The problem is that we can't argue with Cait, because it is impossible to find a double-blind study of children raised without gender cues to see how they were treated. Even infants are reactive to their environments, and if we can't control the environment, we can't prove anything.
I wish I could remember the author of the study, but I read within the past few months about a particular study that explored how different people treat infants - even newborns! - based on presupposed gender. I know there have been a number of these, but tihs one was particularly good about counting waiting time to respond to cries, face contact OFFERED by the adult, adn so on.
All the studies in the world that show that baby girls do X and baby boys do Y don't mean a thing if we are working from the assumption that people are incapable of treating known baby girls and known baby boys the same. Gender socialization happens the minute you are born! No one is going to win a nature vs. nurture argument, given that constraint.
Given the existence of hard-wired biological expressions, including natal sex, sexual orientation, and gender identity, I would find it profoundly bizarre if there wasn't SOME component of gender expression that had hard-wired roots. But I certainly can't prove that, and I don't believe that ANY test can be created to prove it (due both to gender as a social construct AND the statistical outlier problem). Beyond that, I highly suspect that the individual variance would be so large that proving an actual, hard-wired difference in gender expression or roles between the 'average' male and the 'average' female would be virtually meaningless in telling us something about any individual in the data sample.