IQ differences: Nature v Nurture and the Norway Study

Full disclosure: Middle Sibling (3 of 5) here, so perhaps not the brightest candle on my family altar…My eldest sister would agree.

The recent Swedish study, widely reported , e.g., found a difference in IQ of about 3 points for one group (eldest male sibling) versus another group (third-born male sibling). The assumption is that this difference is entirely nurture because the gene pool of the siblings would be the same since by and large they have the same parents. “The findings provide “evidence that the relation between birth order and IQ score is dependent on the social rank in the family and not birth order as such,” they concluded.”

Here is one reaction quoted in the article:
Frank J. Sulloway of the Institute for Personality and Social Research at the University of California, Berkeley, welcomed what he called the Norwegians’ “elegantly designed” analysis.
“These two researchers demonstrate that how study participants were raised, not how they were born, is what actually influences their IQs,” said Sulloway, who was not part of the research team.

Does this Norwegian study support or diminish a corollary conclusion that the contribution of Nurture to IQ is about 3 IQ points and that therefore differences between two groups beyond 3 points should be attributed to genes (Nature)?

No, because the “nurture” factor between siblings is still similar. e.g. All things being equal, you are probably better off as the second child in a solid family with educated parents than the first child in a broken home with an absent father and druggie teenaged mother.

Anyway, you should be aware that there is no absolute pair of percentages that compare nature and nurture for iq differences. Because the differences in the “nurture” factor can change depending on where and when you are.

For example, if society decides to put more resources into making sure that disadvantaged children get a better upbringing, then the percentage of IQ difference that can be ascribed to “nurture” will go down. Thus, in a homogenous, socialistic, wealthy nation, one can expect that the percentage for “nurture” is lower than in other places.

For the record, here is a link to the article itself:

Explaining the Relation Between Birth Order and Intelligence

Note that the study says nothing at all about what the actual difference due to “nurture” might be. They were only interested in birth rank (order of birth) vs social rank in the family (whether the individual was raised as eldest or second eldest due to the death of older siblings). The researchers were careful to control for some other factors that might affect IQ:

From the figure caption:

If you are interested in the details of their Materials and Methods, additional information can be found here:
It should also be noted that using siblings as the basis for comparison controls not only for genetics, but to some degree for a number of “nurture” factors that could be involved in intelligence, including socioeconomic status, family environment (aside from social rank), education, nutrition, and so on. If the figures in the study have any general meaning at all, it would be that a 3 point difference is an absolute minimum for that due to “nurture,” and the actual contribution is much higher.

If you’re getting into The Bell Curve territory here, be advised that the thesis of a hereditary difference in IQ between “races” has been largely discredited by various studies. E.g., a comparison of children sired by American GIs on German women during WWII found no significant IQ difference between the children of white and black fathers; and since WWII, many developing countries have raised their average measured IQ by a whole standard deviation – not by changing their gene pool in any way, but through improved nutrition, health care and public education. You can read about all that in this book.

For that matter, AFAIK, there is no scientific evidence of any innate psychological differences between racial or ethnic groups – not IQ, not personality type, nothing.

Innate psychological differences between men and women remains an open question, but there is no proof of any difference in general IQ.

There is a significant IQ differential between African-Americans and white Americans, that much is well established and not at all controversial, but it can’t be attributed to heredity.

Related Pit thread.

This shows that the already established factoid that first-borns tend to perform slghtly better on standardized IQ tests is a result of their functioning as the eldest child in their household and the “social rank” that such a position engenders, and not due to some biological factor relating to a mother being primiparous vs multiparous. And that is all.

It in no way sets an upper cap on what is a nuture vs a nature influence. Indeed, as already pointed out, one can imagine many “nurture” influences that would have likely much greater influence on intellectual functioning than birth order status.

As a general rule performance on IQ tests (which != intelligence) is indeed highly due to genetic influences. One twin study came up with a figure that genetic influences accounted for about 80% of IQ variation by age 12. (Interestingly enough the degree of genetic influence became larger with increasing age.)

Thanks for all the comments, all.

Colibri, I am having trouble with your conclusion. I agree that using siblings controls for nearly all nurture factors, including even prenatal nutrition. This should leave genes as the only cause left for variation in IQ among siblings unless any other obvious factors are present. (The two obvious candidate ones are sex and birth/social order.) Therefore a study such as this which takes the major candidate nurture factor in siblings (social order) and assigns it a value of 3 IQ points leaves the rest of the variation for male IQ in a given family to genes, absent some other candidate nurture factor.

What is the argument behind saying this 3-point difference is the minimum? Is there another candidate for a nurture factor within a male sibling cluster?

BrainGlutton, it’s not possible to raise any nurture/nature questions without “getting into Bell Curve territory” I suppose, but my personal approach is to try and sort out the puzzle one careful piece of data at a time. While I appreciate your sentiment that genetic inheritence has been discredited as an explanation for IQ differences among separate genetic populations, that road has been traveled elsewhere, as you know, and revisiting it here is not my intention.

The differential between white Americans and black Africans is wider.


The difference is about 15 points between white Americans and black Americans, and about 30 points between whites and black Africans.

On the one hand, black Americans are better off, economically than Africans. On the other, most black Americans have white ancestors. So I’m not sure what that says about the nature/nurture debate.

Here we are talking about variation within families. But in the OP, that is not what your question addressed:

I assume by “two groups” here you are talking about any two groups, not simply within families. There are many other nurture factors that could account for differences between families and between other social groups. These include socioeconomic factors, education, culture, health, nutrition, and others.

If what might be assumed to be a rather minor “nurture” factor such as family rank is related to a 3 point difference in IQ, one might also assume that all the rest of these factors combined are likely to make a much greater contribution to IQ differences between families and between family groups.

Of course, the - entirely appropriate - reason why Science asked Sulloway to write the accompanying commentary piece is that he’s long argued that birth order has a significant effect on personality, essentially due to such environmental differences. He’s not quite a neutral observer in the debate.

Chief Pendant while I wish to avoid posting similar material both here and in the Pit thread (and that thread did come first) it must be noted that sibships, even twin sibships do not in any way control for all nuture variation. Even identical twins raised together have unique experiences that shape their individual outcomes. Genes may be important but so is culture and an individual’s “Unique Environment.”

Thanks. I did see that pit thread and posted a comment there…I did not mean to create a duplicate thread but I am not of an inclination to the kind of language and personal ridicule that are permitted in the Pit.

On the topic of IQ and twins, I note that although (identical) twin sibships do not control for nurture variation (as you point out) there is in fact a very close correlation in IQ between twins. This has been used as evidence for genetic predisposition being the predominant control for phenotypic expression of IQ.

Colibri, I am not doing a good job expressing my question, and I apologize. Within a family, nurture varies less than it does between families, but within families IQ varies quite a lot among siblings–closer to the distribution of IQ among the population to which the family belongs. To ascribe the wide distribution of IQ that exists among siblings to nurture, one needs to find what nurture differences exist within a family unit. Birth/social order has been advanced as one such nurture difference within families, and this study seems to suggest that that difference is real, but quite small (which is why the study was Pitted, I believe).

If the difference from the birth/social order nurture factor is small, then Nurture advocates must find some other Nurture factors to advance their position that IQ is not predominantly a reflection of one’s genetic inheritance. It seems to me that this study, therefore, by showing what a minor contribution one of the only possible nurture differences has, actually advances the position that it is genes and not Nurture which account for IQ differences in sibships.

What this study does not publish is the distribution range of IQ among the siblings studied. I would be interested to see that. Assuming it is a normallly wide distribution, Nurture advocates are going to have to look elsewhere for studies to support their position. For instance, if the IQ variation in sibships in this study were 15 points, you have only found two or three of them to attribute to Nurture unless you can think of some other nurture factors within sibships besides birth/social order.

It’s all smoke and mirrors. What is intelligence, and how is it measured? Is there a good definition of intelligence? Over the years I have heard this discussed in many ways. When I took an IQ test way back there, the feeling was the test held a bias for certain groups of people.

Can an uneducated person, one that can neither read of write have intelligence, and how can it be measured?

CP I had linked to an example of the twin studies, so yes, I am aware of them. Again, those studies show that by adulthood about 80% of the variation in IQ’s is attributable to genes and that about 20% have an environmental factor (at least within one socioeconomic or cultural group). This study shows that of that 20%, the portion that is related to birth order is a function of social ranking and not due to some biological effect like paternal or maternal age or parity status. The twin study was cited to try to answer dircetly your op re how much of IQ should be attributable to genes vs environment as the study that you cited does not actually hav any relevance to that question.

It may be of interest to you how twin studies work. The premise is to generally to compare the concordance of a trait between monozygotic and dizygotic twins. Both twin sets share the “Shared” or family environment to a very similar degree and differ in their “Unique” or non-family environment to similar degrees. Therefore any difference in concordance rates for any particular trait can be largely attributable to a genetic component either directly or indirectly. Another strategy is to compare twins adopted into different families but that provides fewer subjects. And a related technique is to study sibs adopted into different familes to sibs raised together for trait concordance. Of the non-genetic contribution to IQ family of origin is a trivial factor but place in the family and outside of family factors (peer groups, etc.) are of more significance.

You are, btw, a bit mistaken about the variation of IQ within families being close to the variation of IQ within the population as a whole. IQ variation between biologic sibs is generally much less than IQ variation of the population at large whereas IQ variation between siblings in a family formed by adoption of one or both sibs tends to be very similar to that of the general population. As an example I can describe to you a real family with three biological children all of whom had IQ testing. All three had IQs in the 130 to 140 range (which means clustered within 2% of the general population) with a 10 to 15 point discrepency between verbal and non-verbal subsets. Their younger adopted sibling has not been tested but clearly does not have the same discrepency between skill sets (suspicion of which had led to the elder sibs’ evaluations in the first place). This familial clustering is more the norm than the exception.

lekatt, performance on an IQ test does not equal a measure of intelligence (other than by those who define intelligence as performance on an IQ test) and it is very accurate to point out that we have no agreed upon definition of intelligence*. Nevertheless an uneducated individual who can neither read or write can have his/her IQ evaluated using the right instrument. Questions can be read to a subject and the nonverbal portion can be taken with little explanation needed. Hell there are IQ tests for infants (such as the Bayley and the Mullens tests) which can give a sense of intellectual function at that point in time.

*To me intelligence means the ability to solve salient novel problems. Thus a whale is highly intelligent at dealing with problems relating to keeping track of objects and individuals in a large volume of space, which humans cannot do well at all, but may be very unintelligent at solving problems salient to humans. But that is a whole 'nuther thread.

To DSeid

I liked your post, says it well.

People place far to much importance on tests. They tell very little about the person. A very interesting piece on Idiot Savants is at this link, kind of messes with mainstream thought about intelligence.

I know some will not want to go read, so I will quote a small paragraph for fun.

Socrates was right when he said wisdom is knowing that you know nothing. Something like that anyway.

Thanks for the post. I am aware of the reasonably high correlation of full-sibling IQs (0.6 or so) and I realize that they vary much less than the “general population.” By “the population to which the family belongs” I meant their actual genetic pool and not the broader general population which might be quite a mix.

Lekatt, I understand your reluctance to place to great a weight on IQ as a measurement of intelligence. Certainly we have all met our share of high-IQ nitwits. Nevertheless, on average there does exist the ability to test for an assortment of types of intelligence, and with g in particular, you are probably better off if your pilot and neurosurgeon are capable of testing higher rather than lower. On average those who have cured disease, designed rocketships, engineered skyscrapers and built bridges did not come from a random cross-section of IQ.

I think there are different facets to intelligence, which the different types of questions of the tests try to measure. You could have three different people making the same high score on a test…say 130, and yet each will have skipped different questions, or given incorrect answers on them. Some people are better than others at questions on numeric sequences, others are better at word or shape associations, and still others are better at finding hidden kangaroos in pictures.

IQ tests can be administered orally, and an illiterate person should, theoretically, be able to shape questions. On the other hand, it’s possible that even there they would be at a disadvantage; if your culture is at the hunter-gatherer stage, then you won’t have any familiar with the neat geometric shapes that are used in these questions, like exact squares and triangles.

Does the study mean that siblings are almost always within a handful of IQ points of each other? If so, it powerfully supports the common wisdom that people tend to find mates of like intelligence. If parents of greatly differing intelligence randomly pass on their genes to their kids, then we would expect to see wider variations in sibling IQ, reflective of the differences between the parents.

No it doesn’t mean that.

Did the study exclude families where the parents had significantly different IQs?