> it seems like undiagnosed autism is very common How could one tell how common undiagnosed autism is? All you can do, surely, is divide the number of autism d

> it seems like undiagnosed autism is very common How could one tell how common undiagnosed autism is? All you can do, surely, is divide the number of autism diagnoses by the number of - autism diagnoses. OK, so you can presume that an autism diagnosis represents some undiagnosed history of autism. But that's presumption. Or I suppose you can make other presumptions - like, this guy's diagnosed autistic, and he functions well; so many other people who function well are probably also autistic. But that's also presumption. The DSM diagnostic criteria for autism look pretty shaky to me. I'm not saying I don't think autism exists; I think it does exist. But if you want to chuck around statistics about autism, then you have to stick to some strict definition. "Undiagnosed autism" isn't strict, on any criteria. The DSM doesn't define undiagnosed anything. [Edit] I'm not at all surprised that some people are diagnosed in their middle age; 40 years ago, autism was a pretty exotic diagnosis, few doctors knew much about it or were experienced with it.