Abstract
AbstractThis study compared the ability of seven statistical models to distinguish between linked and unlinked crimes. The seven models utilised geographical, temporal, and modus operandi information relating to residential burglaries (n = 180), commercial robberies, (n = 118), and car thefts (n = 376). Model performance was assessed using receiver operating characteristic analysis and by examining the success with which the seven models could successfully prioritise linked over unlinked crimes. The regression‐based and probabilistic models achieved comparable accuracy and were generally more accurate than the tree‐based models tested in this study. The Logistic algorithm achieved the highest area under the curve (AUC) for residential burglary (AUC = 0.903) and commercial robbery (AUC = 0.830) and the SimpleLogistic algorithm achieving the highest for car theft (AUC = 0.820). The findings also indicated that discrimination accuracy is maximised (in some situations) if behavioural domains are utilised rather than individual crime scene behaviours and that the AUC should not be used as the sole measure of accuracy in behavioural crime linkage research.
Highlights
Behavioural crime linkage (BCL) seeks to address serial offending by linking crimes based on the fact they share similar offender crime scene behaviours (Woodhams, Bull, &Hollin, 2007)
The findings indicated that discrimination accuracy is maximized if behavioural domains are utilized rather than individual crime scene behaviours, and that the Area Under the Curve (AUC) should not be used as the sole measure of accuracy in behavioural crime linkage research
Within the BCL literature, it has been suggested that statistical methods such as those tested in the current study could be applied to a police crime dataset and used to rank order crime pairs that are behaviourally similar, thereby providing the analyst with an evidence-based way of focusing their attention on those crimes that are most likely to be linked (e.g., Tonkin et al, 2017)
Summary
Over the last 15 years academic research has sought to develop statistical methods for linking crimes that could support BCL in practice, with the hope that such methods might (in the future) be incorporated within computerized decision-support tools (e.g., see Tonkin et al., 2017) These tools would be able to process vast quantities of crime scene information in a quick and efficient manner, highlighting to law enforcement practitioners those crimes that are most likely to be linked and providing a summary of behavioural similarities and differences between the various crimes (e.g., Canter & Youngs, 2008; Grubin et al, 2001; Oatley, Ewart, & Zeleznikow, 2006). Would this help analysts to process more cases in less time, but it may help to increase BCL accuracy by reducing the cognitive load on analysts.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
