Violence risk assessment is a decision-making task that transpires in numerous legal and clinical settings in which the possibility of a person’s future violent behavior is of concern. Common contexts in which violence risk assessment occurs include involuntary civil commitment, release from prison or forensic hospital, sentencing, transfer of youths to adult court, and sexually violent predator determinations. The formal study of violence risk assessment proliferated after the seminal research studies and legal decisions in the 1960s and 1970s established that despite the legal demands for and constitutionality of risk assessments, little empirical evidence could be garnered in their support. Researchers soon focused on identifying promising risk factors for violence and, more recently have been developing and evaluating structured violence risk assessment measures intended to improve on unstructured clinical prediction.
The Purpose and Application of Violence Risk Assessment
The purpose of violence risk assessment differs somewhat across applications, but at its core, it is the estimation of the likelihood of future violent behavior posed by an individual. In some settings (i.e., treatment discharge planning), risk assessment also includes a specification of the risk factors present in a case and the risk management or intervention strategies that would be necessary to mitigate risk. In other settings (i.e., sexual predator determinations, prison security-level classification decisions), less attention is given to the specific nature of the risk factors that underlie a risk assessment than to the end result (an estimate of level or amount of risk).
Almost invariably, risk assessments transpire within legal contexts. One exception would be the duty to protect that many mental health professionals have in the context of private psychotherapy. Even here, however, a mental health professional is subject to common law and has ethical duties to identify high-risk patients. At some point in their careers, most mental health professionals will be faced with a client or patient who poses a risk of violence.
In terms of applications, violence risk assessment commonly is employed for release decisions from prisons and psychiatric facilities (both civil and forensic). Although specifics differ across jurisdictions, release may be contingent on a decision that a person does not pose an undue risk to public safety. In some such applications, efforts are made to gear the risk assessment toward identifying risk factors that would be important to target in violence-reduction treatments. Violence potential is also relevant in the civil commitment context, in which persons must not only be mentally ill but also pose a risk to others (or self) to be detained involuntarily. In criminal sentencing contexts, risk assessment can be used to inform decisions about whether a person should receive a custodial or a community sentence. In more extreme manifestations, it may be used in the determination of whether a person is sentenced to death. An increasingly common use of risk assessments is to determine whether a person meets statutory criteria for being declared a “sexually violent predator”; that is, a sexual offender who poses a substantial risk to re-offend in a sexual manner may be committed for treatment after serving a term of incarceration. Other countries use risk assessment within the context of indeterminate sentencing statutory provisions. Furthermore, risk assessment often is used when persons are admitted to prisons, youth detention centers, or psychiatric facilities, to determine the security level of their placement and/or their risk-relevant treatment needs. There are numerous other legally relevant contexts in which risk assessment figures prominently (i.e., child abuse evaluations, custody and access evaluations, judicial interim release decisions, immigration hearings).
Risk assessments typically are done by persons within the human services fields, such as psychology, psychiatry, social work, nursing, or substance use counseling. Personnel within prison systems (i.e., correctional classification officers) also engage in a type of risk assessment for offenders admitted to penal systems.
History of the Scientific Study of Violence Risk Assessment
Although the clinical and legal tasks of violence risk assessment have been occurring in some form for centuries, the contemporary scientific study of violence risk assessment has its roots in important legal decisions and empirical studies that occurred between the mid-1960s and late 1970s. During this time, important legal decisions such as Baxstrom v. Herold, a 1966 decision of the U.S. Supreme Court, found the extant process underlying postsentence confinement of the so-called mentally disordered dangerous offenders unconstitutional and ordered hundreds of such persons, all of whom had been declared to be “dangerous,” to be released from confinement or transferred to less secure institutions. The sociologist Henry J. Steadman took advantage of the opportunity to study the fate of these persons and reported that there were very few who subsequently committed new violent offenses, suggesting that the accuracy of the determinations of dangerousness was low. Despite these developments, other legal decisions upheld the constitutionality of clinical predictions of violence. Furthermore, the 1976 groundbreaking case of Tarasoff v. Regents of the University of California went so far as to impose a positive legal duty on psychiatrists and psychologists to forecast the potential violence of patients under therapy in some circumstances.
In the early 1980s, Professor John Monahan summarized what he called the “first generation” of empirical studies on the prediction of violence. He concluded that clinical predictions were not very accurate, in particular leading to an unacceptably high false-positive error rate. He also called for a “second generation” of research that would focus on identifying meaningful risk factors for violence and using empirically based procedures for making risk assessments. Since that time, researchers have indeed identified numerous meaningful violence risk factors, such as substance use problems, psychopathic personality features, anger, impulsivity, antisocial peers, antisocial attitudes, previous violence, young age at first violent act, stress, treatment nonadherence, lack of social support, and some features of mental illness.
Although there has been no well-recognized declaration of a “third generation” of risk assessment research, efforts since the early to mid-1990s have capitalized on what was learned from risk factor research that has led to the construction and evaluation of risk assessment measures that compile and integrate numerous empirically validated risk factors. A further movement has included increased emphasis on the reduction of risk or prevention of violence, as opposed to solely estimating the likelihood of future violence. For this reason, a single definition of violence risk assessment is elusive. The two primary contemporary approaches to risk assessment, discussed below, adopt somewhat differing conceptualizations of the task.
Contemporary Approaches to Violence Risk Assessment
There are two primary approaches to violence risk assessment—structured and unstructured. The structured risk assessment typified first-generation research on risk assessment and remains commonly used today. However, unstructured risk assessment, sometimes called clinical prediction, is based primarily on clinicians’ discretion and lacks rules that guide the risk factor selection or integration process. As such, it is vulnerable to decisional biases and widely varying quality across clinicians. For these reasons, although research supports that it can achieve statistically significant predictive levels, this differs across clinicians. Furthermore, most research indicates that it has lower reliability and predictive validity than more structured approaches to risk assessment. Therefore, a purely unstructured, discretion-based approach to risk assessment cannot form the basis of defensible risk assessment.
To increase the reliability and validity of risk assessments, researchers focused on developing and evaluating structured approaches: actuarial decision making and structured professional judgment (SPJ).
Actuarial Decision Making
The actuarial approach has a long history in psychology in terms of prediction. It is defined by the application of algorithms (equations, score cutoffs, decision rules) to the combination of risk factors to reach a predictive decision, thus improving the consistency and accuracy of such decisions. It also tends to incorporate risk factors that have been selected through empirical means (i.e., those items that add independently and incrementally to the prediction of the outcome). Therefore, at least within development samples, the risk factors used to make decisions have empirical support. Meta-analytic research suggests that actuarial prediction is more accurate than unstructured clinical prediction in approximately 50% of research studies, by about a 10% increase in hit rate.
Problems have been noted with actuarial approaches. Perhaps most important, actuarial measures that rely on statistical selection of risk factors and decision algorithms are subject to predictive degradation when used in new samples, and hence their generalizability is potentially tenuous. Therefore, cross-validation is a vital component of actuarial measure development, and evaluation and must be performed prior to the use of such measures. Furthermore, there is a tendency for actuarial methods to emphasize time-invariant risk factors that are less relevant to violence risk management and reduction aims than time-varying, or dynamic, risk factors. Because there has been a strong conceptual shift in the violence risk assessment field from a purely predictive model to a risk management or harm-reduction model, the predominantly predictive focus of some actuarial methods has been criticized. Although identifying risk level can indicate the intensity of necessary intervention, the failure to include risk factors that can serve as treatment targets reduces the relevance of actuarial prediction for informing the type of treatment that would reduce risk. These limitations have led some researchers to express concern about this method of decision making and to develop another approach—SPJ.
Structured Professional Judgment
The SPJ approach shares some features with typical actuarial approaches, such as specifying risk factors with empirical support that are to be considered in a risk assessment and providing operational definitions and coding rules for risk factors. These features are intended to facilitate its reliability and predictive validity. However, the SPJ approach allows professional judgment at the point of making final decisions. It provides guidelines for such decisions but does not impose strict cutoffs or algorithms derived from a particular sample because of their potential predictive degradation and instability across samples. Evaluators are encouraged to consider the relevance of risk factors to the given individual being evaluated and to make decisions of low, moderate, or high risk based on the number, pattern, and relevance of such risk factors, as well as the probable degree of intervention that would be necessary to mitigate risk. The other distinguishing feature of SPJ is its explicit emphasis on risk management and the importance of bridging assessment and management. It does this through including time-varying risk factors and by providing guidelines for linking risk factors to risk management strategies, the point of which is not only to provide the means of assessing the likelihood of future violence but also to facilitate risk reduction and the prevention of future violence.
The SPJ approach has been criticized for allowing professional judgment to form part of the final decision-making process. Critics argue that this reduces its reliability and validity. However, those studies that have compared actuarial and SPJ measures or the structured judgments produced by SPJ with the numeric use of SPJ measures have generally found that SPJ decisions fare as well or better than actuarial decisions. Criticism also has been directed toward the publication of some SPJ measures prior to the accumulation of evaluation studies. Developers counter that SPJ measures are professional guidelines that are based on the empirical literature. They do not rely on empirically derived, potentially unstable algorithms that are in need of cross-validation, and hence, they do not require the same type of validation. Both arguments have merit. However, specific SPJ measures ought to have empirical support for their interrater reliability and predictive utility (of risk factors and risk judgments) prior to recommendation for use in practice.
- Baxstrom v. Herold, 383 U.S. 197 (1966).
- Douglas, K. S., & Kropp, P. R. (2002). A prevention-based paradigm for violence risk assessment: Clinical and research applications. Criminal Justice and Behavior, 29, 617-658.
- Heilbrun, K. (1997). Prediction versus management models relevant to risk assessment: The importance of legal decision-making context. Law and Human Behavior, 21, 347-359.
- Monahan, J., Steadman, H. J., Silver, E., Appelbaum, P. S., Robbins, P. C., Mulvey, E. P., et al. (2001). Rethinking risk assessment: The MacArthur study of mental disorder and violence. New York: Oxford University Press.
- Mulvey, E. P., & Lidz, C. W. (1995). Conditional prediction: A model for research on dangerousness to others in a new era. International Journal of Law and Psychiatry, 18, 129-143.
- Tarasoff v. Regents of the University of California, 17 Cal.3D 425 (1976).