IQ is substantially genetic, but “how genetic” depends a great deal on when in life you measure it and the conditions a person grew up in. Twin and adoption studies consistently estimate that genes account for roughly 40 to 50 percent of the variation in IQ during childhood, with that figure climbing to around 80 percent by late adolescence and adulthood. That still leaves a large role for environment, and the interplay between genes and surroundings turns out to be far more interesting than a simple percentage split suggests.
Heritability Rises as You Get Older
One of the most robust findings in intelligence research is something called the Wilson Effect: the genetic contribution to IQ differences grows steadily from infancy through early adulthood. In childhood, shared family environment (the home you grow up in, the schools you attend, the neighborhood around you) exerts a strong pull. But by the late teenage years, those shared-environment effects shrink, and genetic influences dominate. A review of key twin and adoption studies, including twins raised apart, concluded that the heritability of IQ reaches about 0.80 by age 18 to 20 and stays at that level well into adulthood.1PubMed. The Wilson Effect: the increase in heritability of IQ with age
Why would genes matter more as you get older? The leading explanation is that as children gain autonomy, they increasingly select and shape environments that match their genetic predispositions. A child with a genetic tilt toward curiosity gravitates toward books, stimulating friendships, and challenging classes. Over years, those choices amplify small initial genetic differences into larger observable ones. Meanwhile, the household environment that parents controlled becomes less relevant once the child leaves home. A study of adoptive and biological families with 30-year-old offspring estimated heritability at 0.42, with genetic influences described as the predominant systematic source of variation in adult intelligence.2PubMed Central. Genetic and environmental contributions to IQ in adoptive and biological families with 30-year-old offspring That figure is somewhat lower than the 0.80 asymptote found in the broader literature, but the study’s confidence interval was wide, and methodological differences across studies (sample age, test used, modeling approach) routinely produce a range rather than a single number.
When researchers model not just twins but extended families, accounting for the fact that people tend to partner with someone of similar intelligence, the picture gets even more nuanced. One extended twin-family analysis attributed 44 percent of adult IQ variance to additive genetic effects, 27 percent to non-additive genetic effects, 18 percent to non-shared environment, and 11 percent to assortative mating itself.3PubMed Central. Reconsidering the heritability of intelligence in adulthood: taking assortative mating and cultural transmission into account In other words, part of what looks like “genetic” in simpler models is actually a consequence of smart people marrying smart people and creating home environments that correlate with their genes.
What DNA Scans Have and Haven’t Found
If IQ is so heritable, you might expect scientists to have pinpointed the genes responsible. They have found thousands of relevant genetic variants, but each one has an almost comically small effect on its own. The largest genome-wide association study of educational attainment (a rough proxy for cognitive ability) analyzed about three million people and identified nearly 4,000 significant genetic variants. A combined polygenic score built from all of those variants explained 12 to 16 percent of the variation in years of schooling.4Nature Genetics. Polygenic prediction of educational attainment within and between families from genome-wide association analyses in 3 million individuals An earlier study of about 1.1 million people found that polygenic scores explained 7 to 10 percent of cognitive performance variance.5Nature Genetics. Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals
Those numbers are impressive for individual genetic prediction, yet they fall far short of the 50 to 80 percent heritability that twin studies imply. This gap, sometimes called “missing heritability,” remains one of the big open questions. Part of the explanation is that many relevant genetic variants are too rare, or their individual effects too tiny, to be detected by current genotyping technology. Another chunk of the heritability likely comes from interactions among genes that polygenic scores do not yet capture well. Rare structural changes in DNA, like large deletions, can also play a role. One study found that the greater the total size of rare deletions in someone’s genome, the lower their measured intelligence, with the effect accounting for a meaningful share of variation in the sample studied.6PLoS ONE. Rare Copy Number Deletions Predict Individual Variation in Intelligence
There is also evidence that the genetics of IQ are not all funneled through one master trait. While a general cognitive factor is highly heritable (one analysis estimated 86 percent of variance in the latent general factor was genetic), specific cognitive abilities like verbal skill, processing speed, and spatial reasoning also show independent genetic influences.7PubMed Central. Genetic and Environmental Influences of General Cognitive Ability: Is g a valid latent construct? A review of specific cognitive abilities found an average heritability of about 56 percent across different skills, though those specific abilities do not show the same increase in heritability with age that general intelligence does.8Intelligence. The genetics of specific cognitive abilities Even within the genetic story, “intelligence” is not one monolithic thing.
Polygenic Scores Do Not Work Equally Across Populations
A practical limitation of polygenic scores is that they were developed overwhelmingly in people of European descent, and their predictive power drops when applied to other ancestry groups. In one analysis using a large U.S. sample of children and adolescents, a cognitive-ability polygenic score retained about 89 percent of its predictive strength in Hispanic/Latino Americans and 66 percent in African Americans relative to European Americans.9Intelligence & Cognitive Abilities. Interpreting Polygenic Prediction of Cognitive Ability: Evidence for Direct, Reliable, and Portable Genetic Effects – Section: Cross-Ancestry Performance Declines in Line with Expectations The decline is not evidence that genetics matters less in those groups. It reflects the technical fact that the genetic markers flagged in European-ancestry data sets do not tag the same causal variants as efficiently in populations with different patterns of genetic linkage. This is a measurement problem, not a biological one, but it matters for anyone tempted to draw group-level conclusions from individual polygenic scores.
Environment Still Moves the Needle
The clearest proof that environment matters for IQ, independent of genes, comes from the Flynn Effect: the steady rise in average IQ test scores observed across the twentieth century. A meta-analysis of 285 studies covering over 14,000 test-score comparisons found gains averaging about 2.3 IQ points per decade, with modern IQ tests (post-1972) showing gains closer to 3 points per decade.10PubMed Central. The Flynn effect: A meta-analysis A broader analysis spanning more than a century of data from 31 countries confirmed that gains were strongest for fluid reasoning (about 0.41 points per year) and that the effect has weakened in more recent decades.11PubMed. One Century of Global IQ Gains: A Formal Meta-Analysis of the Flynn Effect (1909-2013) In some countries the rise has stopped entirely; Norwegian conscript data showed substantial gains from the 1950s through the early 1970s, then a slowdown, and a complete halt by the mid-1990s.12Intelligence. The end of the Flynn effect?: A study of secular trends in mean intelligence test scores of Norwegian conscripts during half a century
Genes in a population do not change meaningfully in a few generations, so the Flynn Effect has to be environmental. Better nutrition, reduced childhood disease, more years of schooling, and greater exposure to abstract problem-solving in daily life are the usual suspects. Education in particular has been shown to have a causal effect on IQ. A meta-analysis of over 600,000 participants found that each additional year of schooling adds roughly 1 to 5 IQ points.13PubMed Central. How Much Does Education Improve Intelligence? A Meta-Analysis A natural experiment in Norway, where compulsory schooling was extended from seven to nine years in the 1960s, confirmed that the extra years of school had a substantial effect on IQ measured at age 19.14PubMed Central. Schooling in adolescence raises IQ scores
Poverty Changes the Genetic Equation
Here is where the “how genetic is IQ” question gets genuinely complicated. The answer depends on the environment. A landmark study of young twins in the United States found that in affluent families, genetic differences accounted for most of the variation in IQ, while shared environment accounted for very little. In impoverished families, the pattern reversed: shared environment explained about 60 percent of IQ variance, and the genetic contribution was close to zero.15PubMed. Socioeconomic status modifies heritability of IQ in young children The idea is intuitive: when children lack adequate nutrition, stimulation, and stability, those environmental deficits overshadow genetic potential. When basic needs are met, genetic differences have room to express themselves.
A German twin study largely replicated this pattern in middle childhood, finding that family socioeconomic status amplified genetic effects on cognitive ability, with shared environment explaining more variance at the lower end of the income spectrum. The effect weakened in adolescence and was not detectable in adulthood in that sample.16PubMed Central. Socioeconomic status amplifies genetic effects in middle childhood in a large German twin sample And in an important twist, a cross-national comparison found that this gene-by-socioeconomic-status interaction was clear in U.S. samples but absent or even reversed in Western Europe and Australia, where social policies provide more uniform access to education and health care.17PubMed Central. Large Cross-National Differences in Gene × Socioeconomic Status Interaction on Intelligence In countries with stronger safety nets, poverty does not suppress genetic expression to the same degree. So the question “is IQ genetic” is not just a biological question; it is partly a question about what kind of society you live in.
What Happens in the Brain
Researchers have looked at the brain itself to understand how genetic and environmental influences translate into cognitive performance. Brain volume shows a modest but real positive relationship with IQ. One study that controlled for family background found that each standard-deviation increase in brain volume was associated with about 0.14 standard deviations higher IQ, a figure that rose to about 0.18 after correcting for measurement noise.18PubMed Central. The causal influence of brain size on human intelligence: Evidence from within-family phenotypic associations and GWAS modeling The same study found a genetic correlation of 0.41 between brain volume and years of education, and the direction of genetic influence ran mainly from brain volume to education rather than the other way around.
Beyond size, the brain’s efficiency and connectivity seem to matter more. Studies using brain imaging have found that higher-scoring individuals tend to show lower overall brain activation when performing the same task as lower-scoring individuals, a pattern called neural efficiency. One study showed that brighter participants used less activity in a specific brain region (the right insula) during tasks of equal difficulty, though the difference vanished when tasks were adjusted to be equally challenging for each person.19PubMed Central. Neural efficiency as a function of task demands Other work found that higher intelligence was associated with lower brain activation overall and stronger short-distance connectivity in the frontal cortex.20Intelligence. Intelligence and neural efficiency: Measures of brain activation versus measures of functional connectivity in the brain A graph-analysis study of whole-brain functional networks found that people with higher IQ scores had more efficiently organized brain networks, with particularly strong effects in frontal and parietal regions.21Journal of Neuroscience. Efficiency of Functional Brain Networks and Intellectual Performance
These brain differences are partly genetic, but they are also shaped by experience. The brain’s wiring is not fixed at birth; it develops in response to the demands placed on it throughout childhood and adolescence. Education itself appears to physically reshape cognitive capacity, not just teach content. One study found that a year of schooling contributed about 0.22 standard deviations to the development of crystallized IQ and 0.14 standard deviations to fluid IQ, with working memory showing even larger schooling effects relative to simple age-related maturation.22Intelligence. Schooling substantially improves intelligence, but neither lessens nor widens the impacts of socioeconomics and genetics
Epigenetic Marks and Cognitive Ability
Between the DNA sequence you inherit and the IQ you score on a test, there is a layer of molecular regulation that responds to experience. Epigenetic modifications, particularly chemical tags on DNA that turn gene activity up or down without changing the underlying sequence, are increasingly linked to cognitive outcomes. One study found that methylation at a specific site in the dopamine D2 receptor gene was negatively associated with IQ: higher methylation (which tends to suppress that gene’s activity) correlated with lower scores, accounting for a small but statistically significant portion of variance.23Translational Psychiatry. Epigenetic variance in dopamine D2 receptor: a marker of IQ malleability?
More broadly, children’s DNA methylation profiles, originally developed to reflect social conditions in adults, turn out to correlate with cognitive and academic performance in children and adolescents. Researchers found that these methylation signatures were associated with processing speed, executive function, reasoning, reading, and math skills, suggesting that the social conditions of early life leave molecular marks that track with cognitive development.24PubMed Central. Socially Stratified Epigenetic Profiles Are Associated With Cognitive Functioning in Children and Adolescents This research is still early, and the effect sizes are modest. But it suggests a concrete biological pathway through which poverty, stress, and other environmental factors can get “under the skin” and influence cognitive outcomes in ways that look biological but are ultimately environmental in origin.
Early Interventions and the Fadeout Problem
If environment shapes IQ, can you boost it with the right program? Yes, temporarily. A meta-analysis of early childhood interventions found that they successfully raised IQ by about a third of a standard deviation (roughly 5 IQ points) by the time the program ended. But the gains faded steadily after the intervention stopped, declining at a rate of about one point per year until they reached zero.25Intelligence. The environment in raising early intelligence: A meta-analysis of the fadeout effect Starting interventions earlier in life did not prevent the fadeout.
Some researchers initially hoped that interventions that raised the “g factor” (the core of general intelligence, as opposed to narrow test-taking skills) might produce more durable gains. A follow-up study tested this directly and found that an intervention did raise the g factor at age three, but the effect had vanished by ages five and eight.26Intelligence. Does the raising IQ-raising g distinction explain the fadeout effect? The fadeout pattern does not mean early interventions are pointless; many produce lasting benefits in health, earnings, and behavior that persist long after the IQ boost disappears. But for cognitive ability specifically, sustaining gains seems to require sustained environmental support rather than a one-time boost.
The Shadow of Eugenics
Any honest discussion of the genetics of intelligence has to acknowledge that this research area carries historical baggage. In the early twentieth century, the eugenics movement used crude and often fraudulent claims about the heritability of intelligence to justify immigration restrictions, forced sterilization, and worse. In the United States alone, tens of thousands of people were sterilized against their will based on judgments about their “genetic fitness,” and the scientific rhetoric of the era lent intellectual cover to racist policies.27PubMed. The social and economic origins of genetic determinism: a case history of the American Eugenics Movement, 1900-1940 and its lessons for today
Modern behavioral genetics bears little resemblance to that era’s pseudoscience, but the history matters for two reasons. First, it explains why many people react with justified suspicion when they hear “IQ is genetic.” Second, it is a reminder that heritability is a population-level statistical concept, not a destiny stamp on an individual. Saying that IQ is 50 or 80 percent heritable in a given population does not mean your IQ was 80 percent determined at conception. It means that, in that population under those conditions, most of the variation between people can be attributed to genetic differences. Change the conditions and the heritability changes too, as the cross-national gene-by-poverty research demonstrates.
A Meta-Analysis That Complicates the Standard Number
Even the headline heritability figures deserve scrutiny. A meta-analysis of 212 studies proposed a model that accounts for maternal womb effects (the prenatal environment twins share), which are often lumped together with genetic effects in standard models. Under this “maternal-effects” model, the prenatal environment accounted for about 20 percent of the similarity between twins and 5 percent of the similarity between non-twin siblings. After adjusting for these effects, heritability estimates fell below 50 percent.28Nature. The heritability of IQ This does not overturn the consensus that genes matter a great deal. But it is a useful corrective against treating any single heritability estimate as settled truth. The number you get depends on what you model, and the field has not fully agreed on what should be in the model.
A twin study focused specifically on high cognitive ability (using over 11,000 twin pairs from six studies across four countries) estimated genetic influences at about 50 percent, shared environment at 28 percent, and non-shared environment at 22 percent.29PubMed Central. A Twin Study of the Genetics of High Cognitive Ability Selected from 11,000 Twin Pairs in Six Studies from Four Countries That shared-environment contribution of nearly 30 percent is larger than what most adult-focused studies find, likely because the sample included younger participants whose home environments still exerted a strong effect. The point is that no single heritability number captures the full picture. The estimate shifts depending on age, socioeconomic context, country, and what statistical model you use. Treating IQ heritability as a fixed fact of nature misses the interesting parts of the science.