Heritability

Lore & Background

Heritability measures the fraction of phenotype variability that can be attributed to genetic variation, not the fraction of an individual phenotype caused by genetics. For example, it is incorrect to say that since the heritability of personality traits is about 0.6, that means 60% of a person's personality is inherited from parents and 40% comes from the environment. Heritability can change without any genetic change occurring, such as when the environment starts contributing to more variation. It is specific to a particular population in a particular environment, and high heritability does not necessarily mean the trait is not susceptible to environmental influences.

Reader's Guide

Heritability is estimated by comparing individual phenotypic variation among related individuals, by examining the association between individual phenotype and genotype data, or by modeling summary-level data from genome-wide association studies (GWAS). It is expressed in two forms: broad-sense heritability (H²), which includes all genetic contributions such as additive, dominant, and epistatic effects, and narrow-sense heritability (h²), which includes only additive genetic variance. Additive variance is important for selection, as the response of a trait to selective pressure is directly related to narrow-sense heritability. Heritability analyses require population variation and cannot account for factors that are invariant in the population. The concept is often misunderstood, leading to the incorrect impression that behavioral traits are 'inherited' or specifically passed down through genes.

Did You Know?

• Heritability can change without any genetic change occurring, such as when the environment starts contributing to more variation.
• Heritability is specific to a particular population in a particular environment.
• High heritability of a trait does not necessarily mean the trait is not very susceptible to environmental influences.
• Heritability cannot take into account the effect of factors which are invariant in the population.

The IQ Heritability Debate and Its Estimates

The question of how much genetic variation accounts for differences in IQ among individuals has stirred academic controversy since the late 1800s. Intelligence within the normal range is not governed by a single gene but is a polygenic trait shaped by at least five hundred distinct genetic factors, which complicates any straightforward attribution. Behavioral genetic research across diverse populations has generally converged on a heritability estimate of roughly 0.5, suggesting that about half the variation in IQ scores can be linked to genetic differences. However, the relationship between age and heritability remains unsettled; most researchers observe that heritability appears to rise across the lifespan, a pattern they attribute to growing gene-environment correlations rather than a simple biological shift. More recent molecular-genetic studies have produced lower estimates than classic twin-based designs, giving rise to what researchers call the 'missing heritability problem'—a gap between the heritability implied by family resemblance and the portion actually explained by identified genetic variants.

Critical Caveats in Interpreting Heritability

Heritability is often misunderstood as a fixed proportion of a trait 'caused' by genes, but it actually measures the share of variation in a trait within a specific population that is attributable to genetic differences. This distinction carries profound implications. If an environmental change affects every member of a population equally—say, improved nutrition raising average height—the mean shifts while the heritability figure stays the same, because individual differences remain unchanged. Conversely, heritability is not immutable: if environmental variation among individuals increases, the heritability estimate drops; if everyone experiences identical conditions, it would approach 1.0. Phenylketonuria illustrates this vividly—once a genetic cause of intellectual disability with a heritability near 100 percent, it can now be mitigated through dietary modification, lowering its heritability. Similarly, vocabulary size is substantially heritable even though every single word a person knows was acquired through learning. These examples underscore that a high heritability figure tells us nothing about how easily a trait can be altered or whether group-level differences stem from genetics.

Individual Variation Versus Group Differences

A point of frequent confusion—and serious scientific importance—is the logical leap from individual-level heritability to explanations of group-level disparities. The research literature makes clear that even though IQ differences among individuals within a population carry a hereditary component, this does not logically extend to mean that average IQ differences between racial or ethnic groups have a genetic basis. The scientific consensus is that genetics does not account for average differences in IQ test performance across racial groups. This distinction matters because heritability is a population-specific statistic: it quantifies how much of the spread in scores around the mean in one particular group is associated with genetic variation. It says nothing about why one group's mean sits higher or lower than another's. Environmental, socioeconomic, educational, and cultural factors that differ systematically between groups can shift means without altering the within-group heritability estimate at all. Conflating these two levels of analysis remains one of the most persistent and consequential errors in public discourse about intelligence and genetics.

Polygenic Inheritance and Regression Toward the Mean

A common intuition is that two parents with high IQ scores will reliably produce children of equal or greater intelligence. The genetics of polygenic traits, however, tells a more nuanced story. While a heritable trait is indeed more likely to appear in the offspring of two parents who both express it at a high level than in the offspring of two randomly selected parents, the more extreme the parental expression, the less likely the child is to match that same extreme. In practice, parents whose IQ sits at either tail of the distribution are more likely to have children whose scores regress toward the population mean than to have children who match or exceed their own scores. The same principle applies to height: the child of two exceptionally tall parents is probable to be taller than the average person, yet unlikely to surpass both parents in stature. What remains is a higher probability that the child will express the trait at all, rather than at the same intensity. This regression toward the mean is a statistical inevitability of polygenic inheritance, where hundreds of small-effect alleles combine, and it tempers any simplistic 'like begets like' expectation for complex traits such as intelligence.