NYC's DOE uses a Gale-Shapley matching algorithm to place ~75,000 eighth-graders a year across 1,600+ schools; 92% got a top-five choice in 2026. A 2025 state audit found no written policies and lapses prioritising homeless and low-income students despite legal requirements.
Students in temporary housing without consistent required priority (2023-24)
31 of 39
Sampled likely low-income students misidentified
Details
Maturity
Established
Promoter
New York City Department of Education
Period
2003–ongoing; state audit 2023–2025; 2026 cycle results
Keywords
school admissions, algorithmic matching, administrative operations, public education
Context
New York City's Department of Education has used a centralised Gale-Shapley deferred-acceptance algorithm to assign students to high schools since 2003-04, now processing roughly 900,000 students across 1,600 schools.
Activities
For 2026-27, the DOE reported that 92% of applicants received an offer from one of their top five ranked choices, up from 90% the previous cycle.
Results
A July 2025 audit by the New York State Comptroller found the DOE had no written policies governing the process; roughly 7,000 students in temporary housing were not consistently given required priority consideration in 2023-24, and in a sample of 39 students, 31 were likely low-income but not correctly identified for priority.
Conclusions
The matching algorithm itself performs as designed at genuine scale, but the surrounding data pipelines, manual-override processes and transparency obligations received far less rigour, a gap surfaced only through independent audit.
Implementation
Indicative cost
Low (< €50k)
Time to results
Long (> 3 years)
Staffing & skills
NYC DOE admissions staff, Manual-override reviewers
Conditions for success
Accurate address/income data pipelines
Documented manual overrides
Local Law 72 transparency compliance
Common failure modes
No written policies for a process serving hundreds of thousands of students
200+ manual overrides improperly documented or missing in one year
Do you run this practice?
Claim it —
verified implementers get a public contact pathway and can propose corrections.
Data sources
Where this practice's information was retrieved from, and when.
Chile's SAE, built with Universidad de Chile researchers using a Nobel Prize-winning Deferred Acceptance algorithm, matched 274,000+ pre-K–12 students to …