For Certain Types of Students, the Finish Line Keeps Receding

Princeton U. Press

September 14, 2009

Ten years ago this month, after the summer of American Pie and The Blair Witch Project, roughly 94,000 students arrived as first-time freshmen at 21 American flagship public universities. Four years later, 49 percent of those students had graduated from the institution where they began. Two years after that, an additional 28 percent had done so, for a total six-year graduation rate of 77 percent.

At less-selective public universities, the numbers are even worse. For one recent cohort, the six-year graduation rate at the University of Cincinnati was 46 percent, according to federal data. At the University of North Carolina at Charlotte, it was 51 percent.

All of those rates will need to improve—a lot—if the nation is going to come remotely close to the Obama administration's goal of restoring America's position as the country with the highest proportion of college graduates in the world.

In Crossing the Finish Line: Completing College at America's Public Universities, released last week by Princeton University Press, three scholars dissect the experiences of the entering class of 1999 at those 21 flagship universities, along with the entire public-university systems of Maryland, North Carolina, Ohio, and Virginia.

"Graduation rates are hugely consequential for what's going to happen to the country down the road," says William G. Bowen, a former president of Princeton University. Mr. Bowen wrote Crossing the Finish Line with Matthew M. Chingos, a graduate student in government at Harvard University, and Michael S. McPherson, president of the Spencer Foundation and a former president of Macalester College.

Among the book's central themes: Large disparities exist in graduation rates by gender, ethnicity, and family income, even after accounting for differences in standardized test scores and high-school preparation. That is not exactly news, but the book grounds those findings in an unusually rich set of data.

Imagine two students with, say, SAT scores of 1050 and B-minus high-school grade-point averages. The only visible difference between them is social class: One of them comes from a family whose income is below the national median, and neither of his parents completed college. The other comes from a family with above-median income, and both of his parents completed college. At all but one of the universities studied in Crossing the Finish Line, the more-privileged student would have been significantly likelier—between 6 and 17 percentage points likelier, depending on the institution—to graduate within six years. (The exception was the State University of New York at Stony Brook, where the graduation rate for less-privileged students is slightly higher than the rate for their better-off peers, after controlling for differences in high-school background.)

The gaps shaped by gender, race, and ethnicity are more complex. At the 21 flagship universities in the study, African-American women and Hispanic women did relatively well. Their six-year graduation rates (72 percent and 76 percent, respectively) were close to the rate for white men (75 percent). But African-American men and Hispanic men lagged far behind, with rates of 59 percent and 66 percent, respectively. Even after controlling for differences in test scores and high-school preparation, black male students' graduation rate was five percentage points behind that of white men.

"What we were struck by," Mr. Bowen says, "is just how pervasive and persistent and substantial these disparities are."

Money Matters, and So Do Grades

So what to make of all this? Mr. Bowen and his colleagues put forward two arguments that are likely to fuel debate for several years (much as Mr. Bowen did with his 1998 study of affirmative action, The Shape of the River, which he wrote with Derek Bok).

The first argument is that money matters. Some analysts have recently suggested that families are so eager for college educations that they will shrug off tuition increases, at least up to a point. But Mr. Bowen and his colleagues found that students from low-income families are quite sensitive to tuition levels, even at relatively inexpensive public universities. For students whose families are in the lowest income quartile, "our estimates imply that an increase in annual net price of $1,000 is associated with a decline of three percentage points in the six-year graduation rate and a decline of 4.5 percentage points in the four-year graduation rate," the authors write.

The second argument is that admissions offices should play down the SAT and ACT, and instead lean heavily on students' high-school grades, even if they come from weak school districts. High-school grades, the authors found, are much stronger predictors of college completion than SAT or other "ability test" scores. (In an aside, the authors suggest that admissions offices should also start to put more weight on Advanced Placement test scores, which also seem to be strong predictors of college graduation.)

The Alchemy of Selective Colleges

The tuition and SAT debates are, of course, evergreens of education policy, and they might still be running long after everyone who reads this article is dead. But some of the most provocative sections of Crossing the Finish Line have to do with a third, less familiar debate: Why, exactly, are graduation rates stronger at selective colleges?

At first glance, the answer might seem obvious. Selective colleges bring in stronger students, so of course they're going to be more likely to graduate. But Crossing the Finish Line, like several other recent studies, finds that even after controlling for students' test scores and high-school preparation, the pattern persists. In general, the more-selective universities in this study had higher graduation rates for students of all backgrounds, including those who might have seemed on paper like poor prospects.

To return to our hypothetical student with a 1050 SAT score and a B-minus average: If that student wants to maximize his odds of completing college, he should enroll in the most-selective institution to which he has been admitted. He should not worry that his classmates there will be better prepared and that he'll be in over his head.

Critics of affirmative action often talk about "mismatch," in which unprepared students flunk out of highly selective colleges. But Mr. Bowen and his colleagues believe that the opposite problem—"undermatch"—is actually more prevalent. Students are more likely to graduate if they enroll at the most-selective college that will let them through the door.

The question is why. "The truth of the matter is that no one understands very well why graduation rates for comparable students are stronger at selective colleges," Mr. Bowen says. "The pattern is relentlessly consistent, but no one has a firm idea why it is the case."

One plausible hypothesis has to do with peer effects. If you're at a selective college surrounded by smart, ambitious kids, you're likely to apply yourself to your studies because that's the local norm. The classic examples here are small liberal-arts colleges. "If you go to a small college, you're part of a quote-unquote class, and everyone expects to graduate," Mr. Bowen says. "There's a lot of pressure to live up to that expectation. That's less true at a big complicated place."

The universities in this study were generally much larger than liberal-arts colleges, but Mr. Bowen and his colleagues found echoes of the same phenomena. All else equal, the universities in this study with a higher proportion of students in residence halls had higher graduation rates. "Honors colleges" within the 21 public flagships also had significantly higher graduation rates than the campuses as a whole.

The Value of Great Expectations

Could less-selective universities somehow bottle some of the selective colleges' intensity, and thereby improve their graduation rates? "Our broad finding," Mr. McPherson says, "is that institutions that create a higher expectation of student success have higher graduation rates. And that seems like something that should operate up and down the line, for anyone who goes to school."

Maybe so—but how to accomplish that remains an open question. One approach that is drawing favor is the creation of "learning communities" for all first-year students, a strategy that has been promoted by, among others, Vincent Tinto of Syracuse University.

This is not the only high-profile study of graduation rates that has emerged this year. This month Washington Monthly released a set of rankings that are based partly on whether a college's graduation rate is unusually high or low, given the number of Pell Grant recipients on campus. And in June the American Enterprise Institute released its own analysis of graduation rates at colleges of various selectivity levels.

All three of these studies concede that graduation rates aren't the final word in college accountability. It would be a disaster, Mr. Bowen says, if colleges responded to government pressure to boost their graduation rates by handing out diplomas like candy. A full accountability system should also include measures of student learning and attention to what happens to students in the labor market, Mr. Bowen says.

But at the same time, he says, policy makers need to pay more attention to graduation rates, and to the many kinds of waste that are involved when people start college but do not finish. College graduation, much more than college attendance alone, seems to transform adults' lives. In the entering class of 1999, for example, students were much more likely to complete college if their parents were college graduates. By contrast, having a parent who started college but did not finish had very little positive effect.

"That implies that the social benefits of crossing the finish line, of getting that degree, are really larger than what happens to that one person," Mr. Bowen says. "These are intergenerational benefits."