Affirmative Action and Minority Enrollments in Medical and Law Schools

Affirmative Action and Minority Enrollments in Medical and Law Schools, by Susan Welch and John Gruhl. Ann Arbor: University of Michigan Press, 1998. 221 pp. $39.50, paper. Despite the controversy generated by affirmative action in higher education, the actual effects of the programs have not been thoroughly analyzed. Instead, the arguments over race-based remedies have been symbolic. With respect to professional training, such as that provided by medical and law schools, the role of affirmative action has been based on anecdotes and assumptions about a few outstanding individuals. These include Bernard Chavis, the African American doctor who graduated from University of California-Davis medical school through the special admissions program Allan Bakke challenged. Initially held up as a success story, Chavis, with later malpractice allegations raised against him, was turned into a cautionary example. Among other examples are Benjamin Carson, the African American who chairs the pediatric neurosurgery department at prestigious Johns Hopkins University, and who separated twins conjoined at the head; Johnny Cochran, who led the "dream team" defense counsel in the 0. J. Simpson murder trial; and Vernon Jordan, one of the most powerful lawyers in the nation's capital, regardless of race. Affirmative Action and Minority Enrollments in Medical and Law Schools presents a longawaited and much-needed empirical basis for the public discourse. Welch and Gruhl's book is a supplement to the massive Bowen and Bok (1998) study, The Shape of the River. It differs from that work in its emphasis on institutional decision making in response to legal doctrines. Although this type of evidence is vital to the affirmative action debate, it is not, as the authors themselves implicitly conclude, sufficient to resolve the matter. As Christopher Edley, Jr. (1996), the Harvard professor who led the Clinton Administration's "mend-it-don't-end-it" review of affirmative action policies, has argued, facts matter but "values matter most" (p. 73). At stake are competing visions of what the institutions of this nation, elite as well as public, should look like and how best to realize those ideals. Both the ends and the means are contested. Using a questionnaire sent in March 1989 to the 118 medical schools and 154 law schools that were operating when the Supreme Court decided Bakke v. Regents of the University of California (1978), Welch and Gruhl describe how these institutions practiced affirmative action, whether they changed in response to legal restrictions, and, most importantly, whether these efforts to remedy racial/ ethnic discrimination worked. Their study deserves consideration by all policymakers and administrators who care about racial/ethnic diversity in higher education. As Welch and Gruhl explain, the Supreme Court was highly divided in Bakke, with four justices voting in favor of affirmative action in theory and four justices voting against affirmative action as it was being practiced. Justice Louis Powell, the Richmond, Virginia, corporate lawyer who had led the American Bar Association, reached a famous compromise. He reasoned that affirmative action could satisfy constitutional standards if it resembled the Harvard College plan, which took race into account as a factor, but it violated constitutional norms if it relied on strict quotas. The Powell opinion has been heavily criticized from all quarters as being impossible to follow. Nonetheless, until the Supreme Court ruled in its 1995 case of Adarand v. Pena (1995) that affirmative action is the legal equivalent of invidious discrimination, the Powell approach effectively governed affirmative action. Welch and Gruhl present many basic findings about the role of legal doctrine in shaping institutional behavior related to affirmative action. Not surprisingly, they note, virtually all those responding to their survey had heard of the Bakke decision. Surprisingly, however, 77% of medical school officials and 63% of law school officials reported that it changed their policies "not at all" (p. …


I. INTRODUCTION
When we began this project in the mid-1980s, affirmative action and the compromises articulated in the Regents of the University of California v. Bakkel decision seemed to be a settled and accepted part of public policy. To be sure, much of the white public had never accepted some of the more far-agreed to hear the case, more than one hundred organizations filed a record number of arnicus curiae briefs. 14 Magazines featured cover stories on the case, 15 and commentators anticipated the result. 16 Commentators called Bakke the most important civil rights case since Brown v. Board of Education. 17 Bakke would determine not only the validity of one school's admissions plan, but the composition of many schools' classes for many subsequent years. Moreover, Bakke was expected to determine the legality of affirmative action and the speed of further civil rights progress in American society.
Even years after the decision, Bakke seemed important. In 1987, when Justice Louis Powell, who authored the main opinion, retired, he was asked which of his opinions was the most important.' 8 "Bakke," he replied without hesitation.19 Today, nearly two decades after the case was decided, its issues and holdings have become a centerpiece of a renewed public debate over affirmative action.
The case began when Allan Bakke, a white engineer who decided to become a doctor, applied for admission to eleven medical schools. 20 Although Bakke had good test scores, he was older than the schools preferred, and he was rejected by all of them. 21 The following year, Bakke applied to UC-Davis and was again rejected. 22 In both years that Bakke applied to UC-Davis, the school admitted through the special admissions program some minority applicants whose test scores and grades were lower than Bakke's. 2 3 In addition, UC-Davis admitted some white applicants whose test scores and grades were lower than 14 See Bakke, 438 U.S. at 268-70; see also O'NWE, supra note 13, at 3-6 (listing organizations that submitted amicus curiae briefs and briefly discussing the strife the organizations experienced as a result of taking a position on affirmative action and quotas). 15  Bakke's through the school's regular admissions procedures. 24 When the medical school at UC-Davis opened its doors in 1968, it had no African-American or Hispanic students. 25 The next year it had just two blacks and one Hispanic student. 2 6 In 1970, the UC-Davis faculty, like many schools' faculties, voted to establish a special admissions program for "economically or educationally disadvantaged" students because of the small number of minority applicants. 2 7 When the school doubled its entering class to 100 in 1971, it reserved 16 of these places for students selected through the special admissions program. 28 Although the program was open to "disadvantaged" students of all races, it was established primarily for blacks and Hispanics. 29 Few Native Americans applied. 3 0 Numerous Asians applied but most were evaluated through the regular admissions procedures. 31 Many whites applied to the medical school, but none were accepted through the special admissions program. 32 The Dean of Admissions later acknowledged that the program was not tailored to take into account lower class whites or middle class minorities. 33 Bakke sued because he had been passed over in favor of minorities with lower scores. He claimed reverse discrimination and asserted the Equal Protection Clause. 34 Although the school passed over Bakke in favor of some whites with lower scores, Bakke had no constitutional or statutory grounds for challenging these admissions.
In June 1978, the Supreme Court announced the Bakke decision. Justice Powell did not exaggerate when he said, "[w]e speak today with a notable lack of unanimity." '3 5 There was no majority opinion. In six separate opinions the Justices split on the two key issues: whether quotas are unconstitutional and whether using race as a positive factor is unconstitutional. 36

OHIO STATE LAW JOURNAL
Justices-Chief Justice Burger and Justices Rehnquist, Stevens, and Stewartapparently concluded that both the quota and any use of race as a positive factor were invalid. 37 Another bloc of four-Justices Blackmsun, Brennan, Marshall, and White-concluded that both quotas and using race as a positive factor were valid. 3 8 Powell was the swing Justice, maintaining that the quota was unconstitutional but the use of race as a positive factor was not. 39 Thus, Powell provided the fifth vote for one issue from each bloc, and his opinion became the controlling opinion. Justice Powell, using "strict scrutiny," considered whether the program served a compelling governmental interest. 4° He rejected all of the university's justifications except for its need to establish a diverse student body. 4 1 However, Justice Powell stated that it was not necessary to set aside a certain number of seats to accomplish this goal. 42 Thus, he proclaimed that the program violated both the Equal Protection Clause of the Fourteenth Amendment and Title VI of the Civil Rights Act of 1964. 4 3 Justice Powell concluded that schools could use race to obtain a more diverse student body and approvingly cited the Harvard Plan. He added an explanation of this plan to the appendix. 44 However, Powell's opinion does not clearly distinguish what schools can and cannot do. They cannot use an obvious quota or set-aside. They should not use separate procedures to evaluate applications from different racial groups. But can they use a flexible "goal" rather than a fixed quota? Powell called this a "semantic distinction" that was "beside the point," 45 but Harvard implied that it set a goal. 46 If schools do not use a flexible "goal," but weight some races differently and enroll about the same number of minorities every year, is this different significantly from using what is forbidden?
Brennan maintained that there was "no constitutional distinction" between 37 See id. at 420-21. Stevens, arguing for this bloc, chided the other five Justices for addressing the issue of race as a positive factor in admissions. He considered these statements dicta. Yet, Stevens seems to address this issue implicitly. Moreover, it is hard to conceive that Stevens and his bloc could reach any conclusion other than that race could not be used at all, given their interpretation of the [Vol. 59:697 the two approaches. 47 The only difference is that the Harvard Plan "proceeds in a manner that is not immediately apparent to the public," whereas the UC-Davis program is more open. 48 Blackmun echoed these views, stating the line was "thin and indistinct." 49 Powell replied that the crucial difference was the "facial intent to discriminate." 50 Yet, Powell acknowledged that a plan that used race as a selection criterion could be "simply a subtle and more sophisticatedbut no less effective-means of according racial preference than the Davis program." 51 Further, in a passage directed toward school officials, Powell assured that "a court would not assume that a university, professing to employ a facially nondiscriminatory admissions policy, would operate it as a cover for the functional equivalent of a quota system. In short, good faith would be presumed .... "52 While the distinction exists in theory, it might not appear in practice.
The Bakke decision was covered extensively by the media, both national and specialized. 5 3 Commentators, reflecting the split ruling, were divided about the likely effect. 54 One study of editorial reaction by nineteen daily newspapers found an emphasis upon the balanced outcome of the decision. 55 Surprisingly, there was relatively little criticism. Perhaps the compromise in the result, or possibly the ambiguity of the ruling, neutralized potential opponents in the press. Yet, activists feared the worst.

OHIO STATE LAW JOURNAL
consequences. The NAACP called the ruling "a major disappointment." 5 6 The Amsterdam News headlined the story, "Bakke: We Lose!" '57 Jesse Jackson, comparing the decision with the withdrawal of federal troops from the South after Reconstruction, proclaimed, "[b]lack people will again be unprotected ... we must not greet this decision with a conspiracy of silence.. . we must rebel. "58 The spokesperson for the Mexican American Legal Defense and Education Fund noted that while Bakke won a symbolic victory, most affirmative action programs would be lawful under the standards of the case. 59 However, the Congressional Black Caucus emphasized that the Court upheld the use of race, and it urged that this point be publicized to shore up support for affirmative action policies. 60 Some opponents of affirmative action applauded the ruling. Some conservative activists, however, predicted that Powell's opinion would drive quota programs "underground;" 61 while universities would purport to adopt the Harvard Plan, they would continue to use de facto quota programs. Thus, Powell's opinion would "produce much the same result" as before but with "an encouragement of duplicity." 62 The Court did not decide any other cases involving school admissions through 1987, but it did sanction affirmative action. 63  off instead of minorities with less seniority. 65 If these rulings had any effect on the impact of Bakke, they should have reinforced the use of race as a positive factor in the decade after the decision.

Im. THE STUDY
Although Bakke affected many types of schools, our research assesses Bakke's impact on applications and admissions to medical and law schools. 66 We chose to examine medical and law schools for several reasons. First, we chose to examine medical schools because Bakke occurred in this context. Second, we chose to examine law schools because the Supreme Court sidestepped a previous case involving law school admissions, 67 and law school faculty are aware of and sensitive to legal decisions. Third, medical and law schools had more special programs to recruit and retain minorities than other professional schools, 68 and admission to these schools had become highly competitive by the time of the Bakke decision. For all of these reasons, Bakke's impact, if any, should be apparent in an examination of these schools.
This study examines both the ultimate targets of the ruling, the students, and those who were responsible for implementing the ruling-the admissions officials. 69 We use three sources of information. First, national application and enrollment trends over a twenty-year period before and after Bakke allow us to examine macro-level year-to-year changes in students' decisions to apply and in enrollments. Second, cross-sectional information on individual schools before and after Bakke highlights micro-level changes in enrollments across schools. Third, 1989 survey responses from admissions officers in medical and law schools reveal their perceptions of Bakke's impact and of their admissions policies. In each of the sections below, we will describe in more detail the data used.

1998]
OHIO STATE LAW JOURNAL

A. Aggregate Enrollment Trends
Perhaps the most direct way to examine the impact of Bakke is to look at nationwide trends in African-American and Hispanic enrollment. 70 We have collected data on several key trends, including minority applications, admission, and enrollment, beginning as early as 1951 for some variables, and as late as 1973 for others. These data can show us how, if at all, both students and admissions committees changed their behavior over time in response to Bakke or to other factors.
We begin by examining the aggregate trends in black and Hispanic first-year medical school enrollment. As Figure 1 shows, African-American enrollment jumped from about 200 in 1965 to almost 1200 in 1975. African-American enrollments were very stagnant through the next decade, increasing only gradually. In the late 1980s, enrollment again began to increase significantly and accelerated in the 1990s to a peak of more than 1500. The trends were generally similar for Hispanics, except for the 1975 to 1985 period, which showed a slow upward growth with a plateau in the late 1980s.
The Bakke decision had little effect on either of these time series. The steep upward climb in enrollment had reached a plateau before Bakke and would not resume until more than a decade after the decision.
Although the longitudinal patterns are somewhat different, the same conclusion can be reached about black and Latino first-year enrollment in law schools. For African-American enrollment, the year of the Bakke decision was only one point in a period of relative stagnation from about 1975 to 1986. This was followed by a rather sharp increase beginning in 1987. Hispanic enrollment rose slowly but fairly consistently over the entire period for which we have data-1969data- to 1995data- . Only in 1975data- , 1983, and 1995 did enrollment decrease from the year before. Again, Bakke seems irrelevant to interpreting these time series data.
We also have data on medical school applications beginning in 1973. Applications by both African-American and Latino students were fairly stagnant from 1973 through 1983. The two patterns then diverged-African-American applications actually declined until they began to increase steeply in the late 1980s. This continued until 1994 when applications began to decline. Applications from Hispanic students stayed very constant until 1990, when they began a slow but perceptible increase through 1996. Again, however, the Bakke 70 Data sources include AAMC (selected years), the Lawyer's Almanac (selected years), and personal communications from Rick Morgan of the American Bar Association. These data are based on individuals' self reports of race and etlnicity. In thousands Data from AAMC, 1973-1997 ,U I t 1973 1975 1977B1979 1981 1983 1985 1987 1989 1991 1993 1995 1S Year -o-Black -Hispanic 2 -

DOES BAKIKE MAT4ER?
In order to calculate the impact of Bakke more precisely than could be done from a visual inspection of the graphs, we utilized the following: annual data on several application, admission, and enrollment variables; a model that included a short term effect and long term effect and several control variables; and time series methodology. Our time series analyses began either in 1968 or as soon thereafter as data were available and ended in 1987. This period allowed us to focus more precisely on Bakke than would a longer time period. The longer the time period examined, the more difficult it is to separate Bakke's effects from others, because "the repercussions of all government actions ramify indefinitely and interrelate with other phenomena, both public and private, many of which simply cannot be quantified and indeed often cannot even be identified." 71 Control variables included a dummy variable for the Reagan Administration years and a measure of federal financial aid available in that year. For the dependent variables of applications and acceptance ratios, we also controlled the aggregate number of applications from all students that year.
We reported each of these trends in more detail elsewhere, 7 2 but Table 1 summarizes these findings. In general, they reinforce our conclusions that Bakke had little specific impact on trends in minority applications, admissions, and first-year enrollment. Though the fit of the models range from good to excellent, only two of the Bakke coefficients are significant.
In examining the longer time series that extends through 1997, it seems clear that the growth in minority applications and enrollments occurred at two points. First, there was a sharp increase after the passage of the Civil Rights Act of 1964, as legal barriers fell and professional schools became aware of the need to recruit minority, particularly African-American, students to their institutions. In addition, competition for admission to professional schools was growing tremendously at this time. 73 Second, there was a significant increase in the late 1980s and early 1990s. This increase may best be attributed to the growing middle class in the African-American community and the large increase in the number of college educated minorities. 74  led to an increase in the number of families who were able to send their children to professional schools. In addition, an economic boom in the early to mid-1990s reduced unemployment and induced some income growth, making it possible for more students to attend college. This growth in minority enrollment began nearly a decade after Bakke and is highly unlikely to have been affected by it.

B. The Effect on Individual Institutions
Another way to explore the impact of Bakke is to look at individual institutions before and after the decisions. Aggregated national data could [Vol. 59:697 DOES BAKKE MATTER? conceivably obscure important gains in individual institutions. If Bakke changed patterns and intensities of minority recruiting, we might expect that effect to be apparent in the distribution of minority gains and losses among different types of schools. For example, we might expect the effect to be less apparent in states with small minority populations. We might also expect a bigger gain in northern schools, rather than southern schools, where resistance to civil rights was more intense. 75 For this analysis, we used the total minority enrollment in the school, rather than just first-year enrollment. 76 The overall growth in minority enrollments in the typical medical or law school between 1978-the time of Bakke-and 1987 was extremely small. Medical schools, for which we have data on blacks and Hispanics, increased their overall enrollment by thirty-five students during this decade, and about one-third of this small growth was in minority students. However, African-American enrollment grew by only .19%, and, on average, Hispanic enrollment grew by .75%. Law schools, in which our data are for "minorities"-a combination of African Americans and Hispanics-actually decreased in overall enrollment during this decade, but the minority proportion of their enrollment increased 2.63 %. Table 2 illustrates that, institution by institution, minority enrollment in 1987 is highly related to minority enrollment a decade earlier. This is most true for African-American students attending medical schools, where earlier enrollment explains over 93 % of the variation in 1987 enrollment (shown in the table as R 2 ). This conclusion is almost as true for African Americans attending law schools, where the variation explained is 89%, and somewhat less true for Hispanics in medical school where the explained variation is only 57%. A high proportion of variation means that those schools with high minority enrollments in 1987 also had high enrollments in 1977, and the same is true for those schools with medium and low enrollments.
If Bakke had an effect on professional schools' treatment of minority applicants, one would expect this effect to vary greatly among institutions, with some moving ahead dramatically and others lagging behind. If so, we should not have expected to find an extremely strong relationship between minority enrollments before and after Bakke. However, the strong relationship found between previous and current minority enrollments suggests that the decision reinforced existing patterns of minority recruiting. Whatever factors promoted 75 See James H. Kuklinski et al., Racial Attitudes and the "New South," 59 J. OF POL. 323-49 (1997). 76 First-year minority enrollment was not available for law schools. Law school data for 1977 were not available, so we used 1979 data. Unfortunately, this captures the first-year post-Bakke class, so the data are partially contaminated. However, two-thirds of the law students would have been accepted before Bakke.

1998]
OHIO STATE LAW JOURNVAL minority enrollment in 1977, also did so in 1987. The possible exception is Hispanic medical school enrollment, in which the variation between the 1987 enrollment and the 1977 enrollment, while substantial, is considerably less than in the other two cases. However, as we will see in the next table, there are other factors in addition to Bakke that help explain changing Hispanic enrollments. In Table 3 we examine the impact of 1977 enrollment on 1987 enrollment, while taking into account several other possible predictors of minority enrollment. 78 The table indicates that none of these other factors affected black enrollment in medical schools. Although the 1977 enrollment remains an extremely strong predictor of 1987 enrollment, neither the type of school or the proportion of blacks in the state population, enrollment, or changing enrollment are related to African-American enrollments in 1987. 79 Indeed, adding these factors to the equation only increases the explained variation in enrollment by a nonsignificant 1%, from 93% to 94%.
Hispanic enrollment continues to be strongly predicted by previous Hispanic enrollment (as we saw in but the proportion of Hispanic population in the state also is an important predictor of that enrollment. Hispanic medical enrollment grew by a larger proportion than did African-American enrollment, and it is clear that much of this growth took place in states with larger Hispanic populations. These states, such as Florida, Texas, and California, are also the site of much of the increase in the Hispanic population. The impact of a state's Hispanic population is largely responsible for the increase in the R 2 from .57-with only the 1977 enrollment figures as a predictor-to .70 in this equation.

OHIO STATE LAW JOURNAL
Hispanics in a state are significantly related to 1987 enrollment as well. Contrary to expectations, minority enrollment in 1987 is higher in those law schools that have decreased their overall enrollment, with all other things being equal. As with minority enrollments in medical school, there are no effects of overall enrollment and type of university.
In both medical and law schools, African-American and Hispanic enrollment a decade after Bakke was very much related to enrollment before Bakke. Though some professional schools moved ahead faster than others in this decade, on the whole, schools that had larger proportions of minorities before the decision had them a decade later. But not all variation in 1987 minority enrollments was explained; we have seen that the Hispanic population of the state affected changes in Hispanic enrollment and, in law schools, overall minority enrollment.

C. The Views of Admissions Officials
Viewpoints of admissions officials, who screen applications and make decisions about whom to admit, provide a third kind of evidence about the impact of Bakke. Through reports in the popular and specialized media, and through the activities and publications of the Association of American Colleges of Medicine and the American Association of Law Schools, it is clear that professors, deans, and admissions officers at both medical and law schools had. ample information about the Bakke decision at the time it was handed down and immediately afterward. 8 ' "Given that attention in professional as well as popular media, it is not surprising that eleven years after the decision, fully 97% of the medical admissions officers and 99% of their law school counterparts had heard of the decision." 82 Most of these officials perceived that Bakke has had minimal impact on admissions policies. For example, officials were asked whether the decision changed their admissions policies at the time Bakke was decided and whether it affected their policies at the time of the survey, which was conducted nearly a decade later. 8 3 Almost no one reported that the decision changed their policies significantly at the time of the decision. However, 11% of the medical school officials and 17% of the law school officials in the survey reported that it changed their policies "somewhat." 84 As indicated in Figure 4, over three- quarters of the surveyed medical school officials and 63 % of the law school officials claim that it affected policies "not at all." One school official said, "[w]e have always upheld affirmative action." '85 Another respondent reported that its affirmative action policy remained essentially unchanged since 1968.86 Thus, only a small minority of schools reported that Bakke changed rather than reaffirmed their admissions policies. Similar to the minority of respondents who asserted that Bakke changed their policies when the decision was handed down, another' minority of respondents reported that the decision had an impact on their decisions in 1989 when we conducted the'survey. One fear of those who opposed the decision was that public attention and professional school actions would focus on the portion their policies are slightly more flexible. One medical official said they ceased using "two lists," one for majority and one for minority applicants. One official said race became less of a factor, and two officials said that it became more of one.

1998]
OHIO STATE LA W JOURNAL of the Court's decision outlawing quotas, rather than the portion of the Court's decision affirming the use of race as a positive factor. 87 "If this emphasis were to occur, these detractors believed, eventually efforts to recruit minorities would cease and opportunities for minorities would diminish." 88 Despite the fears of some officials, the majority of medical and law school officials did not believe that opportunities for minorities were diminished as a result of Bakke. In fact, of those who believed that Bakke has had a continuing impact, more believed it has worked to increase the admissions of minority students than to limit admissions. Indeed, among law school respondents, twice as many believed the decision has opened the doors wider than believed it has closed them tighter. Among medical respondents, only slightly more respondents believed that Bakke improved access for minorities than believed it limited access. Most admissions officials believed that Bakke legitimized existing practices rather than changed them, and where it did change practices, most officials believed Bakke improved them.   There -is still more evidence that the impact of Bakke was to legitimize policies rather than change them. Respondents were asked whether they took race into account before Bakke and whether they take race into account, as a positive factor, in making decisions today. Most schools took race into account in a positive way before the decision, and a few more schools did afterwards. The safest conclusion to reach is that the decision apparently did not deter schools from considering race as a positive factor, and may even have encouraged a handful. Being black, Hispanic, or a Native American earns students positive credit in the admissions process at most law and medical schools. As indicated in Figure 5, by 1989, over 95% of law and 90% of medical schools gave extra consideration to African Americans. Moreover, 93% of law schools and 69% of medical schools did the same for Hispanics. 89

D. Some Conclusions
Many commentators, referring to the "chilling effect of Bakke," predicted that the ruling, by invalidating quotas, would result in fewer applications by minority students, less pressure on schools to admit them, and, consequently, smaller minority enrollments. 90 Over time, supporters of affirmative action came to view Bakke as an essential stimulus to minority enrollment. Based on this systematic study of Bakke's impact, we conclude that its effect on boosting or curtailing minority enrollment was far less than either supporters or opponents predicted. Nonetheless, the decision was significant because it legitimated and institutionalized the practice of affirmative action in admissions decisions. Bakke outlawed racial quotas, but the decision also validated affirmative action practices that give some positive weight to the race or ethnicity of candidates. 9 1 Therefore, although Bakke is increasingly under challenge, it continues to be the Supreme Court decision underpinning affirmative action in higher education admissions.
We are confident in our conclusions, but we realize there are limits to this study. As with any examination of the impact of a Supreme Court case, it is difficult to separate the effect of the Court's ruling from the effects of other factors occurring at the time. Minimizing the period of the series analysis reduces, but does not eliminate, the number of other factors conceivably 89 These proportions are very similar to those reported for African Americans in a 1974 survey of medical schools, but lower for Native Americans, Asian-Americans, and Hispanics. See Wellington & Montero, Equal Educational Opportniity Programs in American Medical Schools, 53 J. oF MID. EDUC., 633-39 (1978). African Americans were the most likely to receive preferential treatment, followed by Native Americans, Hispanics, and Asians. Differences could be in recall of our respondents and different samples, as well as changing practices.

1998]
OHIO STATE LAW JOURNVAL affecting minority enrollment. Moreover, predicting individual behavior from aggregate characteristics is always risky. We have examined decisions made by a few thousand students and a few hundred schools in our aggregate analysis, but we have tried to predict their behavior with variables affecting millions. Ideally, we would have conducted a before and after study of both individuals and institutions. It would have been useful to examine students' reasons for applying or not applying to professional schools to see if their reasons differed before and after Bakke, and indeed to determine if students were even aware of Bakke and its predicted impact. Moreover, such a before and after design could have been profitably used to study the admissions process. However, even at the time we conducted our survey, much time had passed since the decision. Therefore, we have no "before" information from individuals and admissions committees except for responses based on memory. Thus, we assembled the best evidence we could after the decision and relied on a multiplicity of sources. 92 Given this evidence, our conclusions are reasonably straightforward. Admissions officials said in response to our questionnaire that the impact on their decisions was "none" or "slight." Hardly any respondents said the impact was "significant." Our analyses of aggregate applications, acceptances, and enrollments confirm these officials' impressions. In particular, the impact of Bakke on the number of minority applicants or enrollees was minimal. The schools with the most minorities the year before Bakke were the ones with the most minorities a decade later, and those schools with the least number of minorities before the decision were the ones with the least minorities after the decision. Patterns of institutional behavior tend to persist, and professional school enrollment before and after Bakke is certainly an example of that persistence.
Our conclusions reinforce the findings of some previous studies on the impact of Supreme Court decisions. 93 Such studies confirm the impression that the impact of Supreme Court decisions is often different than what would be expected from reading the decision or even from reading media reports about its presumed consequences. Cases considered are rendered do not always have an impact other than the legal and political commentary that follow in their wake. Indeed, the most fundamental finding of these studies is that rulings do not always change behavior in the expected way, to the expected extent, or at all.
To a great extent, the minimal impact Supreme Court decisions produce derives from the refusal of the parties involved and the lower courts to comply with the ruling. If they do comply, the results might be different from the ones intended, or the effect might be far less than the extent anticipated. Moreover, if there is an impact, it often varies from place to place. 94

1998]
and to an attorney before interrogating them. 99 Other officers did advise suspects, but nonetheless found them more willing to confess than to exercise their rights. 10 0 Therefore, contrary to predictions, there was only a small decline in confessions following Miranda.' 0 ' Courts must rely on others to implement their decisions. 1 02 For Bakke, admissions officials and others who affect admissions decisions formed the implementing population. Despite initial fears by civil rights organizations, hindsight indicates that most of these officials were at least somewhat committed to increasing minority enrollment. 103 Because admissions officials believed they should make this effort, they established procedures to increase enrollment. However, to the extent that the Court's ruling made this goal more difficult to achieve, these officials might be less than enthusiastic to implement these procedures. Our findings that the schools with the best record of minority enrollment before Bakke were also the schools with the best record a decade after Bakke suggest the influence of the implementing population.
The Court must rely also upon others to communicate the rule to those who must implement it. 1°4 As we previously noted, law and medical school officials were very aware of the Bakke decision. However, the ruling itself was not clear, and ample communication of an ambiguous message still leaves an 99 See Medalie et al., supra note 97, at 1394. The authors state that after the Miranda decision, police officers did not adopt the Court's policies. For example, half of the defendants studied were not given the silence warning, less than two-thirds reported they were denied the station-house counsel warning, and more than two-thirds reported they were not given any of the formal Miranda warnings. See id.
100 See id. at 1395 (noting defendants were "loathe" to use attorneys, often giving statements because they did not understand how Miranda applied to them); see also Interrogations in New Haven: The Impact of Miranda, 76 YALE L.J. 1519, 1578 (1967) (reporting that 19 out of 37 defendants in their study did not react to the Miranda warnings and incriminated themselves  JOHNSON & CANON, supra note 93, at 77. The authors define an "implementing population" as persons who carry out and translate judicial policies into action. See id. In most cases, the implementing population is "a group effort by some bureaucratic organization." Id. at 77-78. However, members of the population tend to be diverse. See id. at 77. ambiguous message. 10 5 These ambiguities might have been clarified by subsequent rulings, but the Justices did not take any related cases in school admissions. Thus far, the Court has left Bakke as its last words on the subject of affirmative action in education.
Whether a ruling has an impact also depends upon the "consumer population"-those individuals who will gain benefits or suffer losses under the ruling.106 Bakke's "consumer population" consisted of the actual and potential applicants to professional schools, both minorities and nonminorities. We found that the impact for these individuals was relatively small, affecting either their rate of applying to professional schools or their opportunities to be admitted.
Finally, we can examine the impact of the decision on the "secondary population," which includes those not directly affected by the decision. 10 7 The media, of course, are a relevant part of the secondary population. The media clearly covered the Bakke decision, but the coverage was primarily superficial. In fact, the media provided very little information to readers about the nature of affirmative action practices beyond those at UC-Davis. Indeed, the media did not begin to describe the way affirmative action practices work at a variety of institutions until well over a decade after the Bakke decision. In the few years following Bakke, the lack of intense media interest helped defuse the issue and contributed to the legitimization of affirmative action. The lack of media concern, coupled with the lack of further judicial scrutiny in the decade after Bakke, also allowed some institutions to drift back into near-quota systems in their affirmative action practices.1 0 8 Thus, Bakke appears to have solidified the existing practice of affirmative action in admissions. While formal quotas disappeared, actions to maintain and increase minority enrollments became institutionalized. Admissions officials look differently at test scores and grade point averages of minority applicants compared with white, non-Hispanic applicants, and, until the 1990s, most  Q. 38, 58 (1974) (asserting that a vague Supreme Court decision allows a lower court to follow its own policies, but a clear decision requires lower courts to follow the decision at least to its minimum boundary).
106 See JOHNSON & CANNON, supra note 93, at 107 (defining "consumer population" as those persons who are personally affected by the particular decision).
107 See JOHNSON & CANON, supra note 93, at 110 (defining secondary population as those persons "not personally affected by a judicial policy").