Forecasting Harm: The Law and Science of Risk Assessment among Prisoners, Predators, and Patients


Scientifically valid instruments are being used for the first time to assess an individual’s risk of violence in criminal sentencing and in the civil commitment of mental patients and sexual predators. Risk factors on these instruments pertain to what the person is (e.g., gender), what the person has (e.g., personality disorder), what the person has done (e.g., past violence), and what has been done to the person (e.g., past victimization). In this Article, I argue that in criminal law, with its emphasis on blameworthiness for actions taken, the admissibility of scientifically valid risk factors is properly constrained to those that simultaneously index moral blameworthiness, i.e., to the defendant’s prior criminal conduct. In contrast, the civil commitment of people with mental disorder—a determination in which moral blameworthiness plays no part—should not constrain the admissibility of violence risk factors, except for those subject to strict Equal Protection scrutiny.

Finally, in the commitment of sexually violent predators, courts should keep evidentiary issues about the admissibility of violence risk factors apart from substantive questions about the constitutionality of the statutes that trigger risk assessment. If commitment as a sexually violent predator is properly categorized as civil commitment, the admissibility of violence risk factors in implementing such commitments should parallel the admissibility of violence risk factors in the civil commitment of mental patients. Disagreement with the substantive merits of sexually violent predator statutes does not justify depriving decision makers of the only kind of scientific evidence—empirically-validated actuarial violence risk assessment—that can effectuate their statutory goals.


Criminal Law | Criminal Procedure | Evidence | Health Law and Policy

Date of this Version

August 2004