|Book Review: The Nurture Assumption
||[Jul. 24th, 2012|04:11 pm]
The latest book I read was The Nurture Assumption by Judith Rich Harris, which was supposed to argue that parents don't really have much of an effect on how their kids turn out.
This sounded ridiculous when I first heard it, but people I trusted like Steven Pinker kept endorsing it, so I finally picked it up. The thesis might be a little more subtle than that. Parents can still impact their kids' biological development - to take an extreme example, if you malnourish a baby, that's going to hurt brain development. They can still guide them into certain areas by, again to take an extreme example, making them go to music lessons every day starting at age four. But they don't have to worry that by being too strict or not strict enough or just the right amount of strict but at the wrong time they're going to seriously harm their children's adult personalities. The most dutiful helicopter parents probably wouldn't change much by plopping their kids on the couch every day and telling them not to bother them.
The evidence is pretty overwhelming. The best support comes from studies of identical twins vs. identical twins separated at birth vs. fraternal twins vs fraternal twins separated at birth. These find that about 50% of the variation in personality is genetic (actually, pretty much every study on personality seems to converge around this number) and the other half is not-genetic. But the not-genetic half has nothing to do with parenting - identical twins raised by the same parents have just as many not-gentic differences as identical twins raised apart, and the same is true of fraternal twins. So half of the difference in the way kids turn out is genetic, but the other half isn't related to parenting.
Scientists have been slow to accept these findings because they have a bunch of opposing studies that match parenting style to results. But Harris does a beautiful dissection of these studies, a dissection pretty illustrative for anyone who has too much trust in the modern scientific process. For example, studies do show that parents who adhere very meticulously to the standard parenting advice have children who, let's say, do better at school. But Harris points out - what personality trait is necessary to adhere meticulously to the latest parenting fads? Conscientiousness. What personality trait is necessary to do well at school? Conscientiousness. And what personality trait is about 50% heritable (recall that most things are about 50% heritable)? Conscientiousness. So the discovery that parents who adhere to parenting advice have children who adhere to school rules is absolutely worthless until you control for conscientiousness - after which the finding should disappear.
To take another example, studies frequently find that parents with a loving, supportive relationship with their children tend to raise happy and cooperative children, and parents with a confrontational relationship with their children tend to raise bratty, defiant children. Harris turns this on its head and says: if a child is happy and cooperative, parents will probably develop a loving and supportive relationship with them. If a child is bratty and defiant, parents will probably develop a confrontational relationship with them. This is sufficiently obvious that any study that just correlates personality and style will, again, be absolutely worthless. Figure out some way to control for this correlation and the connection between parenting style and personality again should disappear.
Harris thinks that these sorts of problem explain the much-trumpeted findings that kids from single-parent homes and children of divorce tend to turn out worse. After all, what kind of fathers abandon their partners and young children? Low conscientiousness fathers who probably have a lot of personal issues. So what kind of children would we expect them to have, just by genetics alone? Low conscientiousness children who probably have a lot of personal issues. And surprise! Children of single parent homes are low conscientiousness and have lots of personal issues! But - and here's something I had never read before - this is true only of homes that are single parent because the father left. If the father died - in a car accident, of cancer, whatever - those children turn out exactly as well as children of double-parent homes! Exactly what one would expect if the problem were caused by what the split implied about genetics and social situation rather than by the parenting itself.
It's not surprising that children don't model behavior they learn from their parents. Parents are horrible people to learn from. First of all, their role in society is completely different from that of children - if a kid sees her parent driving a car, or arguing with a teacher, that's something the kid shouldn't copy - but much of parent behavior, maybe a majority, is like that. Second, parents' interactions with their children are completely uncharacteristic of any other interaction they should expect to encounter; imagine learning social politics from a parent who ends all her arguments with "because I said so", or social norms from a parent who lets her kid get away with things because she's "so cute".
Instead, Harris thinks that children are mostly socialized by other children. That's why children want, let's say, baseball cards and Pokemon even if their parents collect stamps; more importantly, it's why immigrant children usually grow up speaking most naturally and fluidly the language that they learn in their peer groups rather than the language they learn at home. She backs this up with anthropology, primatology, and evolutionary psychology - in most hunter-gatherer tribes, most chimp bands, and most societies before the Industrial Revolution, parents pretty much just threw their children at the other children in the tribe after age three or so and didn't interact with them much besides feeding them and giving them a place to sleep. The children spent most of their time in mixed-age playgroups that did most of the heavy lifting of socializing them.
In fact, until about 1900, this idea that parents were responsible for raising their children didn't really exist. This bothers me. At this point it's easy for me to believe that things we take for granted in our society are culturally conditioned and may not be true for some godforsaken tribe in the mountains of New Guinea, but to have them be younger than my great-grandmother and still have me think they're the natural state of the human condition is pretty atrocious. I guess all those conservative bloggers are right when they say you've got to read old books or else you won't even realize how trapped in a modern worldview you are.
I'm pretty convinced by her arguments. Which is too bad, because it means our society is expending crazy amounts of effort in completely useless directions. And it also raises some bigger problems. For example, if about a hundred years worth of scientists have been wrong about something as big and as obvious as "Parenting style influences your kids' personalities", then what else is science wrong about?
Take the idea of "major calibration failures". That is, right now I think there's practically no chance that Bigfoot or the yeti exists. But if it were discovered Bigfoot really did exist, then instead of saying "Okay, you were right about Bigfoot, but obviously there's no yeti, that's just crazy", I would have to say "Wow, whatever thought processes I was using for cryptozoology seem to have been completely flawed; for all I know there might be yeti too. Or a Loch Ness monster."
If I were to learn ghosts really existed, that would be even worse - I could at least admit Bigfoot without accepting that the entire physicalist worldview was wrong. If ghosts turned out to exist, I would have to pretty much re-evaluate everything - numerology, reincarnation, God, demons - all would become relatively plausible.
So the bigger a deal I admit I was wrong about, the more I have to accept I might be wrong on a greater number of similar matters. I don't think "parents have no effect on their children's personalities" is as big a deal as "ghosts exist", but it does make me worry how much of (social) science is total bunk.
On the other hand, it's also encouraging. The typical view of scientific controversies is still pretty Galilean: there's this believe that some iconoclast points out that the orthodox establishment is wrong, and then the orthodox establishment spends the next few decades trying to grind them into dust and condemning them as stupid and evil, and their view only comes to be accepted after all the orthodox leaders are dead and a new generation has taken over. That doesn't seem to be what's happening here.
Judith Rich Harris wrote her book from a position mostly outside the field, most of the orthodox developmental psychologists shrugged and said "Huh, we never really thought about that", and although certainly not everyone has come around to her point of view her theories are being discussed widely and respectfully in the community and a new generation of students is already being taught that this is an interesting controversy. She gets her articles published in mainstream journals and apparently won some big prize for best new psychology research.
So although it doesn't look good for scientists' intelligence not to have come up with these sort of critiques before, it seems relatively complimentary to scientists' integrity and open-mindedness. And (I hope) it doesn't necessarily touch hot-button issues like climate change scientists vs. climate change deniers, or academic medicine vs. alternative medicine, because those are all situations where scientists know that people disagree with them, have read the arguments against them, but still continue believing they're right and the other side is stupid.
Overall I've raised my probability that there are important flaws with modern scientific paradigms that no one has really brought up, but decreased my probability that any particular "heretical" community that says a specific science is flawed is correct.