the fact that Central High (male) offered more math than did Girls High, with the availability of Italian in the latter considered by the school district, but not by the court, to be an adequate substitute. Earlier, an equally divided U.S. Supreme Court failed to overturn a Third Circuit decision finding that the Central/Girls pairing withstood scrutiny under the United States Constitution (Vorchheimer v. School District of Philadelphia, 1976).
Another area of higher education that has seen extensive use of statistics, although not particularly successfully, is gender discrimination in employment—illegal under both Title IX and Title VII, which bars employment discrimination on the basis of race, sex, color, national origin, and religion. Differences in the gender composition of faculty have been explained by differences in experience and field of specialization and, less convincingly due to measurement difficulties, productivity; there probably are few colleges and universities that have themselves or through litigation undergone generally ineffective regression studies of faculty salaries.
Here, the complexity of deciding what work is equal and what qualifications are equal might lead one to believe that the field is ripe for statistical analysis. An early (1977) kit developed by Elizabeth Scott has been widely used to detect gender disparities in pay, but the recommended statistically based remedies (Scott & Gray, 1980) have found less support, perhaps reflecting either academics’ lack of faith in evidence-based decisions affecting themselves or a lack of confidence in or understanding of the techniques.
As statistics have found a role in other areas of litigation, perhaps in the future, a technique will be found to level the playing field for academics as well.
Mary Gray is professor of mathematics and statistics at American University in Washington, DC. Her PhD is from the University of Kansas, and her JD is from Washington College of Law at American. A recipient of the Elizabeth Scott Award from the Committee of Presidents of Statistical Societies, she is currently chair of the American Statistical Association Scientific and Public Affairs Advisory Committee. Her research interests include statistics and the law, economic equity, survey sampling, human rights, education, and the history of mathematics.