Long-Term Impacts of a Full-Year Interleaved Mathematics Program (NewGlobe/Bridge, Nigeria)
NewGlobe / Bridge International (EdWorkingPaper 23-876) · 2023
grade Crctdeveloper-ledreplicatednumbers spot-checked
Sample
62 schools, grade 5
Population
Grade-5 students, Nigeria; full-academic-year interleaved vs blocked math program.
Design
Largest and most diverse FIELD test of interleaving — a full-year cluster RCT at scale.
Key findings
The largest at-scale field test of interleaving: a full-year program raised short-term retention only d=0.29 — about 3x smaller than the best controlled-classroom estimate, converging on the meta-analytic math value. Tempers the enthusiastic 'd=0.83' framing for real-world scale.
Genetic confound
Domain-skill outcome.
Replication notes
Largest field test; converges on Brunmair's 0.29-0.34 math estimate, well below Rohrer's controlled-classroom 0.83.
DOI / URL
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Interleaved vs blocked, short-term retention | d | 0.29 (vs Rohrer 2020 lab-adjacent 0.83 — ~3x shrinkage) | researcher-designed | short-term; field | active-alternative | end-of-treatment | domain-skill |
Cited by
- Interleaving (mixing problem types vs blocking)moderate supportconf: mediumgc: low