Your Genes Shape Which Field You Study, Not Just How Far You Go

Educational choices profoundly shape careers, income, health, and social networks — yet genomic research has almost exclusively focused on years of schooling rather than the field chosen. This landmark study addresses that gap, performing GWASs of ten broad educational fields across 463,134 Europeans to reveal that genetics meaningfully influences the direction of education, not merely its level.

Using population-wide registry data from Finland (FinnGen), Norway, and the Netherlands, the team classified participants into ten fields — such as engineering, health, humanities, social sciences, and arts — according to international education classification standards. GWASs were run for each field with effective sample sizes ranging from approximately 40,000 to 317,000. Meta-analyses and within-family designs (using sibling data) were employed to disentangle direct genetic effects from confounding due to population stratification, indirect parental genetic effects, and geographic educational policies.

The study identified 17 independent genome-wide significant loci associated with 7 of the 10 fields. Average SNP-based heritability was approximately 7% per field. Critically, these genetic associations remained significant and largely unchanged after conditioning on polygenic scores for educational attainment (years of schooling), confirming that the signal captures field preference rather than general academic ability. Within-family analyses attenuated effect sizes somewhat but preserved the directional pattern, suggesting genuine direct genetic contributions alongside some environmental confounding.

To characterize the structure of field-specific genetic variation, the researchers applied genomic principal component analysis to the genetic covariance matrix across all ten fields. Two dominant dimensions emerged: a technical-vs.-social axis (STEM and business at one pole, health and education at the other) and a practical-vs.-abstract axis (vocational trades vs. humanities and arts). GWASs of these two components yielded genome-wide significant hits and demonstrated distinct patterns of genetic correlation with external traits — including openness to experience, agreeableness, risk tolerance, educational attainment, and various mental health outcomes.

The findings carry broad implications. They demonstrate that 'horizontal' educational stratification — differences in field choice among people with similar schooling levels — has a partly genetic basis, operating through vocational interests and behavioral dispositions. The two genomic dimensions map onto well-known psychological frameworks of interests (e.g., Holland's RIASEC model) and illuminate how individual genetic predispositions interact with social structures like gender norms and labor markets. The authors caution that heritability estimates reflect a specific historical and social context and do not imply genetic determinism; policy and social environments remain powerful levers for shaping educational trajectories.