In some districts, Gillani said, existing school boundaries are so jumbled that they “exacerbate” segregation, forcing students to travel farther to attend schools with children of their own race. These instances allowed a computer to easily draw a more efficient and integrated map.
Certainly much longer bus routes would be needed for a more dramatic reduction in segregation, as students would have to travel deeper into white, black, or Hispanic enclaves. But in an effort to please families, Gillani and his colleagues have stopped student commute times from growing by more than 50%, say 10 to 15 minutes. This displacement constraint meant that in Atlanta, for example, a city divided between white residents in the north and black residents in the south, racial integration would improve more in the center of the city and less in its northern and southern areas. .
To measure how well their redesigned maps have desegregated schools, Gillani’s team calculated how 98 school districts fared on a dissimilarity index, a 0-to-1 scale of the even distribution of white students among schools. Zero (0) means no segregation; all students go to a school that accurately reflects the make-up of the district. One (1) means complete separation. Imagine a half-white, half-black town with only two schools. If one school is all-white and the other all-black, that’s a 1. The 12% reduction in segregation achieved by the computer scientists in the simulation means that 98 school districts, on average, went from 0.33 on this index. However, this is an average and some neighborhoods remained quite close to 1, highly segregated.
“It’s not a huge change, but it’s still a move towards something more integrated,” Gillani said. To achieve this relatively modest degree of desegregation, about 20% of the three million primary school students in these 98 districts would have to change schools.
Gillani’s tool directly redraws school boundaries based on children’s races. But a 2007 Supreme Court decision, Parents Involved in Community Schools v. Seattle School District No. 1, has limited the ability of districts to consider race when voluntarily incorporating schools. (In contrast, race can and should be considered when enforcing desegregation court orders.) Today, many districts seeking to desegregate have shifted to using socioeconomic proxies for race, such as than family income. Gillani said he could instead adjust the map-drawing tool to optimize for socio-economic diversity.
As part of that work, Gillani and his colleagues created the www.schooldiversity.org website where anyone can see how elementary school boundaries could be improved in 4,000 districts. Almost every district in the country has more than one primary school. For now, the publicly visible tool is limited to seeing the current borders and how they might change under one set of conditions: maximizing racial integration while limiting travel time to 50% and school size to 15%.
When I researched my childhood school district of Simsbury, Connecticut, it was interesting to see the proposed changes. A school with a higher percentage of black and Hispanic children was literally cut in half in order to distribute those students to predominantly white schools. It would become a small school with less than 150 students – not economically practical.
Gillani said he intends to release his code and datasets, allowing other researchers and school districts to explore other parameters and make their own trade-offs. That is expected to happen later in 2022, when her paper, “Redesigning Attendance Boundaries to Promote Racial and Ethnic Diversity in Elementary Schools,” currently undergoing peer review in an academic journal, will be published.
Akeshia Craven-Howell told me she wished her community had access to a tool like this when she was associate superintendent at Charlotte-Mecklenburg Schools in North Carolina, where she helped redefine school boundaries, before he left in April 2022 and joined Bellwether Education Partners, a consultancy.
“What we were missing was a transparent way for the community to see how the boundaries might change under different scenarios,” Craven-Howell said.
“I think it can be helpful to show what’s possible,” Craven-Howell said. “It’s a powerful tool for community engagement. But a lot of messaging and communication work needs to happen in parallel. We need to give families confidence that their children will benefit, not just socially, but also at school.
Ultimately, segregation is a thorny political, cultural and social issue. Gillani and her co-authors acknowledge that their technocratic approach is not “enough” to drive policy change in the face of parents who oppose integration. But, they wrote, “it can help illuminate possible avenues of integration ‘at hand’ that districts and families may not have explored before.”