Many school districts offer families the opportunity to select from different schools through one of three enrollment systems: common application, decentralized, or ranked-choice. A working paper by Christopher Avery, Geoffrey Kocks, and Parag Pathak of Blueprint Labs at MIT finds that ranked-choice systems better match students—particularly disadvantaged students—to their top-choice schools.
Families seeking options beyond their assigned neighborhood school are typically required to apply for admission, a process that requires additional time and resources. In decentralized systems, families apply to each school individually. Common application systems streamline the process by allowing students to apply to multiple schools through a single application, though each school makes its own admissions decision. Ranked-choice systems also use a single application, but ask families to rank their preferred schools, resulting in a single admission decision to their highest possible preference.
To identify which system matches more students to their first-choice, the researchers examined six Boston charter schools from 2015 to 2020 during the district’s transition from a decentralized to a common application system. Using data on school capacity, application volume, and student preferences, they simulated how each system would assign students.
The ranked-choice model performed best, assigning 38.7 percent of students to their top-choice school, compared to 36.1 percent under decentralized admissions and 35.6 percent under common enrollment. The advantage was even larger for disadvantaged students: 41 percent of students receiving free and reduced-price lunch were assigned their top-choice, compared to 36 to 37 percent under the other two systems. Ranked-choice systems also reduced “mismatching”—instances where two students could switch schools and both receive a more preferred option. In Boston in 2017, about 9 percent of students were mismatched, but only 1 percent would have been in a ranked-choice system.
While the researchers took steps to make the findings generalizable beyond the Boston context, they note that the real performance of any enrollment system depends heavily on local context and choice supply and demand.