Link between Asperger's and violence

I suppose this deserves a post of its own.  I recently mused on a personal reminiscence and discussed how those with high-function ASD (Autism Spectrum Disorder) might well harbor special resentments about their isolation and ill treatment, while cautioning that this doesn't mean that they are especially susceptible to punishing random, innocent people on rampages.  Even those whose autism manifests in a violent reaction to environmental stimuli that can't be coped with, won't necessarily be likely to plan and execute some vengeful spree.  It is a different phenomenon.

John Elder Robinson has, naturally, published some short articles on the relevance of an alleged Asperger's diagnosis to the Newtown massacre.  He makes some valid points.  For example, he says,
Correlation does not imply causation.

Correlation does not imply causation. . . . It was a random, irrelevant coincidence. . . .How about this factoid: Most school shooters are Caucasian males. You might find that statement a little more shocking than the previous one. But it’s true. Does that mean every white male Caucasian who enters a school is a potential mass murderer? Of course not.  Suggesting a mass murderer had Asperger’s is much the same – it may be true, but stating the fact does nothing to explain the crime, nor does it help prevent other crimes in the future. What it does do – and this is important – is paint a whole swath of population – Asperger people – with a brush that says “potential mass murderer.”

This is a valid point and it puts me in mind of both the irrelevant reminders in media that the Tea Party movement is "mostly white" as well as the recent eagerness of certain journalists to identify (wrongly, it turned out) a shooter as a Tea Party affiliate, as though that connection would be somehow relevant.  A killer's mental diagnosis is a more plausibly relevant fact, though as Robinson argues, it could unfairly bias the public.  He seems to be implying that the information should be filtered as obviously irrelevant.  Maybe, more appropriately, it should be more properly vetted and put into perspective by an expert. 

Robinson elsewhere says,

There is no scientific evidence linking ASD with homicides or other violent crimes. In fact, studies of court records suggest that people with autism are less likely to engage in criminal behavior of any kind compared with the general population, and people with Asperger syndrome, specifically, are not convicted of crimes at higher rates than the general population.

Another good point, although, as Iohawk cautions on an unrelated subject, Simpson's Paradox can fool us.  It could be that a particular subset of ASD people that exhibit violent behavior are more likely than more "neurotypical" people with violent histories to go on killing sprees, but that this is obscured by the much larger subset of ASD people that are much less likely to do violence than their neurotypical counterparts.  In fact, Robinson's citing of statistics on criminal behavior in general may further obscure significant correlations.

Those caveats aside, I think there is much more provocative data supporting a link between psychotic rampages and psychopharmaceuticals than between them and ASDs: