Blade Runner had an impact on me, both as a film and because it was an introduction to the writings of Philip K Dick, whose whimsical work was based on wondering what it meant to be human. Are we as individuals merely constructions of fundamental genetic coding mechanisms, which create treasured but probably false memories of childhood and delusions of uniqueness which give us a sense of identity, but which serve no purpose other than to keep us going, as our genes require? Dick’s speculative hunt for fake humans cast the Turing test in a new light: not as a proof of the ability of artificial intelligence to “pass for human”, but as a test of humanity itself, which leads to a most severe demotion for those who fail to convince interlocutors that they are really human. All that aside, I revere him for his aphorism: Reality is that which, when you stop believing in it, doesn’t go away.
My text is the recent paper by Plomin, DeFries, Knopik and Neiderhiser “Top 10 Replicated Findings From Behavioral Genetics. Perspectives on Psychological Science” 2016, Vol. 11(1) 3–23. The deeper background is the major textbook in the field by those same authors 2013 Behavioral genetics (6th ed.). New York, NY: Worth.
By the way, 6 editions should give pause, should it not, to those who just want to jump into the field without doing any of the reading. Below I list, in very abbreviated form, and largely based on the text with a few additions and simplifications, 10 genetic findings which replicate:
For example, a review of the world’s literature on intelligence that included 10,000 pairs of twins showed identical twins to be significantly more similar than fraternal twins (twin correlations of about .85 and .60, respectively), with corroborating results from family and adoption studies, implying significant genetic influence.
For example, for psychopathology a meta-analysis of 14 twin studies of schizophrenia showed monozygotic (MZ) concordances of about 50% and dizygotic (DZ) concordances of about 15%, suggesting significant genetic influence; this finding has been corroborated in more recent studies, as well as in adoption studies.
For personality, scores of twin studies over the decades have shown evidence of significant genetic influence for dozens of traits studied using self-report questionnaires; results have been confirmed in meta-analyses with adoption and family data as well as twin data on 24,000 pairs of twins. Traits such as political beliefs, religiosity, altruism, and food preferences also have shown significant genetic influence. A recent meta-analysis of data drawn from 3,000 publications on nearly 18,000 traits of 15 million twin pairs showed that this finding is not limited to psychological traits.
For general intelligence, heritability estimates are typically about 50% in meta-analyses of older family, twin, and adoption studies as well as newer twin studies. For personality, heritability estimates are usually between 30% and 50%. For example, well-being is a relative newcomer in relation to genetic analyses of personality; a meta-analytic review of 10 studies based on 56,000 individuals yielded a heritability estimate of 36%.
Some traits, such as individual differences in height, yield heritability as high as 90%. Behavioural traits are less reliably measured than physical traits such as height, and error of measurement contributes to nonheritable variance. Many have noted that no traits are 100% heritable.
Although this finding might seem obvious and unsurprising, it is crucial because it provides the strongest available evidence for the importance of environmental influence after controlling for genetic influence. Because genetic influence is significant and substantial, one must control for genetic influence when investigating environmental influence. Environmental research using genetically sensitive designs has led to three of the most important discoveries about the way the environment affects behavioural development, presented as Findings 7, 8, and 9.
If only a few genes were responsible for the heritability of a trait, selected lines would separate after a few generations and would not diverge any further in later generations. In contrast, selection studies of complex traits show a linear response to selection even after dozens of generations of selection, as seen, for example (Fig. 1), in one of the largest and longest selection studies of behaviour that included replicate selected and control lines (DeFries, Gervais, & Thomas, 1978).
It is a pity this figure is not in colour, but the findings are a brilliant example of the breeder’s equation in action (R = h2 S. R is the response to selection, S is the selection differential, and h2 is the narrow-sense heritability.)
The top lines are those mouse strains selected for high activity even in stressful situations, which constitute fearless explorers (non-reactives) while the bottom lines are fearful mice, rooted to the spot in stressful settings (reactives) and shitting frequently. (One of my colleagues, the late Gudrun Sartory, charged with counting the boluses of faeces emitted by such mice, posed a methodological dilemma over a tea break in the Institute of Psychiatry canteen: one terrified mouse had eaten a bolus in panic. Should she count, as required by the test protocol, only the boluses at the end of the experiment, or add the eaten one to the final total?)
Without any form of selection, the middle lines reveal average levels of fear. The difference between the fearless and fearful strains is evident after 7 generations and undeniable after 10 generations. It is selective breeding imposed from outside the population which has caused this massive difference over time, with clear implications for humans under strong selection for any particular heritable characteristic, who might be expected to be very different after 10 generations, say 250 years.
Another overlooked point from selection studies is that genetic effects transmitted from parents to offspring can be due only to additive genetic effects (the independent effects of alleles and loci that “add up”) in contrast to non-additive genetic effects in which the effects of alleles and loci interact. This is important information because it would be difficult to identify specific DNA differences responsible for heritability if genetic effects on behavior were caused by interactions among many loci (epistasis).
Phenotypic covariance between traits is significantly and substantially caused by genetic covariance, not just environmentally driven covariance. That is to say, taking genetics into account shows that many traits share genetic pathways.
More than 100 twin studies have addressed the key question of co-morbidity in psychopathology (having more than one diagnosed disorder), and this body of research also consistently shows substantial genetic overlap between common disorders in children and in adults. For example, a review of 23 twin studies and 12 family studies confirmed that anxiety and depression are correlated entirely for genetic reasons. In other words, the same genes affect both disorders, meaning that from a genetic perspective they are the same disorder.
The genetic structure of psychopathology does not map neatly onto current diagnostic classifications. Moreover, correlations between personality dimensions and psychopathological diagnoses also are mediated genetically, most notably between neuroticism and depression.
Shared environmental effects, as from family life and school, decrease with age. Good family lives and good schools are not the essential start in life that many people have always imagined, or at least not crucial in societies where family life and schools are reasonably good.
Longitudinal genetic studies consistently show that phenotypic correlations from age to age are largely due to genetic stability. In other words, genetic effects contribute to continuity (the same genes affect the trait across age), whereas age-to-age change is primarily the provenance of environmental factors.
For intelligence, similar results have been found, for example, in a meta-analysis of 15 longitudinal studies. This finding creates an apparent paradox: How can the heritability of intelligence increase so substantially throughout development if genetic effects are stable? How can the same genes largely affect intelligence across the life course and yet account for more variance as time goes by? Increasing heritability despite genetic stability implies some contribution from what has been called genetic amplification (Plomin & DeFries, 1985). In other words, genetic nudges early in development are magnified as time goes by, increasing heritability, but the same genetic propensities continue to affect behavior throughout the life course.
This amplification model has recently been supported in a meta-analysis of 11,500 twin and sibling pairs with longitudinal data on intelligence, which showed that a genetic amplification model fit the data better than a model in which new genetic influences arise across time. Genotype/environment correlation seems the most likely explanation, in which small genetic differences are amplified as children select, modify, and create environments correlated with their genetic propensities.
Although it might seem a peculiar thing to do, measures of the environment widely used in psychological science—such as parenting, social support, and life events—can be treated as dependent measures in genetic analyses.
If they are truly measures of the environment, they should not show genetic influence. To the contrary, a review of the first 18 studies in which environmental measures were used as dependent measures in genetically sensitive designs and found evidence for genetic influence for these measures of the environment. Significant genetic influence was found for objective measures such as videotaped observations of parenting as well as self-report measures of parenting, social support, and life events. How can measures of the environment show genetic influence?
The reason appears to be that such measures do not assess the environment independent of the person. As noted earlier, humans select, modify, and create environments correlated with their genetic behavioral propensities such as personality and psychopathology. For example, in studies of twin children, parenting has been found to reflect genetic differences in children’s characteristics such as personality and psychopathology.
For example, rather than assuming that correlations between parenting and children’s behavior are caused by the environmental effect of parenting on children’s behavior, one should consider the possibility that the correlation is in part due to genetic factors that influence both parenting and children’s behavior. Individual differences in parenting might reflect genetically driven differences in children’s behaviour, or differences in parenting might be due to genetically driven propensities of parents that are inherited directly by their children. For example, for children aged 2 years, the correlation between the Home Observation for Measurement of the Environment was .44 in nonadoptive families, (in which parents shared nature as well as nurture with their offspring), compared with .29 in adoptive families in which parents and offspring were genetically unrelated.
Disentangling genetic and environmental influences on correlations between environmental and behavioral measures is important because 1) if these correlations are mediated genetically, interpretations that assume environmental causation are wrong, with important implications for intervention 2) genetically sensitive designs can identify causal effects of the environment free of genetic confounds 3) genetic mediation of the association between environmental measures and behavioral traits suggests a general way of thinking about how genotypes develop into phenotypes, moving from a passive model of imposed environments to an active model of shaped experiences in which humans select, modify, and create experiences in part based on their genetic propensities.
This is an extraordinary finding, and overturns many long-held assumptions. It is reasonable to think that growing up in the same family makes brothers and sisters similar psychologically, which is what developmental theorists from Freud onwards have assumed. However, for most behavioral dimensions and disorders, it is genetics that accounts for similarity among siblings. Although environmental effects have a major impact (see Finding 2), the salient environmental influences do not make siblings growing up in the same family similar. The message is not that family experiences are unimportant but rather that the relevant experiences are specific to each child in the family. This finding was ignored when it was first noted and controversial when it was first highlighted, but it is now widely accepted because it has consistently replicated. The acceptance is so complete that the focus now is on finding any shared environmental influence, for example, for personality and some aspects of childhood psychopathology.
For instance, for antisocial behavior in adolescence, shared environment accounts for about 15% of the total phenotypic variance; however, even here non-shared (unique) environment accounts for more of the variance, about 40% in meta-analyses, although this estimate includes variance due to error of measurement. Academic achievement consistently shows some shared environmental influence, presumably due to the effect of schools, although the effect is surprisingly modest in its magnitude (about 15% for English and 10% for mathematics) given that this result is based on siblings growing up in the same family and being taught in the same school. An interesting developmental exception is that shared environmental influence is found for intelligence up until adolescence and then diminishes as adolescents begin to make their own way in the world, as shown in meta-analyses.
Rather than asking whether a monolithic factor like parental control is primarily responsible for non-shared (unique) effects, it might be necessary to consider many seemingly inconsequential experiences that are tipping points in children’s lives. The gloomy prospect is that these could be idiosyncratic stochastic experiences. However, the basic finding that most environmental effects are not shared by children growing up in the same family remains one of the most far-reaching findings from behavioral genetics. It is important to reiterate that the message is not that family experiences are unimportant, but rather that the salient experiences that affect children’s development are specific to each child in the family, not general to all children in the family.
Quantitative genetic methods suggest that common disorders are the extremes of the same genetic factors responsible for heritability throughout the distribution, although the evidence is indirect and the methods are somewhat abstruse.
Research using the DF method has shown consistently that group heritability is substantial for cognitive disability such as language, mathematical, and general learning disability, as well as for reading disability. An interesting exception involves severe intellectual disability (IQ< 70), which DF extremes analysis suggests is etiologically distinct from the normal distribution of intelligence.
On the basis of common SNPs, it seems safe to hypothesize that most common disorders are at the genetic extreme of the spectrum of normal trait variation. This seems a safe hypothesis because heritability of complex traits and common disorders is caused by many genes of small effect (Finding 3), which implies that together these genetic effects will contribute to a quantitative distribution, as Fisher (1918) assumed, even though each gene is inherited in the discrete manner hypothesized by Mendel (1866).
These are not laws, and certainly not commandments, simply findings which are very probably true. Further work will amplify the details. Crucially, we need to understand precisely how many genes of small effect end up having the massive effects we notice in everyday life.