The management of sex offenders risk of committing sex crimes is of paramount importance to the criminal justice system. Criminal justice and treatment professionals assess risk of sexual recidivism using validated risk assessment tools such as the Rapid Risk Assessment for Sex Offender Recidivism (RRASOR), the Structured Anchored Clinical Judgment- Minimum Version (SACJ-min), the STATIC-99, and the STATIC-2002. However, these scales were created using data primarily from sex offenders released from prison or institutions for dangerous sexual predators. Moreover, although these scales are more accurate than clinical or professional judgment, their accuracy still is only modest and needs improvement. The aim of this research was to identify risk factors and how to combine risk factors to improve standardized risk assessment tools. The driving assumption of the current work was that sex offenders sentenced to probation and sex offenders released from prison are very different on important criminal history, offense, mental health, and social characteristics related to risk of sexual recidivism. Therefore if these two populations are different the current assessment scales used may not be appropriate for probation populations or may not predict well certain subgroups of sex offenders such as sex offenders who also commit domestic violence against adult intimate partners. Moreover, prior research has never considered how supervision may modify the behavior of sex offenders and change the risk factors that predict sexual recidivism. Are the risk factors for sexual recidivism the same if sex offenders are under probation or parole supervision or are free in the community without any supervision?