More is known about the influence of genes on social class than upon race and sex; indeed, Murray writes that “the basics have been known for decades.” (209) The technical literature treats socioeconomic status as the sum of heredity (genetic influence) and environmental influence. The latter component can be further divided into shared and nonshared environment. The nonshared environment includes things not shared by people in the same family, like birth order, differential parental treatment, extrafamilial networks, accidents and illnesses. Studying twins, especially monozygotic twins, is a useful technique for reducing the effects of the nonshared environment; usually such twins attend the same schools and have similar social circles. In practice, measurement error must be allowed for as well: e.g., the number of books in a child’s home is one factor sometimes counted as part of its environment, but it is clearly a very imperfect proxy for how much intellectual stimulation the child actually receives in the home.
Hereditability, as it is understood in the technical literature, is “a ratio calculated as the variance attributable to genes divided by the total variance in phenotype.” (210) It is a property of human groups, not individuals:
Suppose that genes explain 70 percent of a population’s variance in height. You can use this information to conclude that “genes probably have a lot to do with how tall Joe is,” but it does not mean that “genes explain 70 percent of how tall Joe is.” (211)
Heredity is not a fixed number for any particular trait. For example, the heritability of IQ rises with age, a result many find counterintuitive. A child’s IQ may be temporarily boosted, e.g., by a preschool educational program, but over time the effect fades and the full effect of genes increasingly apparent.
Heredity varies by population. For example, the heredity of SAT scores at an elite high school will be higher than at an ordinary school where the students were brought up in similarly excellent environments—educated, involved parents and plenty of money. This makes the variation due to the environment less within that sample so that the remaining variation is more likely due to variation in genes.
Heritability is a ratio—the ratio of the variation due to variation in genes to the total variation in the variable being measured. A basic formula is
where H is heritability, Vg is variation due to people having different genes and Ve is variation due to people experiencing different environments (shared and unshared). The narrower range of environments experienced by students at the elite school (i.e., a smaller Ve) means that a smaller denominator is used for calculating that ratio; since Vg remains the same in both the numerator and the denominator, the ratio as a whole is greater; at the extreme, if there is no variation due to the environment (i.e., Ve = 0, as would be the case if all the sample subjects were reared in exactly the same environment), the ratio would equal 1—all of the variation would be due to people having different genes; H=1.
As a result, the more environmental influences are equalized, the higher heredity becomes. As Murray explains:
It is a statistical necessity: the phenotype is the result of genes and environment. In a perfect world where everyone had completely full opportunity to realize their talents, heritability of those talents would converge on 100 percent because the environment relevant to those talents would no longer vary. (212)
As early as 1976, two researchers noted that
a consistent—though perplexing—pattern is emerging from the data. Environment carries substantial weight in determining personality—it appears to account for at least half the variance—but that environment is one for which twin pairs are correlated close to zero. We seem to see environmental effects that operate almost randomly [resulting in nonshared environmental influences]. (219)
As another researcher put it, “theories of socialization had assumed that children’s environments are doled out on a family-by-family basis. In contrast, the point of nonshared environments is that environments are doled out on a child-by-child basis.” (227) For example, within a family, two siblings may have different experiences at school, or they may seek out or elicit different environments (say, music vs. sports) because of genetic differences. Accordingly, Murray’s eighth proposition states: “The shared environment usually plays a minor role in explaining personality, abilities and social behavior.”
In 2015, a group of seven scholars published a meta-analysis of nearly every twin study carried out between 1958 and 2012; it involved 2748 publications, 14,558,903 twin pairs, and explored 17,804 traits. From this enormous assemblage of data, Murray extracted the evidence on thirty traits relevant to personality, abilities and social behavior. Only for two of them was the contribution of the shared environment greater than one third: 36 percent for ‘basic personal interactions’ and 34 percent for ‘problems related to upbringing.’ “Yes, these data seem to say, you can have some effect on your kids’ manners and you can also cause problems.” (223) For all 28 other traits, shared environment accounts for no more than 26 percent of variance. For such important traits as temperament and personality functions, work and employment, intimate relationships, and family relationships, shared environment contributes no more than 6 percent.
One must add the caveat that an extremely bad home environment can make a significant difference: i.e., truly awful parenting which involves severe deprivation and abuse can damage children permanently.
Wealthy parents can give their children a high standard of living. Often, they can get them out of youthful scrapes or into desirable first jobs. But, says Murray, “it’s not so easy for parental influence to get the child promoted. The more competitive the industry and the more cognitively demanding the job, the less influence family wealth has.” (221) And, of course, wealthy and high-status parents pass on to their children the genetic factors which partly explain their own wealth and success. What they cannot do is use their wealth or status to make their “children more than trivially ‘better’ than they would otherwise have been where ‘better’ is defined in terms of personality, abilities, or social behavior.” (221)
Murray’s ninth proposition states that “class structure is importantly based on differences in abilities that have a substantial genetic component.” The basic reasoning behind this was set forth by psychologist Richard Herrnstein in 1973: 1) if differences in mental abilities are inherited, and 2) if success requires those abilities, and 3) if earnings and prestige depend upon success, then 4) social standing (which reflects earnings and prestige) will be based to some extent on inherited differences among people. This argument was set forth with 800 pages of detailed empirical support in The Bell Curve, a book cowritten by Herrnstein and Murray in 1994 and subtitled ‘Intelligence and Class Structure in American Life.’
In the present work, Murray immediately follows up his ninth proposition with the following caveat:
The bulk of the variance in success in life is unexplained by either nature or nurture. Researchers are lucky if they explain half of the variance in educational attainment with measures of abilities and socioeconomic background. They’re lucky if they can explain even a quarter of the variance in earned income with such measures. The takeaway for thinking about our futures as individuals is that we do not live in a deterministic world ruled by either genes or social background, let alone by race or gender. But Proposition #9 is about social classes, not individuals. (228–9)
Time and chance happen to us all, but they do not push us all in the same direction; spread over an entire society, the effect of genes will inevitably tell.
The general factor of intelligence, known as g and measured by IQ tests, is not only the most important heritable trait contributing to success, but far more important than any other individual trait. Recent confirmation comes from a study of 6653 UK twins which correlated scores on the British school-leaving exam known as the GCSE with nine heritable traits. IQ alone statistically explained 34 percent of variation, while the other eight combined explained just 28 percent.
In the US, criticism of testing has focused on the high correlation between socio-economic status (SES) and performance on college admissions tests such as the SAT. The big question concerns the direction of causation: is SES causing high scores, or is inherited intelligence what put these families in the high-SES category?
An exhaustive analysis of this question, along with a comprehensive review of previous studies was published in 2009 by a team of psychologists at the University of Minnesota. They found that controlling for admission test scores reduced the correlation of parental SES and college grades from +.22 to –.01. On the other hand, controlling for measures of parental SES only reduced the correlation between admission test score and grades only from +.53 to +.50. This would seem to leave little room for argument.
Evidence for the influence of IQ and parental SES on success in later life is less clear, but Murray cites thirteen measures based on six databases and in only two cases is the correlation coefficient for SES higher than that for IQ.
The final proposition states that “outside interventions are inherently constrained in the effects they can have on personality, abilities, and social behavior.” In practice, ‘outside interventions’ usually refer to such practices as counseling, tutoring, mentoring, after-school activities and job training. The reasoning behind the proposition is simple: 1) if the shared environment explains little of the variance in cognitive repertoires (as stated in proposition eight); and 2) if the only environmental factors that can be affected by outside interventions are part of the shared environment; then 3) outside interventions are inherently constrained in the effects they can have on cognitive repertoires. In other words, “it is not within our power to do much to change personalities or abilities or social behaviors by design on a large scale.”
The truth of this final proposition mostly follows from what has gone before, so rather than adducing evidence directly in its favor, Murray devotes his discussion to showing why five major objections fail. The first three dispute the first premise above, asserting that it is 1) wrong for some important outcomes, 2) wrong for the early stages of life, or 3) wrong when it comes to changing self-concept.
We saw that shared environment accounted for more than one third of variation for only two out of thirty traits related to personality, abilities and social behavior discussed in a thorough 2015 meta-analysis of twin studies. But there is no a priori cutoff for how much variation a factor must explain to be considered substantial. For six other traits, the shared environment accounted for over 20 percent of variation, including such important items as educational attainment (25 percent) and disorders due to multiple drug use (26 percent). If outside interventions could have an effect on the shared environment factor contributing to these traits, might they not be worth the effort?
The best-case scenario for improving shared environment is adoption:
In effect, adoption at birth to competent parents gives us a glimpse of what would happen if an outside intervention could magically be successful at changing a wide variety of parenting behaviors from bad to good … . But adoption is as good as it gets. (246) If the shared environment explains just 26 percent of the variance, the outside intervention has to be big—boarding school, for example, or moving that family out of the neighborhood, or adoption into a new family. (242)
Short of establishing a police state bent on socializing all children in the same manner (as occurred in ancient Sparta and was aimed at in the USSR), the sorts of outside interventions that can be applied to larger numbers of people later in life generally amount to no more than a few hours a week, and must compete against all sorts of other past and present influences. Social agencies simply do not have the means to apply more radical remedies on a large scale. And of course, all it would do would be to reduce Ve in our ratio, so that a greater percentage of the variation would be due to genetic variation. Getting rid of variation is a difficult task indeed.
It is sometimes suggested that outside interventions can work in the early stages of life before habits have set and the child is more malleable. Murray acknowledges that
if interventions are ever going to work, they’re going to work in infancy and early childhood. But it’s one thing to believe that; it’s another to confront the empirical findings about the difficulties and constraints that have attended a half century of attempts to intervene early in life. (246)
When pre-school programs for disadvantaged youth were instituted in the 1960s, they produced a large effect: 35 percent of a standard deviation, nearly equivalent to half the Black-White kindergarten achievement gap. But subsequent experience showed this effect faded at a rate of 3 percent of a standard deviation per year. After 1980, even the initial effects of such interventions had shrunk to 16 percent, a finding which probably reflected improved conditions for children in the control groups.
In 1998, congress mandated a large and rigorously designed evaluation of Head Start; the report was published in 2010.
After one academic year in the program, effect sizes in six language and literacy areas ranged from .09 to .31 [i.e., 9–31 percent of a standard deviation], but there was negligible impact on math skills or on children’s attention, antisocial [behavior] or mental health problems. The limited effects at exit disappeared within two years. “By the end of the first grade, both achievement and behavioral ratings of treatment group children were essentially similar to control-group children.” (251)
It is sometimes asserted that outside interventions can have a positive effect on a student’s “self-concept.” The original version of this theory led to the self-esteem movement, comprehensively de-bunked in the early 2000s. More recently, a somewhat more plausible variant has been put forward.
Researchers administered a Standard Progressive Matrices test to fifth graders. One group of children was praised for their intelligence, while another was praised for the effort they put into the test.
Children praised for being intelligent subsequently displayed less task persistence and less task enjoyment. They became more concerned about getting a good score than about learning new things. They became protective of their image as “smart” and reluctant to jeopardize it. (256)
As Murray wryly notes, this finding “was especially jarring for a society in which many upper-middle-class parents incessantly tell their children how smart they are.” (256)
These findings have spawned the “growth mindset movement.” Advocates believe an emphasis on intelligence is harmful because it teaches children that their results follow from a fixed trait. They strive to convey to students the efficacy of effort, teaching them to interpret failure as a mere stepping stone to later success.
Common sense suggests such an approach could be beneficial for at least some students, but empirical assessments of growth mindset interventions have yet to reveal large effect sizes. It is also difficult to disentangle the effects of the interventions from pre-existing personality characteristics such as openness and conscientiousness, as well as from cognitive ability.
A fourth objection to the constraints on outside interventions questions whether nonshared environment really cannot be affected by outside interventions.
The best way to study the nonshared environment is by looking at monozygotic twins reared together: they have the same genes and the same home environment, so differences must be due to the nonshared environment.
But it has been found that those differences are not stable over time. Cognitive differences last no more than a few years and personality differences change even more quickly. No identical twin differences are stable over many years. The necessary implication: the nonshared environmental factors are not stable, but more like random noise. (259–260)
Effective interventions, however, could only be based on stable patterns.
The fifth objection is a recently fashionable appeal to “epigenetics.” This is a relatively new field of study dealing with auxiliary mechanisms which document that environmental events can switch off the expression of some genes (by making them less accessible to transcription machinery) or in some cases switch them on or modulate the intensity of their expression. This is something which goes on in the human body all the time, but there does not seem to be any evidence that we will soon be able to control the process for our own ends. Murray warns that
with rare exceptions, the mainstream media’s reporting on the science behind epigenetics bears little resemblance to what’s actually been discovered. (261) … As far as I can tell, no serious epigeneticist is prepared to defend the notion that we are on the verge of learning how to turn genes on and off and thereby alter behavioral traits in disadvantaged children (or anyone else). (268)
Murray points out that his final proposition, viz., that “outside interventions are inherently constrained in the effects they can have on personality, abilities, and social behavior,” may not remain true forever.
Who knows what role future drugs might play in enhancing learning and positively affecting personality traits and social behavior. At some point, the promise of … genetic editing will be realized, and all bets about the ability to change people by design in substantial numbers will be outdated. (269–270)
But we are definitely not there yet.