The AQ test is useful, but you have to remember it's not diagnostic; just a device you can use to see if you're looking in the right direction. The thing about autism is that a lot of its traits occur in the non-autistic; only in autistic people, they're very strong and prominent. The distinction between "weird by typical" and "autistic" can get awfully small. You could actually get a low score on the AQ test and still be autistic, if you had a small number of very strong traits; or else get a high score and not be autistic because you have a large number of very weak traits. (For the record, the test picked me out fine; I scored 38.) A neuropsychologist specializing in adult autism is the best person to go to for a diagnosis; many child psychologists are used to its expression in children and many adult psychologists aren't familiar with "childhood disorders" at all.

As far as anybody can tell, the autism spectrum actually isn't a discrete category at all; you just draw the line between typical and some-kind-of-autistic at the point where it starts to cause some kind of impairment. But on the other side of that line are people who are very close to autistic; who have a lot of autistic traits, but who don't have "significant impairment". I think there are a lot more people with "autistic traits" than with any kind of autism; actually, I'm pretty sure that's where the socially-clumsy nerd stereotype came from. Autism, like many neurological diagnoses, is an extreme expression of traits you find in typical people too. I wouldn't be surprised if maybe one-in-twenty people or even more have identifiable autistic traits that aren't diagnosable as anything but personality quirks.

This makes those of us on the autism spectrum feel a little more connected to rest of the world!