Mismatch

FIGURE 3.1. The Importance of Relative Position

What followed would be one of the first clear examples of a phenomenon we will see often: the unwillingness of the academic and broader intellectual community to engage research that demonstrated a mismatch problem. The psychology and science communities not only failed to make a big deal of Elliott and Strenta’s findings; they pointedly ignored them. Science societies did not hold panels on the new research; science journals did not sponsor symposia. None of the popular science media wrote about the work, nor did the mass media report it. Not a single article, pro or con, appeared on the work in the five years after its publication. The higher education establishment was equally unmoved. Indeed, when in 1998 the former presidents of Princeton and Harvard, William Bowen and Derek Bok, published The Shape of the River, their widely hailed and supposedly definitive 459-page work on the effects of affirmative action, they did not even mention Elliott and Strenta’s work. (We will return to Bowen and Bok in Chapter Six.) This cold reception surprised and disappointed Elliott. On reflection, he wondered if the higher education establishment already intuitively sensed and, thus, instinctively shunned science mismatch. “Everybody knows it but it’s the elephant in the room,” he recalls. “They just won’t talk about it. They see it but they can’t say it. It’s a taboo subject.”

The idea that relative disadvantage might be a bad thing in a college environment had not always been taboo. In the 1960s sociologist James Davis had written the article, “The Campus as a Frog Pond” (alluding to the choice between being a big frog in a little pond or a little frog in a big pond). Davis had found evidence that a student’s aspirations were influenced by his level of achievement relative to his or her college peers. Students’ aspirations for “high-achievement careers” were whetted when they were big frogs and diminished when they were little frogs. Davis’s work generated great interest and discussion and became something of a classic in the field of sociology.

But the simple intuition behind the “frog pond” hypothesis became combustible when race was added to the mix. In 1970, at the dawn of the affirmative action era (only a few years after Davis’s article appeared), law professor Clyde Summers spelled out in detail the potentially self-defeating aspects of aggressive racial preference programs. His article was not well received, and the idea disappeared from discussions in legal academia for decades. Around the same time, economist Thomas Sowell laid out the same idea in the context of elite college affirmative action and gave the problem its most durable name: mismatch. Sowell and Summers not only pointed out the potentially counterproductive effects of racial preferences but also deduced many of the likely characteristics of affirmative action systems, as detailed in Chapter Two. The fact that Sowell was black and that a handful of other black analysts agreed with him only seemed to increase the virulence and unanimity with which establishment leaders dismissed his warnings.

In these very early years of preference programs, amid campus unrest and urgency to admit more minorities, many colleges and professional schools largely set traditional academic measures aside in admitting minorities, and flunk-out rates were high. But administrators argued that they simply needed to fine-tune the programs, and faculty were often loath to criticize preferences, lest they be seen as racially insensitive. In the late 1970s, as the pool of minority students improved somewhat and schools became more selective in choosing “diversity” students, outcomes improved to a point at which the problems were less obvious to a casual observer. But most of those involved in affirmative action programs, from admissions officers to faculty supporters, were unfriendly to close scrutiny, in part for legal reasons, as we shall examine in Chapter Thirteen. This code of silence explains why an entire generation passed before, in the 1990s, studies like Elliott and Strenta’s even began to look at the effects of preferences and helps to explain why there was so little discussion of their findings. The taboo Elliott came to perceive was very real.

In 2004, eight years after Elliott and Strenta’s work appeared, two more psychologists weighed in on the issue of “relative preparation” and success in science. Frederick Smyth and John McArdle of the University of Virginia obtained access to the carefully guarded “College and Beyond” dataset—a longitudinal study assembled by the Mellon Foundation during the 1990s with extensive information on tens of thousands of elite college students. This dataset was not only much larger than the one Elliott and Strenta had used, covering twenty-eight colleges rather than four, but it also covered schools with varying degrees of eliteness. This meant that Smyth and McArdle could directly examine how students with particular credentials performed at schools where those credentials would put them at the middle or nearer the bottom on the academic preparation spectrum of their respective classes.

Smyth and McArdle found, as they expected, that one’s overall level of academic preparation was closely associated with STEM success; students with stronger high school records were more likely to go into the sciences and more likely to graduate with science degrees. But their second hypothesis, that STEM success would be greatest at the most elite schools, “was not supported.” On the contrary, they found that students with low academic indices realtive to their classmates were at much greater risk of not persisting in STEM. Their findings were entirely consistent with Elliott’s and Strenta’s. Indeed, because they had detailed comparative data, they could actually measure the mismatch effect. They found that, had all the black and Hispanic students in their sample enrolled at schools where their credentials were close to the class-wide averages, 45 percent more of the women minorities and 35 percent more of the men minorities would have completed STEM degrees.

Given these large effects, Smyth and McArdle concluded in their paper that “admission officers . . . [should] carefully consider the relative academic preparedness of science-interested students,” and they encouraged students aspiring to science and engineering careers to “compare their academic qualifications to those of successful science students” at the colleges they considered attending. Their advice was not given lightly. Smyth had spent years as both a high school and a college counselor before becoming an academic, and his concerns about how to advise promising minority students being courted with offers from elite schools had partly motivated his interest in the topic.

Smyth and McArdle hoped that their fairly definitive findings would change practices or at least generate significant discussion. It did not. Like Elliott and Strenta’s work, their article “seemed to vanish without a trace,” Smyth noted. His fellow psychologists generated scores of articles on the politically hot question of whether “stereotype threat” undermined minority achievement in college (see Chapter Six), but greeted with glacial indifference the more politically sensitive question of whether relative levels of preparation hurt minority achievement. Even with the mounting and unchallenged evidence of a “science mismatch” problem, no university took seriously the concerns that these scholars raised.

A concrete way to think about the size of the science mismatch effect (although none of these four authors put their findings quite this way) is to recall the academic index mentioned in Chapter Two, in which the student credentials one brings into college are measured on a scale of 0 to 1000; 550 might mark the range at which students frequently attend at least minimally selective schools, and 850 to 900 would capture the median student at the most elite schools. In this range every 10-point increase in an academic index increases by about 1 percent the chance that a student will attain a STEM degree. This is hardly mysterious: academically stronger students tend to cluster more in the sciences and are more likely to complete their intended degree program successfully. But a 10-point increase in the index of a student’s classmates (holding his own index constant) reduces his chance of getting a STEM degree by about one-half of a percent. This reduction is the mismatch effect. Attending a school where your classmates have, on average, an index 140 points higher than you do (a fairly typical racial preference size at selective schools) is equivalent, in terms of your likelihood of achieving a science degree, to knocking 70 points off your index score.

Science mismatch came up again and again in interviews conducted for this book. One interviewee happened to be a Dartmouth graduate, a young black woman who grew up in what she described as a rough neighborhood in a large Midwestern city and overcame significant challenges in her home life—domestic violence, a drug-addicted father, and frequent moving around. In spite of all this, she was sufficiently industrious and bright to succeed in high school, and she entered Dartmouth hoping to major in biochemistry and go on to medical school.

As she described it, things went south almost immediately: “I was with classmates from nationally ranked private and public schools . . . it is not until you meet the best that you realize that you are in a whole other playing field . . . . At first I thought I could handle it, but then as freshman year rolled along, I got hammered academically. People in my class had had science since grammar school, but I wasn’t even introduced to science until my sophomore year of high school . . . . I had never developed any of the skills that I needed to achieve.” She struggled through her freshman year but felt more and more convinced she would not be able to keep up. She switched to a major in the humanities for which she still felt unprepared (“I could never read fast enough”) but that was survivable.

A young black administrator who has worked on minority recruitment and retention at universities in the Northeast spoke to us about the frustrations of many students he counsels: “They have these big dreams of being a doctor, but many pre-med students just aren’t up to the demanding courses, so they get weeded out very early on and become extremely discouraged.” So much so, he says, that many mismatched aspiring scientists and other students at one college where he worked “become miserable and end up leaving. And of those that stay, by the time that they graduate, some students have told me that they don’t want the institution to contact them for at least five years, because they feel that they had been used as a poster child for diversity. And so they are just worn out when they leave.”

This administrator adds, “You’d be lying if you said that mismatch didn’t exist. It’s the elephant in the room that nobody wants to talk about . . . . If an institution is just implementing these programs so that they can say that they have these diversity numbers, and they know that they are bringing in kids who are not prepared academically—yet still don’t provide the necessary resources for them—then they are definitely doing these kids a disservice.”

Dr. William Hunter passed through his own ordeal, although he was not academically mismatched in the same sense as many of the students we discuss. Graduating at the top of his predominantly white public high school class in the small town of Uncasville, Connecticut, he enrolled in Wesleyan. On arrival he found himself woefully unprepared for courses that stressed critical thinking. “I will never forget one of my first science exams,” he recalls. “I studied hard, as I always did. And I was used to it paying off: You memorized these thirty facts that you will be tested on. No problem. Well, this test was handed out, and . . . the first question was, ‘Based on what you know about chemistry and biology, design a theory of life based on silicone instead of carbon.’ . . . It was the first time that I had ever gotten a D-minus.” By his junior year he feared he would fail physics and another course. “I was worried that I would never get into medical school,” he recalls. “What am I going to do with my life? I had what I can only describe as a complete nervous breakdown.”

In the end, after moving out of Wesleyan’s self-segregated “Black House,” Henderson—whose outstanding high school record would probably have qualified him for admission to Wesleyan without any racial preference—overcame his struggles, graduated with honors, won other accolades, and went on to medical school and a highly successful career as an internist at a major medical center. Others, he emphasizes, weren’t so determined—or lucky. “There were four black and Latino students that were pre-med when we started that program. As of today, only two of us became physicians. There were a couple of people that were so far out of their element. And then there were Negro urban kids who clearly, clearly, clearly had no business being there. They were not prepared educationally, emotionally . . . . I think that the quota system was well intended but ultimately misguided in some ways. Because it was designed to give advantages to people who were not always ready to take advantage. So as a result, they were in over their heads, and instead of giving them a path to success, it was a doorway to failure.”

When we explain science mismatch to friends, they sometimes suggest that this might simply reflect an unpleasant but ultimately healthy weeding out of the unprepared. “If these students are unable to compete in a tough science curriculum,” they ask, “would you really be doing them a favor by putting them someplace less demanding? They might get a science degree, but will they be good scientists? Isn’t it better to find out earlier that they aren’t cut out for a science career?”

These questions are often asked by nonscientists who themselves went to elite schools and underestimate the depth of strength in American higher education as well as the many paths that lead from a solid STEM education. Elite scientists we have interviewed agree that although a strong graduate program is important for a high-level science career, at the undergraduate level it is most important that students get a solid education, learn the fundamentals, and discover their passion. If they go to a school where they are well matched, these things are more likely to happen, and if they excel, they will have access to top graduate programs. If they are well matched and their performance is middling, they will still have a solid education and a valuable degree; every source we have examined confirms that across the broad spectrum of American colleges, graduates with STEM degrees have significantly higher earnings than non-STEM graduates with otherwise similar characteristics. These patterns are, so far as we can determine, just as true for blacks as for whites. This would help explain why a disproportionate number of black MacArthur Prize winners have attended historically black colleges; at those schools they avoided mismatch.

Science mismatch tends to reduce learning, damage self-confidence, and divert students from their tentative path for reasons having little to do with their ability to succeed in a STEM career. We can see no silver lining here.

Another side effect of science mismatch is stereotyping. Colleagues tend to be taken aback when we point out that black and Hispanic high school seniors are, all other things being equal, significantly more interested in science careers than their white peers. People are surprised to hear this because they know relatively few minority scientists. But a large dataset from the University of Michigan (UM), shown in Figure 3.1, helps to illustrate the disconnect. UM is a particularly interesting case because, even though it used large racial preferences in selecting its 1999 freshmen (the cohort analyzed here), its outstanding reputation and low cost also attracted an unusually large number of blacks who could be admitted without preferences.

As before, we illustrate the data using the 0-to-1000 academic index, which measures student preparation at the end of high school; the great majority of UM students are in the 700–900 range of the index. We see that blacks are significantly more likely than are whites with the same index scores to obtain science degrees. But at UM, students with credentials below 660 will almost never obtain science degrees (we think because of science mismatch), even though those students are still quite strong and even though the data at Figure 3.1 suggests that at a less elite school they would be quite likely to obtain science degrees. Because of preferences, a disproportionate number of UM blacks are in this “under 660” range; thus, the overall proportion of blacks at UM achieving STEM degrees is lower than the white proportion.

FIGURE 3.2. Preferences Can Hide Black Achievement

Source: Authors’analysis of data provided by the University of Michigan on students matriculating in 1999.

Over the very same years that Elliott and Strenta and Smyth and McArdle were uncovering and diagnosing the problem of science mismatch, an entirely independent strand of research was developing along parallel tracks. This second strand had its origins in the early 1990s with Harvard’s president, Neil Rudenstine. Harvard in the 1990s had a highly diverse student body; its traditional WASP constituency no longer made up even a majority of students. But pursuing diversity on the faculty had been much less successful, there and elsewhere. At most colleges blacks and Hispanics were less than half as represented in regular faculty positions as they were among the undergraduate student body. In the 1970s and 1980s many assumed that once blacks and Hispanics began to catch up with white rates of college attendance, many more would go on to become professors. By the mid-1990s it was evident that this wasn’t happening.

A key problem, it was clear, was that Hispanics and especially blacks were not entering the academic pipeline. Few non-Asian minorities were earning doctorates in most academic fields; indeed, few non-Asian minorities were even entering doctoral programs in the vast majority of fields. Although blacks and Hispanics were attending professional schools in substantial numbers and were getting many doctorates in fields like education and psychology (leading to careers outside the academy), few were pursuing paths that led to academia.

In 1995 black doctorates in all the sciences, engineering, humanities, and social sciences (excluding psychology) accounted for only one-fifth of all black doctorates, compared to three-quarters of white and Asian doctorates. In many fields—not just the sciences—minorities were scarce indeed. For example, blacks made up 8 percent of new law students in 2000, but only 3.5 percent of those earning doctorates in the humanities and less than 1 percent of those earning doctorates in economics.

Rudenstine had discussions about this issue with Elinor Barber, an analyst in the provost’s office at Columbia who had herself conducted research on the issue. They developed the idea of a study that would focus on the pipeline to academia, trying to understand what sort of students became interested in academia, what sort of college environments fostered their interest, and what kinds of support helped them to persevere. They wanted to understand why some aspiring minority scholars failed to get into the pipeline and how others slipped out. As her collaborator, Barber brought aboard Stephen Cole, a distinguished sociologist at SUNY Stony Brook, and Rudenstine arranged a seed grant from the Council of Ivy League Presidents.

Barber and Cole conducted focus groups and pretested a survey of high-achieving undergraduates. Then, with additional funding from the Mellon Foundation, they created a large sample of nearly eight thousand students, drawn from the Ivy League, small liberal arts colleges, flagship state universities, and historically black colleges, all of whom were on course to graduate in the spring of 1996. By design, they chose students with strong academic preparation and “oversampled” strong minority students, so nearly all blacks and Hispanics at the selected institutions who had done well were included in the sample. They wanted more than just the aggregate data, so they followed up the written surveys with personal interviews of over one hundred students.

Barber and Cole were immediately struck by the effect that large racial preferences had on minority students’ choice of college and, hence, on student performance. At some of the large state universities they studied, very few blacks had college GPAs above 2.8, reflecting the fact that the strong black students who would have excelled at the state university had been recruited away by elite college preferences. Barber and Cole suspected and soon verified that low grades translated into low academic self-confidence and that students with low academic self-confidence were very unlikely to pursue a career in academia or to even pursue a doctorate in the field of their college major. Indeed, they tended to lower their overall career aspirations. Barber and Cole confirmed that a very large number of well-prepared, underrepresented minority students came to strong schools hoping to become college professors. But getting mediocre-to-bad grades tended to kill those aspirations. Few students who, in their own eyes, did not do well in college wanted to fight uphill battles to become professors and spend their careers in colleges.

In sharp contrast, minority students who attended schools where their academic index put them at or above the average entering freshman had a very different experience. They tended to get much better grades. Their academic self-confidence tended to grow rather than wither. And they maintained their enthusiasm about pursuing doctorates and academic careers. Controlling for other factors, these “well-matched” students were about twice as likely as their “ill-matched” peers to sustain their plans to join academia.

Though they knew nothing of the science mismatch research, Cole and Barber had thus, by 2000, found another area in which large racial preferences appeared to backfire on intended beneficiaries. Empirical research—that is, the facts—had led the two scholars to conclusions that were profoundly unwelcome to most if not all their sponsors. Cole and Barber used the “frog pond” metaphor to articulate their finding, but the mechanism was essentially identical to the “relative deprivation” concept advanced by Elliott and Strenta and the “fear of failure” identified by Clyde Summers. The intended beneficiaries of preferences in all these situations seemed to face a serious mismatch effect.

Importantly, Cole and Barber (and Elliott and Strenta for that matter) found that Hispanic students at elite schools experienced much less mismatch than did blacks. This might be because Hispanics generally received significantly smaller racial preferences; perhaps there was a “sweet spot” where preferences expanded opportunities without generating mismatch. More work was (and is) necessary on this point.

Cole and Barber’s project stretched a decade from conception to publication (not an unusual period for an academic project involving extensive data collection). Then, before the manuscript was finished, Barber tragically developed cancer and died. Their book, Increasing Faculty Diversity, was published by Harvard University Press in 2003 and contained a host of prescriptions to increase minority participation in the pipeline to academia. But it also contained a clear statement of the problem of mismatch, offering this advice to student counselors:

Instead of recommending that minority students go to the most prestigious school they can get into, high school guidance counselors should recommend that each student go to a school where he or she is likely to do well academically. An HBCU may be such a school. Guidance counselors, in short, should try to reduce some of the lack of fit between the level of academic preparation of minority students and the schools where they enroll.

These words were published at almost the exact moment that Smyth and McArdle were giving the same advice to counselors as a way of reducing science mismatch.

Despite the extraordinary involvement of Ivy League leaders in the conception and funding of the project, despite its publication by the Harvard University Press, and despite the careful, richly documented, dispassionate way in which Cole and Barber presented their conclusions, the arrival of Increasing Faculty Diversity was greeted mostly with silence. Even though the book came at the height of a national debate on racial preferences in higher education, catalyzed by momentous Supreme Court cases decided in the summer of 2003, the book received virtually no mention in the press. One of the few was in the Chronicle of Higher Education, which suggested that the book’s primary funder, the Mellon Foundation, was trying to distance itself from the findings. Mellon need not have bothered; the book sold few copies and generated no discussion. As with Elliott and Strenta, Cole could feel a palpable chilling of academic and funding connections. “It was really ignored,” Cole recalled in an interview. “The leaders of higher education institutions have their political views and they adhere to them regardless of the evidence.”

So far as we know, no scholar has made a case against the findings of Elliott and Strenta, Smyth and McArdle, or Cole and Barber; their work, however challenging to the conventional wisdom, has gone utterly unrefuted. The failure of the academy to take account of the problems these scholars identify is serious. To be sure, none of these works sought to assess affirmative action as a whole or tried to weigh the various benefits of racial preferences against their aggregate costs or to explain whether other forms of preferences, like legacies or athletic scholarships, might produce the same harms as large racial preferences. But these three major studies had each provided compelling evidence that the intuitively logical mechanisms of mismatch were real and were hurting large numbers of minority students. This should have spurred massive concern and discussion about corrective strategies or at least an organized effort by higher education leaders to verify or refute the findings. Instead, higher education leaders along with the reporters covering them chose to pretend the studies did not exist.