Behavioural Genetics

Heritability (h²) is the proportion of observed variance in a trait in a specific population that is attributable to genetic differences between individuals. It ranges from 0 (no genetic contribution to variance) to 1 (all variance is genetic). Critically, heritability is a population statistic, not a property of individuals: it says nothing about how much of any particular person's trait is "due to genes." It is also specific to a population and an environment: heritability of height is ~90% in rich countries (where nutrition is uniform, so most variation is genetic) but lower where nutrition varies dramatically. A highly heritable trait is not immutable: reading ability is substantially heritable, but literacy rates have been transformed by universal education. Heritability does not imply genetic determinism.

Non-shared environment (E) refers to environmental factors that make siblings reared together different from one another — as opposed to shared environment (C), which makes them similar. In most personality and psychopathology measures, E accounts for a very large fraction of variance, often equal to or greater than heritability. Sources of non-shared environment include: differential peer group experiences, birth order effects (though often modest), unique life events, idiosyncratic treatment by parents (they don't treat all children identically), intrauterine variation (MZ twins share a womb but may differ in placentation), and random developmental noise. E in twin studies also includes measurement error, which is why twin-study E estimates are overestimates of true environmental non-sharing. Identifying specific non-shared environmental variables that reliably predict trait variance has proven difficult.

Intelligence is one of the most-studied traits in behavioural genetics. MZ twin correlations for IQ are approximately 0.85; DZ correlations approximately 0.60; adopted siblings converge toward zero correlation by adulthood. Heritability estimates for general intelligence in adulthood consistently cluster around 60–80% in Western samples — and, remarkably, heritability increases from childhood (~40%) to adulthood (~60–80%), while shared environmental effects decrease. Plomin and colleagues argue this reflects gene-environment correlation: as children grow older and gain autonomy, they increasingly select and create environments that fit their genetic propensities — "the child creates their own environment." Shared environment matters more in childhood (when parents control the environment) and for children raised in poverty.

Twin and adoption studies indicate that complex psychological traits like intelligence, personality, and schizophrenia are 50–80% heritable. Modern genome-wide association studies (GWAS) identifying common genetic variants (SNPs) explain only a small fraction of this estimated heritability — often < 20%. This gap is "missing heritability." Several explanations are debated: rare variants not captured by GWAS; gene-gene interactions (epistasis); gene-environment interactions; epigenetic mechanisms; overestimated twin heritability due to violated assumptions (e.g., assortative mating, gene-environment covariance); and the statistical challenge of aggregating many small genetic effects. Polygenic scores — summing thousands of SNP effects — now explain ~10–15% of IQ variance in population samples, a substantial improvement but still far short of heritability estimates.