The Evidence on Teaching

How Much Does Education Improve Intelligence? A Meta-Analysis

Ritchie, S. J., & Tucker-Drob, E. M. · 2018

grade Bmeta-analysisindependentreplicated
Sample
142 effect sizes, 42 datasets, over 600,000 participants
Population
Quasi-experimental studies of education's effect on intelligence (policy changes, control-for-prior-IQ, regression discontinuity).
Design
Meta-analysis across three distinct quasi-experimental designs, all of which converge — the strongest evidence that schooling itself raises measured intelligence.
Key findings
The single most important corrective to a claim this database previously made too strongly. Education DOES durably raise measured IQ — roughly 1-5 points per year of schooling, persisting into old age. The defensible claim is therefore 'no durable gains in g', NOT 'no durable IQ gains'. Schooling is the one environmental intervention with a robust, lasting effect on intelligence test scores.
Genetic confound
Quasi-experimental designs (compulsory-schooling law changes, RD) that break the usual family-selection confound.
Replication notes
CRITICAL companion: Ritchie, Bates & Deary (2015) find the education-cognition association is NOT mediated by g — it consists of direct effects on specific skills. So durable gains occur on IQ TESTS without necessarily reflecting gains in g.
DOI / URL
10.1177/0956797618774253

Effects

OutcomeMetricValueMeasureTimingVsHorizonClass
Effect of one additional year of education on IQIQ points~1 to 5 points per year (central estimate ~1.5-2), and the effects PERSIST across the lifespanstandardizedadultbusiness-as-usualadulthoodg

Cited by