The Baserate Problem and the Prediction of Dangerousness: A Reappraisal

The literature suggests that relatively few patients in maximum security mental hospitals are dangerous to others upon release. Because of the low baserate of post-release violence, it has been argued that future dangerous behavior can never be predicted with accuracy in the release decision context. The present article argues that the low baserate of violent behavior does not preclude, in principle, the development of more accurate predictive methods because each patient is considered for release on multiple occasions.