The Criminally Insane: a Community Follow-up

In a landmark study, Careers of the Criminally Insane, Steadman and Cocozza investigated the behavior of patients who were released from institutions for the criminally insane in the State of New York. Protesting the court decision (Baxstrom v. Herold) mandating a transfer to civil hospitals for evaluation and possible release, state corrections officers claimed their patients were "too dangerous" for that degree of freedom. In the tradition of the Baxstrom study, Thornberry andJacoby's research makes a solid contribution to understanding the institutional decision-making process by which mentally ill criminal offenders the "mad and bad" are confined over long periods of time. Like Steadman and Cocozza's study, the current research takes advantage of a natural experiment resulting from a court deciSion. In 1969 a Pennsylvania court ruled that because of an absence of due process in the commitment of patients to Farview Hospital (a maximum security mental institution), the continued incarceration of Donald Dixon and others whose sentence for criminal conviction had expired was unconstitutional. Again paralleling Baxstrom's outcome, the Dixon court required that sentence-expired mentally ill offenders be transferred to civil mental hospitals where, after re-evaluation, decisions about their release to the community would be made. The transfer was resisted by the staff at Farview who claimed their patients were "too dangerous" for even civil hospitalization. Accuracy of Predictions The project first investigates the accuracy of predictions that the patient will behave "dangerously." Thornberry andJacoby suggest that assessments of patient dangerousness are actually "political predictions" based not on the characteristics of an individual, but on the assumed characteristics of a group to which the individuals belong. Because the group is thought to have a high probability of producing violent or assaultive behavior, each individual member is assumed to be "dangerous." Further, the political context of decision-making allegedly encourages clinicians to "over-predict" dangerousness to avoid public criticism and sanctions resulting from the release of someone who turns out to be harmful to others. Consistent with their observations are prior studies finding assessments of patient dangerousness resulting in very high rates of false positives,/:e., those predicted to be dangerous who are not actually assaultive. While the authors describe two components of "political prediction" evaluator's perception and situational pressures they are unable to address the relative contribution of each to the outcome because neither one is carefully measured. The influence of situational pressures (resulting in decisions to continue confinement) is simply assumed to exist. The attributional process (by which group