University of Liverpool, UK.
University of Central Lancashire, UK.
J Interpers Violence. 2021 Nov;36(21-22):NP11852-NP11876. doi: 10.1177/0886260519888522. Epub 2019 Dec 2.
Offender profiling follows the idea that if offenders' crime scene actions can be empirically linked to their background characteristics, it will be possible to predict one from the other. There is a lack of research exploring whether homicide offenders' crime scene actions are predictive of their criminal histories, despite the potential utility of such information. The current study addresses this gap in the literature. A sample of 213 adult male-on-female homicides with sexual or unknown motive was drawn from a U.K.-wide database. Relationships between 13 preconviction variables and 29 crime scene behaviors were explored using a bivariate statistical approach. Subsequently, binary logistic regression models were used to predict the presence, or absence, of specific preconvictions based on a of offense behaviors. Analyses highlighted 16 statistically significant associations between key offense behaviors and previous convictions, these associations were often "less likely" to result in previous conviction. The analysis failed to find any association for various other variables, most notably sexual preconvictions. Results indicate offenders' criminal histories can be predicted from their offense behaviors, though not all preconvictions may be similarly suited. Implications for practice are discussed.
罪犯侧写遵循这样一种理念，即如果能够将罪犯在犯罪现场的行为与他们的背景特征进行实证联系，那么就有可能从一个特征预测另一个特征。尽管此类信息具有潜在的效用，但缺乏探索凶杀案罪犯的犯罪现场行为是否可以预测其犯罪历史的研究。本研究填补了该文献中的空白。从英国范围内的数据库中抽取了一个 213 名成年男性对女性的凶杀案样本，这些案件具有性动机或动机不明。使用双变量统计方法探讨了 13 项定罪前变量与 29 项犯罪现场行为之间的关系。随后，使用二元逻辑回归模型根据犯罪行为预测特定定罪的存在或不存在。分析突出了关键犯罪行为与先前定罪之间 16 个具有统计学意义的关联，这些关联通常“不太可能”导致先前定罪。分析未能发现各种其他变量（尤其是性犯罪定罪）之间存在任何关联。结果表明，可以根据犯罪行为预测罪犯的犯罪历史，但并非所有定罪都可能具有相似的适用性。讨论了对实践的影响。
