本研究之目的在以檢視信度方式嘗試建立一本土之性侵者動態再犯危險評估量表，以供國內性侵者之治療、處遇、及評估人員能加以運用，使其對性侵者之治療與處遇能發展出有依據的評估與建議。本研究以加拿大發展之 SONAR及美國所發展之 SONTPS為工具，並以嘉義及高雄地檢署觀護人室現有之保護管束性侵者共 21人，每三個月由兩位觀護人對每位性侵者填寫 SONAR與 SONTPS量表一次，並於第二次填寫後以問卷詢問填寫之二量表簡易性與適合性問卷，並詢問相關之建議。並舉行焦點座談，以實際了解建議修改方向。研究結果顯示： (1)兩量表之百分比一致性差異不大，均達到近九成的百分比一致性。 (2)兩量表之明確性與簡易度之比較上 SOTNPS量表之明確性平均值為 1.714，簡易度平均值為 1.702，皆較 SONAR量表之穩定與急性因子明確性與簡易度 (1.500與 1.450)為高。 (3)建議可改用新版之 SOTNPS且可增刪部分題目，如增網路偏差行為、接近本案或潛在相似被害人之機會，刪除心理穩定度與改變階段一致性低題目，將更能掌握社區個案之動態，結合社區處遇的措施，讓觀護人與警政單位更能與輔導員全方位地關注性罪犯之社區生活，並能提高警覺以降低個案之再犯可能性。本研究之建議亦在討論之中。

This study was to establish a local dynamic sex offender risk assessment scale for the purpose of improving supervision and treatment in Taiwan. The SONAR and the SONTPS were used as tools to assess the sex offender on probation or parole. The probation/parole officer (PO) were trained to fill the tools every three months, and were asked to answer the simplicity and fitness of each item and their suggestion to change. The results show as the followings. 1. The consistency of percentage of the two scales was both close to ninety percent, and the difference was not significant. 2. The fitness and simplicity of the SONTPS (1.714 and 1.702, respectively) were higher than the ones of the SONAR (1.500 and 1.450, respectively). 3. It was suggested to add high consistency items, such as deviant internet behavior and having chance to be close to similar or original victim, and to delete low consistency items, such as mental stableness and stage of change. Some other suggestions and limitations were also presented.