Josephine Tey’s first job was to convince her readers that the mystery of what happened to the Little Princes hadn’t already been solved—that the solution everyone had accepted for so long was just a myth. That’s my first job too, though I will make shorter work of it than Tey did. But I’m fighting the same uphill battle. People see what they expect to see; ambiguous results, like ambiguous pictures, are interpreted in accordance with preconceived notions. Sir George Bellew, author of The Kings and Queens of Britain, looked at the portrait of Richard III and saw in Richard’s face clear indications of his “unscrupulous nature.”44
Seven months after I told Andrew Collins and Eleanor Maccoby about the errors in their American Psychologist article, the journal published some letters to the editor commenting on the article, along with the authors’ reply. This would have been their opportunity to set things straight, at the very least by correcting the wrong reference they gave for the Suomi study and their mixup of “calm” and “nurturant.” They didn’t do it. They admitted no errors in their reply.45
What’s worse, Collins is evidently still telling students who take the introductory psychology course at the University of Minnesota about Suomi’s cross-fostered monkeys. The website for this course currently includes the following item among its “learning objectives”:
Describe the findings of the Suomi study that Dr Collins will tell us about. It’s a good piece of evidence in the nature/nurture controversy—the biological infants of reactive rhesus monkey mothers were fostered by ‘calm’ & ‘reactive’ rhesus monkey mothers. What happens?46
Yes, tell us, Dr. Collins. What happens?
Collins, Maccoby, and the three other authors of the American Psychologist article summed up their conclusions as follows:
This new generation of evidence on the role of parenting should add to the conviction, long held by many scholars, that broad, general main effects for either heredity or environment are unlikely in research on behavior and personality. Statistical interactions and moderator effects are the rule, not the exception.47
Main effects are unlikely; interactions are the rule. Maccoby had said pretty much the same thing at the NICHD conference:
If a given kind of household or a given style of parenting has different effects for children with different predispositions, this means that parenting often functions to make children in the same family different rather than alike.48
What they are saying is that a given style of parenting has no predictable effects: the same parental behaviors can cause one child to become more cheerful and another more depressed, one more honest and the other more deceptive. If a given style of parenting had predictable effects even on the average—if certain parental behaviors led to cheerful or honest children more often than not—then we would see main effects of parenting. Since the effects of parenting depend, according to the developmentalists, on the predispositions of the child you happened to get, it would seem pretty clear that no one could give you advice on how to raise that child, unless they had some secret way of finding out about his or her predispositions. Right?
A few months after the conference I got a letter from the director of NICHD. All the participants in the conference were invited to contribute to a booklet of parenting advice aimed at the general public. “Consider it this way,” the letter said. “If you could give parents one piece of advice from your research, what would it be?”49
My advice would be to run the other way if you see an advice-giver coming toward you, but I didn’t think that was what the director wanted to hear so I didn’t bother to reply. The booklet, with NICHD’s imprimatur on it, came out in 2001; Adventures in Parenting, it’s called. Here are two samples of the advice it contains:
Much “acting out” stems from children not knowing how to handle their emotions. Feelings can be so intense that usual methods of expressing them don’t work. Or, because feelings like anger and sadness are viewed as “bad,” your child may not want to express them openly. Encourage your child to express emotions in a healthy and positive way.
Because you can’t be with your child all the time, you should know who is with your child when you’re not. Friends have a big influence on your child, from pre-school well into adulthood. Much of the time, this influence is positive, but not always. With a little effort from you, your child might surround him or herself with friends whose values, interests, and behaviors will be “pluses” in your child’s life.50
According to the acknowledgments page of the booklet, the 1999 NICHD conference on parenting “provided the impetus for this publication.” But a major theme of that conference was that parenting doesn’t have any main effects: only gene-environment interactions. If parenting doesn’t have any main effects, it means that if you encouraged your child to express emotions in a healthy and positive way, it might have a beneficial effect, but then again it might have a deleterious one. If you spent a little effort to see that your child surrounds him- or herself with the right kind of friends, your efforts might pay off, but then again they might backfire. Your results will vary because what you get depends on the kind of kid you started out with, and the writers of the advice booklet have no idea what kind of kid you started out with.
The developmentalists are speaking out of both sides of their mouths. In the American Psychologist article, five of America’s most prominent developmentalists said that “broad, general main effects for either heredity or environment are unlikely in research on behavior and personality.” But all they really want to deny is the existence of broad, general main effects of heredity (the evidence for which is incontrovertible). They don’t really want to deny the existence of broad, general main effects of the home environment. If the developmentalists gave up their belief in broad, general main effects of the home environment, they would be forced to admit that giving advice to parents is useless.
In reality, it is worse than useless. Advice-giving has moral repercussions. If there are good parents—the ones who provide just the right amount of nurturing but who avoid being overprotective, who encourage their children to express their emotions in a healthy way, who take the time and trouble to make sure that their children have the right kind of friends—then there must also be bad parents, the ones who fail to do these things. And if the right kind of parenting produces good kids, then bad kids must be the product of the wrong kind of parenting. If your child is timid, you must have overprotected him or failed to provide sufficient nurturing. If your child acts out or bottles up emotions, it’s because you didn’t encourage her to express emotions in a healthy way. Even if there are no main effects, only interactions, that doesn’t let you off the hook: you could have prevented your high-reactive baby from turning into a timid kid, if only you hadn’t overprotected him!
The detective in The Daughter of Time was up against four centuries of unquestioned belief in Richard III’s culpability. I am up against only fifty years of unquestioned belief in parents’ culpability. Sixty-seven years ago, when I was born, parents didn’t get blamed when their children turned out badly; according to the cultural myths that were prevalent at that time, bad kids were “born that way.” Alice James attributed her depression to heredity, not to her parents’ child-rearing methods.
The unquestioned belief in the good and bad effects of parenting has resulted in a systematic bias in the reporting of research results. In addition to the exaggeration of supporting evidence and the slanted interpretation of ambiguous evidence, there has also been a suppression of negative evidence. I mentioned one case in The Nurture Assumption: a study of the effects of child-rearing practices, done in the 1950s, that failed to produce the expected results. The null outcome went unpublished until Eleanor Maccoby—one of the researchers on the study—mentioned it in print, thirty-five years later.51
Jerome Kagan’s first noteworthy publication was a book titled Birth to Maturity, which came out in 1962.52 It was the final report of a major research project in which almost a hundred subjects had b
een followed from infancy to young adulthood. Periodic assessments were made of the subjects themselves and of the “maternal practices” of their mothers. Of the large number of correlations the researchers calculated between maternal practices and child outcomes, only 6 percent—about the percentage you’d expect to occur by chance—were statistically significant.53 This failure to find meaningful results was apparent, however, only to readers who carefully scrutinized the appendixes of the book, where the correlation coefficients were reported. In fact, it was apparent only to those who carefully scrutinized the first edition of the book. The reprint edition of Birth to Maturity, published in 1983,54 has suffered an appendectomy.
Some years ago I had a brief e-mail interchange with a developmentalist. She told me that in the early 1970s she and a colleague—an older man with an eminent reputation—had carried out a large and elaborate research project of the sort called an “intervention study.” Intervention studies, which I will discuss in a later chapter, are designed to improve child outcomes by training parents to use better child-rearing methods. This study, according to my correspondent, “yielded NO effects…. These data were never published unfortunately.”55
Medical researchers testing the efficacy of a new drug are expected to disclose any financial ties they have to the manufacturer of the drug, but conflicts of interest occur all the time in developmental research and nobody raises an eyebrow. The researchers themselves and the granting agencies that fund their research—NICHD, for example—are united in their faith in the efficacy of parental influence.
A physician commenting on medical research in the Journal of the American Medical Association said, “If possible, the effectiveness of an effort should be determined by someone outside the effort who has nothing to gain by its perpetuation.”56 The corollary of this good advice is that the effectiveness of an effort shouldn’t be determined by those who have a lot to lose if the effort turns out to be a failure.
There’s an old story about two ladies in Victorian England who have just heard the news about the theory of evolution. One exclaims, “According to Mr. Darwin, our ancestors were apes!” The other replies, “That cannot be true! But if it is true, let us pray that it will not become generally known.”
The developmentalists sincerely believe in the efficacy of parental influence. But if parental influence turns out to be a dud—if the best kind of parenting they can think of is really no better than the placebo—they are praying that it will not become generally known. This prayer is not motivated entirely by self-interest. In the Newsweek article, psychologist Frank Farley of Temple University warned of possible consequences:
[Harris’s] thesis is absurd on its face, but consider what might happen if parents believe this stuff! Will it free some to mistreat their kids, since “it doesn’t matter”? Will it tell parents who are tired after a long day that they needn’t bother even paying any attention to their kid since “it doesn’t matter”?57
Farley’s statement is absurd on its face. Thousands of generations of hominid parents took good care of their children (have you any idea how much trouble it was to keep a child alive in the days when there were no permanent homes, no baby carriages or strollers, no jars of baby food, and no disposable diapers?), even though it never occurred to these parents that what they were doing might have long-term effects on their children’s personalities. But what bothers me about this statement is not so much the ignorance of evolutionary psychology it reveals: it’s the arrogance. Farley seems to think that he and his colleagues have the right to decide what American parents should be told about the efficacy of parenting—that he and his colleagues have a right to withhold or suppress information if, in their opinion, it wouldn’t be a good idea to have it become generally known.
In the past, researchers looking for gene-environment interactions often relied on imprecise methods for estimating the contribution of genes. Subjects were presumed to have an inherited tendency to be fearful or irritable on the basis of a pretest given in infancy.
In the future, researchers will look for gene-environment interactions by decoding their subjects’ genes. A few studies of this sort have already been published. One, for example, showed an interaction between stressful life events and a gene associated with anxiety and depression. People who have one variant of this gene are at risk of becoming depressed, but this generally happens only if they experience multiple stressful events, such as the loss of a job or the breakup of a romantic relationship. People who have the alternate version of the gene are unlikely to become depressed no matter how stressful their lives are.58 It’s a sensitivity type of gene-environment interaction: the gene makes its owner more vulnerable to the slings and arrows of outrageous fortune.
There is a paradoxical danger in this kind of research: it sounds so frightfully scientific that the reader is discouraged from looking too closely at it and may simply accept the researchers’ conclusions. But let the buyer beware: a scientific approach to the genetic component of a gene-environment interaction may be coupled with a naive, old-fashioned approach to the environmental component. The example I have in mind is the study—you may have read about it—in which researchers found that a variant of a certain gene, which codes for an enzyme abbreviated as MAOA, greatly increased the likelihood of antisocial behavior in adulthood, but only in males who had been treated badly in childhood.59 The low-MAOA variant of the gene allegedly increases a child’s sensitivity to what the researchers called “maltreatment” but that was generally referred to as “abuse” in the news coverage.
The results, it was claimed, can explain why some, but not all, maltreated children grow up to become violent criminals. The conclusion might be true but the data from this study offer scant support for it. There is a well-known correlation between abuse in childhood and unfavorable outcomes of one sort or another in adulthood, but correlations are uninformative about causes. The study that focused on the low-MAOA gene is simply a correlational study packed together with a genetic study that controlled for only one gene. Maltreatment (which was defined by the researchers as any of a variety of unfavorable experiences,60 such as more than one change of primary caregiver in childhood or a “soiled, unkempt appearance” at age three) is statistically associated with a number of other environmental factors. And complex outcomes such as antisocial behavior are highly unlikely to be influenced by only a single gene—it usually takes many, working together. I can think of several alternative explanations for the results of this study.
For example, the low-MAOA gene might lead to antisocial behavior only if it is coupled with another (unknown) gene that produces a high activity level. Highly active children try their parents’ patience and are more likely to be treated harshly;61 thus the maltreatment could be a consequence of having the high-activity gene, and the antisocial behavior could be a consequence of having both genes. This would be a gene-gene interaction, not a gene-environment interaction. Another possibility is a gene-environment interaction that is due to environmental factors other than maltreatment by parents. The unfavorable experiences included in the researchers’ definition of maltreatment occur more commonly in homes of lower socioeconomic status, and such homes are more likely to be located in neighborhoods where crime and violence—antisocial behavior—are prevalent. As I mentioned earlier, neighborhood effects on antisocial behavior are well known.
Nobody questions that child abuse is cruel and immoral. But evidence that is in accord with deeply held beliefs should be looked at just as critically as evidence that goes against these beliefs.
The term “red herring” comes from the sport of foxhunting. A red herring is a smelly smoked fish (the reddish color appears when it is smoked) that was sometimes dragged across a trail in order to distract the hounds from following the track of the fox. When this was done in training, rather than as a way of sabotaging a hunt, its purpose was to teach the hounds not to be sidetracked from their real goal, which was to find the fox.
My goal is
to find out why people differ from one another in personality. This chapter was about gene-environment interactions, which turned out to be a red herring. That makes three now. Mind you, I’m not claiming that gene-environment interactions don’t occur: just that they are not the perpetrator we are hunting for. They’re fish, not fox.
Gene-environment interactions exist, but the ones that have been convincingly demonstrated are all of the same sort: certain genotypes are more sensitive to certain environmental conditions. Such interactions do not produce null effects; they cannot explain why the behavioral geneticists found that differences in home environments had little or no effect on adult personality.
Nor—to return to a point I made at the beginning of the chapter—can they account for the differences between identical twins reared in the same home. Imagine, for a moment, that Kagan’s high-reactive infants came in sets of two: identical twins, two babies with the same genotype. If the test given at four months is a valid measure of their genetic predisposition to be nervous, both will be judged high-reactive. What happens next? Will their mother be indulgent with one and strict with the other? Will she tell one twin, “Get into that sandbox and play with the other kids, dammit!” and the other, “Better stay here by my side”? It’s possible, of course—I will examine this possibility in the next chapter—but in all likelihood, as the critics of behavioral genetics keep pointing out, she will treat them much alike. So can we attribute their personality differences in adulthood to the fact that “a given style of parenting has different effects for children with different predispositions”? No, because identical twins have the same predispositions. A gene-environment interaction can produce differences between two individuals only if they have different genotypes. Since identical twins have the same genotype, a gene-environment interaction cannot produce a difference between them. It would take a difference in environments—a main effect, not an interaction.
No Two Alike Page 11
