For decades, colleges and universities have pursued geographic diversity in their student bodies. Web pages proudly trumpet that this year the college has students "from all 50 states and from over 80 countries" (Harvard University) or that "the students come from throughout the United States and the world" (Stanford University). Even public universities use these formulations. At the University of Michigan in 2011, students came from "81 of 83 Michigan counties, all 50 states, and 54 countries."
Nobody bats an eye at geographic diversity. What's more, it seems to be on sound constitutional footing. Talent is everywhere, and excellence its magnet. The attraction of students from dispersed geographical origins toward any given campus itself proves that a campus' offerings merit attention.
But a second reason to be enthusiastic about geographic diversity gets less discussion. Social scientists have long distinguished between "bonding ties," which connect people who share similar backgrounds, and "bridging ties," which link people who come from different social spaces.
Since the 1970s, scholars have been aware that bridging ties are especially powerful for generating knowledge transmission; more recently, scholars have argued that teams and communities that emphasize bridging ties and learn how to communicate across their differences outperform more homogenous teams and communities in the development and deployment of useful knowledge.
For this reason, I think it's time to double-down on geographical diversity in college admissions. We should take it to the level of ZIP codes and, in particular, to the level of the ZIP+4 system. This ZIP code system divides the United States into geographic units as small as a city block or group of apartments.
Data management software is now sufficiently powerful that college admissions offices should easily be able to maximize ZIP+4 diversity from within a pool of applicants whose credentials exceed that college's entrance threshold.
And given current residential patterns - with their extremely high degree of socioeconomic, racial, ethnic and ideological segregation, well-described in Bill Bishop's book "The Big Sort" - geographic diversity at the level of ZIP+4 will bring other sorts of valuable diversity along with it. This is all within the context of a very old and unproblematic tradition of geographical diversity.
Some will object that such an approach might lead to the well-off moving strategically into neighborhoods with marginally less good provision of schooling, thereby displacing, as the likely beneficiaries of particular ZIP-code slots, those who are currently at more of a disadvantage in the college sweepstakes.
Indeed, researchers have documented such a phenomenon of strategic choice in Texas since the Top 10 Percent program, which guarantees admission to a public college or university in the state to students who are in the top 10 percent of their graduating high school class, was introduced.
A 2011 paper written for the National Bureau of Economic Research analyzed Texas school transitions between eighth and 10th grade and found, "Among the subset of students with both motive and opportunity for strategic high school choice, as many as 25 percent enroll in a different high school to improve the chances of being in the top 10 percent. Strategic students tend to choose the neighborhood high school in lieu of more competitive magnet schools."
But that, I would counter, is not bad news at all. Just as bridging ties are beneficial on college campuses, they are also valuable in schools and neighborhoods.
As Richard Rothstein of the Economic Policy Institute argued in a recent paper (in a volume I edited), ongoing racial residential segregation is one of the most important causes of low achievement in the public schools that serve disadvantaged children.
Other scholars, including Annette Lareau at the University of Pennsylvania, have made similar points about socioeconomic residential segregation. Getting students with more family and social resources back into neighborhood schools should help those schools.
Could a college admissions process structured around ZIP+4 diversity incentivize the well-to-do to opt for somewhat less exclusive neighborhoods and so on down the income ladder? If so, we would have begun to reverse the incredibly damaging socioeconomic and ethnic residential segregation that is a drag on this nation's ability to once again be first in the world in our level of collective educational attainment. What a relief that would be.
Danielle Allen is a professor of social science at the Institute for Advanced Study in Princeton.