First, I want to express appreciation for the author's sensitive, perceptive, and courageous account of her family's experience. Nothing that any study says can change the essential horrifying and heartbreaking facts described in this article.

That said, I agree with the poster above who suggested that we need more data and less anecdote if we want to understand this disorder. The problem is that the data on autism and violence isn't so easy to come by. That's not - as this article seems to imply - because of some tacit conspiracy amongst parents and/or researches to candy-coat the disorder. I suspect that was a bit of projection on the author's part. There is actually a lot of motivation to understand the relationship between autism and violence in the mental health community, where many of these individuals end up needing lifetime interventions around this issue.

The problem is that researchers then have a pool from which to draw study subjects that is not necessarily representative of the disorder itself. This is for a few reasons. First, the disorder was not diagnosed as frequently or with the same criteria pretty recently, and interventions have changed drastically over the years. So statistics about the entire adult population of people on the autistic spectrum mix a whole lot of apples and oranges. Second, the adult population that researchers are most likely to come into contact with are those that are in need of the most psychiatric intervention; and studies do seem to show that increased rates of violence in adults with autism spectrum disorders may be due to underdiagnosed comorbitiy (= other psychiatric conditions, such as depression or bi-polar disorder). Lastly, this population is already being treated with a number of psychotropic medications, at least some of which (neuroleptics) are associated with increased aggressive behavior in adults with autism when administered in the wrong dosage.

All of this means that we can't really take one good reliable survey of adults with autism spectrum disorders and conclude from there whether that problem is due to the biology of autism, due to emotional or social factors, due to effects of the treatment itself, etc.

That being said, the best study I found was on adult incarceration records of people with autism spectrum disorders in Denmark. The study showed that a small but significant fraction (less than %10, more than the "normal" population) of adults with autism had criminal records, and a larger number (but still less than %20) of adults with Asperger's syndrome had criminal records. But, again, that doesn't tell us why, or how much that has to do with the underlying biology of the disorder. It does remind us to take criminal and aggressive behavior seriously in this population while also warning against generalization and stereotype.

I'm no expert in this area -- I got all of this from an hour of searching and reading. So if there are experts who can say more, please do so. But there is good (if imperfect) research being done on this subject and there are resources out there to learn more about this subject from a data-driven perspective if you know where to look. It's just that, as with anything involving the study of human beings, the answers are rarely clear cut. Which makes anecdotes so appealing, because there is the possibility of clarity in discussing a single person that you can't hope to have when researching an entire group of people.

Sorry to nitpick, but I hate it when we use the word casually that way. In spite of what the root of the term seems to imply (split mind), it isn't about conflicting personalities within a single person. It's tossed around all the time as a cheap and easy metaphor for "self contradictory," and that use of the term contributes to a massive misunderstanding of a disease that affects millions of people. Schizophrenia - actual schizophrenia - means having delusions, hallucinations, and other disturbances of cognition, speech, mood, and behavior. It is not multiple personality disorder/disassociative identity disorder.