In some districts, Gilani said, existing school boundaries are so degraded that they will “increase” isolation, forcing students to go farther to go to school with children of their own caste. These illustrations made it easier for a computer to draw more efficient and integrated maps.
To be sure, a much more dramatic reduction in isolation would require a much longer bus route as students would have to travel deeper into white, black or Hispanic enclaves. But in an attempt to appeal to the family, Gilani and his colleagues limited student travel time to more than 50 percent, say, 10 to 15 minutes. This travel restriction means that in Atlanta, for example, a city that is divided between white residents in the north and black residents in the south, ethnic solidarity will be further developed in the middle of the city and less in its northern and southern regions.
To measure how much their redesigned maps have divided schools, Gilani’s team calculated how 98 school districts worked on a disparity index, a 0-to-1 scale of how white students are evenly distributed among schools. Zero (0) means there is no isolation; All students go to a school that exactly reflects the structure of the district. One (1) means complete separation. Imagine a city that is half white and half black with only two schools. If one school is all white and the other is black, then that is 1. The 12 percent isolation that computer scientists have achieved in the simulation means that 98 school districts, on average, have improved the index from 0.39 to 0.33. However, this is an average and some have remained close to District 1, severely isolated.
“It’s not a big change, but it’s still a movement towards something that is more integrated,” Gilani said. About 20 percent of the three million primary school students in these 98 districts need to change schools to complete this relatively moderate level of isolation.
Geelani’s tool directly redraws the boundaries of the school based on the children’s race. But a 2007 Supreme Court decision, involving parents in Community School vs. Seattle School District No. 1, limited the ability of districts to consider ethnicity when they were voluntarily merging schools. (Conversely, race can and should be considered when complying with a disqualification court order.) Many districts today who want to segregate have switched to using socio-economic proxies for castes such as family income. Gilani said he could instead adjust the map-drawing tool to optimize for socio-economic diversity.
As part of this work, Gilani and his colleagues created the website www.schooldiversity.org where anyone can see how the boundaries of primary schools in 4,000 districts can be improved. It is almost every district in the country where there are multiple primary schools. At the moment, the publicly visible tool is limited to viewing current boundaries and how they may change under one condition: maximizing racial solidarity while limiting travel increases by 50 percent and school size by 15 percent.
When I looked back on my childhood school district of Simsbury, Connecticut, it was interesting to see the proposed changes. A school with a high percentage of Black and Hispanic children was literally halved so that those students could be scattered around the white school. It would be a small school with less than 150 students – not economically viable.
Gilani said he wanted to publish his code and datasets, allowing other researchers and school districts to explore other parameters and tradeoffs on their own. This is expected to happen after 2022, when her research paper, “Restoring Presence Boundaries to Promote Ethnic and Racial Diversity in Primary Schools,” is currently being published in a peer review in an academic journal.
Acacia Craven-Howell told me that she wished her community could have access to such a tool when she was an associate superintendent at Charlotte-Mecklenburg School in North Carolina, where she was involved before and in April 2022 when the school’s boundaries were rebuilt. Belvedere has joined Education Partners, a consulting firm.
“What we lacked was a clear way for the community to see how the boundaries could change in different situations,” said Craven-Howell.
“I think it might be helpful to show what’s possible,” Craven-Howell said. “It is a powerful tool for community participation. But there is also a lot of messaging and communication work to be done. We need to reassure families that their children will benefit not only socially, but also educationally.”
Finally, segregation is a complex political, cultural and social problem. Gilani and his co-authors acknowledge that their technocratic approach is not “sufficient” to change policy in the face of parents who oppose unification. But, they wrote, “it could help illuminate potential avenues of integration ‘within reach’ that both the district and the family have not explored before.”