Allan Michael, Grace Randolph C, Rutherford Bronwyn, Hudson Stephen M
Department of Psychology, University of Canterbury, Private Bag 4800, Christchurch, New Zealand.
Sex Abuse. 2007 Dec;19(4):347-67. doi: 10.1177/107906320701900402. Epub 2007 Sep 14.
To explore the relationship between dynamic risk factors and recidivism in child molesters, we studied a sample of men (N=495) who completed an intensive, prison-based treatment program in New Zealand. During the follow-up period (M=5.8 years), 9.9% were reconvicted for a sexual offense. A self-report psychometric battery was administered at pre-treatment that assessed a range of variables related to sexual attitudes and beliefs, emotional functioning, and interpersonal competency. Factor analysis showed that individual differences in the battery could be described by four dimensions-Social Inadequacy, Sexual Interests, Anger/Hostility, and Pro-Offending Attitudes. Factor scores for each dimension were significantly correlated with sexual recidivism. Logistic regression analyses confirmed that the Sexual Interests and Pro-Offending Attitudes factor scores, as well as an Overall Deviance score which combined the dimensions, provided significant additional validity for predicting recidivism beyond the Static-99 (Hanson and Thornton Law and Human Behavior 24:119-136, 2000). When added to the Static-99, the Overall Deviance score increased the area under the Receiver-Operating Characteristic curve (AUC) from 0.72 to 0.81. These results show that psychometric self-reports can provide valid measures of dynamic risk factors, and that inclusion of such measures can improve risk prediction beyond that achievable by static factors alone.
为了探究儿童性侵犯者的动态风险因素与再犯之间的关系,我们对一组男性(N = 495)进行了研究,这些男性在新西兰完成了一项基于监狱的强化治疗项目。在随访期(平均5.8年)内,9.9%的人因性犯罪再次被定罪。在治疗前进行了一次自陈式心理测评,评估了一系列与性态度和信念、情绪功能及人际能力相关的变量。因素分析表明,该测评中的个体差异可由四个维度来描述——社交不足、性兴趣、愤怒/敌意和支持犯罪态度。每个维度的因素得分与性再犯显著相关。逻辑回归分析证实,性兴趣和支持犯罪态度因素得分,以及一个综合了这些维度的总体偏差得分,在预测再犯方面,除了静态-99量表(汉森和桑顿,《法律与人类行为》24:119 - 136,2000年)外,还提供了显著的额外效度。当将总体偏差得分添加到静态-99量表中时,接受者操作特征曲线(AUC)下的面积从0.72增加到了0.81。这些结果表明,自陈式心理测评能够为动态风险因素提供有效的测量,并且纳入这些测量能够改善仅靠静态因素所实现的风险预测。