To The Nurture Assumption home page
(published by the New Jersey Association of Cognitive-Behavioral Therapists)
Copyright 1999 by Judith Rich Harris
Posted with permission.
I commend the NJ-ACT Newsletter for publishing Milton Spett's intelligent and interesting review of my book The Nurture Assumption. Unlike many reviewers of this book, Dr. Spett read it carefully and did an excellent job of summarizing its main points. Nevertheless, I don't agree with everything he said. I'd like to comment on three points in particular.
1. "Failing to find significance does not prove the null hypothesis."
The null hypothesis is the hypothesis that the "independent variable" of interest (in this case, the way the parents rear the child) has no effect on the "dependent variable" (the child's personality and mental health). As Dr. Spett correctly pointed out, you can't prove the null hypothesis. All you can do is reject it at some level of confidence. My assertion that parenting has no long-term effect on the child's personality is tantamount to saying that the null hypothesis is true, and Dr. Spett is right -- I can't prove it.
But why should I have to? Shouldn't the null hypothesis be our starting point? Shouldn't we begin by assuming nothing -- no effect -- and shouldn't the burden of proof be on those who wish to reject the null hypothesis? In the absence of solid evidence that would enable us to reject it at some level of confidence, shouldn't we proceed on the assumption of no effect?
The problem is that the null hypothesis of zero parental influence has not been treated as the starting point, and that's exactly what I mean by "the nurture assumption." To say, "We will assume that parents do have an effect until you prove they don't" is an example of the nurture assumption in action.
Now that I have identified myself as a staunch defender of the null hypothesis, I can confess that I chuckled at Dr. Spett's joke that, had it not been mathematically impossible, I would have made the percent of adult personality determined by parenting even lower than zero percent!
2. Dr. Spett correctly stated that heredity can account for about 50 percent of the variation in adult personality and that what I set out to do was to account for the other 50 percent.
He was also correct in saying that I have not proved that my theory can account for the other 50 percent (though it is a bit of a stretch to say, about a book with 57 pages of endnotes and references, that "the amount of research she cites supporting her theory is zero, zilch, none, nada"). Where Dr. Spett goes seriously astray is when he throws up his hands and says that "we just don't know what determines the other half of adult personality" -- that it could be experiences in the peer group, experiences at home, or some combination or interaction of the two.
I have not proved my theory can account for all of the unexplained 50 percent, but the evidence I presented makes it highly unlikely that parental influence -- acting either as a main effect or an interaction -- could explain more than a tiny fraction of it. If one looks at the evidence with an unbiased eye, only three slim hopes remain for parental influence.
The first is that parental effects might exist in the portion of the population for which we lack good data -- the same portion that slips through the sieve of the census-takers. However, this is only a small fraction of the population. The majority of the people who consult psychotherapists come from middle- or working-class homes, and we have abundant data on people who grew up in homes of this sort.
The second possibility is that parental influence does have effects, even in the populations that have been most studied, but the effects are very small or of such a subtle nature we haven't figured out how to measure them. Bear in mind, however, that developmental psychologists have spent the last fifty years searching in vain for these effects.
The third possibility is the most interesting: perhaps parents do have effects on their children but these effects are completely unpredictable and unsystematic. Many people find this idea appealing. After all, we know perfectly well that parents treat their children as individuals and neither want nor expect them to turn out all alike. So this notion makes a certain amount of sense.
Until you examine it more closely. When psychologists talk about parental influence, do they really mean that if you want a truthful child, you might do well to set an example of truth-telling, but you might do equally well (or better) to adopt a policy of lying all the time? Do they mean that rigid, unloving discipline might have bad effects on the child but is just as likely to have good effects? If the effects of the parents' behavior are totally unpredictable, isn't it pointless and misleading to talk about "parental influence"?
The conclusions of the studies I surveyed hold equally well for ordinary siblings, for adoptive siblings, and for identical or fraternal twins. Consider a case in which identical twins, for genetic reasons, are both born with a tendency to be "inhibited" or fearful. Jerome Kagan, who has studied children of this sort, believes that a certain style of child-rearing will cause them to become less fearful. Now, if the mother of these twins happened to have a knack for rearing children in this manner, wouldn't she apply it to both her twins? If she takes one to a playgroup, wouldn't she take the other? And shouldn't they both become less fearful?
But the evidence indicates that, if what the mother is doing has any effects at all, the effects must be different for each twin. One might become less fearful because of the mother's style of child-rearing, but the other will remain fearful or become more so. Identical twins reared in the same home are no more similar in personality than those reared in separate homes. What's interesting is not the fact that the ones reared in separate homes are so alike, but the fact that the ones reared in the same home at the same time by the same parents are so different. Their scores on personality tests correlate only about .50. That is the mystery I set out to solve.
3. Finally, I'm afraid that Dr. Spett may have left you with the erroneous impression that I said it doesn't matter whether or not you let your baby cry, or whether or not you are nice to your child.
Of course it matters! Children are thinking, feeling, sensitive human beings who are completely dependent on the older people in their lives. I believe that parents have a moral obligation to provide their children, to the best of their ability, with a happy homelife. They should do this, not because they think it's going to make the child more successful in the long run, but because a child's happiness or unhappiness is important right now, today.
The other reason why a parent's behavior matters is that it affects the parent-child relationship. Although, according to my theory, relationships have no long-term effects on personality, that doesn't mean they are unimportant. By behaving in a cruel or negligent manner, parents can ruin forever their relationship with their child, even though the child may come out of it intact (many do).
But relationships work both ways and parents, too, have a right to be treated fairly. In his review, Dr. Spett mentioned that many people who consult psychotherapists "have a deeply ingrained belief that their neurosis was caused by their parents." "If this type of explanation helps patients feel that their lives make sense," he said, "this is all to the good."
I don't believe it is. First, I don't believe that people should be encouraged to go on believing things that aren't true. Second, it might divert attention from the real source of the patients' problems. Third, it's not fair to the parents. Fourth, it could harm patients in the long run by making it harder for them to maintain or re-establish close ties with their parents.
People seek psychotherapy because they are unhappy. Wouldn't it be
better for them to establish warm and comfortable relationships with
their parents than to go on harboring grudges against them? Aren't
their parents more likely to give them the love and support they are
seeking if they can go to them and say, "Mom, Dad, I realize now
that it wasn't your fault"?
Milton Spett replies:
I would like to thank Judith Rich Harris for her kind words about my review, and I would like to comment on her third point. I agree that we should not send our patients out into the world blaming their parents for their problems. But in my experience, this doesn't happen. As patients improve and become less defensive, most of them spontaneously realize that their parents meant well and did the best they could, given the constraints of their own background. Patients then begin to enjoy a new relationship with their parents in which they feel sympathy, rather than anger, at their parents' limitations.
I do not believe it is fruitful to spend therapy time debating the nurture assumption with patients who blame their parents for their problems. On the other hand, I often quote The Nurture Assumption to patients who are guilt-ridden over their children's imperfections.