Heritability is a statistical measure of how much of the variation in a trait across a population can be traced to genetic differences among individuals, rather than to differences in their environments. It is always a number between zero and one (or expressed as a percentage), and it always describes a population at a particular time and place, never an individual person. A heritability of 0.60 for height, for instance, does not mean 60 percent of your height came from your genes. It means that, in the group studied, about 60 percent of the reason people differ in height is genetic. That distinction sounds subtle, but getting it wrong leads to deep misunderstandings about what genes can and cannot tell us.
A Property of Populations, Not People
The most common mistake people make with heritability is treating it as a statement about a single person. If someone tells you “the heritability of intelligence is about 50 percent,” they are not saying half of your IQ came from your DNA and the other half from how you grew up. They are saying that when researchers look at a large group of people and ask why some scored higher and some lower, roughly half of that spread in scores can be statistically linked to genetic differences among those people.
This matters because the number is entirely dependent on the population and the environment it lives in. If you measured heritability in a group where everyone ate the same food, attended the same school, and breathed the same air, environmental variation would shrink toward zero and heritability would shoot up, not because genes became more powerful but because you removed the competing source of variation. Conversely, in a group with wildly unequal access to nutrition and education, the same trait could show lower heritability because environmental differences are doing more of the work. Heritability is not a fixed property of a trait. It is a snapshot of a population in a context.
Narrow-Sense, Broad-Sense, and Why the Distinction Exists
Geneticists actually use two flavors of heritability. Narrow-sense heritability captures only additive genetic effects, which is the straightforward summing up of contributions from many gene variants, each nudging a trait a little bit in one direction. Broad-sense heritability adds in non-additive effects, including how gene variants interact with each other within or across different genes.1PubMed Central. How to estimate heritability: a guide for genetic epidemiologists For most practical purposes, when you see a heritability estimate in the news, it is referring to the broad-sense version. The narrow-sense version matters more in animal and plant breeding, where breeders care specifically about the portion of genetic variation that parents reliably pass to offspring.
Neither version tells you anything about a single gene “for” a trait. Both describe the combined influence of potentially thousands of genetic variants, each contributing a tiny amount. And both are locked to the specific population being studied. As one epidemiological guide puts it, heritability estimates “cannot necessarily be translated into a wider or more general population.”1PubMed Central. How to estimate heritability: a guide for genetic epidemiologists
How Researchers Estimate Heritability
The workhorse method for decades has been the twin study. Identical twins share virtually all of their DNA; fraternal twins share about half, on average, just like any siblings. By comparing how similar identical twins are on a given trait versus how similar fraternal twins are, researchers can estimate how much of the variation is genetic. Over the past half century, twin studies have produced heritability estimates for more than 17,000 different traits.2PubMed Central. A robust and unified framework for estimating heritability in twin studies using generalized estimating equations
Twin studies rest on a key assumption called the equal environments assumption: the idea that identical and fraternal twins are exposed to equally similar environments when it comes to the trait being measured. If identical twins are actually treated more alike than fraternal twins in ways that affect the trait, the study could overestimate genetic influence.3PubMed Central. An investigation of a measure of twins’ equal environments Critics have raised this concern repeatedly. A thorough evaluation using nationally representative twin data concluded that the assumption is probably not strictly valid for most traits, but the resulting bias is likely modest rather than catastrophic.4PubMed. What can we learn from twin studies? A comprehensive evaluation of the equal environments assumption In other words, twin-study estimates are informative but should be read as useful approximations, not as exact measurements.
More recently, genome-wide association studies have offered a different route. Instead of comparing twins, researchers scan the DNA of large groups of unrelated people, looking for common genetic variants associated with a trait. This approach gives its own heritability estimate, often called “SNP heritability,” based on how much of the trait variation is tagged by the variants on the DNA chip. Comparing the two methods is where things get interesting.
The Missing Heritability Gap
When scientists first compared twin-study estimates with the results of large-scale DNA scans, a striking gap appeared. Twin and family studies suggest that roughly half of the variation in intelligence is attributable to genetics, yet genome-wide association studies could only account for about 10 percent of the variation.5PubMed Central. Three legs of the missing heritability problem This pattern, with DNA scans capturing far less heritability than twin studies predict, showed up for trait after trait and was quickly dubbed the “missing heritability” problem.
Several explanations have been proposed. Part of the gap comes from rare genetic variants that are too uncommon to show up on standard DNA chips. Part may come from complex interactions between genes. And part may reflect that twin studies themselves overestimate heritability to some degree, while DNA scans underestimate it by missing variants that are not well captured by current technology. Including rarer and harder-to-measure variants does recover a meaningful share of the missing signal, though the exact amount depends on statistical assumptions about how effect sizes relate to variant frequency.6PubMed Central. What Does Heritability Mean? The Role of Genes & Environment The gap has narrowed over the years as studies have gotten bigger and more inclusive of rare variants, but it has not closed entirely, and some researchers suspect it never will fully close because the two methods are measuring slightly different things.
When Genes and Environment Refuse to Stay in Their Lanes
Heritability is framed as a partition: some fraction of variation is genetic, some is environmental, and the rest is noise. That framing is convenient but hides the fact that genes and environments constantly entangle with each other in ways that resist clean separation.
Gene-environment interaction is the straightforward version of this entanglement. The same genetic variant can have a big effect in one environment and a negligible effect in another. A genetic predisposition toward high blood pressure, for example, may play out very differently depending on salt intake, stress levels, or access to healthcare. Statistically, these interactions show up when the combined effect of a genetic variant and an environmental exposure departs from what you would predict by just adding the two effects together.7PubMed Central. Gene-environment interactions in human health These interactions can also help explain part of the missing heritability gap, because standard models that assume genes and environments contribute independently may systematically miss the combined effects.8PubMed Central. Twins and the mystery of missing heritability: the contribution of gene-environment interactions
Gene-environment correlation is subtler and, in some ways, more disorienting. It describes situations where a person’s genetic makeup influences which environments they end up in. Researchers distinguish three types: passive, evocative, and active.9PubMed Central. Genotype-environment correlations: implications for determining the relationship between environmental exposures and psychiatric illness Passive correlation happens when parents provide both the genes and the home environment, so a child with genes linked to high verbal ability may also grow up surrounded by books. Evocative correlation occurs when a child’s genetically influenced traits draw out particular responses from other people: a naturally cheerful toddler gets smiled at more. Active correlation is when people seek out environments that match their dispositions, like a musically inclined teenager joining a band.
In all three cases, what looks like “environment” when you measure it is partly shaped by genes, and what looks like “genetic influence” on a trait may partly reflect the environments those genes helped create. This makes the clean partition at the heart of heritability estimates somewhat artificial. The numbers are still useful, but they describe a statistical decomposition, not two truly independent forces.
Heritability Is Not Static Across the Lifespan
One of the more counterintuitive findings in behavioral genetics is that heritability can change dramatically with age. For IQ, multiple twin and adoption studies converge on a pattern now called the Wilson Effect: the heritability of intelligence is modest in early childhood and climbs steadily, reaching roughly 0.80 by age 18 to 20, where it stays well into adulthood.10PubMed. The Wilson Effect: the increase in heritability of IQ with age
This seems paradoxical at first. You might expect that as people accumulate more life experience, environmental factors would matter more, not less. But one plausible explanation ties back to gene-environment correlation. As children grow into adults, they gain more control over their own environments: what they read, who they spend time with, what careers they pursue. Those choices increasingly reflect their genetic predispositions, so genetic influence on the trait appears to grow. The genes have not changed; the person’s ability to select environments that amplify those genes has increased.
How Socioeconomic Conditions Shift Heritability
A long-debated idea known as the Scarr-Rowe hypothesis predicts that heritability of cognitive abilities should be higher in wealthier, more advantaged environments, because privilege removes environmental barriers and lets genetic potential express itself more fully.11Intelligence. Socioeconomic status amplifies genetic effects in middle childhood in a large German twin sample In an impoverished environment full of unmet needs, the thinking goes, those unmet needs dominate the variation and mask genetic differences.
The hypothesis has intuitive appeal, but the evidence is mixed and increasingly tilts against it in its original form. A recent study examining interactions between genetic propensity scores and family socioeconomic status across dozens of educational outcomes found that the vast majority of the interactions were negative, the opposite direction from what Scarr-Rowe predicts. Out of 42 tested interactions, 39 were negative, and the pattern was more consistent with a “compensatory advantage” model, where higher socioeconomic status compensates for genetic disadvantage rather than amplifying genetic advantage.12Research in Social Stratification and Mobility. Interaction of family SES with children’s genetic propensity for cognitive and noncognitive skills: No evidence of the Scarr-Rowe hypothesis for educational outcomes
If that compensatory model is right, it carries a powerful implication: improving environments for disadvantaged children does not just add an environmental boost on top of genetic predispositions. It may actually buffer children whose genetic lottery tickets were less favorable, narrowing the outcome gap more than a simple additive model would predict.
Policy Can Change Heritability
Perhaps the most vivid illustration of how heritability shifts with context comes from education. An analysis of educational attainment across birth cohorts in a large twin sample found dramatic shifts over just a few decades. For people born before 1940, family background explained about 47 percent of the variation in how far they went in school, while genetic factors accounted for 41 percent. For men born after 1940, the picture reversed sharply: the heritability of educational attainment jumped to between 67 and 74 percent, and the environmental impact of family background shrank to just 8 to 10 percent.13PubMed. Education policy and the heritability of educational attainment
The most likely explanation is that postwar expansion of educational access, including public schooling, scholarships, and reduced barriers for working-class families, leveled the environmental playing field. When more people have roughly equal access to education, the remaining differences in how far they go start to reflect genetic variation in aptitude, motivation, and related traits. Heritability went up not because genes became more potent but because one major environmental barrier was reduced. This is a concrete example of how a policy change can literally alter the heritability of a trait.
Why Within-Group Heritability Cannot Explain Between-Group Differences
One of the most consequential misuses of heritability is the leap from “this trait is heritable within a group” to “differences between groups must also be genetic.” The logic feels intuitive but is mathematically unfounded. A 2024 paper in the Proceedings of the National Academy of Sciences demonstrated formally that information about genetic and phenotypic variation within groups simply cannot be used to decompose between-group differences into genetic and environmental components.14PubMed Central. Heritability within groups is uninformative about differences among groups: Cases from behavioral, evolutionary, and statistical genetics
A classic thought experiment helps. Imagine taking seeds from a single, genetically diverse batch of corn and planting half in nutrient-rich soil and half in nutrient-poor soil. Within each plot, the variation in plant height would be highly heritable, because the soil is the same for every plant in the plot and only their genetics differ. But the difference in average height between the two plots would be entirely environmental. High within-group heritability tells you nothing about why the groups differ. This principle applies directly to human populations that have experienced different histories of nutrition, education, discrimination, and opportunity.
Epigenetics and the Blurred Boundary
Epigenetic modifications, changes to how genes are read without altering the DNA sequence itself, add another wrinkle to heritability. These modifications can be influenced by diet, stress, toxin exposure, and other environmental factors. Some have been shown to persist across generations, at least in animal models, creating a form of inheritance that is not captured by standard genetic analysis.15PubMed Central. Epigenetic inheritance and the missing heritability Components of traits linked to stature, fertility, and food metabolism have been shown to respond to environmental conditions and to be epigenetically inherited.
Whether epigenetic inheritance can solve the missing heritability problem is contested, though. A key challenge is duration: epigenetic marks typically reset within one or two generations in mammals. If a mark does persist for many generations, it starts behaving like an ordinary mutation and would likely be picked up by standard DNA scans anyway.16Genetics. Epigenetic Inheritance and the Missing Heritability Problem So epigenetics may help explain why certain traits seem to run in families even when standard genetic markers cannot account for all the resemblance, but the jury is still out on how much of the total missing signal it can explain.
Heritability in the Wild
Most heritability estimates come from human twin studies or laboratory organisms raised in controlled conditions. In wild animal populations, the picture gets messier. A study of a free-living mammal population found that the heritability of total fitness declined in part because environmental variation in longevity increased, inflating the denominator of the heritability fraction without any change in genetic variance itself.17PubMed. Heritability of fitness in a wild mammal population In other words, when the environment gets harsher or more unpredictable, heritability drops, even if the genetic architecture of the trait is identical. Organisms in nature face fluctuating food supplies, predators, parasites, and weather. That noisy environmental backdrop makes heritability estimates lower and less stable than what you see in a lab or even in a wealthy human population.
This reinforces the earlier point that heritability is a property of a population in a given environment. Move the same population to a different environment, or wait a generation for conditions to shift, and the number changes. Researchers studying evolutionary potential in wild populations have to grapple with this instability constantly, because the heritability of a trait today may not predict how that trait will respond to natural selection tomorrow.
Phenotypic Plasticity and Developmental Constraints
Related to heritability is the concept of phenotypic plasticity: the ability of an organism to change its traits in response to environmental conditions. A highly plastic trait can look very different in two environments even when the underlying genetics are identical. The classic example is skin tanning. The genes that control melanin production set a range, and sun exposure determines where within that range a person ends up.
Plasticity itself has a genetic basis and can evolve. Recent theoretical work has explored what happens when plasticity is not unlimited but is bounded by developmental constraints, modeled as a ceiling on how much a trait can change in response to the environment.18PubMed. The genetics of phenotypic plasticity. XVIII. Developmental limits restrict adaptive plasticity When an environment shifts abruptly, populations with constrained plasticity may have to rely more on genetic change across generations, a process called genetic assimilation, rather than flexible adjustment within a single lifetime. For heritability, the implication is that the same trait can have high heritability in a stable environment where plasticity is not being tested and lower heritability in a changing environment where organisms differ in how flexibly they respond.
What Heritability Cannot Do for Prediction
Even a very high heritability estimate does not mean you can predict an individual’s trait from their DNA. Heritability describes how much of the population-level spread is genetic, but any single person’s outcome reflects the accumulated influence of thousands of variants, specific environmental exposures, random developmental noise, and gene-environment entanglements that are unique to them. Current polygenic scores, which add up the effects of known trait-associated variants, capture only a fraction of the genetic variance and even less of the total trait variance. They are useful for research and, in some cases, for identifying people at elevated risk for certain diseases, but they are poor crystal balls for any one person’s future.
Understanding this gap between population-level heritability and individual-level prediction matters for how you interpret consumer genetic tests, risk scores, and media reporting on “the gene for” some trait. A trait can be highly heritable and still be profoundly shaped by environment at the individual level. The two statements are not in tension; they are describing different things.