The use of dangerousness as a criterion for incarceration assumes someone's ability to make accurate predictions about the future occurrence of such behavior. However, a convincing body of literature makes clear the difficulties involved in predicting events with very low base rates. The author indicates, however, that these predictions “are accompanied by rather high rates of ‘false positive’ errors; that is, the great majority of the persons predicted as likely to engage in future violent behavior will not display such behavior.” Despite the possibility of such errors, social and political pressures are often brought to bear to incarcerate potentially dangerous persons on the presumption of “better safe than sorry.” The author notes that while the mentally ill have been targets for preventive detention, other sources of demonstrated dangerousness do not provoke comparable concerns (e.g., repeating drunken drivers).
In citing the work of Dershowitz, the author states that:
almost all organized societies appear to have employed some mechanism such as preventive confinement for the purpose of incapacitating persons who are perceived to be dangerous but who cannot be convicted for a past offense... The more a society circumscribes the formal criminal process with very tight safeguards that make criminal convictions difficult, the greater the likelihood that many dangerous persons will manage to benefit from these safeguards.
In American society, legal procedures delegated as “civil” are used to achieve greater leeway in the uses of preventive confinement. “By resorting to legal word games and verbal gymnastics, that which we would not and could not do to persons convicted of serious crimes and judged to be deserving of punishment, we manage to do quite readily to those designated as the recipients of our benevolence.”