Can Anyone Predict Dangerousness?

Can psychiatrists and other mental health practitioners predict violence?

Because of the shootings on the Virginia Tech campus, in Tucson and, more recently, in Aurora, it’s an important question.

A recent article published in the Association for Psychological Science magazine examines trends in how mental health professions assess risk and predict violence. The article’s authors, John Monahan of the University of Virginia, and Jennifer L. Skeem, of the University of California, Irvine, begin by examining the mostly commonly used methods.

At one end of the “risk assessment” spectrum are predictions based on an examiner’s clinical experience. In other words, it’s a judgment call.

On the opposite end of the spectrum are “analytical risk assessments.” These are tests that usually involve assigning a number to various behaviors and other “risk factors.” For example an evaluator might consider a person’s age, if there is a history of past violence, the number of hospitalizations, and evidence of anger issues or poor self control. Each factor is assigned a numerical point and if the score at the end is high enough, that individual is judged a high risk for being violent.

Before you dismiss the value of these assessments consider this. Clinicians are being asked more and more often in civil, criminal and juvenile justice cases to predict dangerousness and violence. Virginia’s Sexually Violent Predator statute not only mandates the use of a specific risk assessment instrument but also specifies a cutoff score that must be achieved if someone is either going to be committed or released. And all states require individuals to pose a “danger” either to themselves or others before they can be involuntarily committed to a hospital. So how accurate are these different tools?

The authors found that assessments which relied purely on a psychiatrist’s or clinician’s judgment have  the least “empirical support.” 

 “Clinicians are relatively inaccurate predictors of violence,” according to a 1993 study cited in the article.

How about newer analytical tests that have grown in popularity?

The authors answer by citing what they describe as “the revolver analogy” — an example first published by other researchers in 1996.

 “Two revolvers are put on the table and you are informed that one of them has five live rounds with one empty chamber, the other has five empty chambers and one live cartridge, and you are required to play Russian roulette. If you live, you will go free. Which revolver would you choose? Unless you have a death wish, you would choose the one with the five empty chambers. Why? Because you would know that the odds are five-to- one that you will survive if you pick that revolver, whereas the odds are five-to-one you will be dead if you choose the other one. Would you seriously think, ‘Well, it doesn’t make any difference what the odds are? Inasmuch as I’m only going to do this once, there is no aggregate involved, so I might as well pick either one of these revolvers; it doesn’t matter which?”

Based on that logic, the authors write:

 “Our view is that group data theoretically can be, and in many areas empirically are, highly informative when making decisions about individual cases.”

In other words, using anaylitical data is better than not using it. Monahan and Skeem state that clinicians would be foolish to ignore such factors as  poor self-control, antisocial attitudes and past violence.

But are any of these risk assessment tools accurate predictors of violence?

Monahan and Skeem don’t give a difinitive answer. But a 2007 study mentioned in their article does. It’s authors concluded that the “margins of error surrounding individual risk assessments of violence are so wide as to make such predictions ‘virtually meaningless.’”

I agree with that viewpoint.  I’ve known of too many incidents when parents have been assured by psychiatrists that their loved one was not dangerous only to have that child harm someone or hurt themselves. I believe it is virtually impossible for a psychiatrist or clinician  to predict future behavior with any degree of confidence — unless a patient is acting violently when they are being examined.

Despite this, we stubbornly adhere to a legal standard that requires a person to  “pose a danger to themselves or others” before we intervene for their own good and safety. That’s foolish.

I had another thought after reading the article written by Monahan and Skeem.  I was curious if parents and other family members would have any better luck predicting violence than the experts. Because they would know the person being reviewed, I would guess that family members would score higher. That is why I think psychiatrists and clinicians should pay attention to what family members tell them.  But I doubt the difference would be that significant.

Monahan and Skeem end their article with a smart suggestion. Rather than spending more time trying to perfect ways to assess risk, the authors suggest that “forensic psychology shift more of its attention from predicting violence to understanding its causes and preventing its (re)occurrence.”


What’s your view? Who’s better at predicting behavior? Can you cite specific examples? If so, please do.

About the author:

Pete Earley is the bestselling author of such books as The Hot House and Crazy. When he is not spending time with his family, he tours the globe advocating for mental health reform.

Learn more about Pete.