Risk Assessment 1.35Forensic Psychology and Policy ImplicationsExperts have consistently agreed that predicting risk for future violent behavior is
an extremely difficult task (G. C. N. Hall, 1990; McGrath, 1991; Monahan, 1981).
However, it is likely that the courts will continue to turn to psychologists to provide
risk assessments, despite the difficulties noted in providing accurate predictions.
The courts have repeatedly ruled that expert testimony is permissible regarding
the predictions of violent behavior (Barefoot v. Estclle, 1983; Schall v. Martin, 1984;
United States v. Salerno, 1987). With this in mind, it is crucial for the mental health
professionals who provide risk assessments to the courts to uphold their ethical duty
and acknowledge the limitations of their expertise in making such predictions.
As illustrated in the case of Barefoot v. Estelle (1983), there are decisions that
juries make which require the risk assessment to be a primary consideration in
their final decision. Perhaps there are legislative and policy reforms that need to be
considered, given the overwhelming evidence that risk predictions are unreliable.
Suppose for a moment that Mr. Barefoot was in fact sentenced to death because
of the testimony by a psychiatrist who claimed to be able to predict with 100%
accuracy that imposing the death penalty was the only way to keep Mr. Barefoot
from committing another violent crime. There are currently no provisions against
an expert witness providing such testimony in court. Given the weight that the
judge and jurors give to expert testimony regarding predictions of dangerousness,
the criminal justice and the mental health systems would do well to place parameters
around the predictions that can be offered in court.Suggestions for Future ResearchTo date, no predictive models exist which can predict future dangerousness with
a high degree of certainty. Given that risk assessments continue to be commonly
requested of forensic psychologists, it is imperative that research continues to ex-
plore factors which are associated with future dangerous behavior. The research
has grown tremendously in the past 20 years on violence prediction; however, the
accuracy of such predictions remains extremely limited. The predictive models that
have been established thus far need to be tested in longitudinal studies across diverse
populations of offenders.
An additional area of research that has not received as much attention concerns
the jurors' decision-making process regarding expert testimony of risk assessments.
This would shed light on the impact that mental health professionals have when
providing risk assessments to the court. The decisions of jurors could then be
compared to the decisions of judges in this regard in order to establish whether the
judge is better able to consider the limitations of risk assessments when rendering
his or her final decision.