Controversy #3: How Are Inherited Traits Influenced by the Environment?
By jtm|June 29th, 2009
According to search engine trackers, the phrase “are inherited traits influenced by the environment?” is one of the most asked science questions online, with hundreds of hits on Google per day. On the surface though, it appears a needless question; of course traits that have a genetic component are also influenced by the environment. No behavioral genetic study has ever come up with a perfect heritability of 1.00 for a trait that varies within a population, least of all because measurement error limits reliable interpretation above approximately the 0.90 level (heritability runs on a scale of 0.00-1.00 like the correlation coefficient, where the variance accounted for is the coefficient squared). Most commonly, about half the variation of a given behavioral trait is due to environmental influences and half to genes. What the question really should be asking is notwhetherinherited traits are influenced by the environment, but howthey are, because that’s where the answer becomes interesting, and it depends on who you ask.
Since the 1970’s, the field of behavioral genetics has come up with some surprising answers as to what aspects of the environment do influence our traits, but these have not yet been widely accepted in psychology. Not to make it a habit of quoting Steven Pinker, but his words in The Blank Slate were some of the most cogent of any commentator, when he wrote: “The three laws of behavioral genetics may be the most important discoveries in the history of psychology. Yet most psychologists have not come to grips with them, and most intellectuals do not understand them… It is not because the laws are abstruse: each can be stated in a sentence, without mathematical paraphernalia. Rather, it is because the laws run roughshod over the Blank Slate.”1
What Pinker separates into three laws can be boiled down to one for the purposes of this article. What aspects of the environment matter to traits such as personality and intelligence? Not the things psychologists have been pointing to for decades: parenting, the family, the neighborhood, the household itself. These aspects of the shared environment have a small or negligible affect on the majority of psychological traits, increasingly so towards adulthood. What aspects of the environment do matter? Differential experiences of siblings within the family, such as accidents, illnesses, interactions with peers, or any other events that happen to one sibling and not the other. These findings from twin and adoption studies lead to the coining of the phrase The Nurture Assumption by Judith Rich Harris for the default position in developmental psychology that what matters about the environment is parenting, or how parents raise and interact with their children. In sharp contrast, what behavioral genetic methods find is that within a normal range of household and parenting styles, without extreme deprivation, malnutrition, or abuse, what leads to differences in behavioral traits are genes and unique experiences outside the home.
The tidings of behavioral genetics were controversial when they started emerging in the 1970s, and would still be if they were in the news today. Ironically, the problem may be that there is not enough controversy about behavioral genetics in and around psychology, especially in the subdisciplines whose foundations it challenged the most. Some developmental psychologists have clung to the hope that interactions between genes and the home environment might be the savior of the nurture assumption, but most of them have not even addressed the behavioral genetic evidence at that level, if at all.
On the issue of interactions, it must be said that there are multiple lines of evidence that weigh against their importance. First, for there to be an interaction between the shared environment and genetic components of a trait, it would have to be a perfect crossover type, where the effects of different parenting practices would have to exactly cancel each other out on average, without any main effects. This type of interaction is rare in nature, and thus implausible at face value. If the crossover type of interaction did occur though, it would make universal parenting advice, the kind favored by child and clinical psychologists, misguided. Since there is no prior way for the psychologist two know what type of child you have, the best they could say is that following their advice would make your child either better or worse, but do nothing on average across all children. Parent-child interactions have been tested directly as well, and differences in outcomes due to parenting have been found to be nearly null after the child’s genes are factored out. That is, effective differences in the way parents treat their children, when they exist, are mostly the result of the genetic profiles the children are born with, an indirect effect heredity, and not the elusive parent-child interaction.2 Even more pertinently, as Judith Rich Harris pointed out in her second book No Two Alike, gene-environment interactions cannot explain why identical twins, which have the same genetic predispositions and the same home environment, are as different in personality as normal siblings in the nongenetic component.3 The fact that identical twins personalities’ only correlate at the 0.50 level, not even close to the 1.00 their genes would predict, is a testament to the power of either chance events during development or experiences outside the home (personally, I favor the randomness hypothesis, like Thomas Bouchard Jr., a notable behavioral geneticist).
You might ask whether this means that parents don’t matter, but both Harris and Pinker are quick to dispel that misunderstanding. Children have a baseline of physical and emotional needs that must be met, and if these are neglected then this will obviously have consequences for later development, not to mention for the child’s relationships with his or her parents. This idea must be distinguished from that claim, however, that normally variable, non-neglectful, parenting practices will have permanent effects on a child’s stable traits, for which there is little direct evidence, and a wealth of behavioral genetic evidence against.
Typically, the goal of Psych Controversies will be to explore and perhaps resolve contention in psychology, but this case is different in that it makes a call towards generating more debate. Every month psychology articles continued to be published and publicized based on the nurture assumption. This happens whenever a correlation is discovered between parental behavior X and children’s outcome Y, and is taken as prima facie evidence of a causal relationship, without controlling for genes. As Harris observes: “It’s an odd thing. Though many developmentalists are now willing to admit that babies differ from one another at birth and that these differences are largely genetic, they still haven’t come to terms with the fact that babies get their genes from their parents.”4
A good example of this effect is the famous link between breastfeeding and children’s IQ. Searching Google for these two keywords produces hundreds of thousands of results, the majority of which are news stories speculating on how breast milk causes higher intelligence in children. The typical finding—that has been replicated dozens of times—is that children who are breast-fed average about 5 IQ points higher than children who aren’t breast-fed or fed formula from a bottle. Also typical are nurturance explanations from researchers, such as this one: “our best estimates are that maternal bonding and the decision to breast-feed account for about 40 percent of that increase, but that 60 percent—3.2 points—are related to the actual nutritional value of the breast milk.”5
The most obvious oversight of these studies can be spotted a mile away by those with even a cursory knowledge of behavior genetics: the moderate to high heritability of intelligence. What happens when maternal IQ is controlled for? The association between breastfeeding and the child’s IQ all but disappears. The largest study yet performed found that controlling for maternal IQ (but not the IQ of both parents, which would have been an even better control) accounted for 71 to 75% of the apparent IQ advantage of children who were breast-fed.6 When further controls such as maternal education, age, family poverty and home condition were added, the effect of breast-feeding on intelligence lost statistical significance. Those last two factors may seem to be in contradiction to the statements above as to the lack of effect of the shared environment on intelligence, but the shared environment component has been found to have a significant effect on cognitive ability up until adolescence, after which it wanes. This is party because the heritability of intelligence increases with age, with the shared home environment playing more of a role early on. It is also possible to control for paternal IQ and any other aspects of the home environment that might confound the link between breastfeeding and children’s cognitive ability; this is done by comparing the scores of siblings who are discordant for breastfeeding, called a sibling pairs analysis. The two studies that have performed this analysis to date have also found no significant effect of breasfeeding7; in fact the effect was even less close to being significant than studies that controlled for mother’s IQ alone, since the pair analysis controls for all factors shared by siblings: genetic and environmental.
To sum up, yes “inherited traits” are affected by the environment, but not the environment as is usually conceived by the layperson or psychologists who rely on the nurture assumption. It is a largely unknown influence, called the unique or non-shared environment, somewhat akin to dark matter in physics, that makes up the vast majority of the environmental aspect of our traits. Psychologists should now be re-orienting their efforts to understand how genes and the unique environment work to make people the way they are, instead of continuing to hold onto outdated pop psychology assumptions about the importance of parenting.
References
1: p. 372. Pinker, S. (2002). The blank slate: The modern denial of human nature. New York: Penguin Books.
2: ibid.
3: Harris, J. R. (2006) No two alike: Human nature and human individuality. New York: W.W. Norton & Company.
4: p. 71, ibid.
5: Anderson J. W., Johnstone B. M., &Remley D. T. (1999). Breast-feeding and cognitive development: a meta-analysis. American Journal of Clinical Nutrition, 70, 525-535.
6: Der, G., Batty, D., & Deary, I. (2006). Effect of breast feeding on intelligence in children: prospective study, sibling pairs analysis, and meta-analysis. British Medical Journal, Vol 333(7575) Nov 2006, 945.
Controversy #3: How Are Inherited Traits Influenced by the Environment?
According to search engine trackers, the phrase “are inherited traits influenced by the environment?” is one of the most asked science questions online, with hundreds of hits on Google per day. On the surface though, it appears a needless question; of course traits that have a genetic component are also influenced by the environment. No behavioral genetic study has ever come up with a perfect heritability of 1.00 for a trait that varies within a population, least of all because measurement error limits reliable interpretation above approximately the 0.90 level (heritability runs on a scale of 0.00-1.00 like the correlation coefficient, where the variance accounted for is the coefficient squared). Most commonly, about half the variation of a given behavioral trait is due to environmental influences and half to genes. What the question really should be asking is not whether inherited traits are influenced by the environment, but how they are, because that’s where the answer becomes interesting, and it depends on who you ask.
Since the 1970’s, the field of behavioral genetics has come up with some surprising answers as to what aspects of the environment do influence our traits, but these have not yet been widely accepted in psychology. Not to make it a habit of quoting Steven Pinker, but his words in The Blank Slate were some of the most cogent of any commentator, when he wrote: “The three laws of behavioral genetics may be the most important discoveries in the history of psychology. Yet most psychologists have not come to grips with them, and most intellectuals do not understand them… It is not because the laws are abstruse: each can be stated in a sentence, without mathematical paraphernalia. Rather, it is because the laws run roughshod over the Blank Slate.”1
What Pinker separates into three laws can be boiled down to one for the purposes of this article. What aspects of the environment matter to traits such as personality and intelligence? Not the things psychologists have been pointing to for decades: parenting, the family, the neighborhood, the household itself. These aspects of the shared environment have a small or negligible affect on the majority of psychological traits, increasingly so towards adulthood. What aspects of the environment do matter? Differential experiences of siblings within the family, such as accidents, illnesses, interactions with peers, or any other events that happen to one sibling and not the other. These findings from twin and adoption studies lead to the coining of the phrase The Nurture Assumption by Judith Rich Harris for the default position in developmental psychology that what matters about the environment is parenting, or how parents raise and interact with their children. In sharp contrast, what behavioral genetic methods find is that within a normal range of household and parenting styles, without extreme deprivation, malnutrition, or abuse, what leads to differences in behavioral traits are genes and unique experiences outside the home.
The tidings of behavioral genetics were controversial when they started emerging in the 1970s, and would still be if they were in the news today. Ironically, the problem may be that there is not enough controversy about behavioral genetics in and around psychology, especially in the subdisciplines whose foundations it challenged the most. Some developmental psychologists have clung to the hope that interactions between genes and the home environment might be the savior of the nurture assumption, but most of them have not even addressed the behavioral genetic evidence at that level, if at all.
On the issue of interactions, it must be said that there are multiple lines of evidence that weigh against their importance. First, for there to be an interaction between the shared environment and genetic components of a trait, it would have to be a perfect crossover type, where the effects of different parenting practices would have to exactly cancel each other out on average, without any main effects. This type of interaction is rare in nature, and thus implausible at face value. If the crossover type of interaction did occur though, it would make universal parenting advice, the kind favored by child and clinical psychologists, misguided. Since there is no prior way for the psychologist two know what type of child you have, the best they could say is that following their advice would make your child either better or worse, but do nothing on average across all children. Parent-child interactions have been tested directly as well, and differences in outcomes due to parenting have been found to be nearly null after the child’s genes are factored out. That is, effective differences in the way parents treat their children, when they exist, are mostly the result of the genetic profiles the children are born with, an indirect effect heredity, and not the elusive parent-child interaction.2 Even more pertinently, as Judith Rich Harris pointed out in her second book No Two Alike, gene-environment interactions cannot explain why identical twins, which have the same genetic predispositions and the same home environment, are as different in personality as normal siblings in the nongenetic component.3 The fact that identical twins personalities’ only correlate at the 0.50 level, not even close to the 1.00 their genes would predict, is a testament to the power of either chance events during development or experiences outside the home (personally, I favor the randomness hypothesis, like Thomas Bouchard Jr., a notable behavioral geneticist).
You might ask whether this means that parents don’t matter, but both Harris and Pinker are quick to dispel that misunderstanding. Children have a baseline of physical and emotional needs that must be met, and if these are neglected then this will obviously have consequences for later development, not to mention for the child’s relationships with his or her parents. This idea must be distinguished from that claim, however, that normally variable, non-neglectful, parenting practices will have permanent effects on a child’s stable traits, for which there is little direct evidence, and a wealth of behavioral genetic evidence against.
Typically, the goal of Psych Controversies will be to explore and perhaps resolve contention in psychology, but this case is different in that it makes a call towards generating more debate. Every month psychology articles continued to be published and publicized based on the nurture assumption. This happens whenever a correlation is discovered between parental behavior X and children’s outcome Y, and is taken as prima facie evidence of a causal relationship, without controlling for genes. As Harris observes: “It’s an odd thing. Though many developmentalists are now willing to admit that babies differ from one another at birth and that these differences are largely genetic, they still haven’t come to terms with the fact that babies get their genes from their parents.”4
A good example of this effect is the famous link between breastfeeding and children’s IQ. Searching Google for these two keywords produces hundreds of thousands of results, the majority of which are news stories speculating on how breast milk causes higher intelligence in children. The typical finding—that has been replicated dozens of times—is that children who are breast-fed average about 5 IQ points higher than children who aren’t breast-fed or fed formula from a bottle. Also typical are nurturance explanations from researchers, such as this one: “our best estimates are that maternal bonding and the decision to breast-feed account for about 40 percent of that increase, but that 60 percent—3.2 points—are related to the actual nutritional value of the breast milk.”5
The most obvious oversight of these studies can be spotted a mile away by those with even a cursory knowledge of behavior genetics: the moderate to high heritability of intelligence. What happens when maternal IQ is controlled for? The association between breastfeeding and the child’s IQ all but disappears. The largest study yet performed found that controlling for maternal IQ (but not the IQ of both parents, which would have been an even better control) accounted for 71 to 75% of the apparent IQ advantage of children who were breast-fed.6 When further controls such as maternal education, age, family poverty and home condition were added, the effect of breast-feeding on intelligence lost statistical significance. Those last two factors may seem to be in contradiction to the statements above as to the lack of effect of the shared environment on intelligence, but the shared environment component has been found to have a significant effect on cognitive ability up until adolescence, after which it wanes. This is party because the heritability of intelligence increases with age, with the shared home environment playing more of a role early on. It is also possible to control for paternal IQ and any other aspects of the home environment that might confound the link between breastfeeding and children’s cognitive ability; this is done by comparing the scores of siblings who are discordant for breastfeeding, called a sibling pairs analysis. The two studies that have performed this analysis to date have also found no significant effect of breasfeeding7; in fact the effect was even less close to being significant than studies that controlled for mother’s IQ alone, since the pair analysis controls for all factors shared by siblings: genetic and environmental.
To sum up, yes “inherited traits” are affected by the environment, but not the environment as is usually conceived by the layperson or psychologists who rely on the nurture assumption. It is a largely unknown influence, called the unique or non-shared environment, somewhat akin to dark matter in physics, that makes up the vast majority of the environmental aspect of our traits. Psychologists should now be re-orienting their efforts to understand how genes and the unique environment work to make people the way they are, instead of continuing to hold onto outdated pop psychology assumptions about the importance of parenting.
References
1: p. 372. Pinker, S. (2002). The blank slate: The modern denial of human nature. New York: Penguin Books.
2: ibid.
3: Harris, J. R. (2006) No two alike: Human nature and human individuality. New York: W.W. Norton & Company.
4: p. 71, ibid.
5: Anderson J. W., Johnstone B. M., &Remley D. T. (1999). Breast-feeding and cognitive development: a meta-analysis. American Journal of Clinical Nutrition, 70, 525-535.
6: Der, G., Batty, D., & Deary, I. (2006). Effect of breast feeding on intelligence in children: prospective study, sibling pairs analysis, and meta-analysis. British Medical Journal, Vol 333(7575) Nov 2006, 945.