Cold, criminal, incurable?

Daniel Murrie and colleagues paid 108 forensic psychologists and psychiatrists to score the same offender case files. Some were told they were advising the defence; others that they were advising the prosecution. The files were identical, the instruments were established and well researched, the professionals were trained.

Scoring was done with two common, well-studied risk instruments. The scores diverged. Those who believed they were advising the prosecution assigned systematically higher risk scores, those who believed they were advising the defence lower ones — with effect sizes up to d = 0.85. That is the ground on which the Psychopathy Checklist–Revised also stands, the best known of these procedures, which in the Anglo-American legal world helps decide detention, treatment and release.

What the checklist measures

The term comes from the clinic, not from the cinema. In The Mask of Sanity (1941), Hervey Cleckley described patients whose outward ordinariness he read as a mask: sociable, articulate, with no discernible disorder of thought, and at the same time unattached to anything they said. Decades later Robert Hare turned this into a procedure. The Psychopathy Checklist–Revised is scored by trained professionals from an interview and the case records: 20 items, zero to two points each, with a customary cut-off of 30. It is not an online quiz and not a device that discovers a hidden essence.

Hare and Craig Neumann summarised the instrument’s structure in four correlated factors: the interpersonal (charm, grandiosity, manipulation), the affective (lack of remorse, shallow affect, failure to take responsibility), lifestyle (impulsivity, need for stimulation, irresponsibility) and the antisocial (early behaviour problems, rule violations, varied delinquency). And they record what public discussion regularly loses:

Psychopathy is therefore not a condition one has or does not have. There is no leap between those who are and those who are not — the threshold of 30 is a convention, not a natural boundary.

How rare high scores are therefore depends on the measurement. In the 2003 manual, 15.7 per cent of 5,408 male prisoners in US institutions reached a score of 30 or above. For the general population, a meta-analysis by Sanz-García and colleagues arrives at roughly 1.2 per cent when the checklist is the basis, and at higher figures for self-report instruments — so the number describes the instrument as well as the people.

Three confusions

First, psychopathy is not psychosis. Most people described this way have no loss of reality. Second, it is not identical to antisocial personality disorder, whose diagnosis focuses more on observable rule-breaking. The overlap runs one way: depending on the study, 50 to 80 per cent of prisoners meet the criteria for antisocial personality disorder, but only around 15 per cent count as psychopathic on the checklist. Most antisocial people are not psychopaths.

Third, a high score does not inevitably mean violence. It can be statistically related to recidivism risk, but it does not determine what a particular person will do. And it is precisely here that the evidence becomes more interesting than public talk about the test suggests.

What the score predicts, and who assigns it

In 2010 Min Yang, Stephen Wong and Jeremy Coid compared nine common instruments for predicting violence. They drew the effect sizes from 28 studies published between 1999 and 2008, each of which had tested more than one instrument, and set them against one another within the same samples. The result: all nine predict violence about equally well, at a moderate level. Anyone who only wants a forecast can essentially swap one for another.

There was one exception. Factor 1 of the checklist — the interpersonal and the affective, which is to say exactly the traits from which the cinema builds its characters — predicted violence in men at no better than chance. What makes the psychopath on screen is the instrument’s weakest part as a predictor. The predictive power sits in Factor 2, lifestyle and antisocial history; but that is essentially the information that someone has already come to notice many times.

The authors draw a conclusion from this that reaches beyond psychopathy: at this moderate level of accuracy, such instruments should not be the sole basis for decisions that demand very high accuracy — preventive detention, for instance.

There is also the fact that a personality score is static and risk is not. Violence and recidivism risks change with substance use, access to victims, social monitoring, treatment and specific crises. Historical factors cannot be altered; dynamic ones can. Assessments should therefore state the period and the situation to which an estimate applies.

Agreement between evaluators is weakest exactly where most depends on it. In research studies it is high; under the conditions of forensic practice it drops sharply, and the experiment at the start shows why. The shift was not sloppiness but a partiality the participants themselves did not notice. It arises wherever a procedure leaves room for judgement — and with items like “superficial charm” or “lack of remorse”, there is a great deal of room.

For practice, what follows is not a rejection of the instrument but a duty of disclosure: whoever presents the score must state who obtained it, on whose instruction, with what level of agreement, and with what margin of error. A score of 32 is not a property of the person assessed alone.

The question of treatability

That psychopathy is untreatable is a widespread article of faith — and, as Randall Salekin wrote in 2002 after reviewing 42 treatment studies, one with little scientific foundation. Therapeutic pessimism undermines the search for effective methods, even though the evidence does not support it. Salekin gives three reasons why no strong conclusion is possible: there is disagreement about what defines psychopathy; its development is poorly understood; and there are few empirical studies and fewer follow-up observations after treatment.

The most robust positive finding comes from the Mendota Juvenile Treatment Center in Wisconsin, a structured programme that works with incentives rather than punishment. Michael Caldwell and colleagues followed 141 juvenile offenders with high psychopathy scores for two years. The 85 held in ordinary custody reoffended violently more than twice as often as the 56 from the programme. Assignment was not randomised, but the difference held up under statistical control. The label must not become the justification for offering no help at all.

The psychopath in the boardroom

The figure that has made it from forensic psychology into the culture pages is the one in the suit. Paul Babiak, Neumann and Hare had access to an unusual sample: 203 corporate professionals selected by their companies for management-development programmes. High scores occurred there more often than in general-population samples, and the structure of the construct matched that found in prison and community samples.

The interesting part is that the correlates cut both ways. High scores went together with good internal ratings for charisma and presentation — creativity, strategic thinking, communication skills. They went together at the same time with poor ratings for responsibility and performance: being a team player, management quality, overall rating. That explains both the career such people have and the damage they do. It does not explain why someone is a difficult boss — the sample was selected by companies, not drawn at random, and a correlation across 203 cases yields no remote diagnosis of one’s own superior.

What remains

The three words of the title can be answered one at a time. Cold names a real cluster of traits, Factor 1 — and that is precisely the one that predicts violence in men no better than chance. Criminal confuses two things: most people with antisocial personality disorder do not reach high checklist scores, and the instrument’s predictive power lies largely in the history one knows anyway. Incurable is the least supported of the three claims: it rests on a literature too small, too inconsistent and too briefly followed up to carry it.

How to recognise the difference in everyday life: by whether a statement about a person can be translated into behaviour. Does someone deceive repeatedly? Do they take responsibility? Does their behaviour change under consequences? Such observations are usable for protection and for decisions. The word “psychopath” is not. It is a probability measure disguised as a statement about essence — and a person is not guilty because their score is high, nor innocent because it is low. Judgment about a particular act requires proof.