Behavioural Genetics

No — this is one of the most common misunderstandings in behavioural genetics. High heritability estimates the sources of variation within a population under current environmental conditions; it says nothing about what would happen if environments changed dramatically. Height has heritability of ~80% in Western populations, yet average heights increased by ~10 cm across the 20th century due to improved nutrition and healthcare. Phenylketonuria (PKU) is essentially 100% heritable, yet is completely preventable by a dietary intervention (phenylalanine restriction). Reading disability is ~60–70% heritable, yet is highly responsive to intensive phonics instruction. Heritability and malleability are orthogonal.

Twin studies suggest high heritability for many traits (IQ ~60–80%, height ~80%, schizophrenia ~80%). But genome-wide association studies (GWAS) — looking for individual gene variants associated with these traits — typically find many variants each with tiny effects that together explain only a fraction of the twin-estimated heritability. This gap is "missing heritability." Possible explanations include: rare variants not captured by standard GWAS arrays; gene-gene interactions (epistasis) not modelled; copy number variations; epigenetic effects not measured by standard genomics; and possibly twin-method heritability overestimation due to violated assumptions. Increasingly large GWAS (UK Biobank, hundreds of thousands of participants) are recovering more heritability through polygenic scores, but the gap has not fully closed.

Gene-environment correlation (rGE) is a significant challenge in twin methodology. If genetically similar twins also end up in more similar environments (because their genotype influences their environmental exposures — active rGE), this inflates heritability estimates. Twin researchers address this through: (1) Comparing MZ twins raised apart vs together — differences between these groups estimate the effect of shared environment independent of genetics. (2) Using measured environments as phenotypes and examining their heritability directly (showing that even environmental measures like "stressful life events" are partly heritable). (3) Multivariate genetic analyses that simultaneously model genetic and environmental paths. (4) Molecular genetic controls: comparing twin-method heritability to GWAS-based heritability (which does not assume random environmental distribution).

One of behavioural genetics' most counterintuitive findings — Plomin's "non-shared environment paradox" — is that siblings raised in the same family are nearly as different from each other in personality as unrelated individuals, despite sharing genes, parents, home, school, and socioeconomic status. This means the shared family environment (c²) contributes virtually nothing to adult personality (c² ≈ 0 for most Big Five traits), while non-shared experiences account for ~40–50% of variance (alongside genetics). Why? Children in the same family occupy different micro-environments: different friends, teachers, birth-order positions, accidents, peer experiences. They also elicit different reactions from the same parents (evocative rGE). What feels like a "shared" environment is, from each child's developmental trajectory, a quite different experience. Turkheimer (2000) termed this one of the "three laws" of behavioural genetics: the effect of being raised in the same family is negligible.

IQ is one of the most extensively studied traits in behavioural genetics. Key findings: (1) Heritability increases with age — from ~40% in childhood to ~60–80% in adulthood and ~80% in older adults (Plomin et al., 2016). (2) MZ twins raised apart show remarkably similar IQ scores (r ≈ .75), implicating genetic factors even in the absence of shared environment. (3) Shared family environment contributes substantially in childhood (c² ≈ 25–35%) but near zero in adulthood — suggesting early environmental advantage effects fade as adults increasingly select and shape their own environments (active niche-picking). (4) GWAS polygenic scores now explain ~25–30% of IQ variance using measured DNA, partially resolving the "missing heritability" for this trait. These findings do not imply that IQ is fixed or unresponsive to education: the Flynn Effect (IQ gains of ~3 points per decade in the 20th century) shows massive environmental effects on mean IQ levels.