Autism, Asperger's, violence and mass murder

Author's Note:  This article was published in the spring of 2014, but it's just as valid today.  You may also enjoy this piece from Psychology today, one I wrote in response to the Newtown killings.
John Elder Robison

* *

Autistic adults are crying foul at a recent Washington Post article that speculates on a connection between serial killings, mass murder, autism, and head injuries. The premise of the article is that those things may be connected in more than a coincidental way, based on a study of mass killers in various databases.  Neurodiversity advocates are rightfully concerned that the story will turn the public against autistic people for no good reason.

I say, Shame on the Washington Post for running such an inflammatory headline to promote such a poorly thought through article.

The ill-considered piece is based on a newly published study – Neurodevelopmental and psychosocial risk factors in serial killers and mass murderers, by Allely, Minnis, Thompson, et al.  Three of the authors are psychologists at the University of Glasgow, one is at the University of Gothenberg, and one is at the Center for Health Science in Inverness, Scotland. 

While the original paper has some interesting findings, the conclusions they reach, and the way they are presented – in my opinion – verge on irresponsible.  There is an overwhelming body of evidence that autistic people end up victimized far more often that we victimize others.  There’s no prior association between being autistic and premeditated aggressive violence. Articles like the one in the Post lead to further exclusion and victimization from a fearful uninformed public.

We saw that in the aftermath of the Newtown killings with the talk about whether Adam Lanza was on the autism spectrum, and by implication, whether autism was a factor in his crimes.  Young people with autism were bullied and isolated by irrational readers who were stirred up by the media frenzy.

The whole idea of this latest study is flawed.  The issue is simple:  correlation does not imply causation.  Researchers forget this to their detriment as they are led to wildly wrong conclusions when two data sets seem to fit together.  I’ll give you an example.  Right now, if we compare data from the US Census and the USDA for the years 2000-2009 we find a near perfect correlation (.993) between the divorce rate in Maine and the per capita consumption of margarine. 

Who knew margarine consumption predicted divorce with such accuracy?  It (probably) doesn’t.  But an ignorant person, looking at the near-perfect fit, could easily be led to that conclusion if he lacked the common sense to see through it.

The problem with the Post article is that common sense isn’t common, the author does not explain this, and the correlations in the article are not as obviously spurious as margarine consumption and divorce.

Here’s what they claim: Of 239 killers, the authors concede that 133 showed no evidence of autism or brain injury.  But then comes the shocker:  67 (or 28%) are described as having possible, probable, or definite autism.  That makes autistic people sound pretty scary, doesn’t it?

It does not make autistic people scary to me, because I’m informed, and I know the statistic is ridiculous.  But people fear what they don’t know, and there is a lot of ignorance about autism.   The group of serial killers and mass murderers had other common traits, too, but the authors did not see fit to mention them in the same light.  For example:

• ·      Almost all serial killers and mass murderers are male.  Should we be scared of males just because most murderers are male?  That’s actually a pretty well-known fact but males remain pretty acceptable in most societies;
• ·      A majority of serial killers and mass murderers are Caucasian.  Caucasians are welcome most places too – at least in the United States and Canada;
• ·      Of those who are not Caucasian a majority are the predominant race of the land where they killed.  “They eat their own,” is a common proverb, and like most proverbs, there’s an element of truth in it;
• ·      In America, a majority of serial killers and mass murderers came from middle class or affluent backgrounds.  Few came from poor backgrounds.  That’s a surprise to many but it’s true;
• ·      In America, the vast majority of serial killers and mass murderers employ both automobiles and firearms in the commission of their crimes.  Yet there is very little regulation over the ownership or use of either in most of the United States;
• ·      A majority of serial killers targeted victims who were physically weaker than themselves – principally women and children.  With all the talk of bullying today that comes as no surprise, but it’s not clear how you’d use that insight to avoid a serial killer;
• ·      Many serial killers target victims who belong to groups scorned or marginalized by the society in which the killer lives.  That too is not surprising to anyone reading the news.  It’s dangerous to work as a hooker or a drug dealer.  If that’s the career path you choose the risk of death at the hands of serial killers is just one of many hazards.

None of the associations above are worrisome, because they are easily understood and evaluated by a layperson.  Yet every one of them is a stronger match than what the study’s authors claim for autism.

That takes me to the next issue with this study . . . their notion of “possible, probable, or definite autism.”  Let’s look first at the killers who really have autism diagnoses in the studied group.  The authors only identified six diagnosed autistic people.  With a total population of 239 killers – mostly male – the latest autism prevalence statistics suggest autism is about as common in this group as it is in the general population.

Taking that a step further, we could say that the diagnosed prevalence does not suggest autism per se is a factor in whether someone becomes a serial killer or not. There are autistic serial killers, blue-eyed serial killers, and brown-haired serial killers, in similar numbers as will be found in any other human population. Such observations – while true – don’t really tell us anything meaningful about why someone becomes a serial killer. They are just unrelated data points, like the correlation between margarine consumption and divorce.

If we had a theory for why autism, brown hair, or blue eyes might make someone into a serial killer, things would be different.  But there is no such theory, and in the absence of one, it is highly inappropriate to make such a suggestion when it applies to an already-vulnerable population.

Next, let’s look at those “possible or probable” autistic killers.  The authors of the paper described in the Post called a killer “possible or probable autistic” on the basis of Internet speculation, speculation from observers in jail, or speculation based on descriptions of the individuals. 

Armchair diagnosis like that may make for fun conversation, but it has no medical or psychological validity.  It’s speculation, pure and simple.  It wouldn’t be allowed in court and it should not have been allowed here.  The essential problem with speculation in the Aggression and Violent Behavior article is that most of the observed traits the authors associated with autism can also be associated with psychopathy and sociopathy, and I suggest those are more likely explanations given the demonstrated behavior of the individuals.

This study would have had a lot more validity if they had taken their sample of 239 killers and asked, "What psychological, neurological, or psychosocial issues can we definitely attribute to each killer?"  That would have produced a much more nuanced and complex result, but it would have been a result with real meaning.  The present study - looking only at autism and head injury - and its possible-probable-definite language, is hogwash.

So, in conclusion, I will just repeat - There are no studies showing a propensity for aggressive violence on the part of autistic people.  This most recent paper – which has gathered a lot of headlines but little professional praise – does not make that case either. 

John Elder Robison is an autistic adult and advocate for people with neurological differences.  He's the author of Look Me in the Eye, Be Different, Raising Cubby, and the forthcoming Switched On. He serves on the Interagency Autism Coordinating Committee of the US Dept of Health and Human Services and many other autism-related boards.  He's co-founder of the TCS Auto Program (A school for teens with developmental challenges) and he’s Neurodiversity Scholar in Residence at the College of William & Mary.  The opinions expressed here are his own.