Interleaved Practice Improves Mathematics Learning
Rohrer, D., Dedrick, R. F., & Stershic, S. · 2015
grade Brctindependentunreplicated
Sample
126 7th-graders, 3 teachers/9 classes
Population
7th-grade real math classrooms, authentic content (graphing lines, slope).
Design
Classroom study over a 3-month practice phase; test delay (1 vs 30 days) randomly assigned. Proximal aligned test; meets WWC standards.
Key findings
First strong classroom math-interleaving result: benefit grew from d=0.42 at 1 day to d=0.79 at 30 days (near-immunity to forgetting). Caveats stated by authors: test used the same problem FORMAT practiced (proximal), and interleaving inherently confounds with spacing. They caution a small initial block may be optimal for a new skill.
Genetic confound
Domain-skill outcome.
Replication notes
First strong classroom demonstration; authors note effects are much smaller than lab (1.34) and confound with spacing.
DOI / URL
10.1037/edu0000001
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Interleaved vs blocked, 1-day test | d | 0.42 [0.07,0.77]; 80% vs 64% | researcher-designed | delayed 1day; classroom | active-alternative | end-of-treatment | domain-skill |
| Interleaved vs blocked, 30-day test — benefit GROWS | d | 0.79 [0.43,1.15]; 74% vs 42% | researcher-designed | delayed 1mo; classroom | active-alternative | under-1yr | domain-skill |
| Baseline group equivalence on standardized (NAEP) items | d | 0.12 (ns) | standardized | immediate; classroom | none | not-applicable | domain-skill |
Cited by
- Interleaving (mixing problem types vs blocking)moderate supportconf: mediumgc: low