Injecting Charter School Best Practices into Traditional Public Schools (Houston field experiment)
Fryer, R. G. · 2014
grade Brctdeveloper-ledreplicatednumbers spot-checked
Sample
20 HISD schools (9 secondary, 11 elementary); Denver/Chicago replications
Population
Low-performing Houston district schools.
Design
The five-tenet No-Excuses bundle transplanted into district schools.
Key findings
The No-Excuses production function transplants: tutoring + extended time + data-driven instruction + teacher feedback + high expectations delivers charter-sized math gains inside a normal district budget. Reading barely moves anywhere.
Genetic confound
School-level assignment; elementary randomized.
Replication notes
Verified incl. cost corrections. Replicated non-experimentally in Denver/Chicago. Talent supply was the binding constraint (300+ interviews for 19 principals).
DOI / URL
10.1093/qje/qju011
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Math | SD/yr | +0.15-0.18 on state tests; +0.10 on low-stakes Stanford-10 — matching top charter effects | standardized | annual | business-as-usual | end-of-treatment | domain-skill |
| Reading | SD/yr | ~0 (the 4:1 math:reading ratio recurs across the whole charter literature) | standardized | annual | business-as-usual | end-of-treatment | domain-skill |
| Cost | $/student/yr | $355 elementary (IRR 28-29%); $1,837 secondary (IRR 13-13.5%; $1,100 without tutoring, IRR 16.5-17.1%) | mixed | n/a | none | not-applicable | attainment |
Cited by
- School spending, charters, and what school-level choices actually move outcomesmixedconf: highgc: low