Another look at across-crime similarity coefficients for use in behavioural linkage analysis: an attempt to replicate Woodhams, Grant, and Price (2007) Tamara Melnyk a *, Craig Bennell a , Donna J. Gauthier a and Donald Gauthier b a Department of Psychology, Carleton University, Ottawa, Ontario, Canada; b Mechanical and Aerospace Engineering, Carleton University, Ottawa, Ontario, Canada (Received 4 June 2009; final version received 18 August 2009) In the absence of physical evidence, investigators must often rely on offence behaviours when determining whether several crimes are linked to a common offender. A variety of factors can potentially influence the degree to which accurate linking is possible, including the similarity coefficient used to assess across-crime similarity. The current study examines the performance of two similarity coefficients that have recently been compared to one another, Jaccard’s coefficient (J) and the taxonomic similarity index (D s ), using samples of two crime types, serial homicide (N 237) and serial burglary (N 210). In contrast to previous research, the results indicate that D s does not significantly outperform J with respect to linking accuracy. In addition, both coefficients lead to higher levels of linking accuracy in cases of serial homicide compared to serial burglary. Potential explanations for these findings are presented and their implications are discussed. Keywords: linkage analysis; taxonomic similarity index; Jaccard’s coefficient; ROC analysis; serial homicide; serial burglary Introduction Police investigators must frequently determine whether a series of unsolved crimes has been committed by the same offender (Grubin, Kelly, & Brunsdon, 2001). In the absence of physical evidence, such as DNA, links between crimes must often be established through an analysis of behavioural evidence (Woodhams, Hollin, & Bull, 2007). Using an investigative technique known as behavioural linkage analysis (BLA), an attempt is made to identify behavioural patterns across crime scenes to determine if the same offender is responsible for all the crimes. Researchers typically focus on two assumptions when considering whether it is possible to use crime scene behaviours to link crimes (Canter, 1995). First, it is assumed that offenders must exhibit relatively high levels of behavioural stability across their respective crime series, reflecting the degree to which each individual manifests the same behaviours across his or her crimes. Second, it is assumed that offenders must exhibit relatively high levels of behavioural distinctiveness for linking to be possible, whereby the actions that a given serial offender exhibits across his or her crimes differ from those exhibited by other offenders. The extent to which these *Corresponding author. Email: tmelnyk@connect.carleton.ca ISSN 1068-316X print/ISSN 1477-2744 # 2011 Taylor & Francis DOI: 10.1080/10683160903273188 http://www.informaworld.com Psychology, Crime & Law Vol. 17, No. 4, May 2011, 359380