Peter Savelyev, Kegon Teng Kok Tan

We use the high IQ Terman sample to estimate relationships between education, socioemotional skills, and health-related outcomes that include health behaviors, lifestyles, and health measures across the lifecycle. By both focusing on a high IQ sample and controlling for IQ in regression models, we mitigate ability bias due to cognitive skill. In addition, we control for detailed personality measures to account for socioemotional skills. We model skills using factor analysis to address measurement error and adopt a powerful stepdown procedure to account for multiple hypothesis testing. We find that among high IQ subjects, education is linked to better health-related outcomes, in contrast to previous evidence. Conscientiousness, Openness, Extraversion, and Neuroticism are linked to various health-related outcomes across the lifecycle. Furthermore, we find that accounting for a comprehensive set of skills, measurement error, and multiple hypothesis testing not only provides greater confidence in several established relationships but also generates novel results.

JEL Codes
I12: Health Production
J24: Human Capital; Skills; Occupational Choice; Labor Productivity
college education
Big Five
health behavior