A few weeks ago, the Shanker Institute published an analysis of segregation by race and ethnicity in D.C. metro area schools (including D.C. proper, Alexandria City, Arlington and Fairfax County in Virginia, and Montgomery and Prince George’s Counties in Maryland). 

The report, written with my co-author Bilan Jama, presents multiple measures to characterize segregation within each of these six districts and across the entire metro area, but it also focuses on segregation between districts. This is a very important distinction for understanding segregation, particularly in large metropolitan areas. Put simply, students may be systematically sorted into schools within each district (e.g., white students may be concentrated in some schools while African American students are concentrated into others), but they might also be sorted between districts (e.g., some districts may serve mostly black, white, Asian or Latino students, while others serve very few such students). Both of these factors affect the racial and ethnic composition of schools, and so both contribute to or attenuate segregation in the metro area as a whole.

The D.C. metro area is an excellent context for this kind of analysis because it is so racially and ethnically diverse, with relatively strong representation of white (26.5 percent), black (34.7), Hispanic (27.2), and Asian students (11.6). This diversity is the “raw material” for truly diverse schools. Unfortunately, we found this not to be the case, and the underlying reasons why are interesting.

We reach two seemingly contradictory conclusions:

1. Segregation is quite low within all but one of our six D.C. metro area districts (the sole exception being D.C. proper, which exhibits high segregation for virtually all racial and ethnic comparisons);
2. But the metro area as a whole is also quite segregated, particularly when it comes to the separation of black from white and Asian students.

So, how can segregation across the metro area be somewhat high when all but one of its six districts – i.e., D.C. proper, which serves a relatively low percentage of the metro area's students – exhibits low segregation? The answer, as you probably already know, is that there is substantial segregation between these districts.

One easy way to think about within-district versus between-district segregation is to assume you had magical desegregating power. This power would allow you to move students between schools within each district – not having to worry about school zoning laws, parental protests, lawsuits, transportation costs or anything else (that’s the magic). You are not, however, able to move students between districts. That is the price of your power. 

Using your newfound magic, you fully integrate all six D.C. metro area districts internally. That is, you rearrange students within each district such that every school has the same exact racial and ethnic composition as the district as a whole (i.e., each school in each district would all have the same percentage of white, black, Latino and Asian students as all the other schools in the district, but those percentages would be different for each district, since you can’t move students across district boundaries). 

So, if you exercised this power, what effect would you have on segregation in the D.C. metro area? How close could you get to the wonderful situation in which every single school in the metro area has the same racial and ethnic percentages as the metro area overall?

Our results show that you could get about halfway there (53.5 percent).

This would be a massive desegregation victory. Still, imbued as you are with magical powers, you might be a bit disappointed. And you might be even more disappointed if you applied this power only to white and black students – in this case, you could only reduce total black/white metro area segregation by 35.4 percent. Again, a big victory for students, but not exactly a superhero-level result.

Why is your effect so limited? The reason is that your power only allows you to affect within-district segregation. Even with all the magical, obstacle-free shuffling of students between schools within each district, you can’t get close to perfect metro area school integration – i.e., every school has the same percentages as the metro area overall – because students are unequally distributed by race and ethnicity across district lines. For example, Prince George’s County serves about 22 percent of all students in the area, but roughly 40 percent of its African American students. In very crude terms, no matter how much you shuffle students between schools within Prince George’s County, you still won’t not have enough white, Latino, and Asian students to reach the overall metro area percentages.

So, your power to perfectly solve within-district segregation would still leave you far short of perfect metro area multiracial integration (and even shorter if you were only addressing black/white segregation). Without a sidekick who has the between-district desegregation power, you are severely limited. And this is the case not only in the D.C. metro area, but nationally as well (Reardon et al. 2000).

This matters for policy because – for several reasons, both legal and logistical –virtually all modern efforts to desegregate schools have been within-district efforts. They have generally been very successful, and there should be many more of them. But a large part of the problem, and therefore the solution, is also between districts, particularly in large metropolitan areas such as Washington, D.C., where there is the greatest potential for true racial and ethnic diversity in schools.