Tip:
Highlight text to annotate it
X
Lecture 33: Nature, and nurture. So far in this course, we've focused our attention on
the mental processes that are characteristic of mature adult human beings. We've taken
mind and personality as givens, and asked how they work. How does a mind work? How does
the mind mediate the individual's interactions with his or her social environment? Now we
want to take up a new question, which is, where do mind and personality come from?
In addressing questions of mental development psychologists tend to take one of three general
views, the ontogenetic, the phylogenetic, and the cultural.
The ontogenetic view has to do with the development of the individual human being. From the ontogenetic
point of view, we want to trace the development of mind and personality across the lifespan
of the individual, from birth, through infancy, childhood, adolescence, adulthood, right through
to old age and death. Most developmental psychologists focus on infancy and childhood, though with
our aging population, there's an increasing interest in life-span development, extending
into old age. This is a domain of developmental psychology, a major sub-field in psychology,
which often divides itself up into psychologists who are interested in cognitive development,
and those who are interested in social and personality development. These lectures are
going to focus on the ontological view of development, and especially on personality
development. But also I'll have some things to say about trends in the study of cognitive
development.
While the ontogenetic view is concerned with development of mind across the lifespan of
the individual, the phylogenetic point of view is concerned with development across
evolutionary time. To some extent, the phylogenetic point of view was expressed by comparative
psychology -- which, as its name implies, is concerned with studying learning and other
cognitive, emotional, and motivational functions in different species and, more recently, in evolutionary
psychology, an offshoot of sociobiology, which traces how our current human and mental behavioral functions
evolved through natural selection and similar processes. From the point of view of modern
evolutionary psychology, much of the way we currently think and behave was shaped by our
adaptation to what they call the environment of early adaptation -- basically the
African savanna in the Pleistocene Era, where humans first emerged.
And finally, there's a cultural view of development, which is concerned with the effects of various
forms of social development on the development of the individual's mind and behavior. This
is represented, in its early form, in what's known as anthropological psychology, which
tries to describe what the so-called "primitive mind" is like. This actually an offshoot of
19th-century European colonialism, which often sought to contrast between the so-called "advanced"
minds of the European colonists and the so-called "primitive" minds of those they colonized.
But more broadly, it's based on the idea that, that there was a development in modes of thought
as our human ancestors moved from the Stone Age to the Bronze Age to the Iron Age, and
even in historical time as human beings moved from the ancient period through the medieval
period, the early modern period of the 15th through the 18th centuries, the modern period,
and now post-modern thought. There's also literature on the role that literacy
plays in shaping how individuals think and how cognitive development is correlated with
economic and political development. These days, there's quite a bit of interest in what's
known as cultural psychology which recognizes a great diversity of mental life across cultures,
without the implications that one culture is necessarily more developed than another
one.
As I say, these lectures are going to focus on the ontogenetic view of development, which
is the way developmental psychology is usually construed. Throughout its history, developmental
psychology, from the ontogenetic point of view, has been focused on the question of the role of nature and nurture in mental
development. This dichotomy between nature and nurture was originally proposed by Sir
Francis Galton, a cousin of Charles Darwin's, was inspired by a line from Shakespeare's
Tempest. Is the newborn child a kind of blank slate, which is written on by experience -- an
organism that acquires its knowledge and skills through direct experience and vicarious social
learning? Or are some aspects of mental life innate, part of the child's genetic endowment,
acquired through the course of human evolution?
As Galton defined the terms, nature refers to all that a man brings with himself into
the world -- the individual's genetic and hormonal endowment, his or her physiological
constitution and temperament, whereas nurture refers to every influence on the individual
that occurs after his or her birth -- the influence of the physical and social environment,
what the individual gets from social learning, socialization, and all aspects of experiential
history. You can see the relationship, I think, between Galton's dichotomy of nature and nurture,
and that dichotomy between nativism and empiricism that we discussed at the beginning of our
lectures on sensation and perception. The empiricists thought that all knowledge was
acquired through experience, and the nativists thought that some knowledge was innate. Galton
is making very much the same kind of distinction.
Of course, there are lots of different ways in which we can think about the relationship
between nature and nurture. Galton seemed to frame the problem as one of opposition,
in which some characteristic of mind and behavior was due to nature or nurture. But of course
we can also think of nature and nurture working together, either exerting independent influences
on some aspect of mind or behavior, or somehow interacting with each other to generate mind
and behavior.
I bet you can guess where this lecture's going to end up. Well, why keep it a secret? A major
purpose of these lectures on psychological development is to illustrate what I think
of as a developmental corollary to the doctrine of interactionism. In interactionism we saw
that the person is a part of his or her own environment, and shapes that environment at
least as much as the environment shapes him or her. Similarly, I want to stress that development
isn't just something that happens to the individual, with the individual passively on the receiving
end. Rather, the developing child, even the infant fresh out of the womb,
is actively engaged in the developmental process and actively working as an agent of his or
her own development. But we're not there yet.
Let's begin our discussion of the ontogenesis of mind and behavior at the beginning, with
the human genome. The human genetic endowment consists of 23 pairs of chromosomes, each
chromosome containing a large number of genes. Every gene is located at a particular place
on a specific chromosome. And since chromosomes come in pairs, so do genes. Corresponding
pairs of genes contain information about some characteristic, such as eye color, skin pigmentation,
and the like. Some traits are determined by single pairs of genes. Others appear to be
determined by several genes acting together. Two of these chromosomes, the X and the Y
chromosome, determine the individual's sex. If you're XX, you're genetically female. If
you're XY, you're genetically male -- although it will be clear from Lecture 35
that things are actually a little bit more complicated than that.
Usually, we think about genes as having to do with an organism's physical characteristics.
But it's entirely possible that there is some genetic influence on the organism's mental
and behavioral characteristics as well. One of the most important technical advances in
recent biological research has been the decoding of the human genome, determining the precise
location of the 20,000 or so genes that make us as a species different from all other species.
And, along with this advance came gene mapping, which makes it possible to determine specific
genes that put us at risk for various diseases, or disposes to various personality traits.
Here you see Craig Venter, whose private research group contributed greatly to the deciphering
of the human genome. And you also see Venter's own genome. These are his chromosomes along
with the identification of specific genes that, as far as we can tell, dispose him to
heart attacks, give him a preference for evening over morning, genes linked to tobacco addiction,
novelty seeking, Alzheimer's disease, and the like. Now the presence of these genes
doesn't mean that Venter is predestined for a heart attack or Alzheimer's disease any
more than he was genetically pre-destined to be a surfer, which in fact he is. It's just
these genes appear to be more common in people who have these problems than in those who
do not. They're risk factors, but not an irrevocable sentence to heart disease or dementia.
The entire set of genes comprises the organism's genotype, or genetic blueprint. An organism's
genotype represents its biological potential. This potential is actualized within a particular
environmental context to produce the organism's phenotype. So one's genetic endowment interacts
with various environmental factors, to produce the phenotype, or what the organism actually
looks like. Because of the role played by the environment, phenotypes are not necessarily
equivalent to genotypes. For example, two individuals may have the same phenotype, but
quite different genotypes. Similarly, it can happen that two individuals will have the
same genotype, but different phenotypes. Of two brown-eyed individuals, that's the phenotype, one might have
two dominant genes for brown eyes while the other might have one dominant gene for brown
eyes and one recessive gene for blue eyes. Different genotypes, same phenotype. Or two
individuals may have the same dominant genes for dimples, but one might
have his or her dimples removed by plastic surgery. Same genotype, different phenotype.
So the environment matters. And we need to distinguish among three different
kinds of environmental influence. The prenatal environment, the environment of the developing
fetus in the womb during gestation; the perinatal environment, the environment present around
the time of birth; and, the postnatal environment, the environment present after birth and
continually changing throughout the course of the individual's life cycle from birth
to death.
Many basic questions of nature and nurture, particularly concerning personality development,
have been addressed by using the technique of behavior genetics. And perhaps the most
popular method in behavior genetics -- really, the gold standard for behavior genetic research
-- is the twin study -- which, as its name implies, compares two kinds of twins in terms of their
similarity in personality. Monozygotic twins, or identical twins, are the product of an
egg that has been fertilized by a single ***, but which subsequently split into two embryos,
thus yielding two individuals who are genetically identical. Dizygotic or fraternal twins occur
when two different eggs are fertilized by two different ***, thus yielding individuals
who have only about 50 percent of their genes in common. The basic idea behind the twin
study method is that if a trait is inherited, even in part, monozygotic (or identical) twins
should be more alike on that trait than dizygotic (or fraternal) twins, because monozygotic
twins are identical genetically, and dizygotic twins only have about 50% of their genes in
common.
The usual technique in twin studies of personality is to administer some personality inventory,
like the NEO Personality Inventory, to a large sample of monozygotic and dizygotic twins,
thus obtaining scores representing each individual's standing on each personality trait measured
by the inventory. Then we measure the similarity of the twins on each trait by means
of the correlation coefficient. You'll remember that this is a statistic that summarizes the
direction and strength of the relationship between two variables -- for example, between
extraversion in one twin and extraversion in the other. If a trait were wholly inherited,
we'd expect to see a perfect correlation between identical twins, a correlation coefficient
of 1.0, and we'd also expect to see a substantial but not perfect correlation of.50 for dizygotic
twins. If a trait is wholly inherited, then genetically identical twins should be identical
in personality. And again if the trait is wholly inherited you would expect some degree
of similarity and personality between dizygotic twins as well -- because they are genetically
similar to at least some extent. For genetically unrelated individuals we'd expect to see no
correlation. If a gene is wholly inherited, and if there is no genetic overlap, there
should be no similarity in personality. An alternative measure of similarity is the
concordance rate. That is simply the percentage of subjects that have some characteristic
in common. We'll use the concordance rates when we look at behavior genetic studies
of mental illness, but the logic is the same. If some characteristic is wholly inherited,
we'd expect to see a concordance rate of 100% for monozygotic twins because they're genetically
identical: if one has the trait the other must have it too. And we'd expect to see a
concordance rate of about 50% for dizygotic twins, and a concordance rate of 0 for unrelated
individuals. Again, to the extent that a trait is inherited, we expect that monozygotic, identical
twins will be more similar to each other than are dizygotic, fraternal twins, regardless
of whether similarity is measured by the correlation coefficient or the concordance rate.
Here's a quick fictional example: this data is made up. Imagine that we had a set of 10
twins, 20 individuals, and we administered to them a scale of extraversion, a 60-point
scale of extraversion, and simply calculated their scores. You can see just by inspection
that there's a high degree of correspondence between the scores of one individual and the
score of his or her twin. It's not perfect, but in general, individuals with low scores
have twins with low scores. Individuals with high scores have twins with high scores. In
fact, if you calculate the correlation, the correlation is almost, but not quite, perfect,
a correlation of.99.
Now here's some real twin study data taken from a study by Loehlin and his colleagues,
in which monozygotic and dizygotic twins were administered a personality inventory that
contain scales measuring the Big Five personality traits. Studies using other personality inventories
have obtained similar kinds of findings. You can see that on each of the Big Five personality
traits, monozygotic (identical) twins are more alike than dizygotic (fraternal) twins.
Results like this provide prima facie evidence for genetic contribution to individual differences
and personality. The simple fact that monozygotic twins are more alike than dizygotic twins
strongly suggests that there's a genetic component to these individual differences, but the correlations
for monozygotic twins are far from 1.0 that we would expect if a trait were wholly inherited.
So it's very clear that genes aren't the only forces that are determining individual differences
in personality. But if you think about it, data of the sort that I've just shown you
doesn't provide wholly convincing evidence that there is any genetic contribution to
personality. Sure, monozygotic twins are more alike than dizygotic twins, and they are genetically
identical, but monozygotic twins also share an environment that is arguably, anyway, more
alike than it is for dizygotic twins. After all, monozygotic twins are of the same sex
and they also resemble each other physically. So, it wouldn't be surprising to discover
that monozygotic twins are treated more similarly than dizygotic twins are. For this reason,
the very best twin studies, like Loehlin's, employ samples of dizygotic twins that are
of the same sex. This reduces, at least to some degree, the discrepancy in environmental
experience that monozygotic and dizygotic twins have. But still, the problem remains. How
exactly can we tease apart the genetic and environmental contributions to personality?
The question is complicated still further by understanding that, in fact, environmental
influences are very complicated. There are two different kinds of environmental influences
on personality: the shared environment and the non-shared environment. The shared environment,
also known as between-family variance, includes all the factors that are shared by children
raised in the same family -- factors that differentiate them from children raised in
other families. As a rule, children in the same family are raised by the same parents,
share a single racial, ethnic, and cultural heritage, live in the same neighborhood, go
to the same schools, attend the same church, synagogue, or mosque. The shared environment
includes all these things, and more, that siblings within a family have in common.
The non-shared environment, also known as within-family variance, includes all the factors
that differentiate among children raised in the same family. Even within a family, children
can differ in terms of such factors as gender -- whether they're boys or girls; or birth
order -- whether they're the first-born or latter-born. Children may have different interactions
with their parents and develop different networks of friends and acquaintances beyond the family.
Different children within a family are also distinguished by non-systematic factors which
include all the things that happen randomly to one child, but not to his or her brothers
and sisters, chance encounters that can really make a difference in the individual's life.
The non-shared environment is an umbrella term that refers to all the unique experiences
that siblings can have.
As it happens, the relative strength of both environmental components of personality, as
well as the genetic component, can be estimated from the observed pattern of monozygotic and
dizygotic correlations. Consider, first, the entire distribution of a trait within a population
-- from those individuals with the lowest scores on extraversion or neuroticism to those
with the highest scores on these traits. This distribution is typically represented by a
more-or-less normal distribution, the famous bell curve. The entire distribution of individual
scores within a population is called the total variance on the trait in question. This total
variance, on some trait like extraversion, can be further decomposed into genetic variance
-- that is, the amount of total variance that's attributable to individual differences in
genotypes, and environmental variance, that is the amount of the total variance that's
attributable to individual differences and environments. Making the distinction between
shared and non-shared environments, in principle, then, the total variance on a trait is equal
to the sum of the genetic variance, and the environmental variance, with the understanding
that environmental variance includes both variance attributable to the shared environment,
and variance attributable to the nonshared environment.
And, in fact, we can calculate the components of population variance -- the genetic variance
and the environmental variance -- from the data given to us by twin studies. And here's
how we do it. The genetic variance can be estimated by the difference between the monozygotic
and dizygotic correlations, multiplied by two. To estimate genetic variance, you double
the difference between monozygotic and dizygotic twins. Don't worry about where the two comes
from, that's a technical detail. But you can get a sense of it if you just consider that,
if a trait were wholly inherited, the monozygotic correlation would be 1.0; the dizygotic correlation
would be 0.50; so the difference between them is 0.50; double that difference, you get 1.0.
All of the population variance would be attributable to genetic variance. That's the idea.
So what's left? Well, what's left is the environmental variance. Variance due to the non-shared environment
is a function of the MZ correlation. Monozygotic twins share both genes and the shared environment,
so any monozygotic correlation less than a perfect 1.0 must reflect the contribution
of the non-shared environment. It's just logic. If monozygotic twins share both genes and
the shared environment, which they do by definition, then to the extent that they do not perfectly
resemble each other, that difference must be due to what they don't have in common -- and
that's the non-shared environment. So we estimate the non-shared environment by subtracting
the monozygotic correlation from perfect 1.0. And once we've estimated the contributions
of the genes and the non-shared environment, variance due to the shared environment is
all that's left. So, that can be estimated simply by subtracting the genetic variance
and the non-shared environmental variance from 1.0.
Now just to get the idea down firmly, let's work through a couple of numerical examples.
Here's the ideal case -- at least from a genetic point of view. The MZ correlation is a perfect
1.00. The DZ correlation is 0.50. Double the difference between MZ and DZ, you get an estimate
of genetic variance of 1.00. All the population variance is attributable to genetic variance.
Of course there's nothing left over, but let's do it anyway. The non-shared environment is
estimated by 1 minus the MZ correlation, that's 1 - 1 or 0, so there is no contribution from
the non-shared environment. And we can estimate the contribution of the shared environment
by subtracting from 1 the genetic variance and the nonshared environmental variance,
1-1-0 leaves 0, no contribution from the shared environment either. This is exactly what we'd
expect if a trait were wholly inherited.
Now let's cut the MZ correlation in half. Monozygotic twins correlate 0.50. Dizygotic
twins correlate 0.40. Double the difference, you get an estimate of genetic variance of
0.20: 20% of the population variance is attributable to genetic variance. To estimate the contribution
of the nonshared environment, 1 minus the MZ correlation of 0.50: 50% of the population
variance is attributable to the non shared environment. What's left? 1 minus .20 minus.50
equals .30. Thirty percent of the population variance is attributable to the shared environment.
Here's another one. Monozygotic correlation of .80, dizygotic correlation of .70. Double
the difference, so 20% of population variance is attributable to genetic variance. 1 minus
the MZ correlation, 20% of the population variance is attributable to the nonshared
environment. And what's left is .60, 60% of the population variance is attributable to
the shared environment.
And one more. Monozygotic twins, correlation of 0.5, dizygotic twins, correlation of 0.25.
Double the difference, 50% of the population variance is attributable to genes. 1 minus
the MZ correlation, 50% of the population variance is attributable to the nonshared
environment. What's left is nothing. None of the population variance is attributable
to the shared environment.
Now let's look at some real data. This is data derived from family studies of IQ, measured
intelligence, including twins. The general pattern here is that as individuals increase
in genetic relatedness, they also increase in their similarity for IQ. As you can see, monozygotic twins
are more alike in IQ than dizygotic twins, even when those monozygotic twins have been
reared apart. Again, this is a very common pattern in family studies of IQ.
Aggregating results from a large number of studies of this sort we can estimate that
the correlation between monozygotic twins for IQ is about .86. And the correlation for
dizygotic twins is about .60. So now let's plug in those values into our formulas. Double
the difference between monozygotic and dizygotic twins, we find that about 52% of the population
variance in IQ appears to be attributable to genetic variance. Subtract the monozygotic
correlation from 1; about 14% of the variance in IQ is attributable to the non-shared environment.
And what's left? About 34% of the population variance in IQ appears to be attributable
to the shared environment. So genes play a role in determining IQ, but genes are not
the sole determinants of IQ: the environment is also important; and of our two kinds of
environment, the shared environment appears to be more important than the nonshared environment
in determining individual differences in IQ.
Here's similar data from a twin study of educational achievement; how much schooling people get.
Again, monozygotic twins are very alike on this, show a correlation of about .86. Dizygotic
twins show a correlation of .66. Double the difference, we see that about 40% of population
variance in educational achievement is attributable to genetic variance. This makes sense: if
education has something to do with intelligence, and intelligence has something to do with
the genes, then genes ought to play a role in educational achievement. 1 minus the MZ
correlation indicates that the non-shared environment contributes only about 14% of
population variance. But now what's left over, .46, the shared environment now turns out to be
not just more important than the non-shared environment -- but also at least marginally
more important than genes. Some families send their kids to school, make their kids go to
school, make their kids go to college, even if the kids aren't all that smart. So, the
shared environment is much more important as a determinant of educational outcome than
it is as a determinant of IQ.
Here are some findings from a behavior-genetic study of adolescent and young adult behavior,
particularly *** behavior. The investigators measured what they called age of *** debut
-- that is, age of first intercourse. These investigators estimated that about 31% of the
variance in age of *** debut was attributable to the genes. Only about 10% of the variance
is attributable to the shared environment. The overwhelming proportion of variance, almost
60%, was attributable to the non-shared environment. I'll leave it to you to speculate about precisely
what those non-shared environmental factors might be.
Here are the results from another study, this time of suicidal behavior. The investigators
measured both suicidal ideation, the extent to which people thought about committing suicide,
and also suicide attempts, whether they actually tried to do so, including completed and uncompleted
attempts. It turns out that there's a genetic influence on both suicidal ideation and suicidal
attempt, though the genetic influence is stronger for ideation than it is for attempt. The shared
environment played apparently no role in suicidal ideation, and just a little role in suicidal
attempt. But the nonshared environment played an overwhelming role in both determining suicidal
ideation and suicide attempt. Population variance in suicidal behavior, both ideation and attempt,
appears to be overwhelming determined by variance in the nonshared environment. And again, I'd
leave it to you to speculate on exactly what those non-shared environmental factors might
be.
Returning to Loehlin's twin study of the Big Five personality traits, remember that the
basic finding was that for each of the Big Five traits, the MZ correlations were about
twice the size of the DZ correlations. The typical MZ correlation was in the high 40s,
the typical DZ correlation was in the mid 20s. So now let's apply our formulas and try
and disentangle the contributions of genetics, the shared environment, and the non-shared
environment to individual differences in personality.
Let's first look closely at just one big five trait, extraversion. We double the difference
between MZ and DZ twins, and get an estimate of 48% of the variance is attributable to
genetic factors. 1 minus the MZ correlation gives an estimate of 52% of population variance
is attributable to the nonshared environment. And the remainder -- well, there isn't any
remainder. Apparently none of the variance in extraversion, for this group, was attributable
to the shared environment.
And that turns out to be pretty much the case for each of the Big Five personality traits.
In each case we see a substantial genetic component to population variants, about 40% or so; a
substantial nonshared environmental contribution to variance, about 50% or so; but a very small
contribution of the shared environment, typically less than 10%. Genes make a contribution to
individual differences in personality, but genes are not decisive. Nor is the family
environment decisive for adult personality. The non-shared environment is far more important
than the shared environment and is at least as important, probably more important, than
the contribution of the genes. This is the typical finding obtained whenever somebody
does a twin study of personality.
Behavior genetics studies don't have to be confined to things like intelligence and personality
traits. We can also use them to study the origins of attitudes. Here's an example from
the Virginia 30K Twin Study, a study of almost 30,000 twins who lived in the state of Virginia.
As part of this study, these individuals were administered an attitude inventory, which
asked them toy endorse various positions on various socio-political issues: school
prayer, property taxes, busing to relieve racial segregation, abortion rights, and so
on. There were 25 of these stated in a more-or-less liberal way, 25 stated in a more-or-less
conservative way; 28 of the items were expressly political in nature -- that is, they had to
do with proposals for new laws. And the subjects were simply asked to indicate whether they
agreed or disagreed or were uncertain about each of these statements.
Here are the correlations obtained in that study. The survey items were collected into
two somewhat different scales measuring liberal versus conservative attitudes. Remember, that's
the basic dimension of attitudes, liberalism versus conservatism. Then there was the scale
of opinionation -- basically, the number of items in which the subjects actually had an
opinion as opposed to being uncertain. And finally, the subjects were asked about their
political party affiliation -- Democrat, Republican, Independent, whatever. You can see that in
each case the correlations for monozygotic twins are higher than the correlations for
dizygotic twins.
And here is how the components of variance in political attitudes turned
out. For both measures of liberalism-conservatism, there was a significant genetic component,
amounting to 30% or 40% of the variance. And that was also true for the
measure of opinionation or strength of attitude. But once again, the contribution of the nonshared
environment was, in each case, much greater than the contribution of the shared environment.
Things were somewhat different for party affiliation. Here there was no genetic contribution to
speak of: almost all of the action is in the environment and the contributions are about
evenly divided between the shared environment and the non-shared environment. It's very
interesting that there is any genetic contribution at all to liberal vs. conservative attitudes,
but it's not surprising that there's no genetic contribution to party affiliation. After all,
party affiliation is something that can change very radically with changes in historical
circumstances. Between the time of the Civil War and the 1950s, the dominant political
party in Virginia, part of the Old South, was Democratic. Beginning in the 1950s, with
the rise of the civil rights movement, and especially with President Nixon's "southern
strategy", the party affiliation of southerners quickly moved from predominantly Democratic
to predominantly Republican. That's probably why the shared environment is as important
as it is, in the case of party affiliation, compared to liberalism, conservatism, or opinionation.
But even with party affiliation, the role of the non-shared environment is substantial.
There may be a strong tendency for children in the same family to all have the same political
party affiliation, perhaps adopting the political affiliation of their parents. But something
can happen to lead one sibling, or a couple, to switch. And that's the impact of the non-shared
environment. But here again, the outline of the results is generally the same as what
we saw for personality traits. Aside from party affiliation, there is a substantial
genetic contribution to political attitudes, but it's not decisive; the environment is
important. And in the final analysis, as important as the shared environment is to party affiliation,
the nonshared environment is a more important contributor to variance in overall liberalism
and conservatism.
The point is really important, so I want to repeat it. Twin studies reveal genetic influences
on personality, including attitudes, and these are interesting. But by far the most surprising
finding of behavior genetics research is the evident power of the environment, and especially
the non-shared environment. Genes are important, but the non-shared environment is just as
important. This typical finding of behavior genetic studies, that the contribution of
the non-shared environment is more important than that of shared environment, is sometimes
misinterpreted as meaning that parents have no influence on their children, but it doesn't
really mean that at all. What it means, first, is that parents don't
have the same effects on each of their children. Presumably, there are some ways in which parents
treat all their children alike. And whatever that is -- in terms of such things as age
of weaning, or age and severity of toilet training, whether the children have play groups
that they participate in, or go to nursery school before they enroll in kindergarten
-- these things just aren't strongly determinative of personality. What's more important, as
we'll see in the next lecture, is that parents just don't have the same effects on each of
their children. Parents can have tremendous influence on their kids, but they don't have
the same influence on each and every one of them.
In addition, it means that there are other forces at work besides the parents. These become
increasingly important as the child begins to move beyond the family -- for example,
by going to school or joining a soccer league. In fact, Judith Harris has proposed a theory
of group socialization which argues that peer groups and peer cultures, not parents, are
the most powerful socialization forces impinging on a child. In fact, Harris argues that socialization
is context-specific, and that children may behave quite differently, follow quite different
sets of rules, depending on whether they're at home or away and depending on which particular
extra-familial context they're in. As an example she points to what's known as
code switching among bilingual and bicultural children. Children born to Spanish-speaking
parents, for example, may continue to speak Spanish inside the home, when they're with
their parents, even though they'll speak English with their peers, outside the home. Harris's
point is simply that something like code-switching occurs beyond the domain of language, and
may affect all aspects of the child's personality and behavior. A child may be one person at
home with his or her parents, and may be quite another person outside the home, when he or
she is with friends. For example, minority children from middle-class families may get
a lot of pressure from their parents to take school seriously and achieve academically,
but they may succumb to pressure from their peer groups to downplay academic achievement
and other forms of what's sometimes called "acting white".
Peer groups and peer cultures are very important and as the child begins to spend more and
more time outside the home, they become ever more important. As a somewhat silly, I suppose,
but still illuminating example, Harris cites food preferences. She asks, "If parents are
such powerful socialization agents, why do they find it so hard to get their children
to eat what they want them to eat?" This is a real problem that every parent has encountered.
And Harris' answer is that children want to eat what their friends like, not what their
parents like. In eating, as in language, as in behavior, there's a lot of code switching
that goes on. And the more the child is subject to group socialization outside the home, the
less influence the shared home environment is going to have.
A study of academic motivation in 4th and 5th grade children by Kinderman makes this
point nicely. Kinderman found that students in the school she studied tended to group
themselves into cliques such as the "brains" and the "slackers". But she also found that
membership in these groups tended to be somewhat unstable, with individual children moving
back and forth from one group to another. And interestingly, Kinderman discovered that
children's' attitudes toward school changed as they changed cliques. Brains lost interest
in school if they moved to the slacker clique, and slackers gained interest if they moved
to the brain's clique. Why kids would move from one clique to another isn't completely
clear, although we know that happens -- just think about your own childhood. But the fact
of the matter is, the movement from one clique to another preceded the attitude change. It
wasn't caused by the attitude change. It wasn't as if one day a kid in the "brains" decided she wasn't
interested in school anymore and she was going to go join the "slackers". So movement from
one group of kids to another changed the child's attitudes, despite the fact that these kids'
IQs remained constant, obviously, and their parental influences did so as well, presumably.
It was the peer group that caused the changes to occur.
Returning to the behavior genetics research that's the theme of this lecture, we can see the
influence of the parents and the peer group on various aspects of adolescent behavior
in a study from David Rowe. In one analysis, Rowe found that parents who smoke cigarettes
tend to have children who smoke cigarettes too, and his behavior-genetic analysis indicated
that this influence was mediated by heredity -- that is, the genetic component of variance
-- but not by the shared family environment. So parents did have an influence on their
children, but that influence was transmitted genetically, through genetic inheritance, not
though the family environment that the parents created for their children. On the other hand,
adolescents who smoke cigarettes tend to have peers who smoke cigarettes as well. This is
an effect of the nonshared environment -- an effect of the environment outside the family.
Rowe and his colleagues found similar causal patterns for alcohol consumption, juvenile
delinquency, *** behavior, and pregnancy. In each case, the shared family environment
had relatively little impact on behavior. Genetic transmission from parent to child
was important, but so was the non-shared environment. Peer groups had a powerful effect on whether
adolescents experimented with tobacco, alcohol, sex, and general misbehavior.
So one way in which the nonshared environment shapes personality, attitudes, and general
social behavior is through the environment outside the family. Schoolmates, playmates,
teachers, ministers, whatever. But, interestingly, the non-shared environment also exists inside
the household. How that can be -- that's the subject of our next lecture.