We Are Wrong About Psychopaths

A New "Switch Theory" of Psychopathy

It’s estimated that around 1% of the general population could be classified as a clinical psychopath. In prisons, it’s closer to 15-25%. These estimates are based on the Psychopathy Checklist Revised, commonly known as the PCL-R.

Created by forensic psychologist Robert Hare, The PCL-R is a 20-item instrument inspired by psychiatrist Hervey Cleckley’s clinical work reported in The Mask of Sanity and based on Hare’s work in prisons.

The PCL-R isn’t like most psychology surveys you’ve seen. Rather than having the subject take the survey themselves, a licensed psychologist interviews the subject, sometimes for up to several hours, and looks over historical information that might offer insight into their behavior.

The psychologist then rates each of the 20 items on a 3-point scale: 0 if the item doesn’t match, 1 if it’s a partial match, and 2 if it’s a good match. Those with scores of 30/40 or higher are classified as psychopaths.

According to Hare, four dimensions make up the PCL-R: Interpersonal, Affective, Lifestyle, and Antisocial. These four dimensions spring from an older 2-factor interpretation of the PCL-R involving one factor that taps into emotional detachment/callousness (now interpersonal and affective) and second factor that taps into antisocial behavior (now antisocial and lifestyle). Two other items (promiscuous sexual behavior, many short-term relationships) do not load onto any of the four factors, but do contribute to the total PCL-R score.

For decades, the PCL-R has been the gold standard for measurement of psychopathy in forensic populations. Another version, the PCL-SV, has been developed for non-forensic/general populations. The PCL has, in some ways, defined psychopathy.

But what is psychopathy?

The Gold Standard, But of What?

It’s difficult to find a real definition of psychopathy from Hare. Here’s the best one I can find from him:

“The purpose of the PCL‐R was to measure the construct of psychopathy. It describes psychopathy as a constellation of interpersonal, affective, lifestyle, and antisocial traits and behaviors.”

But this confuses measurement and construct, a criticism that has been levied before by Jennifer Skeem and David Cooke, who further stated in a controversial paper that:

“Discovering that this diagnostic tool happened to predict violence and criminal recidivism, researchers may have run too quickly: Predictive utility cannot substitute for construct validity. In our opinion, the field is in danger of crossing a thin line between using the PCL–R as a “common metric for psychopathy” (Hare & Neumann, 2005, p. 57) and equating this metric with the theoretical construct of psychopathy.”

Best I can tell, Hare has a very good description of behaviors that he has labeled psychopathy, but no real theory explaining the observations. Without a theory of what psychopathy is, his logic is sort of circular: Psychopathy is defined as a 30 or higher on the PCL-R, which measures psychopathy… which is defined as a 30 or higher…

To be clear, the modern version of the PCL-R does seem to be measuring something quite well, something to do with deviancy in the modern US. The statistical model of the instrument is quite good, but, as his critics rightfully point out, “fit indices do not address whether the items selected are adequate for defining the construct.” And this is the key issue. There is no good theory for why the items on the PCL-R factorize like they do or why they are related to one another.

The biggest problem with all of this is that those few who have tried develop some theory of psychopathy are mostly relying on Hare’s description of psychopathy. And if Hare’s description of psychopathy isn’t describing a true trait, one that carves nature at its joints, then explanations of that description will be led astray.

So, here’s a new way we might think about psychopathy.