Quote: (10-05-2011 09:23 AM)Samseau Wrote:
Quote:Quote:
By the way, nice job clipping my quote to make it look like I said something I didn't say, especially since my post is right above yours. Was that unintentional or are you arguing in bad faith?
My bad, that's what late night posting can do to you.
No worries. Apparently late-night posting can make you give inadequately brief explanations of how twin studies work too, so allow me to elaborate.
The way you set up a twin study is you look at pairs of monozygotic twins (MZ, or identical twins) and dizygotic twins (DZ, fraternal twins). The MZ twins have all of their genetic material in common and the DZ twins only have half of their genetic material in common since they don't come from the same fertilized egg.
Now define three terms:
A = genetics--the variation attributable to directly inherited traits.
S = shared environment--variation attributable to environmental factors that the twins have in common.
E = non-shared environment--variation attributable to environmental factors that the twins don't have in common.
These three terms account for all of the variation in the trait, i.e.
A+S+E = 1.
Now note that monozygotic twins share all of their genetic traits and all of their shared environment (true by definition), so all of the difference between the identical twins is due to E, the non-shared environment. The opposite of the difference between the twins is the correlation coefficient, r. So we get
r(mz) = A + S for MZ twins.
The analysis is the same for DZ (fraternal) twins except that they only share half of their genetic material, so we get
r(dz) = 1/2A + S for DZ twins.
Subtract the second equation from the first and multiply by two:
2*[r(mz)-r(dz)] = A
So the variation in the trait attributable to genetics is twice the difference between the correlation among the identical twins and the correlation among the fraternal twins.
Next, you can get E by going back to the first equation, A+S+E = 1, and substituting r(mz)=A+S. So
E=1-r(mz).
Finally, get S using the original equation and substituting with A=2[r(mz)-r(dz)]. It comes out to
S=2[r(dz)-r(mz)].
-----------------------------------------------
Alright, that's enough math. So what they did was they measured how well your twin's sexuality predicts your own sexuality when you have an identical twin vs. a fraternal twin, then they used the above derivation to find out how much of the variation is due to genetics, environmental factors the twins have in common, and environmental factors the twins don't have in common. What the results show is that the homosexuality in males is due to about 35% genetics and 64% environmental factors that twins raised in the same household don't have in common. What are the factors that twins raised in the same household have in common?
-Parents and other relatives
-Family structure (age and number of siblings, etc.)
-Surrounding community and implicit values
-Early education
-Exposure to environmental agents, more or less
Now what are factors that they wouldn't have in common?
-Maternal-fetal hormone levels (this is a prime suspect because we have an actual mechanism for how this produces changes in the brain, and when we look at brain scans of the regions affected by these hormones they're different in gay and straight men).
-Epigenetic factors
-Maybe some environmental agents
-A limited number of experiences
-A limited amount of variation in how they're treated
-How they might respond to some environmental stimuli
You're right that it remains possible that there could nonetheless be differences in upbringing that would show up as part of the variation due to nonshared environment, but the idea behind a twin study is that you try to minimize these variations by looking at people who were raised in very similar circumstances. The real gold standard test would be to look at twins who were separated at birth and compare them to twins who were raised in the same household, but unfortunately that's a rare occurrence so we don't have very many opportunities to do studies like that.
The bottom line with this study is that if things like upbringing were really so important in determining someone's sexuality, it surely would have shown a stronger effect from shared environment. The real smoking gun is that the study
didn't show any effect due to shared environment even when looking at children whose upbringing was bound to be very similar, even if not identical for the reasons you've stated.
As for why women score higher in shared environment, I think the authors attributed it to measurement errors in assessing women's sexuality. A lot of women who identify as straight will have some degree of attraction to women while it's a lot more cut and dry for men, which makes it more difficult to place them in one category or the other for the purposes of the study.
------------------------------------------------
Also, to go back to our earlier discussion, I looked up the statistics on sexual orientation of transsexuals and it turns out that it matters whether or not the person with GID is an adult or a child. I couldn't find the primary literature source for this but it's actually about 75% of children with GID who go on to become gay if they don't get a sex change. I said the majority turn out to be attracted to women so I was mistaken about that, but my original point was that sexual orientation and gender identity aren't the same thing, which is true--25% of children with GID will be attracted to women.
Here's a paper that looks at sexual orientation in transsexual adults. You might not have access to the full text so I've uploaded it with this post. They're significantly more likely to be attracted to women than children are and it depends on whether or not they fall into the "autogynephilic" category, i.e. they're sexually aroused by the idea of being a woman.
Finally, to get back to the original point of the article,
here's some data that the hormone-suppression therapy they're putting the kid on seems to lead to good outcomes. Looks like the doctors are practicing solid, evidence-based treatment.