A Really Bad Study of Autism and Violence

A number of posts have discussed bad media coverage linking autism to violence. At Forbes, Emily Willingham kicks a bad academic study to the curb:

According to mental health professionals who personally diagnosed serial killer Jeffrey Dahmer, he had a personality disorder. That didn’t stop the authors of a recent paper attempting to link autism and mass murderers, serial killers, and other homicidal maniacs from listing Dahmer as “highly suspected” of having an autism spectrum disorder, along with a 61 other people who were never diagnosed with one, including Timothy McVeigh, Terry Nichols, and Dylan Klebold.

I’ve seen some recklessness in my wanderings through the world of autism science, but these authors reach depths I cannot fathom. If you doubt that, let me just point out that they use the Daily Mail as one of their citations to demonstrate that a killer not diagnosed with autism might have it and cite an author who very much wants to make up a diagnostic category called “Criminal Autistic Psychopathy” as a subset of Asperger’s. Which no longer exists.

In their paper, which is making a splash, of course, Clare Allely and co-authors claim that 67 of the 239 “eligible killers” they evaluated in their review had “definite, highly probable, or possible” autism spectrum disorder. But a closer look at their numbers shows that of these, only six were in the “definite” category [ETA: details on those six summarized here]. That’s 2.5% of the total of 239 they examined. It’s a percentage that happens to be just slightly less than the 2.6% identified in the most thorough study of autism prevalence in the general population to date, in South Korea.

The evaluations of the South Korean population were thorough, but perhaps no other population receives more expert attention and evaluation than killers like these who survive their crimes, as Dahmer, McVeigh, and Ramirez did, and killers who do not but whose writings and other leavings undergo deep scrutiny from experts involved in their cases. Their evaluations are not a mystery. Psychologists had access to Dahmer and evaluated him. Psychologists had access to Richard Ramirez (the Nightstalker) and evaluated him. Ditto McVeigh. These are professionals who personally evaluated these killers, and “autism” was not on their list of labels.

It is inappropriate for anyone–much less these authors, giving the imprimatur of science and peer review–to diagnose from a distance someone they have never even met and with whose case they are not deeply familiar. Add to that their conflation of mass murderers and serial killers, whose psychic motivations can be very different, and this entire paper is one big, hot, irresponsible mess.