No Two Alike
Page 7
There is one more objection to behavioral genetic methods that I’ll have to deal with: the charge that twin studies are invalid because they’re based on a fatal error the critics call the “equal-environments assumption.” Since behavioral geneticists don’t assume that a pair of twins reared in the same home have equal environments, it should really be called the “equally-similar-environments assumption.” It’s the assumption that the home environments of reared-together fraternal twins are as similar as those of reared-together identical twins. This assumption is wrong, the critics claim, because parents treat identical twins more similarly than fraternal twins.25 If identicals actually have more similar environments than fraternals, then the greater environmental similarity might explain why identicals are more alike in personality than fraternals. This means that the estimates of the heritability of personality might be too high: environmental effects on personality may have been mistaken for genetic effects.
It is true that behavioral geneticists make the equally-similar-environments assumption. The reason is historical: their methods were first used to answer the question of why children tend to resemble their parents in intelligence. How much of this resemblance, the researchers asked, is due to the intellectually enriched or impoverished environment the parents provided and how much to inherited genes? This is where the “nature-nurture controversy” had its roots and why it was so bitter: it involved the politically charged question of the heritability of intelligence. There were people in the 1970s who very much wanted the heritability of intelligence to be zero.26 They have since been reduced to bickering about whether it’s closer to .60 or to .30.
When the variable being studied is intelligence, it is reasonable to assume that identical twins and fraternal twins have equally similar environments. Parents do not, as a rule, provide an intellectually enriched environment for one of their children and not for the other, especially if the children are the same age. The features of the home that are assumed to play a role in the development of mental abilities—the parents’ education and their attitudes toward academic achievement, the vocabulary they use in conversation, the dictionaries and computers that the home does or does not contain, the family excursions to the museum or the ballpark—all are likely to be the same for two siblings reared in the same home, whether or not they are identical twins.
But the equally-similar-environments assumption becomes less tenable when it comes to personality. In fact, there’s no question about it: parents behave in a more similar fashion to identical twins than to fraternal twins, and in a more similar fashion to ordinary siblings than to step- or adoptive siblings. How a parent behaves toward a child is in part a reaction to the child’s behavior; the parent-child relationship is a dialogue to which both parties contribute. That was the point I was making in the quote from The Nurture Assumption that appears at the beginning of this chapter, the passage in which I mentioned that one of my children was adopted. It is impossible to treat two children the same if they behave differently.
Two children who have similar genes are more likely to be similar in behavior, as the behavioral genetic studies showed. Therefore, they are more likely to be treated similarly by their parents.27 So if you believed that the way children are treated by their parents has long-term effects on their personalities, and if you believed that being treated in the same way by their parents makes children more alike, then you would conclude that studies that compare reared-together identical twins with reared-together fraternal twins produce inflated estimates of heritability.
Fortunately, behavioral geneticists don’t have to rely on a single method for calculating heritability; as I mentioned, they generally use two or more methods to home in on a result. Since each method has its own strengths and weaknesses, the critics have had to resort to thinking up a different, ad hoc criticism for each one. But let’s say, for the sake of argument, that they are right in this case: that there might be some overestimation of heritability in twin studies. What would that mean? Why should I care?
The answer is, I don’t. I’m not trying to explain why identical twins are so alike—I’m trying to explain why they are so different. It’s immaterial to me whether heritability accounts for 30 or 60 percent of the variance: I’m interested in the variation in personality that is not due to genes—the unexplained variance. If identical twins are treated so similarly by their parents, how come they’re not more alike? How come a sizable portion of the variance remains unexplained by their shared genes plus the presumed effects of the presumably similar treatment by their parents?
Whether or not the behavioral geneticists have overestimated heritability, they have provided us with valuable information. Their evidence has eliminated from contention all aspects of the environment that siblings who grow up in the same home do in fact share, whether or not they are twins. All the things that siblings do have in common have been shown to be ineffective; these environmental factors can explain no more than a negligible fraction of the variation in personality. But any aspects of the environment that can differ for two siblings raised in the same home, including how they are treated by their parents, are still in the running.
Also still in the running is the possibility that some aspect of the home environment might be the same for two siblings and yet they might react differently to it. If, for instance, their parents are always quarreling, and one child withdraws from the tumult and becomes an introvert while the other reacts by becoming a hail-fellow-well-met type, these dissimilar reactions will contribute to the unexplained variance—the variation in personality I’m trying to explain.
The behavioral geneticists have not ignored these possible sources of differences between siblings. Nor am I planning to sweep them under the rug. I’ll return to them in the next chapter and the chapter after that.
The violation of the equally-similar-environments assumption means that identical twins actually have more similar home environments than fraternal twins do. Since they are genetically identical, they might also be expected to react more similarly to environmental events such as parental quarreling. And yet the unexplained differences between them are as wide as the unexplained differences between other kinds of sibling pairs. They are as wide as the unexplained differences between a pair of unrelated strangers. This is the important clue that I mentioned earlier. What makes it hard to understand is the fact that identical twins are so much more alike. Shouldn’t there be less variation to account for?
It isn’t an illusion; identical twins really are more alike in personality than other kinds of sibling pairs, though the resemblance is a long way from perfect. The correlation between their scores on personality tests is around .45 or .50, whether they were reared together or apart.28 The correlation is considerably lower for fraternal twins and ordinary siblings, and is approximately zero for adult adoptive siblings who were reared in the same home; adoptive siblings aren’t alike at all.29 Heritability—the effects of genes—accounts for virtually all of the resemblances between the biologically related pairs. (The effects, if any, of the greater environmental similarity of identical twins would be included in the heritability estimate.)
What confuses people is that they expect the amount of variance attributed to heritability to equal the correlation, but in most cases it does not. The reason is that correlations represent only similarities between siblings, whereas genes are responsible for differences as well as similarities. Some of the differences between siblings who aren’t identical twins are due to the fact that they have different genes. For example, fraternal twins sometimes differ in eye color. This difference is entirely due to the fact that they inherited different eye-color genes. In a given population, the correlation for eye color in fraternal twins might be only .75, but the amount of variance accounted for by genes is nonetheless 1.00. Whether you’re a identical twin or a fraternal twin, your eye color is completely a function of your genes (that’s what a heritability of 1.00 means), but genes for eye color can vary.
Here’s how the estimate of about .45 for the heritability of personality can hold true for identical twins, fraternal twins, and adoptive siblings alike. For identical twins, heritability accounts for all of their similarities and none of their differences. For fraternal twins, heritability accounts for some of their similarities and for some of their differences. And for adoptive siblings, heritability accounts for none of their similarities (there are no similarities to account for) and for about half of their differences.30 The upshot is that the amount of variance not explained by genes is roughly the same, around 55 percent, in each case. In one major recent study, five different types of sibling pairs—reared-together identical twins, fraternal twins, ordinary siblings, half-siblings, and step-siblings—were all assessed with the same exhaustive battery of tests, and the conclusions held true across the board. There were no important differences among the sibling pairs in the proportion of variance that was unexplained.31
This counterintuitive finding has an important implication. It implies that whatever causes reared-together identical twins to differ in personality also causes ordinary siblings and step-siblings to differ, and to differ just as much.
So the perpetrator I’m looking for is not heredity; nor is it the environment your parents provided for you and your siblings. It doesn’t care where you got your genes and is indifferent to your childhood address. The perpetrator I’m looking for widens the personality differences between identical twins in the same way, and to the same degree, as it widens the differences between ordinary siblings. It produces nongenetic differences between two people raised in the same family that are as big as those between a pair plucked at random from the population.32
One suspect whose modus operandi could fit that description is chance. I’m not even going to count this as one of my red herrings, because if chance—randomness—is responsible, then my mystery would have no solution. It would be like one of those irksome books where it turns out that the victim died of natural causes or faked his death and is living with his mistress on an island in the Bahamas.
Could differences in personality all be due to random things that happen during development? A neuron zigged instead of zagged and you are impulsive rather than cautious? A bully knocked you down in fourth grade and you are shy rather than outgoing?33
If it were true—if the unexplained variance were all due to random perturbations of one sort or another—then my quest for a solution would be in vain. In psychology there’s no good way to study random things because you can’t combine data from different subjects in order to look for overall trends. If each subject turned out the way he or she did for idiosyncratic reasons, there wouldn’t be any overall trends. Whatever made one person impulsive would be different from whatever made the next one impulsive. All trees, no forest.
It isn’t difficult to find evidence that favors the no-solution conclusion. There is a certain amount of randomness—or what appears to be randomness, given our current understanding of biological processes—in all human characteristics. The human genome is not capacious enough to specify every whorl of the fingerprints, every synapse in the brain. To some extent the details of construction have to be worked out on the construction site, which adds variability to the outcome—“developmental noise,” biologists call it. The result is that identical twins aren’t exactly alike physically. They may differ slightly (occasionally more than slightly) in height and weight and in appearance; the family and friends of identical twins generally have no trouble telling them apart. MRI scans show subtle differences between their brains.34
I mentioned in the previous chapter that the identical twin of a person with schizophrenia has only a 48 percent chance of also being diagnosed with the disorder. Similar findings have been reported for physical malfunctions. If one twin falls ill with type 1 (childhood onset) diabetes, the other has only a 40 percent chance of developing the disorder. The heritability of the tendency to get ear infections in childhood is .73; the remainder of the variance is unexplained. As far as physicians can tell, it’s random. Sharing a home does not make children more alike in whether they do or don’t get ear infections, any more than it makes them more alike in aggressiveness or conscientiousness.35
There are biological differences between identical twins, due to unpredictable things that happen before and after they are born. A neuron zigs instead of zags. One fetus occupies a better position in the uterus. One twin falls down the stairs or contracts a virus. Exposure to pathogens has been proposed as a possible etiological trigger both for schizophrenia and for type 1 diabetes.36 The schizophrenic twin may have developed the disorder not because her life was more stressful but because she happened to encounter a virus at the wrong time.
A few pages ago I admitted that heritability of personality characteristics might be somewhat overestimated due to the violation of the equally-similar-environments assumption. It is possible that some of the variance attributed to genes should really have been attributed to shared environment. But now it appears that some of the variance not attributed to genes isn’t environmental at all: it’s biological. Some of it may even be genetic. Medical researchers recently described a case in which one identical twin was born with a cleft lip and palate, the other with a normally formed face. It turned out that the birth defect was due to a genetic mutation, present in one twin and not the other.37 The mutation must have occurred after the fertilized egg that gave rise to both twins split in half and formed two separate embryos.
The biological perturbations that affect the outcomes of identical twins also affect nontwins. A virus or a mutation, a neuron that zigs instead of zags, can happen just as easily to babies born singly. So these random biological events—developmental noise—could explain some of the unexplained variance in personality.
But orderly genetic effects plus developmental noise can’t explain all of it, because that would mean that virtually all the variation in personality was built in, one way or the other. It would mean, in other words, that the environment has no lasting effects on personality—that young humans are incapable of making long-term modifications of their behavior on the basis of their experiences. This is implausible both on theoretical and empirical grounds. Much of the evidence I will discuss in this book has to do with the ways that children modify their behavior in response to their experiences.
What about randomness that isn’t biological? What about environmental randomness? One twin gets knocked down by a bully in fourth grade and the other doesn’t. One gets the upper bunk bed, the other gets the lower. One stays home sick from school on a particular day and that makes all the difference.38
Again, I have to admit that this kind of randomness can’t be ruled out; no doubt it accounts for some portion of the unexplained variance. There are many opportunities for random events to occur in people’s lives, and random events sometimes do have dramatic long-term consequences. It is a familiar theme in real life and in fiction. Here’s Pip, in Charles Dickens’ Great Expectations, describing the consequences of his first visit to Miss Havisham’s house:
That was a memorable day to me, for it made great changes in me. But, it is the same with any life. Imagine one selected day struck out of it, and think how different its course would have been. Pause you who read this, and think for a moment of the long chain of iron or gold, of thorns or flowers, that would never have bound you, but for the formation of the first link on one memorable day.39
It was a random event of a much less memorable sort that changed the course of my life and led me to become a writer of books and articles about psychology. I had a friend who had a dog
she needed to find a home for, and I helped her write an ad for the classified section of her local newspaper. A few months later she needed help rewriting an article that had been turned down by a psychology journal—my friend was an assistant professor of psychology at Rutgers University—and she thought of me. “You have a way with words,” she told me. Her opinion was based on the ad I had helped her write, a classified ad about a dog. The shortest writing job I’ve ever had. It started me on the path that ended up here.
On second thought, perhaps I would have become a writer anyway. At any rate, writing the classified ad was only one link—one zig where I might have zagged—in a long chain of thorns and flowers that may or may not have been random. I probably wouldn’t have become a writer if I hadn’t been rejected by my peers for a crucial four years of my childhood.40 Those years as a social outcast turned me into an introvert; previously I had been a boisterous, outgoing child. If my parents had decided to buy a house in some other community, perhaps my schoolmates wouldn’t have rejected me. When, after four years, my family moved to a different part of the country, I stopped being an outcast. But by then the link had been forged; my personality had changed.
Personality can change; evidence from large numbers of subjects shows that it can change, to some extent, even in adulthood, if one’s circumstances change.41 But it is considerably more plastic, more flexible, in childhood. Children are adaptable. They adapt to their society, their culture—they become socialized. And they also adapt to their individual circumstances, as I adapted to being a social outcast.