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

Summary

Used New York City's DBN-suffix convention as a cohort proxy for the 2002-2008 Bloomberg-era small-schools wave (suffix ≥ 500), validated against MDRC's documented count of 23 large 'failing' high schools closed during the program. Pulled per-cohort survival rates and average four-year graduation-rate trajectories.

Data sources

• Database tableschools

  One row per New York City public school (DBN, name, district, borough, school_type, grade_band, latest_enrollment, proclivity_decile, proclivity_score, closed_at, admission_category).

• Database tableschool_year_metrics

  Long-format per-school per-year metric facts (school_dbn, year, metric_key, subgroup, value, suppressed). Loaded from DOE/NYSED public files via scripts/loaders/*.

• Loader scriptscripts/loaders/_small_schools_query.ts
• Cached outputdata/stories/small-schools-data.json

Steps

1. Split high-school cohort by DBN suffix. Sanity check: cohort with suffix <400 should yield ~23 closed schools, matching the MDRC-documented 2002-2008 closure wave.

   SELECT CASE
            WHEN (SUBSTRING(dbn FROM 4 FOR 3))::int >= 500 THEN 'small_post2002'
            WHEN (SUBSTRING(dbn FROM 4 FOR 3))::int >= 400 THEN 'medium'
            ELSE 'legacy_comprehensive'
          END AS cohort,
          COUNT(*), COUNT(*) FILTER (WHERE closed_at IS NOT NULL)
     FROM schools WHERE grade_band::text='HS'
    GROUP BY cohort;

2. Average per-cohort four-year graduation rate by year, restricted to still-open HS schools.

   SELECT cohort, year, AVG(value), COUNT(DISTINCT school_dbn)
     FROM schools s JOIN school_year_metrics m ON m.school_dbn=s.dbn
    WHERE s.grade_band::text='HS' AND s.closed_at IS NULL
      AND m.metric_key='graduation_rate_4yr' AND m.subgroup='ALL' AND m.suppressed=false
    GROUP BY cohort, year;

References

• Bloom & Unterman, 'Sustained Positive Effects on Graduation Rates Produced by NYC's Small Public High Schools of Choice' (MDRC, 2014)
• MDRC — NYC Small Schools of Choice Evaluation
• Chalkbeat: Bloomberg early school closures benefited future students (2015)

Reproduce

Run `npx tsx scripts/loaders/_small_schools_query.ts > data/stories/small-schools-data.json`. To strengthen the cohort definition: pull MDRC's 123-school study roster (in their 2014 sustained-effects publication) and replace the DBN-suffix heuristic with a name-by-name DBN list. To replicate MDRC's lottery-based effect estimate, the city Department of Education's Office of Enrollment holds the high-school-admissions lottery data; access typically requires a research-data agreement.