Eric Silver

A classification tree approach to the development of actuarial violence risk assessment tools

Since the 1970s, a wide body of research has suggested that the accuracy of clinical risk assessments of violence might be increased if clinicians used actuarial tools. Despite considerable progress in recent years in the development of such tools for violence risk assessment, they remain primarily research instruments, largely ignored in daily clinical practice. We argue that because most existing actuarial tools are based on a main effects regression approach, they do not adequately reflect the contingent nature of the clinical assessment processes. To enhance the use of actuarial violence risk assessment tools, we propose a classification tree rather than a main effects regression approach. In addition, we suggest that by employing two decision thresholds for identifying high- and low-risk cases--instead of the standard single threshold--the use of actuarial tools to make dichotomous risk classification decisions may be further enhanced. These claims are supported with empirical data from the MacArthur Violence Risk Assessment Study.

Neighborhood structural characteristics and mental disorder: Faris and Dunham revisited.

We examined the relationship between neighborhood structural characteristics and mental disorder using data from the National Institute of Mental Healths Epidemiological Catchment (ECA) surveys (n = 11,686). After controlling for individual-level characteristics, we found that neighborhood disadvantage was associated with higher rates of major depression and substance abuse disorder, and that neighborhood residential mobility was associated with higher rates of schizophrenia, major depression, and substance abuse disorder. Implications for future research on the social ecology of mental disorder are discussed.

Assessing violence risk among discharged psychiatric patients: toward an ecological approach.

This paper draws upon data from the Pittsburgh site of the MacArthur Foundations Risk Assessment Study, a large-scale study of violence risk among persons discharged from psychiatric hospitals, to examine the effect of the neighborhood context on risk of violence. This paper has two purposes: (1) to assess the extent to which the inclusion of neighborhood characteristics enhances violence prediction models—models that traditionally only include individual-level characteristics; and (2) to assess the consistency of individual level risk factors across different neighborhood contexts. Results indicate that neighborhood poverty has an impact over and above the effects of individual characteristics in identifying cases with violence. These findings support efforts to include neighborhood context in the assessment and management of violence risk among discharged psychiatric patients.

The Role of Depressed Mood and Anger in the Relationship between Family Conflict and Delinquent Behavior.

Drawing on R. Agnews (Foundation for a general strain theory of crime and delinquency. Criminology 30: 47–87, 1992) general strain theory, this paper examines whether depressed mood and anger mediate the effects of family conflict on delinquency. We examine data on 7758 students, 14–16 years old, attending the compulsory 9th and 10th grades of the Icelandic secondary school system. We use structural equation modeling to show that exposure to arguments and fights at home are positively related to both depressed mood and anger among adolescents. Anger is positively associated with delinquent behavior whereas depressed mood has no effect on delinquency.

Beyond Criminalization: Toward a Criminologically Informed Framework for Mental Health Policy and Services Research

The problems posed by persons with mental illness involved with the criminal justice system are vexing ones that have received attention at the local, state and national levels. The conceptual model currently guiding research and social action around these problems is shaped by the “criminalization” perspective and the associated belief that reconnecting individuals with mental health services will by itself reduce risk for arrest. This paper argues that such efforts are necessary but possibly not sufficient to achieve that reduction. Arguing for the need to develop a services research framework that identifies a broader range of risk factors for arrest, we describe three potentially useful criminological frameworks—the “life course,” “local life circumstances” and “routine activities” perspectives. Their utility as platforms for research in a population of persons with mental illness is discussed and suggestions are provided with regard to how services research guided by these perspectives might inform the development of community-based services aimed at reducing risk of arrest.

A Cautionary Note on the Use of Actuarial Risk Assessment Tools for Social Control

Actuarial risk assessment tools for predicting interpersonal violence and criminal recidivism have proliferated in recent years, promising to enhance their use in the day-to-day operations of the criminal justice and mental health systems. The authors consider the social and political implications of using actuarial tools for social control by offering two specific challenges. First, they argue that actuarial tools are designed primarily to facilitate the efficient management of institutional resources rather than to target individuals or social conditions in need of reform. Second, they argue that the group-based nature of actuarial prediction methods may contribute to the continued marginalization of populations already at the fringes of the economic and political mainstream.

Race, neighborhood disadvantage, and violence among persons with mental disorders: the importance of contextual measurement.

The individualistic fallacy (i.e., the fallacy of assuming that individual-level outcomes can be explained exclusively in terms of individual-level characteristics) is a problem with most research on violence, and is particularly problematic in research on mental disorder and violence. This article illustrates the importance of measuring community context by showing that race is not an important predictor of violence among persons with mental disorders when neighborhood disadvantage is controlled statistically. More generally, these results suggest that researchers run the risk of perpetuating the individualistic fallacy in studies of violence by persons with mental disorders when they use individual-level risk factors as predictors, but do not control for community context.

Assessing the Violent Offending and Violent Victimization Overlap Among Discharged Psychiatric Patients

Prior studies have documented linkages between mental disorder and both offending and victimization. However, few studies have examined the violent offending–violent victimization overlap among mentally disordered individuals and none have examined the factors that are jointly related to their covariation. Here, we assess this overlap during the first ten weeks following hospital discharge among a large sample of psychiatric patients from three large cities. Findings indicate that: (1) violent offending and violent victimization show substantial covariation; (2) although each of the two outcomes were predicted by a few unique risk factors, several risk factors were similarly predictive of both outcomes; and (3) even after adjusting for demographic, clinical, and social risk factors, the correlation between violent offending and violent victimization remained robust. Implications for theory, research, and policy are highlighted.

Constructing actuarial devices for predicting recidivism: a comparison of methods

A recent contribution to the prediction literature by Steadman et al. features a novel “iterative classification” procedure for constructing risk screening devices. In this article, the authors apply the iterative classification procedure to a large recidivism data set, across a range of recidivism outcomes and cross-validation conditions. The purpose of this study is to assess the generalizability of the iterative classification procedure and to draw comparisons with more traditional methods of device construction. Results show the iterative classification procedure to outperform other standard device construction procedures in terms of the percentage of cases classified as high or low risk but not to outperform more traditional device construction procedures on a variety of other performance measures. Implications for future research on the construction and evaluation of risk screening devices are discussed.

A Multiple-Models Approach to Violence Risk Assessment Among People with Mental Disorder

Actuarial models for violencerisk assessment have proliferatedin recent years. In this article, we describe an approach that integrates the predictions of many actuarial risk-assessment models, each of which may capture a different but important facet of the interactive relationship between the measured risk factors and violence. Using this multiple-models approach, we ultimately combined the results of five prediction models generated by the iterative classification tree (ICT) methodology developed in the MacArthur Violence Risk Assessment Study. This combination of models produced results not only superior to those of any of its constituent models, but superior to any other actuarial violence risk-assessment procedure reported in the literature to date.