Linking serial sexual offences: Moving towards an ecologically valid test of the principles of crime linkage

Abstract

Purpose To conduct a test of the principles underpinning crime linkage (behavioural consistency and distinctiveness) with a sample more closely reflecting the volume and nature of sexual crimes with which practitioners work, and to assess whether solved series are characterized by greater behavioural similarity than unsolved series. Method A sample of 3,364 sexual crimes (including 668 series) was collated from five countries. For the first time, the sample included solved and unsolved but linked‐by‐DNA sexual offence series, as well as solved one‐off offences. All possible crime pairings in the data set were created, and the degree of similarity in crime scene behaviour shared by the crimes in each pair was quantified using Jaccard's coefficient. The ability to distinguish same‐offender and different‐offender pairs using similarity in crime scene behaviour was assessed using Receiver Operating Characteristic analysis. The relative amount of behavioural similarity and distinctiveness seen in solved and unsolved crime pairs was assessed. Results An Area Under the Curve of .86 was found, which represents an excellent level of discrimination accuracy. This decreased to .85 when using a data set that contained one‐off offences, and both one‐off offences and unsolved crime series. Discrimination accuracy also decreased when using a sample composed solely of unsolved but linked‐by‐DNA series (AUC = .79). Conclusions Crime linkage is practised by police forces globally, and its use in legal proceedings requires demonstration that its underlying principles are reliable. Support was found for its two underpinning principles with a more ecologically valid sample.

Case linkage in Australian serial stranger rape

“…Many studies have examined case linkage accuracy of sexual assault using Jaccard's coefficients and ROC analysis (Bennell et al, 2009(Bennell et al, , 2010Mokros and Alison, 2002;Slater et al, 2015;Winter et al, 2013;Woodhams et al, 2007bWoodhams et al, , 2019Woodhams and Labuschagne, 2012). When assessing cross-crime similarity coefficients of linked and unlinked offence pairs, these researchers' Jaccard's coefficients for linked pairs ranged from 0.39 to 0.52 and unlinked pairs ranged from 0.17 to 0.34, supporting behavioural consistency.…”

“…The past 15-20 years have seen case linkage research and practice flourish, with support for linkage practice found across high-volume offences, as well as interpersonal offences (Bennell et al, 2010(Bennell et al, , 2014Burrell and Bull, 2011;Grubin et al, 2001;Labuschagne, 2006;Santtila et al, 2008;Slater et al, 2015;Tonkin and Woodhams, 2017;Tonkin et al, 2011Tonkin et al, , 2012Winter et al, 2013;Woodhams et al, 2007bWoodhams et al, , 2019Woodhams and Labuschagne, 2012;Woodhams and Toye, 2007). As research and investigative methods continue to develop, specialised databases have been created to assist in the case linkage process, such as the Canadian Violent Crime Linkage Analysis System (ViCLAS), the Violent Criminal Apprehension Program of the United States of America (USA) and Australia's Violent and Sexual Crimes Database (VSCD).…”

“…They found that the use of multiple behaviours resulted in higher linkage accuracy than reliance on behavioural themes (AUC = 0.84 versus AUC = 0.74, respectively). Furthermore, the inclusion of as much information as possible into a linkage analysis resulted in the highest success (Winter et al, 2013;Woodhams et al, 2019). Bennell et al (2009) and Slater et al (2015) used ROC analysis to examine different Jaccard's coefficients to identify the most appropriate threshold level.…”

“…It was previously argued that including all crimes within a series may lead to bias because of those highly prolific offenders who show either very high or very low consistency across their series, but case linkage practice includes several offences (Bennell et al, 2009). Slater et al (2015), Woodhams and Labuschagne (2012) and Woodhams et al (2019) have addressed this concern and increased ecological validity by incorporating all serial offences, as does this research project.…”

“…This method is not ecologically valid as a case linkage analysis will generally compare serial and non-serial offences, as well as solved and unsolved offences. Slater et al (2015), Winter et al (2013) and Woodhams et al (2019) addressed this by conducting ROC analyses using both serial only data and all offence data. Slater et al (2015) found a slightly higher (but not statistically significant) similarity coefficient for unlinked non-serial offence pairs compared to unlinked serial offence pairs and all studies found moderate to good AUC levels regardless of the sample used.…”