Violence Prediction

Which individuals are apt to commit a single violent act? Who among them will commit more such acts? What degree of certainty can judges, parole boards, mental health institutions, and the general public hold in terms of violence predictions by psychologists and psychiatrists? What level of confidence, beyond the level of pure chance, can evaluators have in their own predictions of dangerousness? Harold V. Hall and Ronald S. Ebert explore the historical foundations of these inquiries and discuss multiple generations of forensic-decision analysis, adding more than a decade of contemporary research to the first edition of this book. They present the most common deliberate and nondeliberate distortions that can influence violence prediction analysis, and they discuss both literature and demonstration studies that are the foundation of prediction accuracy. Additionally, they explore extant systems, the dangerousness prediction decision tree and major pitfalls of its! use, as well as ethical considerations in forensics and the assessments of dangerousness. Throughout this volume, the authors include a wide range of pertinent figures, tables, appendix materials, and references for thoughtful analysts to use as they seek to understand the complex issues of predicting dangerousness and potential levels of human predatory violence.

from the Preface by the author

MOST VIOLENT ACT: "The most violent act I ever committed was beating up a prostitute in a hotel in Atlanta. After I had paid her the money We talked about her service and she pulled a switchblade and demanded my wallet. I was able to grab the lamp from the nightstand and strike her in the head before she could get any momentum in Swinging the knife. I must have hit her pretty bad because a pool of blood was already forming on the carpet. The thing that got me about the incident was that it made me feel good. I left and never looked back."

The above true account comes from a research project by the author called "1000 Violent Americans." Here, adult, employed men and women of all walks of life in this country were asked to write down their most violent act, their "almost" most dangerous deed in order to assess inhibitions to aggress, a faked account of violence, and to fill out several deception detection measures to gauge response set. The results of this research effort, combined with findings from several dozen other violence-related investigations since 1969, are aimed at developing violence prediction systems which explicitly represent the decision path and utilized content of the predictor. Overall, progress toward these goals can be characterized as promising and increasing in momentum. A more rigorous evaluation would point out that violence prediction decision analysis has now progressed beyond the nominal into the ordinal stage of classification. Much work is still qualitative. Some subareas such as plotting severity of violence over time, based on a 60,000 respondent survey by the National Institute of justice, are purely quantitative in application.

This book is designed for the forensic professional who is involved with the assessment, research or training of individual violence prediction. Many different methods are presented in this book so long as violence Prediction accuracy, decision-analysis, or impact can be positively affected as a result. By the end of the book, you should be able to compare and contrast true and false positives and negatives, cite developmental experiences associated with later violence, point out opportunity factors to violence, isolate triggering stimuli, and be cognizant of various inhibitions to aggress. You will become aware of the approximately 25 forensic deception styles used by examinees in presenting previous violence along with a number of techniques to deter-mine the existence, direction, and magnitude of client misrepresentation and distortion. You will become acquainted with why dangerousness is nearly impossible to predict from psychiatric diagnosis and why it is always hazardous to do so. Most importantly, you will be able to proceed through your own violence prediction decision path and yield reasonable, testable, and circumscribed conclusions in regards to human predatory violence, and in doing so, clearly communicate to your referral source the basis for proffered conclusions.

Several caveats are in order: (1) All cases in this book are true but disguised to protect the identification of both victim and perpetrator. Follow-up data to assess prediction accuracy is available in each case; (2) This book is atheoretical and geared toward practice; the behavioral science literature is discussed when appropriate but is not the predominant concern; (3) The state of the art in dangerousness prediction is nominal/ordinal rather than having desired interval/ratio properties. Continuing research will greatly expand the basic concepts in this book, with a breakthrough probably represented by a composite battery of psychometric devices measuring different and overlapping areas of predictee violence, much as in neuropsychology and some other psychological subfields; (4) Overlearning is necessary to ingrain application of dangerousness prediction methods and to prevent predictor regression to baseline levels of decision analysis and accuracy. In this sense, this book is deliberately designed for redundancy; (5) This book requires participants to examine their own decision paths. Yet, many predictors are reluctant to change their biases and values in regards to violence, especially when loss of self-esteem or prestige is equated with giving up cherished beliefs about dangerous behavior; (6) A common complaint is that proper dangerousness prediction requires much time and effort, adding to an already overloaded work schedule. This is true but the author has no sympathy for this problem. Tough! The issues raised in dangerousness prediction are of high importance for both society and the individual. Much time can be saved by use of the recommended procedures and formats. Obsessive-compulsiveness and a strong spirit of the hunt for truth and impact are very helpful.

Many thanks are owed to Dr. Evan Wolfe, consultant to the State of Hawaii Adult Probation Division, who initially inspired much of the early interest in violence prediction. Dr. John (Joe) Blaylock, as the writer's previous supervisor at the State of Hawaii Courts and Corrections Branch, allowed for the opportunity to explore violence to others from a research and training context, Dr. Udo Undeutsch of the University of Cologne, Federal Republic of Germany, probably the finest forensic psychologist in Europe, clearly showed how much we Americans can learn from our international contemporaries in terms of distortion and deception analysis. Dr. John Monahan of the University of Virginia School of Law was very helpful in discussing the corrective factor question for unreported violence. The trek through Markov chains and other group prediction schemes was assisted by Dr. Terrill Holland of the California State Department of Corrections at Chino. Other individuals, too numerous to name, are owed thanks for the development of this book and for this, the author is deeply grateful.
--Harold V. Hall