Can Redesigning Lottery-Based School Choice Promote Diversity of Enrollment Across a City?
A Stanford team works with the San Francisco United School Board in an effort to create more equitable schools
In August 2026, 5,000 new students will begin the school year at 58 elementary schools across one major U.S. city. If diversity, predictability, and proximity are the three most important criteria the city’s school district aims to balance, how might they “best” divide up the kids?
While this might sound like an intimidating word problem from an upleveled SAT exam, it is in fact a real-life challenge currently facing the San Francisco Unified School District (SFUSD).
In 2011, the San Francisco school board implemented something called school choice whereby residents could “choose” (which really means “apply”) to send their child to any school in the city, a system sometimes referred to as a “school lottery.” However, since the selection process was perceived by the public as opaque and unpredictable, parents especially haven’t been entirely happy with the results.
Additionally, by 2018 it was clear San Francisco’s schools were becoming more segregated, a phenomenon seen in other large school districts across the country that had also embraced unrestricted school choice.
The SFUSD school board then began a process to more clearly articulate their priorities and explore policies that would support them. They passed a resolution in 2020 to update the way students are assigned to schools and outlined three priorities: increase racial and economic diversity in the classroom, offer a high degree of predictability of where a child would go to school (thus reducing anxiety for parents), and ensure the school was not too far from a child’s home.
Irene Lo, an assistant professor at Stanford in the department of Management Science & Engineering, who had previously researched student assignment issues for New York City, learned about these efforts and wondered if her expertise might be applicable in this situation. “A colleague and I went to several school board meetings and realized our understanding of the research and our ability to model outcomes could be useful in helping them make an informed decision. We approached a couple of board members and began working with SFUSD in 2019,” said Lo.
This turned out to be fortuitous timing. “Stanford’s involvement has really helped us work iteratively toward solutions that we think will faithfully enact the new policy, help SFUSD work toward its long-term goals, and ultimately benefit students, families, and schools,” said Joseph Monardo, Director of the Enrollment Center at SFUSD.
Modeling the Options: Putting Data to Work
Lo’s team, Equitable Access to Education (a 2022 Stage 2 impact lab), began by analyzing existing demographic data. They found significant overlap between racial and socio-economic segregation in the district. To determine which approach would break up this map and make for a more equitable system, they created a choice model. Using historical choice data, the team developed a simulation tool that predicted how various policies would perform and then presented the results to the board.
The board determined the policy that would best achieve their goals was to divide the district up into zones, restricting students to choose only from the approximately 10-12 schools within their zone. In addition, the policy would include what SFUSD calls “diversity categories,” where spots are set aside for certain populations related to equity priorities, like homeless children or children in the foster care system.
Next, Lo’s team created a zone optimization tool to help the board model the outcomes with various zone maps and diversity criteria. In collaboration with colleagues at the Stanford Graduate School of Education and drawing on research from the Stanford Educational Opportunity Project, they created a suite of measuring tools to help track things like predictability, proximity, diversity, choice, and equity of access. They are now further developing visualization tools so administrators and community members can more easily understand the complex data and the results of the modeling.
How Will the City Respond?
How the public will respond remains an open question. The school board has voted to implement the policy for the 2026-27 school year and is currently working with city partners and parent advisory groups to draft proposals for zones, which they will take to the larger community for their input in 2024.
“The biggest challenge is going to be shifting from a system where choice is relatively unconstrained to one with very obvious constraints,” says Monardo. “Currently, families can apply to any school in the district. With this new policy, which will be implemented at the elementary school level only in 2026, families will be limited to applying to only the schools in their zone, or to the subset of citywide schools, like K-8 schools. This change will be understandably difficult for some people, but those constraints support the policy goals of diversity, proximity, and predictability. We are engaging the community throughout the process, and we know most people want a fair and just system that works for every family in the city.”
As far as whether or not this approach would work for other school districts, Lo says, “Our hope is that the approach is generalizable and scalable. The final decisions of each school district may be different based on their particular priorities, concerns, and constraints, but the tools and process will be very useful for taking a data-driven approach to resolving these complex questions.”
Jenn Brown is a freelance writer and non-fiction book editor.
Open Question is a forum for commentary about social problems and methods of tackling them in the drive for progress.