Spacing Effects in Learning: A Temporal Ridgeline of Optimal Retention
Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. · 2008
grade Bquasi-experimentindependentunreplicatednumbers spot-checked
Sample
1,354 online participants; 32 facts each
Population
Adults, online panel; obscure trivia facts (proximal aligned recall/recognition), not curriculum.
Design
Large parametric experiment — the best source for the optimal-gap parameter. Crossed 26 gap x retention-interval cells; final tests to 1 year.
Key findings
Best quantitative source for spacing parameters: the optimal inter-study gap is non-monotonic (inverted-U) and scales to the target retention interval — roughly 10-20% of the desired retention interval. The spacing benefit is smaller (but reliable) at short intervals and large at long delays (recall gain peaks ~111% at a 70-day interval).
Genetic confound
Low: within-subject gap manipulation isolates scheduling from ability.
Replication notes
Numbers corrected per audit (earlier draft mislabeled recognition gains as recall and aggregate as 350-day). The 10-20%-of-RI rule is widely cited.
DOI / URL
10.1111/j.1467-9280.2008.02209.x
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Optimal gap vs 0-day gap at 350-day (1yr) retention interval | % gain; d | recall +77% (d=0.9); recognition +60% (d=2.1) | researcher-designed | delayed 1yr; lab/online | active-alternative | under-1yr | domain-skill |
| RECALL gain by retention interval (7/35/70/350 days) | % recall gain | 10 / 59 / 111 / 77% — NON-monotonic, peaks at 70 days | researcher-designed | immediate-to-1yr; lab | active-alternative | under-1yr | domain-skill |
| Optimal gap as proportion of retention interval | % of RI | ~20% at 1-week RI declining to ~5-10% (7%) at 1-year RI | researcher-designed | delayed; lab | active-alternative | under-1yr | domain-skill |
Cited by
- Spaced (distributed) practicestrong supportconf: highgc: low