Affirmative Action at Berkeley

We continue the debate on affirmative action in response to Karen Paget's "Diversity at Berkeley: Demogoguery or Demography" (TAP, Spring 1992) and Paul Starr's "Civil Reconstruction: What to Do Without Affirmative Action" (TAP, Winter 1992)


The University of California at Berkeley is one of the most selective large public universities in the country. This is true whether one measures selectivity by the average quality of the student body or by the proportion of applicants admitted.

In recent years, UC Berkeley has also become a symbol of the nation's supposed future, as less than one-half the freshmen admitted each year, for the last several years, have been non-Hispanic white. The media has attributed this phenomenon to changing demographics. However, simple population dynamics fail to explain most of the enrollment statistics. Most of the change is attributable to affirmative action policies designed to be inclusive of members of groups who would otherwise be grossly underrepresented in a demographic sense and, in fact, were until recently.

Karen Paget, in her article "Diversity at Berkeley: Demagoguery or Demography?" (TAP, Spring 1992) gleefully describes the effects of this policy in limiting white enrollment at Berkeley. Her defense of Berkeley's affirmative action begins with the observation that California's demographics are changing. In the future, according to Paget, no ethnic or racial group will represent a numerical majority. So the 1988 entering freshmen class that had no ethnic majority was, I guess, "progressive" in spite of the fact that the demographics had not yet reached the desired point.

Paget goes on to knock down the straw man of nonqualification. She contends that there is a "common misperception" that most minority students do not meet the published criteria. Most of us are not college admissions specialists. This is a somewhat technical and arcane field, and writers like Paget don't do much to illuminate the territory. I submit that "most of us" believe that black applicants are held to lower standards than white applicants and that this difference is often substantial. Yet, Paget purposefully understates the black advantage in admission in observing that "if two hypothetical students, one black and one white, have, say, 3.6 grade point averages and SAT scores of 1100, the black student would have a better chance of admission to the Berkeley campus." This is grossly misleading.

The truth is: the black student, in the top 2 percent of black students, is almost certain to gain admission. The white student, in the top 20 percent of the white universe, is unlikely to earn admission. This is so because of the racial quota system being used by Berkeley. (I use "quota" here advisedly. In recent years the term has become misleading shorthand for a system that gives preference to minorities over whites. Usually, the system is not, strictly speaking, a quota. In the case of Berkeley and UCLA, however, the admissions process operates so as to attain a specified percentage of black and Hispanic enrollment.)

The Berkeley campus has been trying to achieve proportional enrollment since the early 1980s. First, it set out to ensure that the "underrepresented" groups such as blacks and Chicanos were proportionately admitted. However, this caused an apparent anomaly: Because Asian-American youngsters outcompete whites, white enrollment was "too low." The solution was to quota in whites relative to Asians. Of course it soon became apparent to some groups of Asians that it was more difficult for their students to be admitted than it was for whites. They raised a fuss, and whites became truly residual.

Subscribe to The American Prospect

Now it is truly a disadvantage to be a non-Hispanic white in California if one wants to be admitted to Berkeley or UCLA (the other UC campuses do not have the same pattern of enrollment by ethnic group as the two leaders). Paget papers this over with appeals to merit (aren't you whites embarrassed, you can't beat Asians?) and diversity. It does boggle the mind that it is important to have blacks and Chicanos proportionately represented but not whites. But as my table "Ethnicity and Admissions at Berkeley (1988-89)" shows, if you are a black high school graduate in California who attains the minimum necessary high school GPA, you have a 70 percent chance of being admitted to Berkeley. If you are white, your odds are 9 percent.

What aggravates this situation is that even within that universe of the qualified, the white applicants have higher average academic indices than do the blacks or Hispanics. This is bit of reality that Paget ignores. She explains that the demographic dynamic continues to move against white enrollment: "Caucasian graduates are declining both absolutely...and in the numbers who apply to the university." She, of course fails, to draw the obvious inference that, given the now very low probability of admissions success at Berkeley, white students are giving up applying.

In the middle of the table is a derivation of the proportion of eligible students in California by ethnicity. This is the actual universe from which admissions decisions are made, not the total of high school graduates. If students were selected according to their place on the eligible list, only 2.5 percent of the Cal students would be black, while almost 70 percent would be white. Now it is the case that in the universe of acceptable students, Asians are probably the highest ranking, followed by, in order, whites, Chicanos, and blacks. If the admission decisions were made purely on merit, Asians would be slightly more overrepresented than they already are, whites would be slightly overrepresented, and Chicanos and blacks would be grossly underrepresented. If this represents relative academic merit, what's wrong with it? It would be no more demographically unrepresentative than the current setup. We would simply underrepresent Chicanos and blacks instead of whites and end up with a higher quality student body.

Paget also observes that if Berkeley was to admit purely by merit, the student body would be in the top 3 percent or 4 percent of California's high school students and mostly white and Asian. As she explains, roughly half of Berkeley's students are selected purely by merit. I can deduce, then, that these students represent the top 1 percent or 2 percent of California high school graduates. As about 75 percent of Berkeley students are white or Asian, I conclude that this part of the student body represents the top 3 percent or so of California students. The rest are somewhere in the bottom of the remaining 9.5 percent.

In SAT terms, this means students with SAT scores over 1350 are being mixed with those who scores 1150; and the 1350s outnumber the 1150s. No wonder so many more of the black students are flunking calculus compared to the Chinese-American youngsters. As long as the class continues to be geared toward the higher achieving group, the success dichotomy will continue. Andrew Hacker observed in Two Nations, "It remains to be seen how Berkeley's new student body will function in practice. In introductory courses, on one side of the room will be Asians admitted on the academic track. Across from them will be blacks and Hispanics with classroom skills at a rather lower level. It is almost as if two dissimilar colleges were sharing the same campus."

Paget rejects the idea of making admissions on merit alone. "What must one believe to adopt an admissions policy based on numbers alone? One must believe that what tests measure as intelligence is valid and reliable, that grades are objective measures, that both tests and grades predict academic success...and that in some past golden age, universities admitted in accordance with some objective standard on intellectual merit." No, I don't have to believe all that; I do, however, believe that grades and tests (considered together) predict academic success, and thus that "admissions by the numbers" is eminently defensible. The graduation success statistics support this opinion. Finally, what does some golden age have to do with it?

When we get right down to it, the affirmative action admissions policy fails the test implied by Paget herself. Do the tests and grades predict academic success? They do. And the policy of admitting by ethnicity fails it. Thirty-seven percent of black students at Berkeley graduate in five years, while 71 percent of white students graduate. Comparing these results with graduation rates at Berkeley for whites 20 years ago or comparing Berkeley with the entire universe of colleges is illegitimate.

In summary, this defense of an outrageous affirmative action program is awful. Paget attempts to justify grossly unequal treatment of individual students simply because of their race.


I wrote the article on Berkeley's admissions policies with the following goals in mind: 1) to describe factually the situation at one elite research university that mirrors the future demographics of California, a state where no single racial or ethnic group constitutes a majority; 2) to broadcast the essential fact that contrary to the perception of students and the general public alike, most admitted students are drawn from the eligibility pool--the top 12.5 percent of high school graduates; 3) to differentiate policies aimed at enhancing diversity from those more commonly understood as governing affirmative action.

On one fundamental point, Lewis Jones and I agree. If statistical precision (meaning test scores and grade point average) were the only criterion for admission, the University of California would be mostly white and Asian. Its student body would represent the top 4 percent of high school graduates rather than the top 12.5 percent as required by the state. We disagree, however, over the desirability of restricting access to UC to such a narrow slice of graduates, on the meaning and significance of numerical measures, and on the facts of Berkeley's admissions policy.

Demographics are crucial. When UC could accept all eligible high school graduates, none of the current admissions controversies existed. As my article pointed out, historically, admissions policies (with the one exception of Georgia State) were never constructed on a strictly statistical basis. Entering classes were selected with many criteria in mind, including the creation of a geographic mix. That is one reason to put such measures as college graduation rates (retention rates) in a more historical context. When white students took more than four or even five years to graduate from college, no one challenged their ability or their right to a slot.

The explosion of college-bound graduates, which is quickly outstripping UC's commitment to admit all eligible students (just not their choice of campus), is magnifying the often minute differences among these UC-eligible students. UC is not rejecting "merit," as Jones asserts, but the belief that parsing people into the most precise statistical measure is an accurate determination of merit and that access to UC should be made on that basis alone. An unfortunate consequence of these new admission pressures and the equation of merit with statistical precision has been the common but erroneous assumption that minority students have been admitted from "below the line" or outside the eligibility pool.

For those who believe, like Jones, that merit is obvious and there is nothing wrong with restricting access to the top 3 percent or 4 percent of college graduates--the obvious implication of Jones's argument, no amount of statistical data will convince. But for the record, his statistics are not reliable (see Jerome Karabel's chart "A Second Look"). Studies have proven wrong, for example, Jones's assumption that minority retention rates are low (they are indeed lower than white rates) because they can't hack it. The reasons are many and varied, as likely the result of personal or family financial problems. Retention rates do not account for those who return and graduate. If a six-year rather than five-year retention rate is calculated for blacks, the figure rises to 51 percent (based on 1983 data).

Finally, Jones conflates two issues that I tried to tease apart. It is true that Berkeley officials might be concerned about the underrepresentation of whites at the campus, and that is precisely why the rubric of affirmative action--whether you are for it or against it--is inadequate to describe the complexity of these admissions policies.


Lewis Jones's critique of Karen Paget's insightful and sympathetic account of "Diversity at Berkeley" is representative of a growing body of thought hostile to both the theory and the practice of affirmative action. Like much of this literature, Jones's assault on affirmative action is distressingly sloppy about matters of fact--a not uncommon irony in the writings of those ostensibly committed to the highest standards of intellectual rigor. Nevertheless, Jones raises some fundamental issues about the justice and efficacy of affirmative action. After identifying several of the specific errors that mar his analysis of Berkeley, I will address these broader issues concerning affirmative action.

Jones refers to "the racial quota system being used by Berkeley," specifically claiming that "the admissions process operates so as to attain a specific percentage of black and Hispanic students." Yet the fact is that Berkeley does not have--and never has had--a quota for any racial or ethnic group. The figures for blacks, the principal target of Jones's ire, suffice to settle this matter: between 1985 and 1990, the number of African-American admits to Berkeley varied from a low of 448 in 1985 to a high of 916 in 1987, with 681, 845, 831, and 645 admitted in 1986, 1988, 1989, and 1990, respectively. If this is a quota, it is a very fluid one indeed.

Jones also errs in his presentation of statistics on ethnic and racial differences in rates of admission to Berkeley. For example, in his table on ethnicity and admissions at Berkeley, he reports that only 8.9 percent of white applicants were admitted; the correct figure, however, is 28 percent--over three times higher than Jones's figure (see my table "A Second Look at Berkeley.") Jones fails to make a distinction between the proportion of applicants admitted to Berkeley (34.5 percent for the fall of 1988) and the proportion of applicants who register at Berkeley (15.7 percent). This failure to distinguish between the admissions rate, which is within the control of the institution, and what is commonly referred to as the "yield rate" (the ratio of registrants to admits), which is a matter of the choices of individuals, does not inspire confidence about either Jones's grasp of the admissions process or his empirical claims.

Despite these errors, Jones is nonetheless correct in claiming that Berkeley gives preference to black and Hispanic applicants. That Berkeley and other institutions practicing affirmative action also give preference to many other categories of applicants--in Berkeley's case, older students, the disabled, athletes, rural students, students with "special talent," and the socioeconomically disadvantaged--somehow escapes his attention. Jones' singleminded focus on preferences given to historically underrepresented minority groups is hardly unique. Indeed, the Office for Civil Rights of the U.S. Department of Education ruled in 1990, after a two-year investigation, that Harvard University's policy of giving significant preference to the children of alumni was "legitimate" and "legally permissible," even though its impact was to reduce the percentage of Asian-American students in the freshman class. The number of beneficiaries of Harvard's policy was large; had alumni children been admitted at the same rate as other applicants, their number in the freshman class would have dropped by nearly 200--a figure that surpasses the total number of black, Mexican-American, Puerto Rican, and Native-American registrants combined. America in 1992 has thus arrived at a peculiar juncture in its tangled history of race relations: while affirmative action programs for subordinate racial minorities are increasingly subject to the kind of bitter assault leveled by Jones, affirmative action for the privileged is now apparently not only permissible but officially sanctioned by the federal government.

What kinds of students are in fact admitted to Berkeley under what Jones calls "an outrageous affirmative action program?" The first thing that needs to be emphasized is that more than 85 percent of registrants from historically underrepresented minority groups ( blacks, Hispanics, and Native-Americans) are fully qualified under the University of California's stringent admissions criteria. These standards require that eligible students be in the top one-eighth of their graduating high school class--a level reached by under 10 percent of the age group, given California's high school dropout rate of roughly one-third. Such standards are exceptionally high for a public university and make the University of California one of the most selective state universities. In 1991, the average GPAs and SATs of freshmen from historically underrepresented minority groups were 3.44 and 1016, respectively. These students tend to come from less advantaged segments of the class structure than their counterparts from other racial and ethnic backgrounds; the median family income in 1991 was $35,000 for blacks and $36,300 for Hispanics compared with $47,400 for Asians and $75,000 for whites. The lower graduation rates for blacks and Hispanics--which look significantly better after six rather than five years (see my table)--are no doubt attributable in part to their comparative lack of financial resources.

While the majority of Berkeley's black and Hispanic students have met the University of California's eligibility standards and hence are fully qualified, Jones is correct in stating that many of them would not be competitive if admissions were determined solely by grades and test scores. Yet no major college or university in the nation admits its freshman class this way, and Berkeley is already very far towards the purely academic end of the spectrum in allocating 50 percent of its available slots solely on the basis of grades and test scores. (Harvard, in contrast, selects fewer than 10 percent of its admits on this basis.)

Jones is aware that total reliance on grades and test scores would lead to a drastic reduction in black and Chicano enrollments, but he is untroubled. "If this represents relative academic merit," he asks, "what's wrong with it?" But whether such an outcome would in fact reflect genuine academic merit is precisely the question. In the face of a formidable body of empirical evidence documenting the relatively modest predictive power of high school grades and especially the SAT (see, for example, James Crouse and Dale Trusheim's book, The Case Against the SAT ), Jones continues to believe that the identification of "merit" in a heterogeneous society is a cut and dried affair rather than a matter that calls for--as the best admissions officers well know--high levels of flexibility and humility. Yet does Jones truly believe that the 33 percent of California's population that is Hispanic and black is so lacking in academic talent and potential that it should occupy fewer than 5 percent of the places in the freshman class of the state's flagship public university? If so, then the rigid and mechanistic admissions criteria he favors have an inexorable logic to them.

To be sure, Jones does identify some anomalies in Berkeley's affirmative action policy. Amidst all the concern about enrolling black and Latino students in rough proportion to their percentage of the population, for example, there has been little discussion about the growing underrepresentation of whites. In the fall of 1991, the proportion of whites in the freshman class reached a historic low of 30 percent, though Caucasians still comprised 52 percent of California high school graduates. Because of the relatively close link between social class and academic achievement among whites, the Caucasian students who are not only eligible for admission to Berkeley but also competitive for one of the slots allocated strictly on the basis of grades and test scores tend (in contrast to the Asian students) to come largely from affluent backgrounds. Thus the mean family income of white students in 1991 was more than $85,000, with 29.3 percent over $100,000 and only 18.1 percent under $40,000. And even a new policy that gives preference to UC-eligible students from socioeconomically disadvantaged backgrounds, regardless of race, has failed to increase significantly the number of students from California's large population of working-class and poor Caucasians.

While Berkeley's affirmative action policy has undeniably produced some genuine political and moral dilemmas, Jones's critique ultimately falls short because of its fundamentally ahistorical character. To understand why race-based programs in education and employment were necessary (and remain so), it suffices to review briefly the system of racial domination that prevailed prior to the advent of affirmative action in 1965.

Though Lacking formally exclusionary policies, America's leading universities in the early 1960s were in reality quasi-segregated institutions. While no precise figures are available, Berkeley itself--the very symbol of 1960s liberalism (and of radicalism)--enrolled a student body at the time of the free speech movement in 1964 that was well under 5 percent black and Hispanic. The situation at the leading Ivy League colleges more closely approximated a system of de facto segregation; data derived from my own study of the freshman registers and senior yearbooks at Harvard, Yale, and Princeton yield the conclusion that blacks in 1960 probably comprised no more than 1 percent of the student body at each of the three institutions, with Hispanics even less represented. This de facto racial exclusion, visible at other colleges and universities in the early 1960s as well, was still firmly in place almost a decade after the Brown v. Board of Education decision of 1954 and the Montgomery bus boycott of 1955-1956.

Yet by the early 1970s, for the first time in the nation's history, America's leading institutions of higher education had finally opened their doors to substantial numbers of black and (to a somewhat lesser extent) Hispanic students. By 1972, for example, black students at Harvard, Yale, and Princeton had come to comprise, respectively, 7, 8, and 10 percent of the freshman class. Of all students enrolled nationwide at the undergraduate level (including those at traditionally black institutions), the percentage of blacks rose from 4.3 percent in 1960 to 9.8 percent in 1975 (see David Karen's article in American Journal of Education, July 1991). The most dramatic gains were made after the urban racial upheavals of 1965-1968, which led, especially after the wave of riots following Martin Luther King's assassination, to the institutionalization of race-based programs. The rise of these programs coincided with unprecedented gains for blacks and Hispanics in gaining access not only to the more selective colleges but also to graduate and professional schools. This historic democratization of educational opportunity would not have occurred without affirmative action programs.

The effects of affirmative action on employment opportunities are less clear, though evidence suggests that it has had a positive effect. It is easy to forget just how exclusionary many labor markets were in 1965 not only in high-status professions such as law, medicine, and academe but also in strategic working-class domains such as construction unions and police and fire departments. The increase in the number of black and Hispanic lawyers, physicians, and professors in roughly the past two decades is well-known, but the record in certain blue-collar jobs is just as impressive. For example, between 1970 and 1990, the number of black electricians more than tripled (from 14,145 to 43,276) and the number of black police officers increased almost as rapidly (from 23,796 to 63,855).

Less certain, however, is whether this increase can be attributed to affirmative action programs per se rather than to other factors such as the antidiscrimination legislation that was part of the Civil Rights Act of 1964 or more general shifts in the climate of race relations. This is a complex and rather technical matter, so some of the conclusions of the National Academy of Sciences' Committee on the Status of Black Americans are of special interest. In its summary report, "Common Destiny: Blacks and American Society," the committee reports that "nearly all of the cross-sectional research finds that blacks' employment share increased more in contractor firms subject to affirmative action requirements than in firms without federal contracts." The report cites as an example of the impact of federal affirmative action programs and antidiscrimination legislation the "large increase in black employment during the 1960s" in the South Carolina textile industry, which "previously had barred almost all black workers."

There is thus a powerful body of evidence suggesting that affirmative action in both education and employment has in fact succeeded in expanding opportunities for minorities, including members of the black and Hispanic working class. Indeed, it is this very success in redistributing scarce and valued resources that has made affirmative action so politically controversial and aroused such intense opposition among certain segments of the population, especially among whites. It is primarily for this reason, I believe, that many liberals--including Paul Starr and other contributors to The American Prospect who are sincerely committed to racial equality--have reluctantly concluded that the time has come to distance themselves from affirmative action, if not to repudiate it outright.

In his article "Civil Reconstruction: What to Do Without Affirmative Action," (TAP, Winter 1992), Starr's point that affirmative action has taken a political toll from progressives and may have damaged their capacity to build broad coalitions is a serious one; redistributing educational and economic opportunity is guaranteed to be controversial even in the best of times, and since 1973 affirmative action has been pursued in the context of a declining economy and increasingly bitter zero-sum conflicts. Yet one can acknowledge that affirmative action may have extracted significant political costs without making the dubious claim--comforting to those who wish to abandon race-based programs--that its "direct effects on the structure of opportunity have been modest." Nor does Starr's conclusion that the liberal commitment to affirmative action may have made the task of building political coalitions across racial lines more difficult require that one believe that "affirmative action policies have helped to perpetuate [white] racism"; in the face of a serious challenge from below to the racial status quo, white racism would, I suspect, have done quite well in any event.

Should the Supreme Court sharply curtail affirmative action, Starr concludes, "it could prove a blessing in disguise." It is not at all clear to me that the Supreme Court will, as he predicts, overturn the 1978 decision Regents of the University of California v. Bakke , but of one thing I am sure: that the elimination of affirmative action would be a serious setback to racial equality and an unfortunate step backward into our nation's tragic racial past.


My old friend Jerry Karabel has neglected some relevant distinctions in my article and its more general argument. I wrote in the aftermath of the confirmation of Clarence Thomas, the departure from the Supreme Court of Thurgood Marshall and William Brennan, and the Court's decision in Croson, which restricted affirmative action in public sector contracting and seemed to prefigure further restrictions by an even wider conservative majority. I raised the question, if the Court does narrow affirmative action, what other measures are available that will help us achieve racial justice?

I made a distinction, however, between affirmative action adopted as a remedy for historically demonstrated discrimination and affirmative action adopted to improve racial balance and representation. The Court seems much less likely to restrict the former, and I emphasized that I was in no way arguing against strong enforcement of antidiscrimination laws. And when looking for examples of the clear-cut successes of affirmative action, Karabel points to exactly the kind of cases, like the electricians and textile mills, that are not at issue between us. It is simply not clear, from the research I have read, that affirmative action in the second sense has had the effects its advocates believe it should have.

In any event, I suggested two other avenues for those concerned about improving the social and economic conditions of minorities -- universal social and economic programs, for which there are broader constituencies of support; and a concerted effort to rebuild the social infrastructure of minority communities, that is, independent institutions such as schools, community services, nonprofit economic development agencies, and other "mediating structures" that are vital to both civic life and prosperity. I didn't oppose the concept of compensatory remedies. I advocated more emphasis on institutional rather than individual remedies: an infusion of capital into minority institutions that have never had much to work with before.

You may also like