The New York Times has an interesting article by Emily Bazelon on that topic. I have a few reservations about it, though. First, whenever someone has a plan to write a piece on how the sexes differ, well, the piece tends to focus on that and not on how the sexes are the same. In most cases, we are not given any information about the relative magnitudes of the differences vis-a-vis the similarities. Because nobody ever writes articles on how the sexes don't differ, we mostly go away with the idea that the differences are humongous.
Second, a very common aspect of much of the writing on sex roles and so on is the implicit assumption that the sex roles in the American society are the natural sex roles. It isn't just the Americans who do this, but because American writing in this area is predominant, a lot of the psychological and sociological research in this area has to do with American college students (because so many studies use college students as cheap subjects). I sensed this bias in the autism piece in the sense that what was assumed to be "natural" for girls is what the American culture regards as natural.
Third, the article mentions Simon Baron-Cohen as an expert on gender and autism, but fails to mention that much of his work is controversial. I have written about the problems with his test for systematizing vs. empathizing behavior earlier on this blog, and I have also written about the fairly obvious gender bias in his book The Essential Difference .
But mostly I have trouble with the basic argument of the piece which is that girls suffer more than boys from autism because social interaction is more important for girls than boys, and because autism makes social interaction so difficult. Yet, at the same time the article quotes one expert as saying that perhaps we see fewer high-functioning autistic girls because girls are better at social interaction and that this serves to hide them from the autistic label. The article tries to explain how both of these arguments could be true at the same time:
But based on their clinical experience, Lainhart and also Skuse see autism as a heterogeneous disorder. Its profile may change and expand as more is understood about girls, whose autism, they worry, often goes undiagnosed. That is partly, Skuse posits, because girls' general aptitude for communication and their social competence helps some Asperger's girls "pass" — they pick up on their difference and carefully mask it by mimicking other girls' speech and manner and dress. In a sense, their femaleness allows some girls to seem less autistic. It is as if they start off with a social advantage — Skuse sees this as a 20-point bonus on a scale of 100 — that helps counter the disorder. This idea isn't necessarily at odds with the findings that show girls to be more seriously affected by autism, Skuse says, because the girls who succeed in masking their deficit wouldn't be included in studies. And so they are missing from the picture. "There is no doubt in my mind that the way we have defined autism currently biases our assessments strongly in the direction of identifying a male stereotype," he says. The C.D.C. agrees and says that as a result the estimate for the number of girls with autism and normal intelligence may be low.
I'm not convinced. It's as if we are saying that autism hurts girls because they can't satisfy the cultural expectations of being good at social interactions, but on the other hand girls are good at social interactions which hides their autism. I guess it could be true, but shouldn't we then find the average level of autism in the girls who are diagnosed to be much, much higher than the average level in boys?