Affirmative Action ’ s Fate : Are 20 More Years Enough ?

1 As policy researchers whose work has centered on affirmative action and related policies, with a focus on the links between affirmative action and diversity, equality, and justice, we approach this paper from the assumption that legal affirmative action policy plays a singular role in fostering diversity and equal opportunity, and that abolishing it would have (and has had) negative results for higher education. We thank Lauren Saenz for her able help with background research for the paper, as well as Catherine Horn and Angelo Ancheta for their insightful feedback on earlier drafts. Education Policy Analysis Archives Vol. 17 No. 17 2 And second, given that current status, what is the likelihood that by 2028 affirmative action will no longer be needed to further the goals of diversity and equality on college and university campuses? Our methods include analyses of prominent court decisions and state legislation, as well as analyses of enrollment data from the Integrated Postsecondary Education Data System to examine the racial/ethnic diversity of first-time, degree-seeking undergraduates at several state flagship institutions from 1994 to 2005 and compare that to the racial/ethnic diversity of the same undergraduate students attending all 2and 4-year degree granting public and private institutions in the flagship’s state. In states potentially facing anti-affirmative action ballot initiatives, these analyses provide one approach to examining the current need for affirmative action in admissions and the likely consequences if the ballot initiatives pass.

Despite significant legal and political challenges, affirmative action remains a legal method of increasing opportunities for higher education in 45 states. 2   Answering the first question requires an analysis of the major federal court decisions regarding affirmative action in higher education admissions, beginning with Bakke and continuing through Gratz and Grutter: Regents of the University of California v. Bakke (1978), Hopwood v. Texas (1996), Smith v. University of Washington (2000), Johnson v. Board of Regents of the University of Georgia (2001), Gratz v. Bollinger (2003), and Grutter v. Bollinger (2003).
The U.S.Supreme Court has ruled on several important cases since its Regents of the University of California v. Bakke (1978) decision and in those cases has maintained its position on the constitutionality of using race/ethnicity as one qualifying factor in college and university admissions.It has also narrowed the use of race-conscious affirmative action in education, most recently at the K-12 level, but overall the practice remains legal and viable.Bakke has not been overturned by the U.S.Supreme Court even though two different conservative-leaning courts have had the chance to do so.Nevertheless, the debate and even confusion about affirmative action in higher education continues.
The primary aim of this article is to provide an overview of the current status of affirmative action in postsecondary admissions through a concurrent review of recent court rulings, state legislation, and higher education enrollment data.It is a particularly opportune time to do so, coming as it does over 30 years after the landmark Bakke (1978) ruling, six years after the University of Michigan affirmative action lawsuits-Gratz v. Bollinger (2003) and Grutter v. Bollinger (2003)-and less than a year after additional states voted on new anti-affirmative action ballot initiatives.While others have provided analyses of a single component, such as the court rulings (e.g., Greene, 2004;Korrell, 2007), analyzing each of these factors in the context of the others is important because the court decisions and state-level legislation interact to affect enrollment outcomes.From a broad philosophical and legal perspective, then, we ask two questions.First, what is the current legal and political status of race-conscious policy in higher education admissions?Second, given that current status, what is the likelihood that by 2028 affirmative action will no longer be needed to further the goals of diversity and equality on college and university campuses?The 2028 guideline stems from Justice Sandra Day O'Connor's admonition in the Grutter (2003) majority opinion in which she expressed the Court's hope that in 25 years-2028-affirmative action would no longer be necessary to achieve the diversity that enriches college life and learning.For the purposes of this article, we focus on affirmative action in education, specifically in higher education admissions, examining K-12 issues where relevant to the higher education context.
In addition, one must consider state political and legal events: California's Proposition 209 (1996), Washington's Initiative 200 (1998), Florida's Executive Order 99-281, the One Florida Initiative (1999), Michigan's Proposal 2 (2006); and the 2008 ballot initiative campaigns in Arizona, Colorado, Missouri, Nebraska, and Oklahoma.We incorporate research findings and analysis of institutional-level data to understand the outcomes of affirmative action case law and ballot initiatives.In addition, we use secondary sources to examine the political implications of the rulings and legislation related to affirmative action policy.
To answer the second question-what is the likelihood that by 2028 affirmative action will no longer be needed to further the goals of diversity and equality on college and university campuses?-thisarticle uses enrollment data from the Integrated Postsecondary Education Data System (IPEDS) to examine the racial/ethnic diversity of first-time, degree-seeking undergraduates at several state flagship institutions from 1994 to 2005 and compare those patterns to the racial/ethnic diversity of the same first-time, degree-seeking undergraduate students attending all 2and 4-year degree granting public and private institutions in the flagship's state.This comparison allows us to examine the patterns of college and university admission of racially/ethnically diverse classes in several select institutions and consider whether it seems viable that in the remaining 20 years such institutions will no longer need affirmative action in admissions decisions.In particular, in the states potentially facing anti-affirmative action ballot initiatives in coming years, these analyses provide one approach to examining the current need for affirmative action in admissions and the likely consequences if the ballot initiatives pass.Finally, we conclude with reflections on what the evolving legal and sociopolitical climate mean for education research, policy, and practice concerned with diversity and equality of educational opportunity.Given the recent changes in affirmative action law, our analyses offer an understanding of affirmative action policies today and for the immediate future that will have both theoretical and practical uses for educators, researchers, administrators, and policy makers.Sandel (1991) has written that affirmative action causes a "conflicted public mind" (p.13) due to the simultaneous desire for racial equality and color-blind policies.The public mind has become no less conflicted in the last dozen or so years as affirmative action policy has been tested and contested in state and federal courts, and through ballot initiatives that have passed in four states with the latest just last year.In addition to the initiative passed in Nebraska in Fall 2008, initiatives were attempted in four other states.Of these four initiatives, three ultimately failed prior to reaching the ballot (in Arizona, Missouri, and Oklahoma), and the fourth was defeated by voters in Colorado, as discussed later.With all of the legal and political activity surrounding affirmative action, it can be difficult to tease out the race-conscious policies and programs that are currently allowable by law and, beyond that, which ones will be able to withstand what are sure to be future challenges.This section attempts to clarify the current landscape of affirmative action policy related to admissions to colleges and universities.

From the Court of Public Opinion to State and Federal Courts
Prior to the 1978 U.S.Supreme Court case Regents of the University of California v. Bakke, the national mood had been tilting in favor of policies and programs designed to support equality of educational opportunity (Gill, 1980).Great Society and War on Poverty programs were indicative of this mood.In the 1970s, however, the U.S. witnessed a change in the desire to address social inequalities that has been characterized as "a spreading mania within American society, a mania increasingly adamant against governmental and societal efforts to help blacks, other minorities and the poor" (Gill, 1980, p. 1).For over 30 years, affirmative action programs have weathered a persistent backlash, combined with legal and political challenges (Moses, 2002).These challenges have taken the form of court cases, state-level legislation, and state ballot measures aimed at curbing or eliminating affirmative action.The landmark Bakke case was the first in which U.S. Supreme Court Justices ruled on affirmative action in higher education admissions.

R eg ents of the Univers ity of Ca lifornia v. B a k k e
Bakke's legacy in affirmative action law is strong (Marin & Horn, 2008).Perhaps its most lasting effect was to clarify that the use of numeric quotas and set-aside places within admission programs seeking to promote diversity in higher education is forbidden because it violates the 14th Amendment to the U.S. Constitution.Even though the legal legacy of the Bakke case is significant, the decision itself was the result of a fractured Supreme Court.The Justices held 4-1-4 that "(a) the minority-admissions program of the University of California Medical School in Davis had discriminated illegally against a white male applicant, but (b) that universities could legally consider race as a factor in admissions" (Sobel, 1980, p. 145).Justices Warren Burger, John Paul Stevens, Potter Stewart, and William Rehnquist decided in favor of Allan Bakke on both counts; Justices William Brennan, Byron White, Thurgood Marshall, and Harry Blackmun decided in favor of the University of California at Davis Medical School on both counts, and, in the swing vote, Justice Lewis Powell decided against the Davis policy, but in favor of universities' ability to use race as a plus factor in admissions decisions.Although Powell was the only justice to use the educational benefits of diversity as his rationale in favor of race-conscious admissions policies, his became the court's controlling opinion since he cast the deciding vote for each side.Divided though it was, the Supreme Court indicated to the nation that affirmative action programs were constitutional and could be implemented legally.
The dispute over the nature of affirmative action and its consideration of race/ethnicity did not end with the Bakke ruling.The next important court case regarding higher education admissions was decided in 1996 with Hopwood v. Texas.

H opwood v. T e xa s
In deciding Hopwood, the Court of Appeals for the 5th Circuit ruled against race-conscious affirmative action policies in higher education admissions, thus nullifying the U.S. Supreme Court's Bakke ruling in the three states in the 5th Circuit: Texas, Louisiana, and Mississippi. 55 Much legal debate ensued after the 5th Circuit issued this opinion.Many legal scholars argued that the 5th Circuit did not have the authority to overrule the U.S. Supreme Court (Torres, 2003).
The White, female plaintiff in the case, Cheryl Hopwood, argued that she had been discriminated against by the University of Texas Law School's admissions system.Hopwood's three-judge panel prohibited the use of race-conscious admissions criteria to achieve diversity at the law school.Their decision concluded that a state's interest in acquiring a diverse student body was not legally compelling enough to justify an admissions program like the one at the law school.However, even though the 5th Circuit struck down the diversity rationale for affirmative action, it maintained that a remedial justification still could serve a compelling interest.Race/ethnicity could be used in admissions decisions only when colleges and universities were trying to remedy the present effects of past institutional discrimination (Greve, 1999).The University of Texas appealed the case to the U.S. Supreme Court.Because the Justices declined to review the case, the ruling was upheld, but only in the 5th Circuit.Hopwood was the first successful challenge to an affirmative action admissions program since Bakke.One year later, then-Texas Attorney General Dan Morales offered clarification on the Hopwood decision for the state, maintaining that its reach extended to programs outside of admissions, including financial aid, recruitment, and scholarships.Subsequent research on the impact of the decision concluded that Hopwood had a chilling effect on college access for Black and Hispanic 6At the flagship University of Texas at Austin (UT Austin), officials put a great deal of effort into implementing various policies and programs to mitigate the effect of losing affirmative action.In addition, the Texas legislature passed House Bill 508-the Top Ten Percent Plan-guaranteeing "admission to the top 10% of a high school graduating class to any public higher education institution in the state" (Marin & Flores,p. 226).high school graduates in Texas (Dickson, 2004;Kain & O'Brien, 2001). 7However, once the U.S. Supreme Court issued its 2003 rulings effectively overturning Hopwood, then-President Larry Faulkner indicated that the university would work with the Texas legislature to resume affirmative action policies (University of Texas at Austin, 2003a).Having had the experience of operating without affirmative action, President Faulkner was eager to re-institute the policy.Dr. Bruce Walker, vice provost and director of admissions at UT Austin said at the time, "We have used race-neutral policies for seven years and still do not have a critical mass of African American or Hispanic students in our classrooms" (University of Texas at Austin, 2003b).Currently UT Austin continues to use both race-conscious admissions policies as well as the Top Ten Percent Plan (Chapa & Horn, 2007). 8  Sm ith v. Univers ity of Wa s hing ton In Smith v. University of Washington (2000), three White applicants who were not accepted to the University of Washington Law School sued the university.Even though Initiative 200 (passed in 1998) required the university to abandon the admissions program under question, the case moved forward.A symbolic victory came for affirmative action supporters in December 2000 when the 9th Circuit Court of Appeals ruled that the Law School's affirmative action program was constitutional.In 2001, the U.S. Supreme Court let stand the lower court's ruling in Smith (Gose & Schmidt, 2001).
Johns on v. B oa rd of R e g e nts of the U nivers ity of Ge org ia Another important affirmative action case is Johnson v. Board of Regents of the University of Georgia (2001).At the time of the ruling, 6% of the University of Georgia's students were African American, in a state in which African Americans made up 25% of the state population (Walsh, 2001).In 2001, the three-judge panel of the 11th Circuit Court of Appeals 9 Gra tz v. B olling e r and Gru tte r v. B olli ng e r upheld a district court ruling in favor of three White women against University of Georgia's affirmative action plan (Firestone, 2001).The University was using a point-based system that automatically awarded "bonus" points to non-White and male applicants rather than conducting an individualized review of each application, which the 11th Circuit Court of Appeals found to be unconstitutional.The University of Georgia decided not to appeal the ruling.Johnson struck down the University's pointbased system that awarded a fixed amount of points for students of color as well as for nearly a dozen other factors such as first-generation college student status (Walsh, 2001); this seemed to pave the way for the U.S. Supreme Court's decisions in the University of Michigan cases.
Soon after Johnson, the U.S. Supreme Court finally agreed to hear a case about affirmative action in higher education, or rather two: Gratz v. Bollinger (2003) and Grutter v. Bollinger (2003).
The plaintiffs in Gratz and Grutter were White applicants who felt that they would have been admitted to the University of Michigan had it not been for the consideration of race/ethnicity in the institution's admissions decisions; Jennifer Gratz sued regarding the undergraduate admissions policy and Barbara Grutter regarding the law school admissions policy.Overall, the June 2003 decisions in Gratz and Grutter reaffirmed the Court's ruling in Bakke, upheld the constitutionality of using race/ethnicity as a plus factor in higher education admissions decisions, and emphasized the importance of individualized, holistic reviews of applications.In Gratz the Justices struck down the University of Michigan's race-conscious undergraduate admissions program and made clear that any type of quota or numerical point system that automatically awards points to minority applicants does not fall under the permissible standards regarding the use of race/ethnicity in admissions decisions.In its Grutter ruling the Court affirmed that the educational benefits flowing from a diverse student body served a compelling state interest.The diversity rationale was the central justification in upholding the constitutionality of affirmative action.Writing for the majority in Grutter, Justice O'Connor explained: "The Law School's educational judgment that such diversity is essential to its educational mission is one to which we defer.The Law School's assessment that diversity will, in fact, yield educational benefits is substantiated by respondents and their amici" (Grutter, 2003, p. 328).In addition, the Grutter decision highlighted Justice O'Connor's idea that affirmative action should no longer be necessary in 25 years. 10Ultimately, the ruling in Grutter invalidated the 5th Circuit's ruling in Hopwood.The Grutter decision thus underscored the importance and legal viability of the diversity rationale for affirmative action in college and university admissions.This justification seems to have wider appeal than the remedial justification, as even those who oppose affirmative action sometimes support the idea of diversity (see e.g., Bush, 2003; Deardorff & Jones,  2007 11   9 The 11th Circuit includes Alabama, Florida, and Georgia. 10In O'Connor's words: "We expect that 25 years from now, the use of racial preferences will no longer be necessary to further the interest approved today" (Grutter, 2003, p. 343).
11 Deardorff & Jones' (2007) survey of southern and midwestern colleges showed that among the southern schools, although administrators generally did not support or agree with the University of Michigan decisions, they all agreed that race plays a significant role in our society, and administrators at all schools reported that diversity was important.
).In fact, even the plaintiffs in Gratz v. Bollinger did not contest the importance of diversity to higher education.In addition, the Grutter court emphasized that institutions should engage in holistic review of applicants, within which they consider both quantitative (e.g., high school grade point average) and qualitative (e.g., extracurricular activities) assessments of the applicant's qualifications for admission.
Pa rents Involved in Com m u nity Schools v. Sea ttle School Dis trict N o. 1, et a l.
Extending outside of higher education are several court cases involving K-12 race-conscious student assignment plans in public schools.Although these cases are important primarily for the Parents Involved in Community Schools v. Seattle School District No. 1,et al. is  The Court had changed significantly in its composition since the 2003 University of Michigan cases.Chief Justice William Rehnquist had died and been replaced by similarly conservative Chief Justice John Roberts, who was appointed by President George W. Bush.Perhaps more significantly, Justice Sandra Day O'Connor retired, and her position was filled by Justice Samuel Alito.Justice O'Connor had been known as a moderate within a divided court and was often the swing vote in contentious cases such as Grutter.By contrast, Justice Alito was expected to align with his conservative colleagues on the high court.It was not surprising, then, when the Supreme Court ruled that voluntary racial integration plans in place in school districts in Seattle and Louisville not narrowly tailored and thus were unconstitutional (Korrell, 2007).However, it left Grutter standing and supported the idea that diversity is a compelling interest in higher education.Nevertheless, efforts at the state-level have been undermining and continue to undermine Grutter.It is to that issue that the article now turns.
The state shall not discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity, or national origin in the operation of public employment, public education, or public contracting.
The above language formed the primary text of state ballot initiatives seeking to curb the use of affirmative action in public institutions in California (1996), Washington (1998), Michigan (2006), andNebraska (2008).The initiative passed in all four states. 13imilar ballot initiatives recently were proposed in four additional states: Arizona, Colorado, Missouri, and Oklahoma.In Arizona, Missouri, and Oklahoma the initiatives did not make it onto the 2008 ballot (Blank, 2008;Hoberock, 2008).Perhaps the most interesting development was in Colorado where the initiative was defeated by a narrow margin.Those behind these campaigns, however, intend to pursue similar ballot initiative campaigns in these and other states in the coming years (Bassett, 2008).Regardless of the official standing of these campaigns, examination of the status of the racial representation in the state systems of higher education is important to understand the context of the battle in these targeted states.
In this section we examine how the initiatives that have passed in California, Washington, and Michigan, as well as the Executive Order banning affirmative action in Florida, have affected or may affect access and equity in higher education.14

California
We then discuss the five states that were targeted by anti-affirmative action campaigns in 2008.
Spurred on by then-University of California (UC) Regent Ward Connerly, in 1995 the Regents of the University of California voted to bar the consideration of race/ethnicity in admissions decisions in the UC system by approving system guideline SP-1. 15California's Proposition 209 soon followed.Known by proponents as the California Civil Rights Initiative (CCRI), Proposition 209 was a ballot initiative for a constitutional amendment to abolish all "preferences" based on race, ethnicity, and sex. 16ven though the Proposition never mentioned affirmative action by name, its effect was to eliminate affirmative action in higher education admissions (as well as in other state programs).Proposition 209 passed with 54% of the vote (Chávez, 1998).Opponents of the amendment challenged its constitutionality in court, but in 1997 the U.S. Supreme Court let stand the ruling by the 9th Circuit Court of Appeals that upheld Proposition 209 (Lederman, 1997).The impact on California's public college student population was felt almost immediately and is visible to this day.In fall 1998, the flagship University of California campus, Berkeley, reported a 52% decrease in the number of Black and Hispanic first-year students for the first class admitted without affirmative action.Because of this, Black and Hispanic students made up only 9.9% of the first-year class, well below the 20.7% of first-year enrollees the previous year (Healy, 1998).At Berkeley's law school, there was only one Black student in the entering class of 1997-1998.In partial response to the negative attention given to the University of California system, in 2001 the Regents voted to rescind their ban on affirmative action (Schevitz, 2001).Because of Proposition 209, however, the Regents' change of heart was entirely symbolic and did little to stem the rollback of students of color in the University of California system, especially the most selective campuses.Contreras (2005) examined the effects of Proposition 209 on college access at three University of California campuses: Los Angeles, Davis, and Riverside.Using parity as a measure of access (a ratio comparing admissions rates to proportional representation in the K-12 system), Contreras found that although this ratio did not change for Asian American and White students, significant declines were experienced by African American, Chicano, and Latino students.That is, she found significant underrepresentation of these groups at all three campuses studied.
Further, consider that UCLA's first-year class in the fall of 1997 had 221 African American students; by fall 2006, it included fewer than half that amount-100 (Leonhardt, 2007, p. 78).
These examples from previous research set the groundwork for our examination of the admissions patterns in the states without affirmative action.Since the state of California has had an affirmative action ban in place the longest, it allows us to examine potential impacts of the loss of race-conscious admissions policies as well as institutional attempts to address such a loss.Examining enrollment data for first-time, degree-seeking undergraduates from 1994 to 2005 for the two most selective UC campuses-Berkeley and Los Angeles-yields trends that corroborate evidence of a clear impact in 1998, the first year the full class was admitted without considering the race/ethnicity of applicants (figures 1 and 2).At the same time we observed significant decreases in UC-Berkeley Black and Hispanic enrollment (noted previously), White and Asian/Pacific Islander (API) student enrollment increased, but by less than 2 percentage points.Interestingly, the proportion of students declining to state their race/ethnicity more than doubled-from 7% in 1997 to 16% in 1998; by 2005 this number was back to 1997 levels.While difficult to say with absolute certainty who made up the "race unknown" group during this critical time, this increase does suggest a negative impact on applicant perceptions of how race/ethnicity is viewed by the institution. 17t UCLA, 1998 was also a critical year, with Hispanic and Black enrollments between 1997 and 1998 decreasing by 30% and 40%, respectively.The enrollment pattern for Hispanic students at UCLA is similar to that at Berkeley.In both cases the decreases in enrollments began prior to 1997, with the sharpest decreases at UCLA beginning in 1995 and the decreases at Berkeley beginning a year later in 1996. 18hen measured from this point, the enrollment share for Hispanic students dropped by 50% prior to the loss of affirmative action.White students showed a loss of fewer than 2 percentage points and the proportion of APIs increased by 1 percentage point.Once again, however, an important part of the story is the increase in the share of those students declining to state their race/ethnicity-at UCLA, in a single year, that group increased by over 2.5 times, from 5.5% to over 14% in 1998.This share was approaching 1994 levels by 2005.
While these enrollment changes provide us with examples of the impact of losing raceconscious affirmative action in higher education, the overall trends are also significant.Although Berkeley and UCLA have made attempts to address the loss of race-conscious admissions policies (Chapa & Horn, 2007), it is clear that their efforts have not yielded significant results.At both institutions, Hispanic and Black enrollments have yet to return to the levels observed prior to the implementation of SP-1 and Proposition 209.This is both in the face of increased programmatic efforts on the part of the institutions and increases in these populations in the state.

Washington
The next state-level challenge to affirmative action policy came in Washington two years after Proposition 209 passed in California.In 1998, 59% of Washington's voters approved Initiative 200 (I-200), a CCRI-like referendum banning the consideration of race and sex in public hiring, contracting, and college and university admissions.In assessing the impact that I-200 had on the college application and enrollment decisions of high school seniors in Washington, Brown and Hirschman (2006) found that in the year following the passage of I-200, there was a significant decrease in the numbers of students of color applying to and enrolling in the University of Washington (the flagship public institution, which bore almost the entire brunt of the effects of the initiative).It is this political context within which White applicants to the University of Washington Law School contested their rejection in the Smith v. University of Washington case (discussed previously).Our analysis suggests similar results and reveals additional patterns.Examining enrollment data for first-time, degree-seeking undergraduates at the University of Washington from 1994 to 2005, we observe interesting changes (figure 3) that are different from those observed at the California flagship institutions, given a similar state-level ban on affirmative action.While the fall of 1999 class was the first admitted under Initiative 200, important enrollment changes occurred as early as 1996.For example, White student enrollment during this period was at its maximum of 68% in 1995 but began to decrease as early as 1996 and has yet to return to the 1995 levels.During approximately the same time, the "race unknown" category witnessed large increases that began in 1997, two years before a race-neutral admissions policy was required by the University of Washington.Interestingly it appears that increases and decreases in the "race unknown" category since 1997 are mirrored by the changes in proportion White enrollment.While it is impossible to know whether the students in the unknown category are White students, these patterns support this possibility, suggesting that White enrollment decreases were not as large as changes in the proportion White might suggest since some White students may have "declined to state." Similar to the patterns exhibited in California, starting in 2000, Asian and Pacific Islander (or API) student enrollment increased after the implementation of I-200 (from 25% to a high of 30%) and Black and Hispanic student shares dropped by over one-third.Black enrollment from 1999 through 2005 stayed at the 1999 levels, and the Hispanic share recovered slightly by 2003 to pre-1999 levels.What is different at the University of Washington are the consistently lower shares overall (and over time) for both Black and Hispanic students compared with the California institutions.

Florida
Florida was the next state to have an anti-affirmative action measure ballot campaign.Connerly's group had begun collecting petition signatures to get such an initiative on the 2000 ballot.However, these plans were derailed in November 1999, when Governor Jeb Bush issued Executive Order 99-281, entitled One Florida.With regard to higher education, this order ended the consideration of race/ethnicity in state college and university admissions.As an alternative, Florida adopted the Talented 20 program, a plan similar to Texas's post-Hopwood Top Ten Percent Plan, ensuring admission of the top 20% of public high school graduates to Florida's state colleges and universities.Like state ballot initiatives that create law or amend state constitutions, Governor Bush's Executive Order is unaffected by the Supreme Court rulings in Gratz and Grutter.However, an important distinction between Florida's executive order and the ballot initiatives passed in other states is that in Florida, although public institutions of higher education cannot consider race/ethnicity in admissions decisions, they still can consider race/ethnicity in non-admissions practices and programs including recruitment, the awarding of scholarships, and outreach programs.As with California, enrollment changes were observed in Florida's flagship universities immediately following the implementation of One Florida (Marin & Lee, 2003).
Examining enrollment data for first-time, degree-seeking undergraduates from 1994 to 2005 at Florida's flagship institution-the University of Florida (UF)-we are afforded another example of the impact of the loss of race-conscious admissions policies.The first class admitted without affirmative action showed enrollment gains for White students and the greatest losses for Black students (figure 4).At this institution we did not observe the same increase to the "race unknown" group upon the loss of affirmative action as we did at UCLA, Berkeley, and the University of Washington.However in 2005 we see a three-fold increase in the "race unknown" category.The reasons for this change are also unknown, but potentially important if this trend continues over time.These examples demonstrate that each state/institution can yield unique results in the face of losing affirmative action.Similar to UCLA and Berkeley, however, Black enrollment at UF has yet to return to the 12% high it saw in 2000 despite the institution's ability to use race/ethnicity in nonadmissions policies and practices (e.g., recruitment, scholarships, etc.) and the implementation of the Talented 20 program in place of race-conscious affirmative action in higher education admissions.Just as troubling as the failure of the recovery of the Black enrollment share to pre-2001 levels is the flat enrollment share of Hispanic students despite a growing share of Hispanic students enrolling in Florida institutions of higher education, up from around 16% in 2001 to near 20% in 2005.

Michigan
The next state vote on affirmative action was in Michigan.The political campaign for Proposal 2, once again spearheaded by Connerly, was announced the day that the U.S. Supreme Court issued its decisions in the University of Michigan cases.On the ballot in 2006, Proposal 2 passed with 58% of the vote.The impact it is having on Michigan's public institutions of higher education will become clearer in time, as they observe application, admission, and enrollment rates.However, both the University of Michigan and Michigan State University reported decreased percentages in freshmen of color for fall 2007 (Baker, 2007).In addition, higher education officials in Michigan are struggling to conduct admissions and still keep diversity in mind.As Wayne State University Law School Dean Frank Wu explained, What do we do if we're serious about racial integration, diversity and the competitiveness of this nation in a global economy?What Prop 2 did was eliminate one method of dealing with these issues, but it doesn't take away the urgency of the issue.(quoted in Erb, 2007, ¶5) Although opponents of Proposal 2 brought a lawsuit challenging its constitutionality, U.S. District Court Judge David Lawson dismissed the lawsuit in March 2008 because, he argued, the plaintiffs did not make their case that the initiative intended to discriminate against people of color (Jaschik, 2008).

Ballot Initiatives in Five Targeted States
Continuing the trend of challenges to affirmative action at the state level, five states debated state ballot initiative proposals for November 2008-Arizona, Colorado, Missouri, Nebraska, and Oklahoma.Some commentators pointed out that the timing of these initiative campaigns was strategic, with the aim of capitalizing on the immigration debates and influencing the outcome of the presidential election (Bello, 2007).Arizona was one of the states targeted by proponents of an antiaffirmative action ballot initiative.Opponents of the initiative believe that Connerly chose Arizona because it has such a large immigrant population and he had hoped the initiative would capitalize on related racial and ethnic divisions (Bello, 2007).For his part, Connerly maintained that racial tensions already existed in Arizona regardless of the ballot initiative campaign (Benson, 2007).Late in August 2008, Proposition 104's supporters withdrew the initiative proposal.Many signatures on the petition to put Proposition 104 on the ballot had been declared invalid by the Arizona Secretary of State, and initiative proponents were unable to validate the signatures before the ballot deadline.Max McPhail, director of the initiative campaign, pledged to return this issue to voters in 2010 (Fischer, 2008).Indeed, the initiative will be on Arizona's November 2010 ballot.
In Colorado, Amendment 46 was defeated by a narrow margin; 50.7% against and 49.2% in favor (Denverpost.com, 2008).During the campaign, sponsors of Amendment 46 collected over 128,000 petition signatures, well over the 76,047 valid signatures needed to get the initiative on the ballot (Gandy, 2008).As in other targeted states, Colorado has relatively few students of color at its state colleges and universities, with about 72% of its 2005 first-time, degree-seeking college enrollment identifying as White. 20This is especially true of the flagship institution of the University of Colorado at Boulder, where in 2005, only 1.4% of first-time, degree-seeking college students were African American, 6.3% Latino, and 6.5% Asian. 21The defeat of Amendment 46 in Colorado marked the first time such an initiative failed to pass at the state level.There are several theories as to why Coloradans voted to defeat Amendment 46, including the Governor's public opposition to the measure, President Obama's strong support in Colorado, the state's large Latino population, the confusing language of the ballot initiative, an unprecedented grassroots effort against it, television and radio advertisements, and a state ballot that included 13 other ballot measures.However, until there are studies of voter beliefs and attitudes about Amendment 46 and affirmative action in Colorado, it is not possible to specify the reasons (Slevin, 2008).While the chilling effect of the initiative campaign itself in Colorado will be difficult to assess fully, one direct result of the vote is that public institutions in Colorado can continue to have race-and sex-conscious equal opportunity programs in public education, employment, and contracting. 22issouri took an unusual approach to challenging the anti-affirmative action ballot initiative proposed there.Secretary of State Robin Carnahan and Attorney General Jay Nixon challenged the language of the proposed initiative.They argued that the language should reflect its actual purpose, which was to ban affirmative action.Had the initiative made it onto the ballot, it is likely that Missouri's ballot initiative language would have been different than the other states.Although the language suggested by Carnahan was challenged in court by the initiative's proponents and the County Circuit Judge struck down Carnahan's language, the language the judge put in place was still more descriptive in mentioning affirmative action than the originally proposed initiative language (Schmidt, 2007).The new ballot initiative language asked voters if the Missouri constitution should be amended to: Ban state and local government affirmative action programs that give preferential treatment in public contracting, employment or education based on race, sex, ethnicity or national origin, unless such programs are necessary to establish or maintain eligibility for federal funding or to comply with a court order.(Lieb, 2008, ¶13) The issue of language is especially important because polls show that the wording of such initiatives affects how members of the public feel about them (Inside Higher Ed, 2008).However, observers did not have a chance to assess whether the different language of the Missouri initiative made a difference for voters in 2008.Although initiative proponents collected almost 170,000 petition signatures by the May 4th deadline-a sufficient number to make the ballot-they did not submit the signatures because it was likely that many of those signatures would not have been validated (Blank, 2008;Darnell, 2008;Franey, 2008).Despite this setback, Connerly was not dissuaded from pursuing this route in Missouri stating that "this is a marathon not a sprint, and it's far from over" (quoted in Blank, 2008, ¶7).In fact, Connerly and Tim Asher, director of the anti-affirmative action ballot initiative campaign in Missouri, are in the midst of efforts to place the initiative on Missouri's ballot in 2010 (St. Louis American, 2008).
In Nebraska an anti-affirmative action initiative campaign led by Doug Tietz was successful in November 2008, passing with 58% of the vote (Gewertz, 2008).The Board of Regents at the state's flagship institution, University of Nebraska, had voted unanimously to oppose the initiative, citing concerns about the university's ability to increase diversity should the initiative pass (Associated Press, 2008).Proponents of affirmative action in Nebraska are worried that now women and people of color will not continue to make progress in public education and hiring in the state.For example, schools and universities may need to adjust both hiring and admissions practices.Yet, overall, only 9% of the enrollment of the University of Nebraska's four campuses are students of color (Gewertz, 2008).
The fifth state that had been involved in the anti-affirmative action initiative process was Oklahoma, led by attorney Devin Resides.Petition signature gatherers worked steadily during fall 2007 in Oklahoma and believed they had collected a sufficient number of valid signatures to get the initiative on the ballot.Oklahoma requires approximately 138,970 signatures, which needed to be filed with the secretary of state by December 10, 2007 (Hoberock, 2008).Initiative proponents submitted 141,184 signatures, but in February, Oklahoma's Secretary of State, Susan Savage, reported that the petition signatures included many duplicates.As a result, proponents withdrew the petition, and Oklahoma did not have an anti-affirmative action initiative on the ballot in 2008 (Schmidt, 2008).

The Effect of Losing Affirmative Action on Student Enrollment
Given the end of affirmative action in another state in 2008-and the targeting of four others that same year-one can now review data from several key states to examine the distinct challenges states may face if they are unable to use affirmative action yet they wish to provide diverse experiences and environments for students attending their flagship institutions of higher education.These institutions and states are important to examine because of their different contexts.Texas was barred from using race/ethnicity under the Hopwood decision, moved to a "race-neutral" percentage plan for admissions, and then resumed using race/ethnicity as a factor (in conjunction with the percentage plan) after the Gratz and Grutter decisions allowed its use.Michigan, the site of the Gratz and Grutter decisions, is now banned from using race/ethnicity after the passage of Proposal 2. Finally, Oklahoma is one of those states that will likely be targeted again for a future proposition.So while their particular circumstances are very different from one another, these states provide a range of examples of bans or potential bans of affirmative action.
The following analysis again considers student diversity at the flagship institutions and speculates on the impact, or potential impact, of losing race-conscious admissions and/or other practices such as outreach and recruitment, financial aid, and other support services.This analysis is similar to the previous examples (California, Washington, and Florida); however, the following includes a direct comparison of changes in each state's college enrollment with changes in the flagship university's enrollment shares, comparing these states' flagship institutional enrollment of first-time, degree-seeking undergraduates by race/ethnicity from 1994 to 2005 to the same category of first-time students enrolled in the public and private 2-and 4-year higher education institutions in the state.The total state 2-and 4-year enrollment is termed the "state-level enrollment." 23hile there are several limitations to this comparison group, it is suitable for these purposes.Those in the reference group have demonstrated they are college-ready and have a strong enough desire to receive some level of postsecondary education that they have enrolled during that year.This is an advantage over the more traditional comparison group of high school graduates in the state since many high school graduates may not have the desire or the basic qualifications to enroll in a postsecondary institution.In addition, the public and private institutions in the state are likely to enroll both in-state and out-of-state students, which is a pool likely to be similar to the flagship institution and unlike the high school pool, which consists entirely of state residents. 24owever, by using this comparison group over the pool of high school graduates, we lose the opportunity to compare the incoming freshman cohorts to the state pool as a whole, and the state pool as a whole may incorporate both potential recruitment opportunities for the flagship institution or shortcomings in preparation by the primary and secondary school systems.In effect, the following analysis compares the enrollment of the flagship public institution in the state to the pool of students who are most likely to be eligible for admission to the flagship, particularly in states where the flagship is nonselective (which is the case in some states that have witnessed ballot initiative campaigns).Thus, any observed changes in the enrollments are only indirectly affected by demographic or educational changes in the secondary school system but are ultimately the direct result of the aggregate admissions decisions of the states' institutions of higher education.

T exa s
The first case is the University of Texas at Austin (UT Austin). Figure 5 shows the enrollment of first-time, degree-seeking undergraduates, by race/ethnicity, at UT Austin from 1994 to 2005.Notice that the overall share of White students at UT Austin peaked in 1997, the year that the Hopwood decision outlawed the use of race/ethnicity in admissions, but the proportion of students who are White has declined steadily since to about 55% in 2005.Hispanic enrollments dropped nearly 2 percentage points in 1997 but have increased steadily to about 18% in 2005.During this same time Asian enrollments increased to a maximum of about 19% in 2001 and subsequently leveled out to shares equal to their Hispanic peers at about 18%.Black student shares dropped to an all time low of just under 3% in 1997, but have nearly recovered to the levels seen in 1994 (about 5%).
However, when one examines enrollments at UT Austin relative to state-level enrollment, a picture emerges that is slightly different from a steady decline of White student shares and a steady increase of Asian and Hispanic shares.Figure 6 plots the percentage of each racial group's statewide enrollment on the horizontal axis and UT's enrollment percentages on the vertical axis.The angled line (marked by the solid line bisecting the figure) describes the point where state-level enrollment is mirrored in the state flagship, henceforth called the equal representation angle.All points above and to the left of the equal representation line correspond to overrepresentation of a particular racial/ethnic group in the state flagship relative to their presence in state-level enrollment, and points below and to the right of the equal representation line corresponds to underrepresentation of a particular racial/ethnic group in the state flagship relative to their presence in state-level enrollment.In addition, within each racial/ethnic group one can fit regression lines fitting the relationship between state and flagship enrollment for that group; those within-group lines appear in Figure 6 with the direction of the arrow on the fitted line representing the enrollment trends (i.e., the arrow head for each line points to the end of the trendline).Fitted regression lines parallel to the equal representation line suggest that the racial/ethnic group's over/underrepresentation is consistent even as their shares in the population and the flagship change, since they are changing similarly in each group.Where the fitted regression line for a group diverges from the equal representation line, that group has over/underrepresentation that is increasing, while a regression line pointing toward the equal representation line has over/underrepresentation that is decreasing.For example, with White enrollment in UT Austin and statewide in Texas, the fitted regression line for White students is above the equal representation line but parallel to it and steadily decreasing.This suggests that White students at UT Austin are consistently overrepresented relative to their share of the higher education enrollment pool despite their falling share overall.The angled line marks the theoretical point of equal representation in the flagship institution and the state's public and private higher education systems overall.Source: IPEDS 1994-2005.
One can contrast this trend to that of Hispanic students at UT Austin, whose fitted regression line is below the equal representation line and diverging from it, suggesting a relatively large and slightly increasing degree of underrepresentation at UT Austin relative to their state-level enrollment.The line for Asian students is equally dramatic with a relatively large overrepresentation that is increasing dramatically as a share of their state-level enrollment.The trend for African American students is not nearly as dramatic, but there is a clear underrepresentation that appears to be holding steady relative to their state-level enrollment.
While Figure 6 describes the nature and direction of the over/underrepresentation of various racial/ethnic groups at UT, the magnitudes of these differences are left imprecise.This can be addressed by the use of another calculation metric-the odds ratio. 2525 Odds ratios are calculated by simply taking the odds of an occurrence for one group expressed by the ratio where pn is the percent enrollment of a particular racial group, and dividing it by the odds of a different group.Thus, an odds ratio of 1 suggests that the odds of a particular event are the same for one racial/ethnic group compared to another.An odds ratio of greater than 1 (1.5 for instance) suggests that one group's odds of an event are 50% higher than the other group's odds.Conversely an odds ratio of less than 1 (0.5 for instance) suggests that the odds of an event for one group are half that of the other group's odds of an event taking place.
This metric allow one to quantify how much over/underrepresentation is experienced by various racial/ethnic groups.The odds ratio for first-time, degree-seeking White student enrollment in UT Austin relative to their presence in the higher education population in Texas as a whole was quite stable, ranging between 1.08 to 1.10 with no clear trend between 1994 and 2005.This suggests that for first-time, degreeseeking students in the flagship institution their odds of being White were between 8% and 10% higher when compared to their odds of being White in the first-time, degree-seeking state-level enrollment pool.For Hispanic students the gap of underrepresentation is quite large-with odds ratios ranging from 0.65 in both 1994 and 2005 to 0.55 in 2000.Thus, during this time period, firsttime, degree-seeking students at UT Austin had odds of being Hispanic between 35% and 45% lower than their counterparts in Texas as a whole.These odds ratios look even worse for Black students whose odds ratios ranged from 0.45 in 1995 to 0.24 in 1997 (the first year of Hopwood) back up to 0.41 in 2005.Thus, during this period, the odds of first-time, degree-seeking students being Black at UT Austin were 55% to 76% lower than the odds of being Black in the first-time, degree-seeking state-level enrollment pool.The story for Asians is far different with their odds ratio ranging from 3.5 to 4.4, which suggests that the odds of being Asian at UT Austin were 250% to 340% higher than their odds in Texas higher education overall.
Figures 5 and 6 and these odds ratios suggest that the underrepresentation of Black and Latino students does not seem to have been a result of the elimination of affirmative action since it existed during the time that race-conscious policies were in place and persisted through the extensive institutional efforts implemented by UT (see Marin & Flores, 2008) to mitigate the impact of Hopwood.This suggests that there may be more at work than simply the loss of race-conscious policies.One possibility may be that the University is simply achieving the outcomes that it intends-the overrepresentation of White and Asian students and an underrepresentation of Black and Latino students-regardless of the available mechanisms.Another possibility is that as affirmative action was taken away as a tool for increasing access, UT Austin deployed all possible resources simply to maintain the status quo.Thus, by focusing attention on the issue of race, the loss of affirmative action forced UT Austin to find other ways to mitigate the potential loss of minority enrollment share.

M ichig a n
The University of Michigan-Ann Arbor (UM) is the state's flagship institution and as with UT Austin, it has demonstrated its commitment to admitting a diverse student body through its defense of its race-conscious admissions policies.Despite this commitment, however, the University of Michigan is facing challenging trends.Figure 7 demonstrates that the share of first-time, degreeseeking undergraduates for most of the racial/ethnic groups at UM has been relatively stable, with White students showing stable shares of about 65% from 1994 to 1999 with a drop in 2000 to approximately 60% and a slow recovery back to approximately 65% by 2005.Asian enrollment shares have been relatively stable at around 12%, Black shares have decreased slightly during this time period from about 9% to approximately 7%, and Hispanic shares have remained stable at approximately 5%.The only other group to show a sustained trend is the "unknown" category, which has increased from approximately 4.5% pre-1998 to between 6.5% to 7% after 1998, with most of that increase appearing in that single year.
The relative stability in enrollment shares hides important trends in Michigan as a whole.Figure 8 shows the first-time, degree-seeking enrollment shares at UM and at the state-level, by race/ethnicity and indicates that White students have been historically underrepresented at UM relative to their presence in state-level enrollment.However, the fitted regression line suggests that lower enrollment shares in the state as a whole and stable enrollment shares at UM are resulting in White student shares moving closer to parity.At the same time, Black students' shares are moving away from the parity line at a rapid pace, with enrollment shares of Black students increasing in Michigan as a whole and falling at the flagship.The trend for Asian and Hispanic students is not quite as clear, but from the fitted regression lines for Hispanic and Asian students it appears that they are both maintaining their shares at UM relative to their shares in the state as a whole.When one quantifies these changes by using odds ratios (again using the odds of a first-time enrolled student at UM being of a particular racial/ethnic group compared to the odds of her/him being the same race/ethnicity in the state first-time enrolled higher education pool), one can see the impressions from figure 8 confirmed.The odds ratios for White students range from approximately 0.80 to approximately 0.90 from 1994 to 2005 and the odds ratios for Black students range from approximately 1.00 in 1994 to approximately 0.50 in 2005.These numbers suggest that the odds of being White at UM compared to the state first-time enrollment pool have increased by approximately 10%.Simultaneously, the odds of being a Black student at UM have decreased by almost 50% compared to the odds of being a Black student in the state first-time enrollment pool.In addition, the odds ratio of being an Asian student at UM versus being Asian in the state first-time enrollment pool have been relatively stable-ranging from 4.00 to 4.75, which is a very high level of overrepresentation.The odds ratios of Hispanic students show a modest downward trend from 2.08 in 1994 to approximately 1.55 in 2004.This suggests a move towards parity, although Hispanic students remain overrepresented.
What do these patterns mean?At the very least, this analysis suggests that institutions do have the ability to alter their enrollments relative to the available pool.This ability may be influenced by several possible factors including their admissions, recruitment, and financial aid policies as well as the composition of the state's qualified applicant pool.At UM this control has resulted in the consistent underrepresentation of Whites, which is a rare finding among state flagship universities.In addition, we find that relative to the available pool, Black students were at one point enrolled proportionately to their population overall, but this proportionality has moved to underrepresentation in recent years.The overrepresentation of Asian and Pacific Islander (API) and Hispanic students relative to that same pool has been a consistent finding, with API students enrolled at much higher rates at UM than in the state first-time enrollment pool.This situation could be thought of as analogous to many Southern K-12 school districts that are enforcing desegregation orders and show much lower levels of schooling segregation when compared to the residential segregation in the district by simply redistributing available students to schools with larger or fewer numbers of minority or White students (Reardon & Yun, 2005).Here, UM is deviating from the state-level enrollment shares by disproportionately enrolling API and Hispanic students when compared to their Black and White peers.The pursuit of such differing racial compositions can have many possible explanations, one of which is the "critical mass" hypothesis, that for the benefits of diverse interactions to be realized, enough individuals of a particular group must be present to avoid tokenism and create the environment necessary for such benefits to be self-sustaining (Regents of the University of Michigan, 2003).
In the case of UM, these data cannot give us insight into why such compositions have been consistently maintained (e.g., critical mass).However, it appears that there may well be a pool of college-ready students already seeking to attend school in-state (those represented by state-level enrollment data) who the institution could recruit and admit to achieve the diversity for which they are searching.Over the next few years we will need to track enrollment changes at UM to examine the impact of Proposal 2 on its student body, particularly to see if the underrepresentation of White students continues to decline and the underrepresentation of Black students continues to increase.
Ok la hom a By contrast, in Oklahoma, a state that was targeted for an anti-affirmative action ballot initiative in 2008, the flagship University of Oklahoma (OU) has experienced (with small variations) an increase in the share of White undergraduates enrolled, a decrease in the enrollment of Black students, and very little change in Asian and Hispanic enrollment (see Figure 9).However, when one contrasts these enrollments at OU with the state-level enrollment, a very different picture emerges.Figure 10 compares first-time, degree-seeking enrollments by race/ethnicity at OU to the state firsttime, degree-seeking enrollment population, and this comparison illustrates the sharp movement of White enrollment to overrepresentation at the state flagship, when early in the 1990s White students were underrepresented.This trend is quite clear and rapid and is caused by both the previously mentioned increase in share of White students at OU and a concomitant decrease in the White share of their counterparts in the state.The story for Black students in Oklahoma is the opposite, with Black students being overrepresented at OU in the early 1990s and quickly moving towards underrepresentation.The fitted regression lines for Black and White students are nearly perpendicular to the parity line-this can only happen if the increases in one sector are exactly matched by the decreases in the other.In other words, this situation represents the fastest path to changing the composition of one sector relative to the other.For Hispanic and Asian students the trends are not nearly as clear.There appears to be a shift for Hispanic students towards underrepresentation; for Asian students there is no discernable trend, but they remain overrepresented.However, when one quantifies the level of over-or underrepresentation using odds ratios, the changes become clearer and the magnitude of the changes are more troubling. 26  26 One reason that the changes are clearer when using odds ratios is because of the small initial shares of many minority groups in these schools.A 2% change in enrollment is very small on a graph that represents percentages from 0% to 90%.But if the initial enrollment share is 4%, that represents a large change of 50% from the initial values.Odds ratios highlight this change in a way that graphical representations cannot.
The odds ratios for White students range from a minimum of 0.95 in 1994 to a maximum of 1.13 in 2004.This suggests that the odds of being White for a first-time enrolled student at OU in 1994 were 5% lower than their counterparts in the state, but in 2005 the odds shifted to 13% higher at the flagship.For Black students the odds ratios moved from a high of 1.07 in 1994 to a low of 0.49 in 2005, suggesting that a first-time enrolled student's odds of being Black at OU when compared to their state counterparts were 51% lower in 2005 when they were 7% higher in 1994.These figures confirm the rapid move towards underrepresentation for Black students illustrated in Figure 10.However, for first-time enrolled Hispanic and Asian students the odds ratios provide more insight.OU students' odds of being Hispanic compared to statewide enrollment were a maximum of nearly 1.60 in 1998 and reached a minimum of 0.85 by 2004.This represents a swing for Hispanic students from over-to underrepresentation.For Asian students the trend is the opposite with their odds ratios rising from 2.27 in 1994 to 2.72 in 2005, so that a student's odds of being Asian at OU were 127% higher than the odds of their state counterparts, and these odds were increased during the observed period.
These trends should be of concern to individuals who think about equity and access to state flagships, not necessarily because it is a deviation from proportionality, but because it signals an important departure from previous practice that is not clearly motivated by any change in policy or any sharp changes in state composition.In fact, these enrollment changes are exactly contrary to the large-scale changes in state composition.In general, where differential enrollment in state flagship institutions becomes most apparent is either in the under/overrepresentation of the dominant racial minority group or in instances when the state share of a particular group is changing relatively quickly, particularly when those changes are not reflected in flagship enrollments.It is important to note that these enrollment changes are occurring despite the fact that the University of Oklahoma is not a particularly selective institution, currently has the ability to use affirmative action, and has a larger pool of minority students to choose from in the state at large.That the University of Oklahoma is currently unable to enroll Black and Hispanic students in the same proportion as they are represented in the rest of the state, even though they were able to do so just 11 years ago, suggests a more complicated story behind these numbers that may be related to institutional decision making, state demographics, and political considerations.Regardless of the specific reason for these changes, if the University is unable or unwilling to maintain the shares of minority students that they had just a few years ago with race-conscious policies available to them, what impact would losing those tools have on the potential to increase access for students of color?This question is particularly critical in light of the fact that the loss of affirmative action would extend to race-conscious scholarships, as well as outreach and recruitment practices, just a few of the practices on which the university might rely to recruit and retain underrepresented students.
In some ways Oklahoma is a good exemplar for the situation found in the states that were targeted for (Arizona, Colorado, Missouri) or passed (Nebraska) ballot initiatives in 2008.In each of these states, White students make up the largest share of students (the overwhelming share in Colorado, Missouri, and Nebraska), and in each of the states White students are overrepresented in their flagship institutions by varying amounts.Where the demographics of the state higher education enrollments are changing (most in Arizona and least in Nebraska), the overrepresentation of Whites and underrepresentation of the major racial minority (Hispanic, Black, and/or American Indian students) at their flagship institutions is becoming exacerbated.In general, Asian students in all these states remain either proportionally represented or overrepresented.
Ultimately, given our review of the current status of affirmative action in higher education admissions and the enrollment data we present, we are skeptical that affirmative action will no longer be needed to further the goals of diversity and equality on college and university campuses by 2028.In our examination of how several colleges and universities have fared with regard to admitting racially/ethnically diverse classes when affirmative action was eliminated as an option, the results are not encouraging.In particular, underrepresented student access seems to be hurt the most by the absence of these policies, thereby affecting the extent to which these institutions can create the conditions necessary to foster the educational benefits of diversity.This is true despite the fact that minority enrollment shares in some of these institutions eventually returned to earlier levels because during that period enrollment shares in the college-eligible population have generally increased, leaving minority students as far (or even farther) behind in representation than before the ban.
It is also important to note that in many of the states considering bans, the enrollment shares of underrepresented minority students are already extremely small and changing towards greater underrepresentation for minority students and greater overrepresentation for White and Asian students, suggesting that even proportional representation may not create the critical mass necessary to support the benefits of diversity.

Guidelines for Institutions of Higher Education
Given the large variation in institutional enrollment changes after the loss of affirmative action policies, it is difficult to predict the results of losing affirmative action on any particular institution.This suggests the need for further research in this area that should include institutionallevel case studies, analyses that disaggregate racial/ethnic groups into subgroups, analyses of graduate and professional school enrollments, and examination of the changes to the "race unknown" category.Given this uncertainty and the lack of data upon which institutions can base their responses, it becomes necessary for institutions to seek out a variety of responses within the legal and policy frameworks both under conditions when the use of race/ethnicity is allowable and when it is not.
Colleges and universities operating under state bans on affirmative action need to find alternative ways to recruit and admit students if they are interested in educating a diverse student body.For example, Leonhardt (2007) noted that after it witnessed such large declines in its enrollments of students of color after Proposition 209, UCLA turned to a special program sponsored and run through a private Black alumni group, wholly separate from the university.As a result, UCLA had increased success enrolling more students of color for fall 2007; it experienced a 13% increase from 2006 in applications by Black students and the rate of acceptance of Black applicants went from 11.5% to 16.2 %.Such programs seem to circumvent the state bans and appear to be legal.Nevertheless, opponents of affirmative action likely will scrutinize such efforts very closely.Law professor and affirmative action critic Richard Sander has already requested UCLA's fall 2007 admissions records through the Freedom of Information Act (Leonhardt, 2007), for example.Nevertheless, as Karabel (1999) has written, "alterations in admissions criterion [sic] and in process can mitigate the effects of measures such as Proposition 209, but they cannot eliminate them" (p.112).
The Supreme Court rulings in Gratz and Grutter and the ongoing experiences of California, Washington, Florida, and Michigan have helped to delineate how states and institutions of higher education can craft their affirmative action policies or respond to bans on affirmative action.In response to the Seattle and Louisville decisions, the American Council on Education (ACE) pointed out that in writing for the majority, Chief Justice Roberts identified just two viable reasons for raceconscious admissions: to remedy the current effects of past institutional discrimination in K-12 systems and higher education, and to advance diversity interests in higher education (Przypyszny & Tromble, 2007).Given these parameters, we have culled guiding points for institutional practice from numerous sources (e.g., Grutter, 2003;Guess, 2007;O'Neil, 2008;Przypyszny & Tromble, 2007).Although the legal landscape surrounding affirmative action is often unpredictable, these loose guidelines follow current law.
Regarding institutions in the 45 states that have not banned affirmative action programs in higher education, we know from Gratz and Grutter that in choosing to consider race/ethnicity in admissions decisions, institutions need to review student applications individually, taking into account quantitative as well as qualitative measures of student academic, social, and personal merit.Diversity may be taken into account when it is part of an institution's mission; race/ethnicity can be considered as one qualification among many-that is, not as the sole or predominating factor (although it can be a "plus" factor between two equally qualified applicants); admissions policy cannot unreasonably constrain the rights of non-minority applicants; institutions need to seriously (though not necessarily exhaustively) consider race-neutral alternatives before settling on admissions policies that take race/ethnicity into account; institutions need to be cognizant about achieving results related to race-conscious policies, but at the same time be careful about how they use the concept of "critical mass," making sure to connect it to the educational benefits of diversity; and institutions need to have a periodic process to review the policy or set up a sunset provision for the policy (Grutter, 2003;Joint Statement of Constitutional Law Scholars, 2003;O'Neil, 2008).
Although these guidelines will not work for the states that currently have affirmative action bans in place, institutions in these states still have a number of options available for recruitment and outreach that are both legal and conscious of institutional missions regarding diversity, equality, and justice.Of course, to comply with state law, policies first need to be changed to make sure that race, ethnicity, color, sex, and national origin are not employed in admissions practices.Then institutions can focus on creative ways of reaching out to diverse students.Such methods could involve recruiting students earlier in their school careers, focusing on socioeconomic status, increasing student aid, and making sure that curricular offerings reflect diverse fields and scholarship.As already mentioned, private, external organizations, such as alumni groups, can work to recruit and fund diverse, underrepresented students.Although they are wary of such external programs, affirmative action opponents seem to be grudgingly accepting of them (Guess, 2007).Finally, institutions that remain interested in increasing racial and ethnic diversity on campus even in the face of a state affirmative action ban will need to take actions to combat the negative symbolism that results when such an initiative is approved by voters.As the figures above illustrate, enrollment of students of color tends to decrease substantially after an anti-affirmative action initiative passes.Students of color may find the vote to be symbolic of an unwelcoming environment for diversity.

Looking to the Future
Von Drehle (2007) examined what he called "the incredibly shrinking role of the Supreme Court" (p.42) under Chief Justice John Roberts.As he reported, under Roberts' conservative leadership, the Supreme Court has heard fewer cases in one year than it has in each of the last 50 years.In addition, as the editorial board of The New York Times (2008) pointed out, the Supreme Court is slowly rolling back civil rights era anti-discrimination policies related to race and sex: "in recent years, the court's majority has been reading federal anti-discrimination laws far more narrowly than Congress intended" (The New York Times, 2008, ¶9).It is a bitterly divided court, and the trend has been for the Justices to consider cases with a narrower scope than they had in years past.It was within this legal and political context that this article presents analyses to make sense of the sometimes conflicting federal-and state-level affirmative action policies, especially in light of the latest state-level threats to affirmative action.
If anti-affirmative action ballot initiatives pass in increasing numbers, colleges and universities likely will face tremendous struggles to maintain or increase student diversity, as is evident in states that have banned affirmative action.Epple, Romano, and Sieg (2008) posited that without affirmative action, the enrollment of underrepresented students of color at highly selective colleges and universities could drop by as much as 35%.Although state ballot initiatives on affirmative action may further curb affirmative action programs in states, there remains the larger societal dispute about whether affirmative action is a policy that fosters diversity and equality of opportunity in the service of all Americans.In itself, the existence of state votes determining the fate of affirmative action will not go very far in deciding the larger moral and political questions surrounding whether affirmative action policy is indeed right or wrong (Moses, 2006).Further, simply allowing members of the public to vote on the issue does not mean that Americans have resolved the moral questions regarding affirmative action, nor does it mean that voters and others have learned what they need to know to make an informed vote: [T]he very perception that this trio of anti-affirmative initiatives constitutes a mortal wound to affirmative action is premised on the appearance of a fair and legitimate process by which affirmative action has been presented, evaluated, and repudiated not only in the court of popular opinion but in courts of law as well.(Crenshaw, 2007, p. 124) 27

References
Race-conscious affirmative action is legal in 45 states and it is still needed.Will it still be needed in 20 years?Our hope is that widespread, meaningful equality of educational opportunity will exist then, but our fear is that as state ballot initiatives seeking to curb race-conscious affirmative action programs become increasingly prevalent, it will in fact become even harder to reach that goal.Perhaps the most salient point is this: These initiatives significantly decrease the chance for public educational institutions to reach the ideal place about which Justice O'Connor wrote.Having even 20 more years to achieve a diverse student body becomes a moot issue if affirmative action is banned widely now.The data presented herein show that by and large the institutions of higher education examined are not achieving meaningful racial and ethnic diversity among their student populations and the loss or threatened loss of affirmative action has made this more difficult for them.Consequently, all students' college and university educational experiences suffer and equity in higher educational opportunity remains elusive.

12 Curbing Affirmative Action State by State: Ballot Measures and Executive Orders
probably the most relevant to higher education.In 2007 the U.S. Supreme Court took up the issue of voluntary race-conscious student assignment in public schools when it agreed to hear Parents Involved in Community Schools v. Seattle School District No. 1 and Meredith v. Jefferson County Board of Education.