Wow, just finished reading this entire thread... My 2 cents:
I agree with Cordelia Fine to the extent that there is a lot of neurosexism and biases in which data to present, making these studies completely worthless for actually saying, "Girls are like X, boys are like Y."
In my opinion, it is entirely social gender roles that dictate the differences between people based on what's between their legs. (This isn't counting transfolks, who don't identify with what's between their legs. I don't know enough about genderqueer vs. trans to comment intelligently here.)
I read Fine's book about a year and a half ago, and I was kind of irked that there was no mention of trans traits.
How I would say it is as follows: gender is socially imprinted on people via stereotypes. Trying to get every trans person to act out the stereotypes of their preferred gender is ridiculous, because cisgendered people don't fit all those boxes perfectly either.
A personal gripe of mine is that I read somewhere that, in Australia at least, you need to explain to a psychiatrist why you are the opposite gender without relying on "I want to be." But if it's all socially defined, then the only way to do so is to shoehorn yourself into stereotypes. (This was touched on by someone else on a previous page.)
Honestly? I think the only differences caused by estrogen vs. testosterone are bodily - not on the mysterious realm of mental processes.
There's a problem in philosophy in that we don't know was consciousness really is. There's nowhere in the brain where we can point and go, "Look, it's a sentient being!" Mind does not equal brain is a common explanation of this. The problem then becomes how you define this nonphysical mind... As yet, there have been no reasonable, let alone testable, theories, for how to define the mind if it is not just the brain, but the corollary is that there have also been no reasonable, testable theories to pinpoint somewhere in the brain for consciousness.
It's a personal interest of mine to come up with an answer for that, but that's a very hard thing to do.
But if we can't even define somewhere in the mind that is our conscious thoughts and sentient observances, then how can any brain scan, regardless of which scan it is, or what software you use, or what statistics model there is available, how can that tell you anything about "men vs. women"?
If we were all just a bag of chemicals, then more testosterone would make people act like stereotypical men (if testosterone had any bearing on that, of course) and likewise for estrogen and stereotypical women. Because we don't satisfy the end result (I sure don't, and I'm part of the "human" subset) then the initial cause is inaccurate. It's a case of A->B. A always leads to B - that's the theory. So if you don't have B, then you don't have A either. (Note that if you did have B, it wouldn't necessarily mean A either, as you could have C->B as well.) But without B (which we don't have) then A (being a bag of chemicals only) is wrong. Hence, sentient decision-making.
And like I said, we can't pinpoint that in the brain, so any theories trying to say what gender means will only ever be accurate by pure coincidence, with no way of knowing which one is accurate, if any, because a part of gender must be within sentience.
It's an unsolvable mathematical equation at present. And even if we could pinpoint sentience, there'd be a hell of a battle in trying to say what those chemicals actually do - my hunch is that it'd all be probabilities. I mean, if you had a 100% part, then the interlocking nature of things would presumably mean that every other aspect would come down to either 100% or 0% - drawing us back to "2 people with same chemical levels would be identical" - in theory.
Okay, I'm rambling... But yeah, these are the things I've been thinking about while reading this thread.