Skip over navigation

Journal Issue: Opportunity in America Volume 16 Number 2 Fall 2006

The Role of Higher Education in Social Mobility
Robert Haveman Timothy Smeeding

On Higher Education and Social Mobility: What Do We Know?

One of the stated objectives of the nation's colleges and universities is to be a meritocratic filter between the economic position of the families in which children grow up and those children's economic position as adults. Higher education is expected to promote the goal of social mobility and to make it possible for anyone with ability and motivation to succeed. To be effective in this role, colleges and universities must seek out ability, motivation, and preparedness wherever it lies and then provide high-quality educational services to their students. The labor market will do the rest, rewarding those who acquire the skills that the nation's postsecondary system has to offer.18

How well are college and university admission, training, and completion fostering this meritocratic goal? If true “merit” could be measured, answering that question would be easy. One could simply assess the extent to which the most meritorious youth were being sought out, admitted, and trained. Indeed, if merit—ability, motivation, and preparedness— were equally distributed among youth regardless of family income or economic position, an effective higher education sector would offer an equal chance of admission and graduation to all—high-income and low-income youth alike. But ability, motivation, and preparedness are all linked to the economic position of the children's families. Children from well-to-do families tend, on average, to have more of all three traits; children from disadvantaged families, to have less. Genetics plays a role in the allocation of ability and motivation, as do the choices made by and the environment created by families of differing incomes. As for preparedness, the nation's primary and secondary school systems train youth from various economic backgrounds for postsecondary schooling. Other articles in this volume address these precollege patterns.19

The absence of a reliable merit marker makes it more difficult to assess how well higher education promotes social mobility. One would be surprised if rates of college admission, matriculation, and graduation were equal regardless of families' varying economic circumstances, and as we will show, they are not. The question, then, becomes whether the inequality in the provision of higher education services is consistent with a pattern of training being offered to those with the most merit. Even more relevant, perhaps, is whether the inequality in higher educational attainment is increasing or decreasing.

Levels and Trends in Economic Inequality in Higher Education
Table 1 presents an overview of some of the findings of David Ellwood and Thomas Kane in their review of early research on the relationship between schooling and economic background over time. The type of schooling described in the table, college-going, says little about total years of completed schooling or college graduation. For students who graduated from high school during 1980–82, the overall rate of college-going is 80 percent for youth from the top income quartile of families, as against 57 percent for youth from the bottom quartile. Youth from the poorest families were concentrated in vocational and technical institutions, while those from the richest families tended to enroll in four-year colleges.20

Between 1980–82 and 1992, the overall college enrollment rate rose 7 percentage points. But the rate for the highest-income youth increased 10 points, while the rate for the lowest-income youth increased only 3 points. In terms of attendance at four-year colleges, the gap between the highest- and lowest-income youth widened far more during this period. While the share of most disadvantaged youth enrolled in four-year colleges fell slightly (from 29 to 28 percent), that for the most well-to-do youth rose substantially (from 55 to 66 percent). The gap between the two groups widened from 26 percentage points to 38 percentage points.21

Inequality and the Quality of Colleges and Universities
The patterns revealed by Ellwood and Kane are consistent with tabulations of Anthony Carnevale and Stephen Rose, who analyzed detailed data from the High School and Beyond study and from the NELS of 1988.22 They divided all four-year colleges and universities into four tiers by quality, based on the Barron index of college selectivity, putting community colleges into a separate category; and divided all families into four socioeconomic status categories, based on their income and parental education and occupation. 23 Their findings are summarized in table 2.

In the 146 top-tier colleges and universities (accounting for about 10 percent of all college students), 74 percent of the entering class is from the highest socioeconomic quartile and only 3 percent from the lowest quartile. In the 253 colleges in the second tier (accounting for about 18 percent of all college students), the shares are 46 and 7 percent, respectively. Only in community colleges is the composition of entering students by family socioeconomic status similar to the composition of all youth of college age.24

Patterns of Educational Attainment by Family Permanent Income
These family income–related gaps in higher education attainment rely on estimates of income that are somewhat difficult to interpret, and in some cases are suspect. First, among the national data collected, income values are sometimes for the households in which students reside, and hence do not necessarily pertain to the parents of these children.25 Second, for some data sources, parental income is supplied by the students themselves in response to survey questions, and these responses are suspect.26 Third, none of these studies allows for the “income needs” of the families of the youth being studied. It clearly matters whether a student from a family with $50,000 a year of income is an only child or has several siblings who are also competing for family resources. Finally, and most important, the parental or family income data are one-year “snapshot” (or transitory) values and hence fail to reflect the long-term (or “permanent”) economic position of students' families.27

Robert Haveman and Kathryn Wilson proceeded in a somewhat different way to get a reliable picture of inequalities in higher education attainment for a specific cohort of youth. Using the Michigan Panel Survey of Income Dynamics (PSID), they selected a nationally representative sample of 1,210 children who were born between 1966 and 1970 and followed them from 1968, the first year of the PSID (or their year of birth, if later), until 1999. This cohort would be expected to graduate from high school in the late 1980s and from college in the early 1990s. The authors measured educational outcomes—high school graduation, college attendance, college graduation, and years of schooling—at age twenty-five. For each individual, they also calculated permanent income relative to “needs” and the wealth of the family in which he or she grew up. The ratio of income to needs is the average real value of the family's income while the youths were aged two to fifteen, divided by the national poverty line (for a family of that size) and the average wealth (net worth) of the family in 1984, when the youths ranged in age from fourteen to eighteen.28

Table 3 summarizes the educational attainment of youth from the bottom and the top quartiles and deciles of family “permanent” income-to-needs ratios.29 While only about 22 percent of youth from the bottom quartile of families attended college, 71 percent from families in the top quartile at least entered a college or university. The gap is nearly 50 percentage points. Among the youth from the top quartile, 42–44 percent graduated from college, as against only 6–9 percent of youth in the bottom quartile, a gap of more than 35 percentage points. Transitions from high school graduation to college attendance and from college attendance to college graduation are also shown. Again, substantial gaps exist between youth from the highest and lowest quartiles in the probability of making these transitions. The gaps between the attainment levels of youth from the top and bottom deciles are even greater, suggesting a continuous relationship between economic status and educational attainment.

The pattern of extreme inequality between youth from the top and bottom quartiles of the family income-to-needs ratio is similar in terms of the allocation of educational services. Table 4 shows the distribution of all high school graduates, college attendees, and college graduates in this cohort of youth, by decile and quartile of family income-to-needs ratio. Among high school graduates, nearly 30 percent are from the top income quartile, while about 20 percent are from the bottom quartile. At least in terms of attainment— though not necessarily in terms of quality-adjusted attainment—high school educational services are distributed relatively evenly among children from various economic backgrounds. The pattern for college graduates, however, is quite different. Among all college graduates in this cohort, more than 50 percent are from families with income-to-needs ratios in the top quarter of the nation, while only 7 percent are from the lowest quarter of families. Similarly, the 10 percent of families in the lowest income-to-needs decile yield less than 3 percent of college graduates. Put differently, half of all higher educational services necessary for attaining a college degree are allocated to youth from the richest quarter of the nation's families, as against only 7 percent allocated to youth from the poorest 25 percent of families and only 3 percent to youth from the poorest 10 percent of families.

How Large Is the Pool of Qualified Low-Income Students?
The question of whether colleges and universities have been making enough effort to admit and enroll qualified students is difficult to answer. The definition of “qualified” is directly related to the selection standards that schools themselves define and impose. Two studies have tried to answer this question for the highest-quality and most selective U.S. colleges and universities, and both have concluded that the available pool of qualified youth is far greater than the group of students admitted and enrolled at these institutions.

The first of these studies, by Carnevale and Rose, uses a simulation approach for 146 toptier colleges and universities (again, accounting for about 10 percent of all college students). They consider an “SAT equivalent” score above 1,000 as evidence of ability to succeed at these first-tier schools, and then compare the share of low-income students who are qualified with the share of these students who are enrolled. Among students with scores above the cutoff, 5 percent were from the bottom socioeconomic quarter (3 percent of comparable students were enrolled), as against 21 percent from the bottom half (10 percent of comparable students were enrolled). More than 800,000 students had an SAT equivalent score of more than 1,000— four-and-a-half times the total number of student slots at the first-tier schools.30

More recently, Gordon Winston and Catharine Hill have used a similar approach to determine whether the nation's most prestigious colleges and universities (twenty-eight of the private colleges participating in the Consortium on Financing Higher Education) could increase their enrollment of low-income students without sacrificing academic standards. Using an SAT equivalent score of 1,420 as the cutoff for “high ability,” they show that 12.8 percent of all high-ability students are from the bottom two income quintiles, a total of about 4,300 students. Today these colleges matriculate only about 2,750 such students, leading the authors to conclude that the colleges could enroll more such students without decreasing selection standards.31

In focusing on the top-quality colleges and universities, these studies do not address the larger problem of lower-scoring but nevertheless qualified low-income students who attend less selective schools. Indeed, more than three-quarters of all college students attend colleges and universities that do not impose high selectivity standards.32 Hence, even if the most selective colleges and universities admitted qualified low-income youth, there would still be a nontrivial attendance gap between the rich and the poor.

Indeed, part of the gap between low-income students' population share and their enrollment in colleges and universities is due to low test scores and other indicators of ability that are indirectly related to family income. For example, although 36 percent of low-income students at high-income high schools were in the top half of the test score distribution, only 24 percent of low-income students at low-income high schools scored at this level.

Although this evidence regarding the effectiveness of higher education's meritocratic filter is not decisive, these gaps are large. More significant, they appear to be growing. Colleges and universities may aspire to weaken the link between family socioeconomic class and life prospects, but their efforts have been discouraging—particularly in the case of the four-year colleges and universities, the traditional heart of the higher education system, producing the highest-quality educational services. In sum, the allocation of educational services (especially services of the highest quality) is concentrated among youth from families with the highest economic status, and the concentration appears to be increasing. This trend has been reinforced by the erosion in state financial support for public higher education over past years, as spending on other priorities, such as medical care for low-income families, criminal justice, and K–12 education has been substituted for support of public colleges and universities.33

Slow Growth in College Graduation Rates: Some International Evidence
At a time when the links between U.S. students' economic origins and their attainment of higher education are strengthening, progress in increasing the number of U.S. college graduates has stalled.34 Indeed, for any given cohort, there has been virtually no change over the past two decades in the share of youth who have been awarded a postsecondary degree. Figure 1 compares schooling for two cohorts observed in 2002— one aged twenty-five to thirty-four (born 1966–75), the other aged forty-five to fifty-four (born 1946–55)—in fourteen industrialized nations.With two exceptions—reunified Germany and the United States—the share of adults with a postsecondary degree has increased in every country. Although the older U.S. cohort ranked second in the share of adults with a postsecondary degree (about 40 percent), the younger cohort ranked fifth. Four countries had gained parity with the United States or forged ahead, with Canada and Japan outpacing the United States by 10 percentage points. Another five countries had closed the gap to less than 5 percentage points. Only Italy trailed behind by more than 15 percentage points. If U.S. colleges and universities had been able to increase the rate of college graduation over this period, they would likely have been able to serve greater shares of youth from lower-income families, thus weakening the link between family economic origins and postsecondary attainment. The increased concentration of youth from higher-income families in America's colleges and universities, together with the constant rate of college completion, seems consistent with a trend toward zero- sum competition among institutions for a relatively constant stock of the best qualified students—who also are concentrated in the nation's highest-income families.35