Pérez Ramírez Meritxell, Redondo Illescas Santiago, Martínez García Marian, García Forero Carlos, Andrés Pueyo Antonio
Universidad de Barcelona,Barcelona, Spain.
Psicothema. 2008 May;20(2):205-10.
Violent behaviour risk assessment is one of the most relevant research areas in current Psychology of Crime. Various scales for violence risk assessment have recently been developed from research about crime careers and risk factors. One of these instruments is the Sexual Violence Risk Assessment-20 (SVR-20), translated and adapted to Spanish by the Group of Advanced Studies in Violence of the University of Barcelona. The goal of this study is to verify the predictive capacity of the SVR-20 to predict sexual violence recidivism in a Spanish sample of sexual offender inmates. The method used was a retrospective study based in 163 sexual offender files and a 4-year time lag. The data were analysed with the logistic regression technique. Of the sample, 79.9% non-recidivist individuals were correctly classified, and 70.8% recidivist individuals. The ROC curve obtained for the model shows a very good discriminant capacity for the SVR-20, with a 0.83 AUC value. The main conclusion of this study is that the Spanish adaptation of SVR-20 is a good instrument to predict the risk of sexual violence.
暴力行为风险评估是当前犯罪心理学中最相关的研究领域之一。最近,基于对犯罪生涯和风险因素的研究,开发了各种暴力风险评估量表。其中一种工具是《性暴力风险评估量表-20》(SVR-20),由巴塞罗那大学暴力高级研究小组翻译并改编为西班牙语。本研究的目的是验证SVR-20在西班牙性犯罪囚犯样本中预测性暴力再犯的能力。所采用的方法是一项基于163份性犯罪档案和4年时间跨度的回顾性研究。数据采用逻辑回归技术进行分析。在样本中,79.9%的非再犯个体被正确分类,70.8%的再犯个体被正确分类。该模型得到的ROC曲线显示SVR-20具有非常好的判别能力,AUC值为0.83。本研究的主要结论是,SVR-20的西班牙语改编版是预测性暴力风险的良好工具。