Examining the Role of Similarity Coefficients and the Value of Behavioural Themes in Attempts to Link Serial Arson Offences

Abstract

When relying on crime scene behaviours to link serial crimes, linking accuracy may be influenced by the measure used to assess across‐crime similarity and the types of behaviours included in the analysis. To examine these issues, the present study compared the level of linking accuracy achieved by using the simple matching index (S) to that of the commonly used Jaccard's coefficient (J) across themes of arson behaviour. The data consisted of 42 crime scene behaviours, separated into three behavioural themes, which were exhibited by 37 offenders across 114 solved arsons. The results of logistic regression and receiver operating characteristic analysis indicate that, with the exception of one theme where S was more effective than J at discriminating between linked and unlinked crimes, no significant differences emerged between the two similarity measures. In addition, our results suggest that thematically unrelated behaviours can be used to link crimes with the same degree of accuracy as thematically related behaviours, potentially calling into the question the importance of theme‐based approaches to behavioural linkage analysis. Copyright © 2012 John Wiley & Sons, Ltd.

Using offender crime scene behavior to link stranger sexual assaults: A comparison of three statistical approaches

“…The inclusion of unsolved and apparent one-off crimes was important because, when practitioners are searching for linked crimes in practice, the databases they search contain a mixture of solved, unsolved, serial, and one-off offenses. By including such offenses in our research, this helped to ensure that the findings were more ecologically valid than those produced in the majority of previous research (which failed to include unsolved and one-off offenses; e.g., Bennell & Jones, 2005;Burrell et al, 2012;Ellingwood et al, 2013;Santtila et al, 2005Santtila et al, , 2008Tonkin et al, 2008;Woodhams & Toye, 2007).…”

“…Over the last decade, a growing body of research has sought to develop statistical methods that might underpin computerized crime linkage support tools (e.g., Bennell & Jones, 2005;Burrell et al, 2012;Ellingwood, Mugford, Bennell, Melnyk, & Fritzon, 2013;Santtila, Junkkila, & Sandnabba, 2005;Santtila et al, 2008;Tonkin, Grant, & Bond, 2008;Winter et al, 2013;Woodhams & Labuschagne, 2012;Woodhams & Toye, 2007;Yokota, Fujita, Watanabe, Yoshimoto, & Wachi, 2007). These studies have found support for the two theoretical assumptions that underpin crime linkage (behavioral consistency and distinctiveness 5 ) and have demonstrated moderate to high levels of accuracy when using offender crime scene behavior to distinguish between linked and unlinked offenses (see Bennell, Mugford, Ellingwood, & Woodhams, 2014, for a review).…”

“…The current study aims to overcome this fundamental limitation by comparing a variety of statistical methods in terms of their ability to distinguish between linked and unlinked crimes (referred to hereafter as discrimination accuracy). This follows a methodology originally developed by Bennell (2002), which has since been adopted in numerous peer-reviewed studies (e.g., Bennell & Jones, 2005;Burrell et al, 2012;Ellingwood et al, 2013;Tonkin et al, 2008;Woodhams & Toye, 2007). The methodology involves creating linked crime pairs (which contain two crimes committed by the same offender) and unlinked crime pairs (which contain two crimes committed by different offenders).…”

The practice of crime linkage: A review of the literature

“…Theoretical research has reflected the broad application of crime linkage, finding support for the principles of behavioural consistency and distinctiveness in burglary (Bennell & Canter, ; Markson, Woodhams, & Bond, ), personal and commercial robbery (Burrell, Bull, & Bond, ; Woodhams & Toye, ), sexual assault (Harbers, Deslauriers‐Varin, Beauregard, & van der Kemp, ; Santtila, Junkkila, & Sandnabba, ; Woodhams & Labuschagne, ), car theft (Davies, Tonkin, Bull, & Bond, ; Tonkin, Grant, & Bond, ), arson (Santtila, Fritzon, & Tamelander, ), and homicide (Salfati & Bateman, ; Santtila et al, ). Support for these principles has also been found using samples containing several crime types (Tonkin & Woodhams, ), both unsolved and solved offences (Woodhams & Labuschagne, ), one‐off and a series of offences (Tonkin, Santtila, & Bull, ), and using different methodologies and data from different countries (Ellingwood, Mugford, Bennell, Melnyk, & Fritzon, ; Tonkin et al, ). Although these empirical studies have found support for the theories of behavioural consistency and distinctiveness, this is often caveated with the notion that these theories do not hold for all offenders, and within all series, to the same extent (Woodhams & Labuschagne, ).…”

A Statistical Approach to Crime Linkage

“…There are many additional transformation functions that are appropriate for linkage data (e.g., groups of behavioral crime variables can be compared with Jaccard's coefficient (Melnyk et al, 2011) or a matching index (Ellingwood et al, 2012)). The selection of appropriate transformation functions is an important step in the linkage process as different transformations will provide different linkage performance (Bennell et al, 2010;Ellingwood et al, 2012;Melnyk et al, 2011;Salo et al, 2012) and can impact the class of models available for linkage. The next section shows how the resulting evidence variables are used to estimate the Bayes factor.…”