The legacy of the small-schools movement: 20 years later, did it work?

We don't have a clean year-opened field for each school, so we use New York City's DBN-suffix convention as a proxy. Legacy comprehensive high schools predate the modernization and carry suffixes under 400 (e.g., DeWitt Clinton 10X440 is itself a legacy school despite the 440 suffix — the convention is approximate). Most schools created in the 2002-2008 small-schools wave have suffixes from 500 to 700. Sanity check: the heuristic identifies 23 closed schools with suffixes under 400, which matches MDRC's documented number of large "failing" high schools closed during 2002-2008 (William Howard Taft, Evander Childs, Walton, Christopher Columbus, John F. Kennedy, South Bronx, Stevenson, and others).

Howard Bloom and Rebecca Unterman's 2010-2014 lottery-based studies are the canonical evidence. They used New York City's high-school admissions lottery as a natural randomization: lottery winners admitted to small schools of choice graduated at rates roughly 8 percentage points higher than lottery losers who attended other schools. The effect held at six years out. The What Works Clearinghouse rated this evidence as meeting its highest standard. The studies focused on 123 academically nonselective small schools — a subset of the broader post-2002 wave, not all 216 — created to serve some of the city's most disadvantaged students.

At the cohort level, the post-2002 small schools post graduation rates 4-6 percentage points below legacy comprehensive schools across every year we have data. In 2022-23 (most recent year with comparable data), small schools graduated 80.2% of their seniors on time versus 85.0% at legacy comprehensive schools. Both cohorts trended upward through the COVID years — citywide graduation rates rose at most schools, partly because of grading-policy adjustments — but the gap stayed steady. This aggregate comparison does NOT contradict MDRC's lottery-based finding: small schools serve harder-to-graduate students, so a positive 8 pp lottery effect can coexist with a 5 pp aggregate deficit, simply because the small schools' student bodies are different.

Of the 242 post-2002 small high schools we identify, 207 are still open in 2024-25. The 35 that closed represent about 14% of the cohort, slightly higher than the 12% closure rate among legacy comprehensive schools but well below the 22% closure rate of the middle-era cohort. The small-schools wave proved more durable than the city's pre-2002 specialty-school additions but less durable than the legacy comprehensive schools — which are themselves the survivors of the 23 large closures during the program. The Urban Assembly network alone accounts for at least 20 schools across boroughs, most still operating. New Visions, Internationals Network, and Outward Bound were among the larger "intermediary" operators that survived.

We can show the cohort-level trajectories and survival rates. We cannot replicate MDRC's lottery-based identification, which is the only way to get a credible program-effect estimate. A reporter who wants to update MDRC's analysis would need the 2002-2024 admissions-lottery data, which the Department of Education holds but doesn't release publicly. Without it, the strongest claim we can make is that the program did not collapse — most small schools created in the wave are still operating, and their aggregate trend tracks the citywide post-COVID rise. The more interesting newsroom question may be which specific small schools became stable institutions (Urban Assembly schools, certain Bronx International Network campuses) versus which became their own next round of low-performers (the schools that account for the 35 closures since).