[ GLOSSARY ]
What is an effect size?
Definition
EvidenceEffect size
An effect size is a standardized measure of how big a difference is, usually in standard deviations (SD), reported as SD, Cohen's d or Hedges' g.
Why it matters for schools
It lets you compare results across studies. Some benchmarks: human tutoring raised PreK-12 learning by 0.288 SD across randomized trials; classroom quizzing gave g = 0.50 across 222 studies; a two-year Khanmigo trial found about 0.06-0.08 SD a year. A six-week pilot with 200 students can only detect about 0.40 SD, larger than published AI-tutoring effects.
[ FAQ ]
Questions and answers
What is a good effect size in education?
It depends on the outcome and test. As a benchmark, human tutoring averaged 0.288 SD across PreK-12 randomized trials.
What is the difference between d and g?
Both are standardized differences; g adds a small-sample correction.
Why can't a small pilot prove an effect?
With 200 students, only effects of about 0.40 SD or more are reliably detectable.
[ SOURCES ]
Sources
Facts checked Oct 18, 2026.
- 1
- 2Yang et al., classroom quizzing meta-analysis, Psychological BulletinMar 8, 2021 · doi.org
- 3Oreopoulos and Low, One Click Away, NBER working paper w356202026-08 · nber.org
[ AI FOR SCHOOLS ]
The AI your school can approve.