For those interested in studies of identical twins reared apart, see IQ Similarity of Twins Reared Apart: findings and responses to critics, Bouchard (1997). A question sometimes raised by critics concerns the extent to which twins reared apart actually experienced different environments. Norms of reaction, GxE interactions, and other (conceptually trivial but potentially obfuscatory) topics are also discussed.
Susan Farber wrote a 1981 book that was critical of twin studies. Click below for larger image.
See also Heritability 2.0.
Readers see also:
ReplyDeleteMore Behavioral Genetic Facts | JayMan's Blog
An old but very interesting study on twin IQ:
ReplyDeleteThe effects of chorion type on variation in IQ in the NCPP twin population
The main takeaway is that differences in monozygotic twin IQ are mostly due to prenatal environment.
The link is broken (fails outright, and the apparent embedded link leads to Bouchard's chapter).
ReplyDeleteI assume this is it? http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1685620/
Thanks for the pointer.
Thanks, that's the paper. Fixed the link above. It now leads to a pdf of the paper.
ReplyDeleteDoing some back of the envelope calculations, this paper on IQ correlation and chorionicity combined with other sibling IQ studies seems to indicate that even among brothers raised apart, pre-natal environment accounts for about 40% of the environment/noise component. This is a neglected area in psychometrics. That 40% is probably a lower bound.
Here goes. I'm sure you'll let me know if I made any untenable assumptions or glaring errors.
ReplyDeletecorr(IQ of MC twin 1, IQ of MC twin 2) = cov( MC twin 1, MC twin 2)/(sd1 * sd 2)
sd1=sd2= the population standard deviation so we'll set it to 1 and ignore it for the remaining calculations.
IQn = Gn+Pn+En
G = genes, P = pre-natal environment, E = post-natal environment
I ignore GxE and other such interactions. Wrong, but I had to start somewhere.
In that case corr_MC(IQ1,IQ2) = cov(G1,G2) + cov(P1,P2) + cov(E1,E2) = cov(G1,G1) + cov(P1,P1) + cov(E1,E2) = 0.9 i.e. assume genes and pre-natal environment are identical for MC twins.
Now look at dichorionic twins. We get a very similar equation except that the Ps are not equal.
corr_DC(IQ1,IQ2) = cov(G1,G1) + cov(P1,P2) + cov(E1,E2) = 0.7
The difference between these two equations is 0.2 and represents the difference in Pearson's correlation due to having different placenta.
Next we look at fraternal twins vs non-fraternal siblings . In the case of fraternal twins, they share a womb, but not a placenta. Non-fraternal siblings do not share a womb, at least not at the same time, so their pre-natal environments differ more than you would expect if they were in the womb at the same time.
corr(fraternal twins raised together) = 0.55
corr(non-fraternal siblings raised together) = 0.47
corr(fraternal twins raised apart) = 0.35
corr(non-fraternal siblings raised apart) = 0.24
From these correlations we see that the difference in correlation caused by being in the same woman's womb at different times is about 0.1. This assumes that cov(E1,E2) is the same for fraternal twins and non-fraternal twins.
Adding this to the 0.2 we got from having different placenta within the same womb, we get a 0.3 contribution from the pre-natal environment.
The portion of non-fraternal siblings raised apart IQ that is not explained by genes is 0.76 (1-0.24). I assume that 0.3 of this difference is due to the pre-natal factors above. So 0.3/0.76 = 0.39
I want to go back through and see what GxE effects do to this. The fact that the correlation in IQs for twins raised together vs. apart drops less than the correlation for siblings raised together vs. apart does give some indication that GxE effects are important and gives one estimate of its magnitude.
Their similarity can be mostly due to prenatal environment too though, because prenatal effects the genes directly.
ReplyDeleteHave you seen the Aguti mice experiment? Clone mothers with clone babies in the womb were fed different food during prenatal and the babies turned out different enough to be considered separate sub-species. They passed it on through generations too.
http://discovermagazine.com/~/media/Images/Issues/2006/Oct/agoutimice425.jpg