No Two Alike Page 14
The same predisposition to behave in a nurturant and/or dominant manner toward younger members of their species has been described in chimpanzees. According to Jane Goodall, “Chimpanzees (like other group-living mammals) tend to behave in varied and predictable ways to members of the several age-sex classes.” Age and sex distinctions are made on the basis of physical cues; thus, “small body size, pale skin, and white tail tuft indicate infancy.” Juvenile chimpanzees of both sexes are attracted to infants and are usually gentle with them. Play gets rougher as the younger animal grows and the older one begins to assert its dominance.38
In our own species, studies have shown that children as young as four “talk down” to still younger children, just the way parents do, in an effort to make their speech understandable to the younger child. The four-year-olds use shorter, simpler sentences when conversing with two-year-olds than with children of their own age. They do this whether or not the younger child is their sibling and whether or not their parents are present. They do it even if they don’t have a younger sibling.39
Firstborns spend much of their time at home in the company of their younger siblings and they talk to them in short, simple sentences.40 No one, however, expects firstborns to simplify their speech when conversing with adults or with children of their own age: it is taken for granted that children adjust their style of speech to their interlocutor, and this is in fact what they do. Yet many people also take it for granted that other patterns of behavior that children acquire through interactions with their siblings will be generalized to other social partners and other social contexts. It is assumed that these behaviors, used habitually at home, become so ingrained or “internalized” that they become a permanent part of the personality.
This assumption, which is seldom stated explicitly, is the fourth problem for Sulloway’s birth order theory, and it applies to other birth order theories as well. Sulloway doesn’t think that firstborns are more bossy and aggressive only at home: he thinks they’re like that everywhere they go, all through their lives. As a result of their interactions with their younger siblings, they’ve grown accustomed to getting their way by being bossy and aggressive, so they apply this strategy in other situations and continue to do so in adulthood. Of course, they don’t make a conscious decision to behave like this: it happens automatically.
Once stated plainly, this assumption becomes testable. If birth order has effects on personality, these effects must occur in childhood, because that’s when people have the experiences that birth order theorists believe are important. So if behaviors acquired at home are generalized to other social contexts, we should be able to see that happening in childhood. If anything, the effects should show up more clearly in childhood than in adulthood, because it’s possible that firstborns learn to keep their aggressiveness under wraps as they get older.
This is where a well-designed birth order study comes in handy. Most evidence on children’s behavior comes from studies that looked at only one child per family (due to a failure to heed Robert Plomin’s advice) and hence are useless; the method provides no way of controlling for genetic influences on behavior. Genetic influences on behavior can look exactly like generalization and may be responsible for many of the modest correlations that developmentalists find between a child’s behavior in two different contexts.
For example, a group of developmentalists found a weak (but statistically significant) correlation of .19 between children’s behavior with their parents and with their peers: the children who behaved in a bossy or uncooperative manner with parents were slightly more likely to behave in a bossy or uncooperative manner with peers.41 The researchers assumed that the obnoxious behavior in both contexts resulted from the parents’ inept child-rearing practices: the child had learned that he could get his way at home by being obnoxious, so he was obnoxious on the playground too. But the researchers didn’t consider the possibility that some children may inherit a predisposition to behave in an aggressive or disagreeable fashion wherever they go. The heritability of these traits—around .45—is more than enough to account for the weak correlation between obnoxious behavior with parents and obnoxious behavior with peers.
One cannot conclude that behavior has generalized from the home to outside the home unless one has a way of controlling for genetic influences on behavior. Birth order studies can do that because, as I said, there are no systematic genetic differences between firstborns and laterborns. The genes that might predispose a child to be aggressive or disagreeable are dealt out evenhandedly to firstborns and laterborns. So if firstborns behave more aggressively than laterborns both at home and outside the home, that would suggest that generalization does occur—that learning to behave in a certain way in one context causes children to behave that way in other contexts as well. It wouldn’t be proof of generalization, however—the similarity in behavior could be the result of similar learning processes occurring separately in the two contexts. On the other hand, if firstborns behave one way at home and a different way outside the home, that would be good evidence against generalization.
Two well-designed studies provide useful data; in both, the researchers looked at two children from the same family and assessed their behavior both at home and outside the home. The first was carried out by the developmentalist Rona Abramovitch and her colleagues. The researchers observed children playing with their siblings and noted that firstborns do tend to dominate their younger siblings, often by behaving aggressively. Then the researchers observed the same children playing with their peers. They saw “no evidence of individual differences in sibling interactions carrying over into peer interactions.” Firstborns who dominated their siblings were not more likely to try to dominate their peers; laterborns were not more likely to allow their peers to dominate them. As the researchers put it, “Even the second-born child, who has experienced years in a subordinate role with an older sibling, can step into a dominant role when the situation permits.”42
The second study, by Kirby Deater-Deckard and Robert Plomin, used behavioral genetic methods to look at aggressive behavior in pairs of adoptive and biological siblings. The aggressiveness of both children was judged five times by their parents and by five different teachers, over a six-year period. What the parents were judging, as the researchers explained, was mainly their children’s behavior at home; what the teachers were judging was behavior at school. The results showed again that older siblings are more aggressive at home but not at school: parents judged the older one to be more aggressive, teachers judged the two siblings to be about the same.43
Based on the teachers’ reports, the heritability of aggressiveness in Deater-Deckard and Plomin’s study was .49; the estimate for shared environment was not significantly different from zero. As usual, about half the variance could be attributed to the effects of genes and none to the home environment, which means that half the variance remained unaccounted for.
Even in studies that provide no control for genetic influences, when researchers look at children’s behavior in more than one social context, the reported correlations are very low—often statistically insignificant, occasionally negative. In particular, children who fight like cats and dogs with their siblings are not at greater risk of getting along poorly with their peers. Nor are children who have no siblings in jeopardy of social failure: only children are not handicapped (nor helped) in their peer relationships by their lack of experience in interacting with siblings.44
The studies I’ve reviewed show that birth order definitely does affect behavior at home; older siblings dominate younger ones and behave more aggressively. Dominance hierarchies have predictable effects on their members: dominant individuals are more aggressive because they can afford to be. Those lower on the totem pole behave in a more conciliatory manner because they don’t want to risk evoking the anger of someone who could beat them up. The personality characteristics Sulloway attributes to firstborns and laterborns—nasty and nice—are the predictable behavioral consequences of
being the dominator or the dominated. However, the evidence shows that these behaviors are not generalized to social contexts outside the home.
And why should they be? The child who is the oldest and biggest at home might turn out to be the smallest on the playground. It wouldn’t make sense for this child to try to dominate his peers the way he dominates his younger siblings. A strategy that works at home might be useless or even hazardous elsewhere. Children are capable of generalizing—of learning something in one context and applying it in another—but they do not do it blindly. The question of when and how generalization occurs is an important one and relevant to my quest; I will come back to it in the next chapter.
Earlier I said that two major attempts have been made in the last few years to find the source of the unexplained variance: Turkheimer and Waldron’s meta-analysis and the NEAD study. Turkheimer and Waldron reported that, of the environmental factors that had been studied, family constellation variables, including birth order, accounted for the smallest amount of variance: only 1 percent. The NEAD researchers didn’t look specifically at birth order (many of their sibling pairs were twins), but they did look at “asymmetrical relationships” between siblings. These would include any in which one dominated the other and any in which sibling A was nasty to sibling B but B was nice to A. In summarizing the results of the NEAD study, David Reiss reported that their data had ruled out asymmetrical relationships between siblings as a source of the personality differences between them.45
These are the results for children and adolescents. Next I will look at data from adults. Actually, the subjects in most of these studies were college students. Researchers classify them as adults even if their parents don’t.
Designing a birth order study is not as simple as it appears at first glance. Small families produce as many firstborns as large families, but fewer laterborns; and small families tend to be higher in socioeconomic status. This means that a firstborn is more likely than a laterborn to come from an upper-class family. Most of the birth order studies you read about in the newspapers are worthless because the researchers failed to control for family size and/or socioeconomic class. Without these controls, statistical differences between firstborns and laterborns that are really due to social class are likely to be mistaken for birth order effects. It takes a very large N to do a properly controlled birth order study.
Studies that do control for family size and/or socioeconomic class, and that use standard methods of assessing personality in adults, seldom find significant differences between firstborns and laterborns. Frank Sulloway agrees with me on this; here’s what he said in an article published three years after Born to Rebel:
When assessed by self-report questionnaires, birth-order effects are typically modest and nonsignificant. Yet systematic differences by birth order are generally found when parents rate their own offspring or when siblings compare themselves with one another.46
Most of the personality data we have on adults come from standard self-report personality questionnaires. When researchers give these tests to large numbers of subjects, they generally find that firstborns are not significantly more (or less) open, conscientious, extraverted, agreeable, or neurotic than laterborns. Since a lot of researchers have done this, and each one typically looks for birth order effects in several different dimensions of personality, significant differences do turn up from time to time—they’re bound to. But there’s no consistency in these occasional successes (as such they are generally regarded): in one study there’s a significant result for conscientiousness; in another, it’s something else. Moreover, in many cases the significant result involved only some of the subjects, not all of them, or showed up only in one test out of two or three the researchers administered. For example, a significant result might be found for female subjects but not for males, or for subjects from large families but not from small families. And again there’s no consistency: in one study a birth order effect might be found only for females; in the next it might be only for males.
Sulloway has attempted to explain away the null or inconsistent outcomes found with self-report personality questionnaires by calling into question the validity of the tests themselves. He has claimed that they’re unreliable and potentially misleading.47 But casting doubt on self-report personality tests is a counterproductive strategy for Sulloway, because he can’t get along without them. These tests are responsible for producing the very finding he is trying to explain: the mystery of the unexplained variance in personality. “The crux of my argument,” he said in the introduction of Born to Rebel, “stems from a remarkable discovery. Siblings raised together are almost as different in their personalities as people from different families. This finding [was] firmly established by studies in personality psychology.”48 Perfectly true, and almost all of these studies used self-report personality questionnaires.
Most of the personality data that led Plomin and Daniels to ask why siblings are so different came from standard self-report personality tests. If there are no differences between firstborns and laterborns on standard self-report personality tests, then birth order cannot account for any of the variance in studies that used those tests. Which means that birth order cannot be the answer to Plomin and Daniels’ question about why siblings are so different.
Sulloway was quite correct when he pointed out that the likelihood of finding significant birth order effects depends on the method used to assess personality: consistent and statistically significant birth order effects are usually found when subjects’ personalities are judged by their parents or siblings, or when they’re asked to compare themselves with their siblings. The difference in outcome doesn’t depend on whether or not self-report tests are used (the subjects who compared themselves with their siblings did so on self-report tests) but on whether the behavior is assessed within the family context or outside of it. Judgments by family members are necessarily based on behavior within the family context and do show birth order effects. But no birth order effects were found when children’s behavior with their peers was observed by researchers or judged by their teachers. Nor are birth order effects found when researchers look at school grades or high school graduation rates (we have voluminous data on such things). If firstborns are more conscientious, as Sulloway claims, it doesn’t help them do better in their schoolwork. If laterborns are more rebellious, it doesn’t make them more likely to drop out of high school or less likely to go to college.49
Then why, when people are asked to name the academic achiever in their family, do they more often than not name the firstborn? And why, when they’re asked to name the rebel in their family, do they more often than not name a laterborn?50
The first question is easy to answer. Firstborns evidently give their siblings the impression that they are more conscientious about their schoolwork. The probable reason is that firstborns are older than their siblings. Older children tend to be more responsible than younger ones and generally spend more time doing schoolwork. In judgments made by family members, age differences are inevitably confounded with birth order.
The answer to the question about rebelliousness is a little more complicated. In fact, research does confirm that laterborns are more rebellious in one respect: they are more likely than firstborns to engage in “problem” teenage behaviors such as early onset of sexual activity, drinking, or drug use. In general, birth order effects are not found when personality or behavior is assessed outside the family context, so this is an exception to the rule. It’s one of those exceptions that turns out to be informative, because there’s also an exception to the behavioral geneticists’ rule that shared environment accounts for little or no variance in behavior: there are shared environment effects for adolescent delinquency. Notice that both these exceptions involve teenage problem behaviors, and that none of the behaviors first occurs at home. Shoplifting, vandalism, drug use, underage sex and drinking—these are things that teenagers do outside the home, in the company of their friends.
T
he explanation for both exceptions is the same: teenagers like to hang around with their older siblings and their older siblings’ friends. The same heightened risk for problem teenage behaviors has been observed in teenage girls who are physically mature for their age and whose friends consequently tend to be older than themselves.51
The fact that laterborns are often exposed to the temptations of teenage life at an earlier age than firstborns, and sometimes succumb to these temptations, is probably the reason why the members of their family tend to characterize them as rebellious. But evidently their rebelliousness doesn’t extend to other types of behavior, such as doing their schoolwork, and it doesn’t show up on personality tests given to adults.
Many people, including some evolutionary psychologists, were impressed by the tables and graphs in Sulloway’s Born to Rebel and by the magnitude of the birth order effects he reported. No one before (or since) found such striking effects, such gaping differences between firstborns and laterborns in personality and behavior. Firstborns, reported Sulloway, were far less likely to accept new scientific theories. Laterborns were far more likely to be politically liberal and rebellious.
But doubting voices soon began to be heard. Many of the reviewers had the same criticism: that despite all the graphic presentation of data, the book failed to provide the information a reader needs in order to evaluate the evidence. “No tables of descriptive statistics,” complained the psychologist Toni Falbo in the book review journal of the American Psychological Association. “No simple presentation of the frequency of people falling into key categories.” And, she said, there were other omissions: