Mathematics Mastery (Ark): EEF primary and secondary cluster RCTs
Jerrim, J., & Vignoles, A. (independent evaluators) · 2015
grade Brctindependentreplicatednumbers spot-checked
Sample
10,114 pupils / 127 schools (two parallel RCTs)
Population
English Year 1 and Year 7 pupils; independent standardized math tests.
Design
Singapore-inspired 'teaching for mastery' curriculum; first year of adoption. NOTE: no strict mastery gating (topics not delayed), so tests the curriculum variant, not Bloom LFM.
Key findings
The modern at-scale test: mastery-branded curriculum on independent standardized tests = +0.07 pooled — 5-10x below the 1980s meta claims, exactly where Slavin's analysis predicts. Cheap (£50-131/pupil), so possibly still cost-effective.
Genetic confound
Minimal (cluster RCTs).
Replication notes
Two parallel RCTs, consistent small positives. EEF: findings 'substantially lower than the average effects seen in the existing literature on mastery approaches.'
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Primary (Year 1) math | ES | +0.10 (CI -0.01 to +0.21; 10%-level only) | standardized | end of year 1 | business-as-usual | end-of-treatment | domain-skill |
| Secondary (Year 7) math | ES | +0.06 (ns) | standardized | end of year 1 | business-as-usual | end-of-treatment | domain-skill |
| Pooled (both trials) | ES | +0.073 (CI 0.004-0.142; p<.05) | standardized | end of year 1 | business-as-usual | end-of-treatment | domain-skill |
Cited by
- Curriculum choice as a school-level levermixedconf: highgc: low
- Mastery learning (teach → test → reteach to criterion → advance)mixedconf: highgc: low
- Math curriculum choice — how much does the textbook matter?mixedconf: highgc: low