The Evidence on Teaching

The Distributed Practice Effect on Classroom Learning: A Meta-Analytic Review of Applied Research

Mawson, W., & Kang, S. H. K. · 2025

grade Cmeta-analysisindependentunreplicatednumbers spot-checked
Sample
22 reports / 31 effect sizes / N>3,000
Population
K-12 and university classrooms; curriculum content.
Design
Classroom-ONLY spacing meta — the key lab-to-classroom translation source. Very high heterogeneity (I2=92%); retention intervals capped at 42 days.
Key findings
Real-classroom spacing effect d~0.54, well below lab estimates, with huge heterogeneity. Practical translation: fixed ~7-day gaps are most reliable, expanding intervals no better and sometimes negative, ~3 spaced exposures suffice ('less is more'), and effects can go NEGATIVE in primary grades.
Genetic confound
Low-moderate: mostly within-class/randomized; heterogeneity reflects design/population, not genetics.
Replication notes
Best estimate of the real-classroom spacing effect; below lab-inclusive estimates.
DOI / URL
10.3390/bs15060771

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Distributed vs massed in real classrooms (longest RI)d0.54 [0.31,0.77] (vs lab-inclusive ~0.85)mixeddelayed up to 42d; classroomactive-alternativeunder-1yrdomain-skill
Schedule moderatordirectionfixed ~7-day gaps most consistent; EXPANDING intervals smaller/sometimes negativemixeddelayed; classroomactive-alternativeunder-1yrdomain-skill
Age/level moderatordirectionsecondary > primary; PRIMARY grades produced the most NEGATIVE effectsmixeddelayed; classroomactive-alternativeunder-1yrdomain-skill

Cited by