Violent offenders had significantly more chaotic lifestyles and were diagnosed with greater psychopathology than sex or general offenders; they were also most likely to reoffend, with just over 25 percent committing additional violent offenses. The sexual offenders had personality characteristics similar to those of general offenders, with similar reoffending rates. More of the violent offenders were rated at medium risk or high risk for reoffending compared with the other two groups. The SHAPS was able to distinguish personality characteristics among the three groups. The Sexual Violence Risk-20 (SVR-20) significantly predicted any reconvictions over 5-year and 10-year followup periods. Study participants were recruited from referrals to a United Kingdom Regional Secure Unit for assessment between 1992 and 1995. Assessment protocols used during this period included the SHAPS. Of the sample of 153 male offenders, 139 completed the SHAPS. Using a retrospective archival research design, assessment protocols and psychology reports were examined by a forensic research psychologist. None of the participants was diagnosed as suffering from a psychotic illness at the time of assessment. Violent offenders had been convicted of a violent offense (actual or grievous bodily harm, murder, manslaughter, wounding, and common assault), with no history of sexual offenses or a sexual element to their offending. Sexual offenders had been convicted of a contact sexual offense (rape, attempted rape, or indecent assault); and general offenders were those convicted of a nonsexual, nonviolent offense. 4 tables and 41 references