Heritability vs Environment: What "IQ Is 80% Genetic" Actually Means

"Intelligence is 80% heritable" is a real finding. It is also, in almost every conversation where it appears, being used to mean something it does not mean.

Heritability is a technical term with a narrow definition: the proportion of variance in a trait, within a particular population under particular conditions, that tracks genetic differences between people. Every clause in that sentence is load-bearing. Strip any one of them out and you get a claim the statistic cannot support — which is exactly what happens when the number escapes into public argument.

The number is real. The number is also not what you think it is. Here's what the evidence actually shows, and — as with test bias — the popular argument goes wrong in both directions at once.

Your score isn't a genetic readout — it's a snapshot of developed ability, right now, against a current reference sample. Our free IQ test takes ~20 minutes and gives you that snapshot instantly, no email required.

What heritability actually measures

Start with what it is not: heritability is not a statement about you.

Consider height, which is highly heritable. Now ask: what percentage of your height is genetic? The question is malformed. You wouldn't have any height at all without your genes, and you wouldn't have any without food either. There's no meaningful way to slice an individual's trait into a genetic portion and an environmental portion. They aren't ingredients in a recipe; they're more like the length and width of a rectangle — you can't say which one contributes more to the area of this rectangle.

What heritability describes is variance across a population. If everyone in a country ate identically well, remaining height differences would be almost entirely genetic, and heritability would approach 1.0. If nutrition varied wildly, heritability would drop — not because genes stopped mattering, but because environmental variation now accounts for more of the spread.

This is why the same trait has different heritabilities in different places and at different times. Heritability is a property of a population in a context — not a fixed constant of the trait itself.

The numbers, by age

With the definition in place, the findings are genuinely interesting. Twin and adoption designs — including the classic studies of twins reared apart — converge on a clear developmental pattern:

Two things in that table deserve attention. The first is the rise — heritability roughly doubles from childhood to adulthood. The second is quieter and, to many people, more startling: shared environment shrinks to about 0.10 by ages 18–20. Shared environment means everything siblings raised together have in common — the household, the parents' style, the neighbourhood, the family bookshelf. By adulthood, in these samples, it accounts for very little of the remaining variance in IQ.

That finding is often misreported as "parenting doesn't matter." It doesn't say that. It says that within the range of environments sampled in these studies — mostly ordinary homes in wealthy countries — differences between those homes don't produce much lasting spread in adult IQ. Which brings us to the crucial caveat later in this article.

The Wilson effect: the finding that runs backwards

Almost everyone's intuition is that environment should accumulate — that as life piles on experiences, schooling, luck, and circumstance, environment should matter more with age and genes less. The data say the reverse. Ronald Wilson first documented this clearly, and Bouchard (2013) named it the Wilson effect: the heritability of IQ climbs with age, hitting an asymptote near 0.80 around ages 18–20 and staying there well into adulthood.

Why would genetic influence grow? The leading explanation is gene–environment correlation. As people gain autonomy, they stop simply receiving environments and start choosing them. A child with a propensity toward verbal ability gets read to, then seeks out books, then picks the friends and courses and career that keep feeding that propensity. Genes don't operate instead of environment here — they operate through it, by steering people toward environments that amplify their existing tendencies. Small initial differences compound into large ones.

So the Wilson effect isn't evidence that environment stops mattering. It's evidence that by adulthood, much of the environment people are in is one their own dispositions helped select. That's a subtler and more interesting claim — and it means the neat partition between "genes" and "environment" was always somewhat artificial. Related developmental patterns show up in IQ across the lifespan.

Heritability isn't a constant — it depends on the environment

Here's the finding that proves heritability is context-dependent rather than a fact about the trait. Turkheimer and colleagues (2003) analysed WISC scores from 320 pairs of 7-year-old twins in the National Collaborative Perinatal Project, a sample that included many families living near or below the poverty line. They let the variance components interact with socioeconomic status rather than assuming they were fixed.

The result was striking: the proportions varied nonlinearly with SES. In impoverished families, roughly 60% of IQ variance was attributable to shared environment and the genetic contribution was close to zero. In affluent families, almost exactly the reverse.

The interpretation is intuitive once you see it. In a deprived environment, whether a child gets adequate nutrition, a stable home, and a functioning school swamps everything else — environmental variation dominates, so heritability collapses. In an environment where those basics are universally met, the remaining differences between children are increasingly genetic, because the environmental floor has been raised for everyone. High heritability can be a marker of environmental equality, not of genetic determinism. A society that gave every child an excellent upbringing would increase measured heritability, not decrease it.

Honesty requires the next part. This interaction has not replicated uniformly. Tucker-Drob and Bates (2016) examined it across countries and found large cross-national differences — the Gene × SES interaction appears in US samples but is largely absent in Western Europe and Australia. The plausible reading is that stronger social safety nets compress environmental variation at the bottom, so there's no deprivation gradient left for heritability to interact with. That's a coherent story, but it remains an inference, and this is an area where the evidence is still moving.

High heritability does not mean unchangeable

This deserves its own section because it's the inference people most want to draw and it is simply invalid. Heritability describes the sources of existing variation. It is not a measure of malleability, and it makes no prediction about what would happen if you changed the environment — because it was calculated inside the environments that already exist.

The empirical rebuttals are strong:

• Education causally raises IQ. Ritchie and Tucker-Drob (2018) meta-analysed 142 effect sizes from 42 datasets covering over 600,000 participants, restricting themselves to quasi-experimental designs — controlling for prior intelligence, exploiting compulsory-schooling policy changes as instrumental variables, and using regression discontinuity at school-entry cutoffs. Their conclusion: roughly 1 to 5 IQ points per additional year of education, and education is the most consistent, robust, and durable method yet identified for raising intelligence. These are causal designs, not correlations.
• The Flynn effect. Whole populations gained ~3 IQ points per decade for a century — environmental change alone, far too fast for genetics. Height is roughly as heritable as adult IQ, and populations still grew dramatically taller when nutrition improved. High heritability didn't stop it. See the Flynn effect for the full picture.
• Deprivation reversal. Remove a severe environmental deficit — malnutrition, lead exposure, absent schooling — and scores move. That's the Turkheimer logic running forward.

The realistic version matters too, though. None of this supports the brain-training industry's claims, and the effect sizes here are meaningful but not magical. Our guide to how to increase IQ covers what actually has evidence behind it — and nutrition and IQ covers where the deprivation-reversal logic applies most clearly.

Genetic nurture: where the dichotomy breaks down

The most conceptually important result in this area is recent, and it quietly dismantles the framing of the whole debate.

Kong and colleagues (2018), working with Icelandic genealogical and genomic data, asked an ingenious question. Each parent passes on half their alleles; the other half are not transmitted. Those non-transmitted alleles are, by definition, not in the child. So they should have no effect on the child whatsoever.

They did. Parental alleles that were never passed to the child still predicted the child's educational attainment — at roughly 30% of the effect size of the transmitted polygenic score. The only possible pathway is environmental: those alleles shaped the parents, the parents shaped the home, and the home shaped the child. Kong called it genetic nurture.

Sit with the implication. A substantial chunk of what gets counted as "genetic effect" in standard designs is genes acting as environment, one generation removed. The tidy partition — genes over here, environment over there — describes a distinction that doesn't cleanly exist in the causal world. This is also why within-family designs (comparing siblings, who share a home and half their genome) have become the gold standard: they strip out exactly this confound.

The missing heritability gap

If twin studies say adult IQ is ~70–80% heritable, you'd expect that reading DNA directly would let us predict IQ nearly that well. It doesn't come close.

Polygenic scores — built by summing thousands of tiny variant effects from genome-wide association studies — have improved substantially, but the best current scores for cognitive ability explain something in the range of roughly 10–16% of variance in independent samples, against twin estimates two to five times higher. That shortfall is the missing heritability gap, and it's a live research problem. General cognitive ability is massively polygenic: influenced by thousands of variants each of vanishingly small effect, which is why finding them requires samples in the hundreds of thousands to millions.

Explanations for the gap include rare variants that common-SNP arrays miss, gene–gene and gene–environment interactions the additive model doesn't capture, assortative mating, and — per Kong — the fact that some twin "heritability" was environmental transmission all along. Whichever combination is right, the practical point stands: nobody can currently predict your intelligence from your genome with anything approaching the precision the 80% figure implies. Anyone selling that is selling something the science can't deliver.

The limit that matters most: within-group ≠ between-group

This is the single most important boundary on the statistic, and it is not a matter of opinion — it's arithmetic.

Heritability is estimated within a population. It carries no information whatsoever about the causes of average differences between populations. The two are formally unrelated quantities.

The standard demonstration: take a sack of genetically varied seed. Plant half in rich soil, half in depleted soil. Within each plot, height differences among plants will be almost entirely genetic — heritability near 1.0, because the soil is uniform within each plot. Yet the difference in average height between the two plots is entirely environmental. High within-group heritability, zero genetic contribution to the between-group gap. Both facts, simultaneously, with no contradiction.

This is why the heritability figure cannot be used to argue anything about group differences in IQ — not between countries, not between any demographic groups. The scientific position on such gaps attributes them to environment: nutrition, healthcare, schooling quality, early cognitive stimulation, toxin exposure, and the accumulated effects of unequal opportunity. The 1996 APA task force found no empirical support for a genetic interpretation of group gaps, and the Flynn effect demonstrates that environmentally-driven population differences can exceed most observed gaps within a few decades.

It's also why we treat national IQ rankings as contested environmental indicators rather than measures of innate ability — see criticism of Lynn and Vanhanen and average IQ by country.

The honest verdict

Four claims, each well-supported, each routinely mangled:

What's left after all those corrections is still substantial and still interesting: genetic differences do account for a large share of why adults in wealthy countries differ from one another in cognitive test scores, and that share grows with age largely because people curate environments that match their dispositions. That's a real finding about a real thing.

It just isn't a verdict on anyone. Your score is a measure of developed ability at one moment, carrying real error bars — see IQ test accuracy — and the heritability literature, read correctly, is one of the strongest arguments against treating any single number as destiny. As with the g factor, the science is more modest and more interesting than the slogans built on top of it.

Frequently asked questions

Related reading

← Browse all articles on the blog

References

1. Bouchard, T. J., & McGue, M. (1981). Familial studies of intelligence: A review. Science, 212(4498), 1055–1059.
2. Bouchard, T. J., Lykken, D. T., McGue, M., Segal, N. L., & Tellegen, A. (1990). Sources of human psychological differences: The Minnesota Study of Twins Reared Apart. Science, 250(4978), 223–228.
3. Neisser, U., Boodoo, G., Bouchard, T. J., Boykin, A. W., Brody, N., Ceci, S. J., Halpern, D. F., Loehlin, J. C., Perloff, R., Sternberg, R. J., & Urbina, S. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51(2), 77–101.
4. Devlin, B., Daniels, M., & Roeder, K. (1997). The heritability of IQ. Nature, 388(6641), 468–471.
5. 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.
6. Bouchard, T. J., & McGue, M. (2003). Genetic and environmental influences on human psychological differences. Journal of Neurobiology, 54(1), 4–45.
7. Davis, O. S. P., Haworth, C. M. A., & Plomin, R. (2009). Dramatic increase in heritability of cognitive development from early to middle childhood: An 8-year longitudinal study of 8,700 pairs of twins. Psychological Science, 20(10), 1301–1308.
8. Haworth, C. M. A., Wright, M. J., Luciano, M., Martin, N. G., de Geus, E. J. C., van Beijsterveldt, C. E. M., … Plomin, R. (2010). The heritability of general cognitive ability increases linearly from childhood to young adulthood. Molecular Psychiatry, 15(11), 1112–1120.
9. Nisbett, R. E., Aronson, J., Blair, C., Dickens, W., Flynn, J., Halpern, D. F., & Turkheimer, E. (2012). Intelligence: New findings and theoretical developments. American Psychologist, 67(2), 130–159.
10. Briley, D. A., & Tucker-Drob, E. M. (2013). Explaining the increasing heritability of cognitive ability across development: A meta-analysis of longitudinal twin and adoption studies. Psychological Science, 24(9), 1704–1713.
11. Bouchard, T. J. (2013). The Wilson Effect: The increase in heritability of IQ with age. Twin Research and Human Genetics, 16(5), 923–930.
12. Polderman, T. J. C., Benyamin, B., de Leeuw, C. A., Sullivan, P. F., van Bochoven, A., Visscher, P. M., & Posthuma, D. (2015). Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nature Genetics, 47(7), 702–709.
13. Plomin, R., & Deary, I. J. (2015). Genetics and intelligence differences: Five special findings. Molecular Psychiatry, 20(1), 98–108.
14. Tucker-Drob, E. M., & Bates, T. C. (2016). Large cross-national differences in Gene × Socioeconomic Status interaction on intelligence. Psychological Science, 27(2), 138–149.
15. Kong, A., Thorleifsson, G., Frigge, M. L., Vilhjalmsson, B. J., Young, A. I., Thorgeirsson, T. E., … Stefansson, K. (2018). The nature of nurture: Effects of parental genotypes. Science, 359(6374), 424–428.
16. Ritchie, S. J., & Tucker-Drob, E. M. (2018). How much does education improve intelligence? A meta-analysis. Psychological Science, 29(8), 1358–1369.
17. Plomin, R., & von Stumm, S. (2018). The new genetics of intelligence. Nature Reviews Genetics, 19(3), 148–159.
18. Savage, J. E., et al. (2018). Genome-wide association meta-analysis in 269,867 individuals identifies new genetic and functional links to intelligence. Nature Genetics, 50(7), 912–919.
19. Plomin, R., & von Stumm, S. (2022). Polygenic scores: Prediction versus explanation. Molecular Psychiatry, 27(1), 49–52.
20. Reis, A., & Spinath, F. M. (2025). The genetics of intelligence. Deutsches Ärzteblatt International, 122, 38–42.