Predicting custodial suicides: Problems with the use of profiles

Publisher Website
Google Scholar
Abstract
Although the problem of jail and lockup suicides is nationwide, this social problem has begun only recently to receive the scholarly attention it warrants. A number of studies have advanced profiles of the typical jail suicide, but noncritical applications of these profiles may lead to problems if practitioners are unaware of their limitations. Different profiles often are mutually contradictory and seem to be more postdictive than predictive. Other problems include the potential for false positives, selective inattention, and the exposure of correctional authorities to inappropriate tort liability. Jail and lockup suicide data from Michigan for 1980–1985 were used to examine the implications of reliance on published suicide profiles. Only seven of 80 suicides were found to fit a national profile completely, while nine suicides would have been totally unexpected. It was concluded that because of the differences between various jail environments and inmate populations, it is unlikely that even a methodologically sound profile would be generalizable from one jail to the next. Some consequences of reliance on profiles are suggested.