Risk Assessment: How Psychologists Predict Future Violent Behavior

Ironically, waiting until the last minute to conduct a
risk assessment may actually increase its accuracy.

Let me get this statement out of the way right at the start: psychologists are pretty bad at predicting the potential for violent behavior. I must also say this: we are a lot better at it than we used to be.

In 1983, the United States Supreme Court decided the landmark case of Barefoot v. Estelle. In this case, the court decided behavioral scientists could competently predict that a specific criminal would engage in future violent behavior "with an acceptable degree of reliability." What was that acceptable degree of reliability in 1983? About 33%.

33% of the time, behavioral scientists could accurately predict whether or not a specific criminal would be violent again. Two-thirds of the time, psychologists got it wrong.

Since that time, there have been huge improvements in the way risk assessments are conducted. Psychologists examine a number of different factors in the individual's life, weigh those factors based on how good (or bad) of a predictor each is on its own, add them all together, and come up with a risk level for the individual. The risk levels are "low," "medium," and "high." Not too shocking.

So what are these risk factors? Generally speaking, they fall into two different camps: static risk factors and dynamic risk factors.

Static risk factors are, as the name implies, pieces of information about a person that do not change over time. These are typically issues from the person's past, such as where he grew up, if he went through any type of abuse as a child, and if he was raised by both parents. There are also some static factors that deal with current information about the individual, such as age (which changes over time, but fairly slowly), and personality structure. We know from research that there are certain life circumstances that correlate highly with future violent acts. By examining these factors, it gives clinicians a reasonable idea of the potential level of risk a person poses.

By far, past violence is the best static predictor of future violence. On its own, it is still not great at predicting a future violent act (for example, a person who has murdered someone in the past is actually not all that likely to murder someone again in the future). But, when past violence is added to a host of other static factors, a clinician's predictive ability rises dramatically. In fact, at the present time, conducting a risk assessment using a static factors method can predict violence (or non-violence) about 65-70% of the time. Still not great, but much better than the 33% that the Barefoot v. Estelle Court thought was acceptable.

When dynamic risk factors are added to static factors, predictive ability rises dramatically. Dynamic factors are those that can change from day to day. Whether the person is in the midst of an acute psychosis or manic phase. If the person's partner just left him and took his kids away. Whether he is actively abusing methamphetamines. These are all factors that can change from day to day or month to month and can increase a person's risk level. But, there is a catch with dynamic factors: as you might imagine, they can change rapidly.

With that rapid change comes great unpredictability. Psychologists cannot look into the future and know what a person's current life circumstances are going to be. Therefore, we cannot rely on dynamic risk factors to help us predict whether a person will act in a violent manner at some unknown point in the future. But, we do know there are a number of dynamic factors that can predict whether a person is going to engage in violence sometime in the next few days (or even the next few weeks, depending on the factors being assessed).

So, here is what we know so far about risk assessment:

1. Psychologists used to be absolutely terrible at predicting future violence. We used to get it wrong two out of three times.

2. We have improved dramatically in this arena. We can accurately predict future violence about 65-70% of the time.

3. We are still not all that great at it. We get it wrong about one in three times.

4. The best current method for assessing the risk of future violence is by examining an individual's static risk factors (Wording in a risk assessment report might look something like this: "Based on a review of Mr. Jones' static risk factors, he is at moderate to high risk for engaging in a violent act sometime within the next seven years").

5. In order to assess for the risk of violence in the short-term (days to weeks), a review of both static and dynamic risk factors is important.

Clearly, we need to get better at predicting violence. The methods we currently use are much better than they used to be, but there is still a lot of room for error. Recent events in Colorado and Wisconsin remind us of the dangers of not catching dangerous individuals before they act. On the flip side, it would be a travesty to take away an innocent person's rights because he may pose a danger at some unknown point in the next decade. I am encouraged by the progress that has been made since 1983, and psychologists are continuing to work hard to refine our approach to risk assessment in order to improve it beyond the 65-70% accuracy threshold. I am sure that, as research progresses, we will get better and better at assessing an individual's risk and protecting society from the its most dangerous citizens.

Thanks for reading-- Max Wachtel, Ph.D.
www.CherryCreekPsychology.com
www.Facebook.com/drmaxwachtel
www.Twitter.com/mwachtel

Posted in: Barefoot v. Estelle (1983),risk assessment,violence