From: Rafal Smigrodzki (rafal@smigrodzki.org)
Date: Fri Jul 25 2003 - 13:46:25 MDT
Mez wrote:
> I'm interested in feedback on this line of reasoning:
>
> Imagine a trait which has a 0.5 heritability. That is to say that
> there's a 0.5 correlation between the trait and one's genes.
> Specifically, let's use IQ. Estimates of the genetic component of IQ
> vary, but 0.5 is not an unreasonable number.
>
> Now let's imagine you wish to bring into this world children with high
> IQs. Using the DNA of someone known to have a high IQ, you create a
> number of clones. In this case, let's pick Albert Einstein as the DNA
> donor. And let's posit that his IQ was 160.
>
> Alternately, if science has identified every IQ-affecting gene (hah!),
> you may start with any human embryos and modify their IQ-related genes
> to match Einstein's.
>
> Here are my suppositions about this thought experiment:
>
> 1) We do not know the relative contribution of genes and environment
> to Einstein's high IQ. In the absence of such specific knowledge, our
> best choice is to use the statistical contribution of genes and
> environment across the population. The 0.5 heritability in the
> population would suggest that half of Einstein's deviation from the
> norm in IQ was due to his deviation from typical human genes, and the
> other half was due to environmental effects.
>
> 2) If we created many children who shared Einstein's IQ-related genes
> and raised them in average environments (or a set of environments with
> a distribution mirroring the social average), then we'd expect the
> average IQ of such children to be 130.
### Well, not necessarily - if the genetic contribution to our particular
Einstein's IQ minus average IQ was different from .5, then the mean of the
clones' IQ would differ from 130. Say, he was a really, really genetically
smart guy, but his particular environment stimulated his development less
than average, then you could get a population of clones with average IQ 150.
It's a bit tricky to apply regression to the mean to a single person,
although if you apply it to a population of IQ 160 donors, you indeed get
the results you describe.
-----------------------------
>
> 3) The distribution of IQs of these children would form a normal bell
> curve centered on 130.
### A normal distribution, yes.
---------------------------
>
> 4) The standard deviation of their IQs would be 10 points. (I'm
> particularly interested in feedback on this point. Would the standard
> deviation remain 10 points given that the genetic contributor to IQ is
> fixed?)
### The SD of most IQ tests (WAIS) is usually 15 with a mean of 100. The SD
would be reduced because of removal of the genetic component of variance.
-----------------------------
>
> 5) As an extension of point 3, as many of these children would have
> IQs of 100 (dead average) as would have IQs of 160 (Einstein)
>
### Yes, with qualifications as in 1).
----------------------------------
> Point 5, to me, is a pretty unexpected conclusion. But it seems be an
> inevitable consequence of a 0.5 heritability of IQ, and the guess that
> Einstein's deviation from the norm in particular was 50% accounted for
> by genetics and 50% by environment.
>
### Buying gametes from an IQ 160 donor will give your offspring only about
11 IQ points over average, they say. Yes, IQ is a finicky trait.
Rafal
This archive was generated by hypermail 2.1.5 : Fri Jul 25 2003 - 10:55:31 MDT