Three Cohorts
DataNLSY79
N = 12,686 · Birth years 1957–64
10 ASVAB subtests administered in 1980. 30 survey waves from 1979 to 2022 provide decades of outcome data spanning early career through retirement age.
NLSY97
N = 8,984 · Birth years 1980–84
12 CAT-ASVAB subtests administered in 1997–98. 20 survey waves from 1997 to 2022 track outcomes from adolescence through the late 30s.
CNLSY
N = 11,545 children of NLSY79 mothers
3 PIAT subtests (math, reading recognition, reading comprehension). 16 waves from 1986 to 2014 covering childhood through early adulthood.
Key Findings
ResultsEarnings return tripled
Per-SD cognitive return on earnings rose from ~$4,900 in NLSY79 to $21–27K in NLSY97 at comparable ages, reflecting the growing economic premium on cognitive ability across cohorts.
Outcome modelsBroad outcome prediction
2–5× higher odds of BA+ degree per SD of g. 1.9–2.1× employment odds. OR 0.65 for arrest and 0.48 for incarceration. Associations span education, labor, health, and family domains.
Full resultsGreater male variability
Males are more variable in cognitive scores across all three cohorts. Observed variance ratio VR = 1.32–1.36, consistent across both ASVAB and PIAT batteries.
Measurement detailsAt a Glance
VisualizationPer-SD Earnings Return by Cohort (Age-Matched)
Dollar increase in annual earnings per one standard deviation of gproxy, at overlapping ages (36–40).
NLSY79 measured at year 2000 (ages 36–43). NLSY97 measured at 2019 (ages 35–39). Age-matched window: 36–40. Difference: +$14,921 (p < 1e-41).
Explore
NavigationMeasurement
Factor extraction, invariance testing, and score construction across ASVAB and PIAT batteries.
View measurementOutcomes
Education, income, employment, wealth, health, and criminal-justice outcomes predicted by cognitive ability.
View outcomesSubgroups
Sex differences, variability ratios, and subgroup comparisons across demographic categories.
View subgroupsFamily
Intergenerational transmission of cognitive ability from NLSY79 mothers to CNLSY children.
View familyMethods
Technical details on statistical methods, model specifications, and analytic decisions.
View methodsFAQ
Frequently asked questions about the project, data sources, and interpretation of results.
View FAQ