My family is in the process of leaving Fairfax County, Virginia, and schools are part of the reason. Naturally, as we look for new places to live, we’re paying a lot of attention to the schools elsewhere, in both Covid and academic terms. As any parent who has moved knows, the major real-estate sites make it easy to get information about local schools. When you look at a house listing, you can immediately see what schools the home is zoned to, along with the schools’ “ratings”—typically from a third-party site such as GreatSchools, which contains more detailed information about school performance and even raw racial demographics.

Two distinct new studies examine how this type of information about schools works: what it actually means and how parents use it. The studies are complicated, and their results raise complicated questions.

In a National Bureau of Economic Research working paper, Joshua Angrist and several coauthors ask why school ratings correlate with racial demographics. Key to the paper is that there are (at least) three different ways to think about a school’s overall academic quality. Two are commonly used in the various rating systems today: “level” measures simply quantify how well kids are doing on standardized tests and the like, while “progress” measures calculate how much kids improve each year that they attend a school. However, such measures don’t necessarily capture a school’s effect on student outcomes. More-advantaged kids tend to get higher test scores, and also to make progress somewhat more quickly, regardless of which school they go to. A “good school,” by a level or progress measure, might just be a school full of rich white kids, not a school that gets the best out of its students.

The third way of grading schools, though, considers the “value-added” they create—above and beyond the advantages a student comes in with. To capture this, an observer would ideally conduct an experiment, randomly assigning kids to different schools and measuring how well they did. Of course, many U.S. jurisdictions simply send kids to school based on where they live, but New York City and Denver let students rank their own preferences for schools before assigning them to one based on an algorithm. Using complicated statistical methods, Angrist, et al. study situations in these cities where similar kids were sent to different middle schools in basically random fashion, in a way that resembles what would happen in a real experiment.

The upshot is that schools’ nonwhite shares are not measurably correlated with the actual value the schools add to student outcomes. This is a single working paper, using highly complex methods on data from just two places. But, taken at face value, that result has compelling implications—some supportive of narratives of “systemic racism” and others decidedly not.

As the authors note, if true school quality has no connection with schools’ racial compositions—and yet common quality measures give lower grades to heavily minority schools—then the standard measures are discouraging integration by steering parents away from minority schools with misleading information. The results would also seem to imply, however, that cities (or at least these cities) are doing a better job than commonly appreciated of providing schools of similar quality to student populations of varying racial compositions.

The paper further shows that progress measures can be adjusted so that they no longer correlate with race but still correlate with value-added. But it’s not clear that this approach is ideal, either—at least if the goal is to give parents the information they need, rather than to engineer the decisions they make. As the authors themselves note, parents “may respond more to peer characteristics than to value-added.” And it’s reasonable for parents to want their kids to attend school with peers who perform highly and make lots of progress each year, both of which correlate with race.

What about racial demographics themselves, which are also easily available to parents? That’s the subject of another recent study, from Chantal A. Hailey and published in Sociology of Education.

Hailey ran an experiment on parents and students in the process of choosing a New York City high school. Her survey provided respondents with several bits of randomly generated information about hypothetical schools, including the graduation rate and some measures of safety—though, notably, not any measure of test scores—as well as whether the schools were majority-white, majority-black, majority-Hispanic, or racially diverse. She then asked her subjects how willing they’d be to attend these institutions.

The bottom line was that “white and Asian families preferred white schools over black and [Hispanic] schools, [Hispanic] families preferred [Hispanic] schools over black schools, and black families preferred black schools over white schools.” Many of these racial preferences, on average, amounted to half a point or more on a seven-point scale, though the preferences of whites and Asians were generally stronger than those of blacks and Hispanics. Moreover, white and Hispanic parents seemed to care about race more than kids from those groups did, yet the opposite was true for black respondents. (The full breakdown of the parents-versus-kids results is especially intriguing: white students preferred white schools without distinguishing much among schools dominated by the various other groups the way their parents did; Hispanic parents wanted to avoid black schools, while Hispanic students gravitated to same-race schools; and black parents didn’t have much racial preference, while black students wanted to avoid white schools.)

In general, then, parents and kids alike care about the racial composition of schools. That undoubtedly contributes to segregation, whether these preferences are put into effect when parents choose houses or when they choose schools. This raises the uncomfortable question of what kinds of preferences among parents and students are permissible, at least under modern racial mores. Rank bigotry aside, what if parents just want their kids to attend school with a “critical mass” of same-race peers so they don’t feel isolated? What if they want a school whose demographics reflect those of the city or country as a whole? What if they use race as a proxy for other school characteristics that aren’t easily available?

There are no easy answers. But the politically incorrect details about the places we can choose to live, and their schools, are only a click or two away.

Photo by Jon Cherry/Getty Images


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next