The analysis covered the demographics, psychiatric histories during incarceration and in the community, previous convictions and incarcerations, the current offense, parole history, subsequent offenses, and the results of Hare's Psychopathy Checklist. A disproportionate fraction of DSOs were working in the very low end of the socioeconomic scale, compared both to the general inmate population and the Canadian male population. DSO designations were also used disproportionately for child molesters. Findings from both sexual assessments and clinical opinions about individual DSOs indicated low reliability for the assessment process. Although predictive instruments suggested that DSOs were low-risk offenders, those who were incarcerated at the time of the study were uniformly high-risk. A simple retrospective psychometric clinical prediction strategy was 77-percent accurate in predicting postrelease recidivism. Twenty-three percent of the subjects received no regular therapy following their conviction. Findings confirm the conclusions of social scientists that preventive detention laws are unrealistic in their goals and assume results beyond current predictive capabilities. Figures, data tables, and 71 references. (Author abstract modified)