Heritability of Intelligence: What Twin and Genome Studies Show

Abstract

Intelligence, measured by IQ tests and related cognitive assessments, has one of the highest heritabilities of any behavioral trait — estimates from twin studies consistently range from 50–80% in adults. But heritability is a population statistic, not a deterministic statement about individuals, and it says nothing about whether group differences in intelligence are genetic in origin. Recent genome-wide association studies have identified thousands of genetic variants associated with educational attainment, confirming polygenic architecture but explaining only a fraction of the variance.

Background

Few topics in science are as politically fraught as the genetics of intelligence. The research exists, is substantial, and produces consistent findings — but those findings are routinely misunderstood in both directions: overclaimed by advocates of genetic determinism and dismissed entirely by those who fear genetic explanations for social inequality. Understanding what heritability actually means is essential to evaluating the evidence.

Heritability is defined as the proportion of variance in a trait within a population that is associated with genetic differences in that population. It is not a fixed property of the trait; it depends on the range of genetic and environmental variation in the population studied. It says nothing about how much of an individual's trait is "due to" genes. And it says nothing about whether differences between populations reflect genetic differences.

The Evidence

Twin studies consistently find high heritability of adult intelligence. Studies comparing identical twins (monozygotic, MZ) who share nearly all their DNA to fraternal twins (dizygotic, DZ) who share approximately half their DNA allow estimation of genetic influence. For intelligence, MZ twins are far more similar than DZ twins, even when raised apart. A meta-analysis of 11,000 twin pairs found average heritability of general cognitive ability of about 0.50 in children, rising to about 0.80 in adults (Polderman et al., 2015, Nature Genetics). The increase with age is one of the most replicated findings in behavioral genetics — genetic influences on intelligence become more, not less, prominent as individuals age and select environments that reflect their genetic dispositions.

Shared environment matters more in childhood than adulthood. In children from disadvantaged backgrounds, shared environment (family and home environment) explains a larger proportion of intelligence variation. Turkheimer et al. (2003, Psychological Science) found that in low-income families, shared environment explained most of the variance in IQ; in high-income families, genetics explained most. This interaction — heritability is lower in impoverished environments — is theoretically and practically important: when environmental constraints are severe, genetics has less room to express itself.

Genome-wide association studies confirm polygenic architecture. Rather than a few genes of large effect, intelligence is influenced by thousands of genetic variants each with tiny individual effects. Lee et al. (2018, Nature Genetics) conducted the largest GWAS of educational attainment (a proxy for cognitive ability) in 1.1 million individuals and identified 1,271 independent genetic variants. The combined "polygenic score" explained only about 11–13% of variance in educational attainment — far less than the heritability estimates from twin studies. The gap (the "missing heritability") reflects limitations of current GWAS methods, not a contradiction of the twin findings.

Heritability estimates cannot be extrapolated to group differences. The most important and most frequently misapplied limitation of heritability evidence: even if intelligence is 70% heritable within populations, this tells us nothing about whether differences between populations are genetic in origin. Lewontin's famous corn analogy illustrates this: genetically identical seeds grown in poor soil will be shorter than those grown in good soil, even if within-group variation is largely genetic. Persistent environmental differences between racial and socioeconomic groups — including access to nutrition, education, health care, and freedom from discrimination — confound any between-group comparison.

Flynn effect shows substantial environmentally-driven change. IQ scores have increased substantially across generations in all tested countries — the Flynn effect — at a rate far too fast to be explained by genetic change. Average IQ in the US increased approximately 30 points over the 20th century. This demonstrates that whatever is measured by IQ tests is highly responsive to environmental influences at the population level, even if within-generation variation is substantially genetic.

Counterarguments

Twin studies may overestimate heritability. The equal environments assumption of twin studies — that MZ and DZ twins experience equally similar environments — has been questioned. If MZ twins create more similar environments for themselves due to their genetic similarity, heritability estimates may be inflated. This concern is reduced but not eliminated by adoption studies, which also show high heritability.

Intelligence tests measure specific skills, not fixed cognitive capacity. IQ tests measure particular cognitive abilities that are valued in contemporary Western educational contexts. The degree to which they measure something biologically fundamental versus culturally constructed is debated. Test scores can be influenced by test-taking familiarity, stereotype threat, and motivation.

What We Can Conclude

The heritability of intelligence within populations is robustly established by converging evidence from twin, adoption, and genome-wide association studies. Adult intelligence is substantially heritable, with genetic influences becoming stronger as individuals age and select environments. This heritability does not imply fixed cognitive capacity, as the Flynn effect demonstrates. It does not imply that between-group differences in cognitive test scores are genetic in origin — the causal architecture of group differences is unknown and cannot be inferred from within-group heritability. The finding is scientifically real and practically significant; its misapplication to justify social inequality is a category error that the evidence does not support.

References

• Polderman, T. J. C., Benyamin, B., de Leeuw, C. A., et al. (2015). Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nature Genetics, 47, 702–709.
• Turkheimer, E., Haley, A., Waldron, M., D'Onofrio, B., & Gottesman, I. I. (2003). Socioeconomic status modifies heritability of IQ in young children. Psychological Science, 14(6), 623–628.
• Lee, J. J., Wedow, R., Okbay, A., et al. (2018). Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nature Genetics, 50, 1112–1121.
• Flynn, J. R. (1987). Massive IQ gains in 14 nations: What IQ tests really measure. Psychological Bulletin, 101(2), 171–191.
• Plomin, R., DeFries, J. C., Knopik, V. S., & Neiderhiser, J. M. (2016). Top 10 replicated findings from behavioral genetics. Perspectives on Psychological Science, 11(1), 3–23.