The Persistence of Early Childhood Maturity: International Evidence of Long-Run Age Effects
Bédard, K., & Dhuey, E. · 2006
grade Bnatural-experimentindependentmixed
Sample
TIMSS grade-4 and grade-8 cohorts across OECD countries; Canada/US attainment data
Population
International student cohorts with varying school-entry cutoffs.
Design
Exploits entry-cutoff variation across OECD countries. LIMITATION: confounds age-at-test with cohort position — precisely what Black/Devereux/Salvanes later resolves.
Key findings
The classic relative-age result: being youngest in the class costs 4-12 percentiles at grade 4, and the gap roughly halves by grade 8. It cannot distinguish a real starting-age effect from being younger when tested.
Counterfactual
Youngest versus oldest children within the same grade cohort.
Genetic confound
Cutoff-based; the concern is measurement artifact rather than genetic confounding.
Replication notes
Widely replicated as a relative-age pattern; its causal interpretation is superseded by the Norwegian decomposition.
DOI / URL
10.1093/qje/121.4.1437
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Grade-4 test scores, youngest vs oldest in cohort | percentiles | -4 to -12 percentiles | standardized | grade-3 | none | not-applicable | domain-skill |
| Grade-8 test scores | percentiles | -2 to -9 percentiles — the gap roughly HALVES from grade 4 to 8 | standardized | grade-6+ | none | not-applicable | domain-skill |
| University attendance | direction | youngest cohort members less likely (Canada/US) | mixed | adult | none | adulthood | attainment |
Cited by
- School starting age, relative age, and academic redshirtingmixedconf: highgc: low