The Evidence on Teaching

Testing (Quizzing) Boosts Classroom Learning: A Systematic and Meta-Analytic Review

Yang, C., Luo, L., Vadillo, M. A., Yu, R., & Shanks, D. R. · 2021

grade Bmeta-analysisindependentreplicatednumbers spot-checked
Sample
222 studies / 48,478 students
Population
K-12 through higher-ed and professional; authentic classroom content and assessments.
Design
Largest CLASSROOM-only testing-effect meta. Mixed RCT + quasi designs; mostly researcher-aligned course assessments.
Key findings
Strongest classroom anchor: g=0.499 across 222 studies/48k students. Effect partly transfers to related untested material (0.32) but ~35% smaller; aligned formats inflate it (0.53 vs 0.40); quizzing barely beats active elaborative controls (big gains are vs no-activity); feedback helps. Real in classrooms but on aligned course tests.
Genetic confound
Achievement/retention outcome; no g or ability-interaction claim.
Replication notes
Numbers verified. Converges with Adesope, Rowland, Agarwal 2021.
DOI / URL
10.1037/bul0000309

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Overall achievement gain from quizzingg0.499 (medium)researcher-designedmostly delayed course exams; classroombusiness-as-usualunder-1yrdomain-skill
Transfer to related but UNTESTED materialg0.321 (~35% smaller than tested material)researcher-designeddelayed; near-transferbusiness-as-usualunder-1yrnear-transfer
Format-match (aligned vs mismatched)galigned 0.531 vs mismatched 0.399 (measure-alignment inflation)researcher-designeddelayedactive-alternativeunder-1yrdomain-skill
Feedback vs no feedbackg0.537 vs 0.374 (feedback helps)researcher-designeddelayedactive-alternativeunder-1yrdomain-skill
Vs active elaborative controls (concept mapping, notes)directiononly MINIMAL advantage — big gains are mostly vs no-activityresearcher-designeddelayedactive-alternativeunder-1yrdomain-skill

Cited by