Twin studies compare identical twins (who share virtually all of their DNA) with fraternal twins (who share roughly half) to estimate how much of the variation in a given trait comes from genetics versus the environment. The basic logic is elegantly simple: if identical twins are more alike on some trait than fraternal twins are, the extra similarity points toward a genetic influence. This method has been the workhorse of nature-versus-nurture research for decades, producing headline findings on everything from IQ and personality to body weight and political beliefs. But the method rests on assumptions that have been debated almost as long as the studies themselves, and modern genomics has revealed some uncomfortable gaps between what twin studies predict and what DNA scans actually find.
How the Comparison Works
Identical (monozygotic, or MZ) twins develop from a single fertilized egg that splits, so they share all of their additive genetic effects. Fraternal (dizygotic, or DZ) twins develop from two separately fertilized eggs and share, on average, about half of their additive genetic variation and a quarter of any non-additive effects (interactions between gene variants).1PubMed Central. What Twin Studies Tell Us About the Heritability of Brain Development, Morphology, and Function: A Review Both types of twins typically grow up in the same household, attend similar schools, and experience roughly the same family environment. When researchers measure a trait in large samples of both kinds, they can statistically partition the variation into three buckets: the portion linked to genetics, the portion linked to the environment shared by both twins (like household income or parenting style), and the portion linked to environmental factors unique to each individual (like a sports injury one twin had and the other did not).2npj Science of Learning. Conventional twin studies overestimate the environmental differences between families relevant to educational attainment
The result is a “heritability estimate,” a number between zero and one that represents how much of the variation in a trait across a population can be attributed to genetic differences. A heritability of 0.50 for body weight does not mean half of your weight is determined by your genes. It means that, in the population studied, about half of the reason people differ from one another in weight traces to genetic variation rather than environmental variation. That distinction matters: heritability describes a population, not an individual, and it can change if the environment changes.
What Twin Studies Have Found About Intelligence
The Minnesota Study of Twins Reared Apart, one of the landmark projects in the field, found that roughly 70% of the variation in IQ was associated with genetic differences. Strikingly, identical twins raised in completely separate homes scored about as similarly on cognitive tests as identical twins raised together.3PubMed. Sources of human psychological differences: the Minnesota Study of Twins Reared Apart That finding was provocative because it suggested that the shared family environment, things like how many books are in the home or which school a child attends, explained surprisingly little of the long-term variation in cognitive ability.
The picture gets more interesting with age. The heritability of IQ is relatively low in early childhood, when shared environment matters a lot, but climbs steadily as people grow up. A review of twin and adoption studies across different ages and countries documented that heritability reaches about 0.80 by late adolescence and stays there into adulthood.4PubMed. The Wilson Effect: the increase in heritability of IQ with age One way to think about this: as children grow older, they increasingly select and shape their own environments in ways that align with their genetic predispositions. A child with a genetic tilt toward curiosity gravitates toward more books, more stimulating friendships, more academic challenges, amplifying the genetic signal over time.
Personality, Weight, and Disease
Twin studies have also been central to understanding personality. The “Big Five” traits, openness, conscientiousness, extraversion, agreeableness, and neuroticism, show heritabilities in the range of about 40 to 60%.5PubMed Central. Heritability estimates of the Big Five personality traits based on common genetic variants One early twin study estimated the genetic influence on each dimension individually, finding values of 41% for neuroticism, 53% for extraversion, 61% for openness, 41% for agreeableness, and 44% for conscientiousness.6PubMed. Heritability of the big five personality dimensions and their facets: a twin study Later studies using different populations have produced somewhat variable numbers, as you would expect; a Danish study found estimates ranging from about 0.15 to 0.56 depending on the specific trait and sex of the participants.7Personality and Individual Differences. The five factor model of personality and heritability: Evidence from Denmark The general pattern, though, holds up: personality is moderately heritable, with the rest of the variation largely attributable to non-shared environmental factors. The Minnesota reared-apart data showed a similar story, with identical twins separated at birth resembling each other on personality inventories about as much as identical twins raised together did.8PubMed. Normal and abnormal personality traits: evidence for genetic and environmental relationships in the Minnesota Study of Twins Reared Apart
Body weight follows a broadly similar pattern. A classic study of identical twins reared apart found intrapair correlations for body mass index of about 0.70 for men and 0.66 for women, with the shared childhood environment contributing essentially nothing to adult weight similarity.9PubMed. The body-mass index of twins who have been reared apart An international study of over 12,000 twin pairs added an important developmental wrinkle: heritability for body weight and BMI was low at birth (under 9%) but increased sharply after the first few months of life, while common environmental influences, which accounted for 74–87% of variation at birth, faded markedly over time.10PLoS ONE. Genetic and Environmental Contributions to Weight, Height, and BMI from Birth to 19 Years of Age: An International Study of Over 12,000 Twin Pairs
In disease research, twin concordance rates have provided some of the sharpest evidence for genetic components. Schizophrenia is a striking example: one study found concordance of 48% in identical twins versus just 4% in fraternal twins.11PubMed. Twin concordance for DSM-III-R schizophrenia That gap clearly points to genetics, but the fact that half of identical co-twins are discordant (one has schizophrenia, the other does not) equally clearly shows that genes are not the whole story. Other conditions tell different tales. A study of gout in twins found that the disease itself was driven primarily by shared environmental factors rather than genetics, even though uric acid levels (the underlying metabolic marker) were substantially heritable.12PubMed. Nature versus nurture in gout: a twin study In other words, having the genetic tendency toward high uric acid did not automatically mean a person would develop gout; environmental factors like diet tipped the balance.
The Equal Environments Assumption
The whole twin method depends on a key assumption: that identical and fraternal twins share their environments to the same degree. If identical twins are treated more similarly than fraternal twins, say dressed alike more often, placed in the same classroom, or encouraged to pursue the same hobbies, then some of their extra resemblance on a measured trait might come from their more similar upbringing, not their shared DNA. Researchers call this the “equal environments assumption,” and it has been contested almost since twin studies began.
Most empirical tests of the assumption have supported it, at least approximately. A comprehensive evaluation using nationally representative twin data found that the assumption is probably not perfectly valid for most outcomes, but that the resulting bias is likely modest.13PubMed. What can we learn from twin studies? A comprehensive evaluation of the equal environments assumption Another study examining the psychometric properties of an “equal environments” questionnaire found that fraternal twins reported significantly less similar childhood and teenage environments than identical twins, which is exactly what critics worry about.14PubMed Central. An investigation of a measure of twins’ equal environments The debate is far from resolved, but the mainstream view is that violations of the assumption push heritability estimates up by a small amount, not enough to invalidate the method but enough to warrant caution about treating any single estimate as perfectly precise.
Why Identical Twins Are Not Truly Identical
A separate complication is that identical twins are not genetically identical in the strict sense. After the fertilized egg splits, each twin accumulates independent mutations as their cells divide. These postzygotic somatic mutations are DNA changes that occur in body cells after fertilization, caused by errors during replication or environmental exposures like ultraviolet radiation.15DNA Research. Functional landscape of genome-wide postzygotic somatic mutations between monozygotic twins Some of these differences may be relevant to disease: whole-genome sequencing of identical twins discordant for psychosis has found DNA variants that differ between the pair, raising the possibility that post-twinning mutations contribute to conditions like schizophrenia.16Translational Psychiatry. Whole genome sequencing study of identical twins discordant for psychosis
Beyond raw DNA sequence, identical twins also diverge at the epigenetic level, meaning in the chemical tags on their DNA that influence which genes are active without changing the underlying code. A landmark study found that while young identical twins were epigenetically indistinguishable, older pairs showed striking differences in DNA methylation and histone modification, changes that altered their patterns of gene activity.17PubMed Central. Epigenetic differences arise during the lifetime of monozygotic twins These accumulating epigenetic differences help explain why identical twins can diverge more as they age, and why one twin might develop a disease the other never gets.
The Prenatal Environment Problem
Even before birth, identical twins do not necessarily share exactly the same environment. About two-thirds of identical twin pairs are monochorionic, meaning they share a single placenta. The rest are dichorionic, with separate placentas. Sharing a placenta can create unequal blood supply, making one twin grow more slowly or creating differences in nutrient delivery. Research on this question has found that monochorionic twins tend to be more discordant for birth weight and related measures than dichorionic twins, and that failing to account for chorionicity can bias heritability estimates for traits like weight and height in early childhood.18PubMed Central. The Prenatal Environment in Twin Studies: A Review on Chorionicity
That said, the impact appears to be limited. A large study testing 66 different traits, including cognitive function, personality, and child behavior, found that chorionicity affected twin similarity for only about 10 of those traits, mostly ones closely tied to birth weight in young children.19PubMed Central. Chorionicity and Heritability Estimates from Twin Studies: The Prenatal Environment of Twins and Their Resemblance Across a Large Number of Traits For most psychological and behavioral outcomes later in life, the prenatal environment does not appear to substantially distort twin-based heritability estimates.
The Missing Heritability Gap
One of the more uncomfortable discoveries of the genomics era is that when researchers scan the entire genome for specific DNA variants associated with a trait, the variants they find collectively explain far less variation than twin studies predicted. This discrepancy has been called “missing heritability.” For intelligence, twin and family studies suggest about 50% of the variation is genetic, yet genome-wide association studies initially accounted for only about 10%.20PubMed Central. Three legs of the missing heritability problem The gap has narrowed with larger DNA datasets, but it has not closed.
Several explanations have been proposed. Twin studies may inflate heritability somewhat due to violations of the equal environments assumption or by absorbing gene-by-environment interactions into the “genetic” bucket. There may also be many rare genetic variants with small individual effects that current DNA scans are underpowered to detect. And some of the twin-study heritability may reflect non-additive genetic effects (interactions between genes at different locations) that genome-wide studies are not designed to capture. Another factor is assortative mating: people tend to pair up with partners who resemble them on traits like intelligence, which inflates genetic similarity between fraternal twins and, if unmodeled, can distort heritability estimates. One analysis found that when assortative mating was properly accounted for, the heritability of adult intelligence dropped from the often-quoted ~80% to about 44%, with an additional 27% attributable to non-additive genetic effects and 11% to assortative mating itself.21PubMed Central. Reconsidering the heritability of intelligence in adulthood: taking assortative mating and cultural transmission into account
Extending the Classic Design
Researchers have responded to some of these concerns by building more sophisticated study designs. The “extended twin family design” adds parents, spouses, siblings, and sometimes children of twins into the analysis, which allows the separation of more parameters than the classic twin-pair comparison can handle on its own. These designs require less stringent assumptions and can produce less biased estimates than the traditional approach.22PubMed Central. Are extended twin family designs worth the trouble? A comparison of the bias, precision, and accuracy of parameters estimated in four twin family models For example, an extended kinship model was used to examine smoking initiation, allowing researchers to simultaneously test for additive and non-additive genetic effects, shared and individual-specific environmental influences, assortative mating, and cultural transmission within a single analysis.23PubMed. Genetic and cultural transmission of smoking initiation: an extended twin kinship model
Another creative approach involves “virtual twins,” unrelated children of the same age raised together in the same family (often through adoption). Because they share their family environment but none of their genes, they provide a direct estimate of how much the shared environment contributes to a trait, something that is harder to isolate cleanly from the standard MZ-DZ comparison alone.24PubMed Central. Genetic and environmental contributions to body mass index: comparative analysis of monozygotic twins, dizygotic twins and same-age unrelated siblings Large longitudinal twin registries, now operating in multiple countries, have also enabled researchers to track how heritability shifts across the lifespan for a wide range of traits.25Twin Research and Human Genetics. An Update on Longitudinal Twin and Family Studies
Surprising Findings in Politics and Social Attitudes
Perhaps the most counterintuitive results from twin research involve political beliefs. It is natural to assume that political ideology is shaped by upbringing, media consumption, and life experience. And those factors do matter. But a combined analysis of over 12,000 twin pairs from five democracies, sampled across four decades, found genetic influences on political ideology regardless of how ideology was measured, the era, or the country.26PubMed Central. Genetic influences on political ideologies: twin analyses of 19 measures of political ideologies from five democracies and genome-wide findings from three populations Heritability for political orientation is estimated at about 40%.27Current Opinion in Behavioral Sciences. On the genetic basis of political orientation Even specific foreign-policy preferences, like the degree to which a person leans hawkish versus dovish, show a moderate genetic component and appear to share a common genetic source of variation with broader political ideology.28PubMed. The heritability of foreign policy preferences
These findings do not mean there is a “conservative gene” or a “liberal gene.” What the data suggest is that genetically influenced personality traits, things like openness to experience, threat sensitivity, and comfort with ambiguity, nudge people toward certain political orientations. The genes are acting on temperament, and temperament shapes the lens through which people evaluate political questions. The environment still accounts for the majority of the variation, and obviously no one is born with an opinion about tax policy.
Ethical Shadows
The history of twin research is not entirely clean. The most notorious modern case involves a study directed by Dr. Peter Neubauer beginning in the 1960s, in which identical siblings were deliberately separated at adoption and placed into different families without the knowledge of the adoptive parents, specifically so researchers could study nature versus nurture. The study ran for decades, and the research records were sealed at Yale University with the condition that they not be opened until 2065.29PubMed. Experiment on identical siblings separated at birth: ethical implications for researchers, universities, and archives today The case drew wide attention after several of the separated twins discovered one another as adults, and it remains a cautionary example of how the drive for scientific knowledge can override the welfare of research subjects. The twins involved were never told they had siblings, and some of the reunited sets reported significant emotional harm from the separation.
Modern twin studies rely on voluntary registries and standard ethics oversight, and nothing like the Neubauer study would be approved by a contemporary review board. But the episode underscores why the public sometimes views behavior genetics with suspicion, and why researchers in the field have worked to distinguish their current practice from its more exploitative past.
What Twin Studies Cannot Tell You
A common misreading of twin research is to treat heritability as destiny. A trait can be 80% heritable and still be dramatically changeable by the environment. Height is one of the most heritable human traits, yet average height has increased substantially in many countries over the past century as nutrition improved. Heritability describes sources of variation within a specific population at a specific time; it says nothing about the limits of what interventions can achieve. A society that equalizes environments for everyone would, paradoxically, see heritability for most traits go up, because the remaining variation would be almost entirely genetic.
Twin studies also cannot identify specific genes. They tell you that genetic variation matters for a trait, but they do not tell you which variants, or how many, or how they work. That is the job of molecular genetics, and as the missing-heritability gap illustrates, the two approaches have not always converged smoothly. Twin studies are a starting point, not an endpoint: they map out the broad territory of genetic and environmental influence so that more targeted research knows where to look.
Another limitation is that most large twin registries have been assembled in wealthy, predominantly white Western countries. Heritability estimates derived from these populations do not automatically apply elsewhere. A trait that is highly heritable in a society with good nutrition and universal schooling might show lower heritability in a society where access to those resources varies widely, because environmental variation would then account for a larger share of the differences between people. The numbers from twin studies are always context-dependent, even if they are sometimes reported as though they are universal constants.