Risk Assessment and Violence Prediction

Another topic related to psychopathy and central to clinical forensic psychology is risk assessment. Risk assessment was more commonly referred to as violence prediction because the primary focus for forensic psychologists was predicting whether a particular person would become violent. However, risk assessment is not simply about making a forced choice about whether a particular individual is going to become violent or not. It is about identifying the factors that are likely to increase the risk for violence, the factors that are likely to reduce the risk for violence, the immediacy of the violence, the severity of any likely violence, and the ways in which the violence can be managed. The comparison has been made that risk assessment is similar to weather forecasting. The local meteorologist often forecasts the probability of precipitation for the next day, 10 days from now, or even further into the future. That same meteorologist may not only assess the probability but also assign different categorical labels to tornadoes, like watch and warning. The meteorologist makes these assessments based on complicated statistical models and historical weather patterns. All of these routine practices that we are exposed to on a regular basis in watching local television also take place in very similar ways when forensic psychologists attempt to assess someone’s risk of violence.

When risk assessments are undertaken, they usually focus on specific forms of violence that rise to the level of some legal charge or conviction. Risk assessments normally assess the risk for “actual, attempted, or threatened physical harm that is deliberate and nonconsenting” (Hart, 2005, p. 4). This violence could consist of violence between a man and a wife, the sexual assault of a child, the stalking of a woman by a former coworker, or even murder. Risk assessment may be necessary for institutionalizing someone who is mentally ill, a defendant awaiting criminal sentencing, a juvenile who is facing transfer to an adult court, or when determining whether someone should be released from prison. Although there are significant differences in assessing violence across different situations, there also are clear similarities.

Risk assessment routinely involves the identification of risk factors. Risk factors are different variables that increase the risk for violence. Common risk factors for violence across different types of violence would include previous violence, substance abuse, psychopathy, employment instability, and early behavioral problems. These risk factors are normally referred to as either static or dynamic. Static factors are ones that do not change or are extremely unlikely to change. For example, previous violence will not change for someone who has a criminal conviction for a violent crime. However, we would expect other things, like the current severity of mental illness symptoms, to wax and wane. Risk factors that are likely to change over time are normally referred to as dynamic factors. It is increasingly important for forensic psychologists to be aware of both the static and dynamic risk factors in assessing risk.

Risk assessments use these risk factors in either clinical approaches or actuarial approaches. Clinical approaches involve the judgments of forensic psychologists that have been formed through their psychological education and professional experience. The traditional approach to risk assessment has been based in clinical methods. A forensic psychologist may sit down with a patient who has recently been admitted to an emergency room and perform a clinical interview; review his criminal, mental health, and social history; and then determine whether he needs further treatment because he is likely to harm someone when released from the hospital. This format has been the norm over the years and is in keeping with the clinical approach for risk assessment (Litwack, 2001). Actuarial approaches differ from clinical approaches in that they are not based on professional judgment. Actuarial approaches rely on formal, objective, and, often, statistical information. As mentioned before, meteorologists rely on mathematical models to predict the weather for the next day. Forensic psychologists may rely on similar types of statistical approaches to assess risk. Actuarial risk assessments consist of determining the presence or absence of a list of predetermined risk factors from an actuarial measure that has been designed to assess risk among a certain group of people, like sex offenders. Based on the number and types of risk factors present, sex offenders receive a score that relates to their likelihood of committing violence in the future.

Clinical and actuarial approaches have been criticized for a variety of reasons. Clinical approaches are often seen as subjective and more open to idiosyncratic biases of the individual forensic psychologist. As a result, clinical approaches are frequently deemed less accurate and more prone to error than actuarial approaches (Quinsey, Harris, Rice, & Cormier, 2006). However, actuarial approaches are criticized as inflexible, time consuming, and unworkable in many clinical decisions. For example, it is difficult if not impossible to perform an actuarial risk assessment in a setting like an emergency room, where there may be very little information available about a patient and a decision is needed immediately (Elbogen, Huss, Tomkins, & Scalora, 2005). Although actuarial decisions may be more accurate in general, their accuracy for a given individual is questionable. Car insurance rates are based on actuarial formulas. An insurance company determines the amount they charge you for car insurance based in part on your past driving record, which seems very reasonable to most people. The more accidents you have had, the more car insurance costs you. However, car insurance rates are also determined based on things you have little if any control over, like sex; women pay significantly less for car insurance than do men. As a man, I may have a spotless driving record and be very responsible but pay more for insurance than someone else simply because of my sex. From the actuarial perspective, the statistics clearly suggest that women will cost less to insure than men, so it seems reasonable. But it seems unfair to an individual male driver who has a spotless record. This situation is similar to the use of individual actuarial risk assessments. They may be accurate in general as applied to a large group, but their accuracy is questionable when applied to a specific person (Hart, Michie, & Cooke, in press).

Although the debate about the superiority and application of actuarial and clinical approaches to risk assessment continues, it is clear that risk assessments have advanced over the past 30 years. Early research in the accuracy of risk assessment was marked by the common belief that forensic psychologists were no more accurate than the flip of a coin. Today, a number of risk factors have been empirically identified for generalized violence, sexual violence, and domestic violence that were unclear two decades ago. As a result, numerous formal or actuarial approaches have been developed that have improved the assessment of risk significantly over the coin-flip analogy commonly referred to years ago. Forensic psychologists not only predict the occurrence of violence but also devise ways to reduce or manage the violence so that it never occurs. There also is greater attention paid to communicating the risk of violence so that simple probabilities (e.g., 80 percent chance of violence) may not be used but instead categorical labels such as low, medium, or high risk are used in a way that is similar to the meteorological terms of watch and warning.

Return to overview of Clinical Forensic Psychology.