Thursday, March 06, 2008

Getting some Flak over Nature and Nurture

Over the last few months, some correspondents have been unsympathetic towards my contentions that nurture, upbringing, schooling and other environmental factors appear to have much less effect on how children develop than might be expected. I accept that this is a radical idea, but it is also one where the evidence is quite conclusive. But perhaps I have been remiss in presenting that evidence.

A few posts ago, I mentioned studies on the Chicago state school system where it was found that there was no direct correlation between an individual’s results and how ‘good’ the school they went to was. As explained in Freakonomics, that is not quite the full story. The system in Chicago is that you can either accept the local school or, if you are unhappy with it, you can enter the lottery. The academics found that children whose parents entered them into the lottery did better than those who just went to the school to which they were originally allocated – even after allowing for the fact that children entered for the lottery might end up at a better school than otherwise. They also found that middle class parents were much more likely to enter their children for the lottery than just accept the allocation. It was a classic case of middle class children doing better wherever they ended up at school.

So was it the middle class upbringing that led to children performing better or the middle class genes? The academics dug deeper and found that there was not a single factor in the children’s upbringing (whether they were read to at bed time, or there were plenty of books in the house etc etc) that correlated to exam results. It looked like it was genetic.

Here’s another example, taken from Steven Pinker’s The Blank Slate. It is uncontroversial that the children brought up by single parents have worse school results, worse outcomes and are more likely to be single parents themselves. Traditionally, this has been assumed to be due to the lack of a father-figure, poverty or the stress involved in living in a broken family. In other words, it is a classic case of how nurture effects the way people turn out. But when you split the figures up between families where the husband has absconded and where he has died, you get different results. Children brought up by widows do not have worse outcomes than in general and are as likely to stay married. Children brought up by divorced or never-married mothers are more likely to be divorced themselves, do less well at school and have lower lifetime outcomes.

How can you explain this? It is certainly not stress. Upsetting as it is for a father to leave home, it is nothing compared to a bereavement. It isn’t poverty either. Widows are not richer than divorcees. The only correlating factor appears to be the parents themselves, in which case the differences must be caused by genes.

This is not just an academic question. All our social and education policies are based on the idea that nurture matters and is something we can change. It is generally agreed that our policies are not working as they should. I would suggest that much of the reason for this is that they are based on an axiom that is untrue.

Click here to read the first chapter of God's Philosophers: How the Medieval World Laid the Foundations of Modern Science absolutely free.


Elliot said...

Interesting! But in that second-last paragraph, is it necessary to jump straight to genes? It's sounds a little like the God of the Gaps - the Genes-of-the-Gaps - if the social sciences haven't figured out some complex social reality, then it must be due to the genes! In this case, white-trash-divorce-getting genes! (And in the American context, you'd start hearing all sorts of sketchy arguments about race, too.)

Just playing devil's advocate.

Unknown said...

Weird, I read ^ his blog too.

In any case, I think you're ignoring the key difference between widows and divorcees/abandoned single parents in society's view. A widow(er) is never blamed for the loss of her(his) partner, but a divorced person or a single parent is immediately hit with the stigma of "What did you do wrong?" The parent who left is considered to be even worse. In addition, when a parent dies, the child doesn't blame the parent in most situations (suicide excepted, and I wouldn't be surprised if the statistics for that kind of child are even worse), and certainly doesn't blame himself or the living parent in many situations either. In divorce or abandonment, however, the child can begin to hate the one who left or hate himself for being somehow responsible - and this sense of victimization is anathema to the properly understood personal responsibility that results in performing well (what do we mean here, anyway? test scores? college admissions?). This is a moral distinction - divorce and abandonment are wrong, societally or religiously speaking; dying, on the other hand, is not blameworthy.

Bereavement is more immediately stressful, and I've certainly seen kids act out, but it doesn't compare to the long-term problems that come from a broken home (again, a term never applied to widow's families). Heck, the dead parent can even become idealized, and serve as a stimulus for the child to do well.

I'd be curious to see how adoption plays out. I'd figure it's somewhere in between, providing the child was adopted early on as opposed to stuck in foster homes for a long period.

Since there's an obvious and relevant societal cause here, I don't see how it can be genes. If you want to argue that people with certain genetic dispositions are less likely to remain committed, that's possible. But the psychological effect on the children seems to me to come from a societal understanding that the two situations are morally distinct.

James said...

Michael and Elliot,

Thanks for the posts. Michael, your theory isn't convincing to me. I know lots of divorced people and I have never heard them complain they are blamed by anyone other than their ex-partners. No one ever blames the children and society blames the absent men, not the single mothers who are treated as victims.

I agree that more work needs doing but genes are in pole position right now.

Have a look at this article by his Pinkerness himself which deals with several of the objections I've been trying to handle here:

James said...

Here's a link that hopefully works:

Why Nature and Nurture Won't Go Away

Unknown said...

Thank you for the Pinker article - the last half or so was very interesting. I especially liked the distinction he made between "human nature" and genes - they're too often treated the same, yet clearly aren't.

I know many divorced people as well, and they certainly feel, as you say, like victims (which would cause problems not extant in widows) but with it the concomitant problem that they feel less worthy than their steadily married counterparts - and even worse if they have kids. This may be a difference between America and Britain, however, and I don't know whether the Freakonomics statistics hold in Europe. There are all sorts of problems with basing these arguments solely on anecdotal evidence, however, so that really only gives a reason why I believe what I do, not why you should.

Even if my reasons don't hold up, I still take issue with the same gap idea Elliot had - it could be some unknown gene, but it could also be some unknown social factor. I'm not denying necessarily that it might be genes, only that I'd need to see a mechanism first. After all, Pinker's examples in his article show that there's a large degree of variance among what is nurture and what is nature and what is mixed - but in every case, he has a reason for why it is what it is, rather than an assumption based on what it's known not to be. That was perhaps the best thing about the article - the push for more testing, more study, and more refinement.

Niall said...

While I'd certainly accept that genes play a very important part in determining social outcomes, I find the argument James presents pretty flimsy.

It is a mistake to assume that genes are responsible for different outcomes simply because those environmental factors typically assumed to play an important role were not show to account for the different outcomes experienced by the participants. However the main problem I have is that the above argument ignores the many, many studies that show the importance of environmental factors.

Just as an example, take expectations. In one experiment, teachers were tricked into believing that certain students were "late bloomers" and that their academic results should improve. The experimenters found that the academic performance of these students improved significantly overtime in comparison to their peers. This was in spite of the fact that the students labeled "late bloomers" were selected at random.

Teachers usually know about the backgrounds their students come from. It's probable that it in most cases, teachers have higher expectations of those from working class areas and/or difficult family settings so in effect, they can create self-fulfilling prophecies.

I'm not suggesting that a teacher's expectations are the most important determinant of a student's academic outcomes, I'm just using it as an example to highlight the potential effects of environmental factors.

Anonymous said...

The piece that's missing in this discussion is something I pointed out last time the topic came up here.

To even begin to understand this one has to be solidly grounded in the field of Attachment Theory (the field of psychology that studies relationships).

The science of relationships can address most of the 'gaps' that are identified in the original post and has much more explanative power (not to mention having much more actual evidence behind it) than the genetic answer.

Just a tiny bit of background and a few examples to illustrate...

Broadly speaking, the development of an individual (socially, morally and intellectually) is profoundly impacted by the level of secure, deep attachments a child has to their adult caregiver(s).

Equally important is the understanding that attachment is polar in nature. This means that in relationships, an individual will choose (usually unconsciously) between ‘competing’ attachments and be drawn towards one, while at the same time pushing away from the competitors (think in terms of magnetism).

The classic example is the young person who finds a new love interest and begins eschewing past friendships. This is a totally natural thing that is occurring in the brain(s) of the pursuer.

If the friendships are at all detected to be competing, the polar nature of the attachment function in humans will begin pushing away from them in order to pursue the new attachment. To the extent that they are not competing, this will not happen.

This is a defense mechanism that is built in. Because those to whom we are attached exert such a powerful influence over how we behave (i.e. we want to be like, to dress like, act like, share the values, etc. those to whom we are attached), individuals who are seen as a threat to our primary attachment are pushed away so that their influence will be decreased.

Now, apply this to the case of divorce vs. loss of a parent to death.

The major difference here is that in the case of death of a spouse, despite the profundity of the loss, no competing attachment is set up with the remaining parent.

As long as the remaining parent can recover to the extent that they can allow the child to have a deep loving attachment to them, in terms of the relationship and its effect on their development, they will be far less adversely affected than the child or divorce.

In the case of divorce, the primary attachment unit (the parents) is cleaved in two and the child is left with two competing attachments.

The attachments don’t even have to be consciously competing (i.e. the mother and father working towards different ends), the mere fact that the one unit (parents) are now two (mom over here, dad over there) sets up the competition.

This state will make it MUCH more difficult for the child to establish a deep, secure attachment to either as they are forced to move back and forth between the two.

This causes several defense mechanisms to engage in order to protect the child from being as vulnerable to the same type of trauma. These defenses will generally make it more difficult for the child to attach to anyone (including the parents).

All of this does tremendous harm to the child’s development.

Another example involves the always sensitive topic (among parents) of how to deal with children and sleep…

Over the past few years, the evidence has become overwhelming that allowing a child to cry unattended for prolonged periods significantly raises cortisol levels in the brain.

It’s also been well established that heightened levels of cortisol actually cause brain growth to be stunted in comparison to children who are not left to cry.

The studies actually refer to this as ‘brain damage’.

So in this case, you have a choice of parenting styles that have a profound impact on what’s happening physically, which in turn has a profound impact on intellectual capacity (among other things).

Same with breastfeeding…which has been shown to have beneficial impact on intelligence (children who are breastfed are less sick, and have higher IQ’s).

I could go on and on with these types of examples, but the point is that the parenting choices that are made do indeed have a profound impact on the development of the child.

There’s really no debate on this any longer. When it comes to the question of nature vs. nurture the answer is clearly…BOTH.

Anyone who discusses this topic (nature vs. nurture) as it relates to development of children is missing a huge piece of the puzzle if they are not at least familiar with attachment theory.

Weekend Fisher said...


I'm really glad that you've never heard that divorced parents have any blame factor going. But speaking as someone who has been divorced: I get it all the time. When someone finds our you're a single parent, the dirty looks abound; for years my son blamed himself even though he was 2.5 years old at the time my husband decided to leave; just the other day I was trying to comfort a friend of mine whose daughter is getting a divorce, only to have the first words out of her mouth be that usually both people are to blame. (I hope she doesn't try that line on her daughter; I doubt she'd find it particularly supportive.) I was once at my mother's when she took a phone call, and the person on the other end of the line obviously asked after her children. She went on at some length about my brother, but didn't mention me.

The blame/fault/guilt/shame thing is always an issue. Please don't underestimate the effects of living with the stigma. It's real.

Take care & God bless
Anne / WF

Anonymous said...

On the Pinker article linked....

I can see why Michael says the second half is interesting...based on the fact that I couldn't go further than halfway. His treatment is so clumsy by that point that I began to lose interest...

In trying to argue that in some cases it's not both, but purely one or the other, he provides some very poor examples.

First, speaking of language, he says ...

Not true. Why do people in England speak English and people in Japan speak Japanese? The ‘reasonable compromise’ would be that the people in England have genes that make it easier to learn English and the people in Japan have genes that make it easier to learn Japanese, but that both groups must be exposed to a language to acquire it at all...

Which is just plain silly. Is anyone REALLY suggesting such a thing? Of course not. What's suggested by saying BOTH is what he points out next...

Though people may be genetically predisposed to learn language, they are not genetically predisposed,even in part, to learn a particular language;the explanation for why people in different countries speak differently is 100 percent environmental.

So he's tacked on the 'particular language' caveat to get around the obvious answer of BOTH, but in the bold part he's already conceded that it's both to the fundamental question. Both nature and nurture answer the how and why of language acquisition (regardless of the particular language).

Later in trying to cite an example at the other extreme(genetics only), he uses autism and schizophrenia.

In the case of autism, while the hereditary predisposition is clear, the cause is as yet undetermined and the ruling out of environmental factors has not been established.

With schizophrenia, the usage is even more inappropriate. Several social factors are known to significantly increase the incidence of schizophrenia.

With his examples of the extremes being so wrong, I found it difficult to keep reading.

James said...

Anne, I’m sorry it has been so tough for you. Clearly some people do attach stigmas to divorced people.

On the other hand, apparently divorcees are generally happier than widows and I would suggest that this is a quite important point. It means that outcomes are better for those brought up in the generally less happy families. That is the opposite of what we might expect if nurture had a big effect. Again, more testing is needed, but the evidence is all pointing towards genes, not environment.

Steve, attachment theory is exactly the sort of thing that the work referenced by Pinker has disproved. We find no correlations at all once we factor in genetics. Attachment theory all sounds convincing enough but when we test it we get zilch. It’s a good example of why common sense can be the enemy of science. Breast feeding is an exception but that is a dietary rather than nurturing question.

Clearly, I’m not convincing anyone. That’s OK. I think we’ll see attitudes shifting slowly over the next decade or two as the evidence piles up and the political implications sing in. I’ll turn to one important policy debate in my next post.

Anonymous said...

Steve, attachment theory is exactly the sort of thing that the work referenced by Pinker has disproved. We find no correlations at all once we factor in genetics. Attachment theory all sounds convincing enough but when we test it we get zilch.

Could you elaborate on that because I am just not seeing it. Maybe provide an example of where he does this specifically?

Beyond that I think your supposing that in the field of psychology Attachment Theory is still ‘just’ theory.

It’s really not. At this point its theory in the same sense that evolutionary theory is theory.

Attachment Theory is not a ‘touchy feely’ set of hypothesis based on common sense. It’s a well developed, well establish (and yes, well tested) set of positions that is broadly accepted.

The testing far from providing zilch is confirming it. I gave a specific example that’s been well established regarding cortisol levels and brain development, and the effect of ‘cry it out’ methods of parenting have.

The parent’s choice of how to raise a child with regard to sleep (responding or not responding) has a real physical impact on brain development.

So you have a situation where a child has a potential intellectual capacity which is most likely genetically determined, but the potential is profoundly impacted by the choice of how to parent. That’s nature and nurture.

It’s a good example of why common sense can be the enemy of science.

I mean this very respectfully, but I think you are simply mistaken here, and I think that’s based on a lack of knowledge of all the disciplines involved in the equation. You are far too quickly dismissing Attachment Theory.

Breast feeding is an exception but that is a dietary rather than nurturing question.

It’s neither the exception (I’ve given another example above, and can provide others), nor is it an exclusively dietary factor.

The benefits to brain development are based on dietary factors…but you are not recognizing the fact that the choice to breastfeed is a question of nurture. You really can’t get around that.

By saying both, I am not saying that breastfeeding is ‘purely’ a nurture issue; I am saying they go hand in hand.

Of course we’d expect the impact of the choice in parenting to have a physical component, but the two are inextricably tied together, as is the case with increased cortisol levels.