Is IQ Genetic? What Twin Studies Actually Show

The short version: in adults, roughly 80% of the variation between people in measured IQ tracks genetic variation, and the influence of the shared family environment falls to around 10%. Both of those numbers move a great deal with age, and in opposite directions to what most people expect.

That finding is easy to state and remarkably easy to misread, so this page spends as much time on what it does not mean as on what it does. It also settles the single most persistent myth in the subject, which is the claim that intelligence comes from your mother.

How much of IQ is genetic?

Twin and adoption studies estimate how much of the variation in IQ within a population tracks genetic variation. The headline finding, summarised by Bouchard (2013) under the name the Wilson effect, is that this proportion changes systematically as people grow up, and it changes in the direction almost nobody predicts.

The intuition most people carry is that childhood shapes you and the imprint lasts. What the data shows is close to the reverse. The household you grew up in has a substantial, measurable effect on your IQ as a child, and that effect fades as you age, until by adulthood it accounts for very little of the difference between people.

The usual explanation is that as people gain independence they increasingly select their own environments. A child reads what is in the house; an adult chooses what to read, who to work with and what to spend years getting good at. Those choices correlate with disposition, so what looks like a genetic effect partly operates through the environments people build around themselves.

Is it nature or nurture?

The question is slightly malformed, and that is not a dodge.

Genes and environment are not two taps filling the same bucket, with percentages that sum to a hundred. They interact. A genetic predisposition affects which environments you seek, how people respond to you, what you find rewarding, and therefore what you practise. The environment then affects how that predisposition expresses itself. When a study reports a heritability of 0.80, it is not saying that 80% of intelligence is genes and 20% is upbringing. It is saying that, in that population under those conditions, 80% of the differences between people statistically track genetic differences.

Change the population or the conditions and the number changes. Where environments vary enormously, such as when some children are malnourished and others are not, the environmental share rises and heritability falls. Where nearly everyone has adequate food, schooling and healthcare, the remaining differences are more genetic almost by definition, because the environmental differences that would have mattered have been removed.

This is worth holding onto, because it means a high heritability estimate is partly a statement about how equal a society is.

Does IQ come from your mother or your father?

From both, in roughly equal measure. The claim that intelligence is inherited only from the mother is false, and it is worth knowing where it came from because the origin explains the confusion.

The idea circulates in the form of a specific-sounding genetic argument: that intelligence-related genes sit on the X chromosome, which sons receive only from their mother, and that “conditioned genes” from fathers are switched off. It traces back to a single 1997 study concerning one X-linked gene relating to social cognition in girls with Turner syndrome. That study was not about general intelligence, did not claim to be, and cannot support the conclusion that has been built on it.

The genomic evidence since then points clearly the other way. Large genome-wide studies find cognition-linked variants distributed across the autosomal chromosomes, the 22 pairs that are not the X or Y, with no concentration on the X. Neither parent contributes disproportionately.

The claim has been fact-checked and rejected repeatedly by Snopes, Africa Check and the Genetic Literacy Project, among others. It keeps circulating because it is a satisfying story with a mechanism attached, which is exactly the shape of a claim that spreads faster than its correction.

Is there a gene for intelligence?

No, and not remotely. Intelligence is polygenic, meaning it reflects the combined contribution of a very large number of genetic variants, each of which does almost nothing on its own.

Davies and colleagues (2018) analysed over 300,000 people and identified 148 independent genetic loci associated with general cognitive function. That sounds like a lot until you consider the effect sizes: individually, these variants account for fractions of a single IQ point. No one of them is a gene “for” intelligence in any useful sense, and the full set still explains only part of the heritability that twin studies imply.

This matters practically. It means there is no test that reads your genome and tells you your cognitive potential, and anyone selling one is overstating what the science supports.

What affects IQ besides genes?

Several things, with education having the best-evidenced effect in well-resourced populations and early-life factors mattering most where conditions are worst.

• Education. Ritchie and Tucker-Drob (2018) meta-analysed the question and estimated a gain of roughly 1 to 5 points per additional year of schooling, using designs such as changes to school leaving ages that separate causation from selection. The effect appears to persist rather than fade.
• Early nutrition and health. Severe malnutrition in early childhood causes real and lasting reductions. So does serious illness during development.
• Lead exposure. One of the clearest environmental effects on population-level cognitive ability, and one of the reasons removing lead from petrol and paint is considered among the highest-value public health interventions of the last century.
• Test-day condition. Not a change in ability, but a real change in the score. Sleep deprivation, alcohol, illness and stress all depress performance on the day, which is covered in our guide to how IQ changes over a lifetime.

What has notably weak evidence is commercial brain training. Gains tend to be specific to the trained task and transfer poorly to general ability, as set out in our guide to improving your IQ.

Will your children inherit your IQ?

They will tend to land above average if you are above average, and closer to the population mean than you are. Francis Galton identified this pattern in the nineteenth century studying height and named it regression toward the mean.

Reed and Rich (1982) quantified it for IQ using over 2,000 parent-offspring pairs. The regression coefficient of a child’s IQ on the mid-parental average was about 0.61, falling to roughly 0.50 where the mid-parental average was above 114. In rough terms, two parents averaging 130 would expect children averaging somewhere in the high 110s.

Two things to hold alongside that. It is an expected average across many families, and the scatter is large, with siblings raised together routinely differing by more than ten points. And it works in both directions: parents well below the mean tend to have children closer to it from the other side. Regression toward the mean is not a story about decline. It is what happens to any trait measured imperfectly and influenced by many partly independent factors.

What heritability does not mean

This is the part that gets mangled most often, so it is worth being explicit. A heritability estimate of 0.80 does not mean any of the following.

• It does not mean IQ is fixed. Heritability describes the sources of variation under the conditions studied. It is not a measure of how changeable a trait is. Height is highly heritable and average height still rose dramatically across the twentieth century with better nutrition.
• It does not apply to you as an individual. There is no meaningful sense in which 80% of your score is genetic. The statistic describes variance across a population, and variance is not a property any single person has.
• It says nothing about differences between groups. Within-group heritability carries no implication for what causes differences between populations. This is a long-established point in behaviour genetics and it is routinely misrepresented by people who should know better.
• It is not a fixed constant of nature. The figure depends on the population and the conditions. Estimates come overwhelmingly from wealthy industrialised democracies, and where environments vary more widely, heritability comes out lower.

Frequently asked questions

The bottom line

IQ is substantially heritable in adults, with roughly 80% of the variation between people in a studied population tracking genetic variation, and the shared family environment falling to around 10% by adulthood. The genetic contribution grows with age rather than shrinking, which is the opposite of what most people assume.

It does not come from your mother. It does not come from one gene, or a hundred and forty-eight genes in any simple way. And heritable has never meant fixed: the estimate describes where differences come from in a particular population under particular conditions, and both of those can change.