A Theoretical and Practical Test of Geographical Profiling with Serial Vehicle Theft in a U.K. Context Matthew Tonkin, B.Sc., M.Sc. * , Jessica Woodhams, B.Sc., M.Sc., Ph.D. y , John W. Bond, B.Sc., D.Phil. z and Trudy Loe z Geographical profiling is an investigative methodology sometimes employed by the police to predict the residence of an unknown offender from the locations of his/her crimes. The validity of geographical profiling, however, has not been fully explored for certain crime types. This study, therefore, presents a preliminary test of the potential for geographical profiling with a sample of 145 serial vehicle thieves from the U.K. The behavioural assumptions underlying geographical profiling (distance decay and domo- centricity) are tested and a simple practical test of profiling using the spatial mean is presented. There is evidence for distance decay but not domocentricity among the spatial behaviour of car thieves from the U.K. A degree of success was achieved when applying the spatial mean on a case-by-case basis. The level of success varied, however, and neither series length in days nor number of crimes could account for the variation. The findings question previously held assumptions regarding geographical profiling and have potential theoretical and practical implications for the study and investigation of vehicle theft in the U.K. Copyright # 2009 John Wiley & Sons, Ltd. INTRODUCTION The locations of criminal events are one of the first and foremost pieces of evidence available to the police in their investigations. Not only are they relatively unambiguous pieces of information, but they can reveal important details about the perpetrator of the crime (Rossmo, 1997). Indeed, a substantial body of research now exists to suggest that offenders do not travel far to commit their crimes, with them preferring to offend in areas with which they are familiar (e.g., Canter & Gregory, 1994; Canter & Larkin, 1993; LeBeau, 1987; Phillips, 1980; Rhodes & Conly, 1991; Snook, 2004; Wiles & Costello, 2000). These findings have prompted researchers to explore whether the locations of crime can reliably predict the home base of a given offender (referred to as geographical profiling; see Canter, Coffey, Huntley, & Missen, 2000). Evidence has now been amassed to suggest that geographical profiling is possible with a number of crime types, including rape (e.g., Santtila, Zappala ´, Laukkanen, & Picozzi, 2003), murder (e.g., Canter et al., 2000), robbery (e.g., Harries, 1999), burglary (e.g., Sarangi Behavioral Sciences and the Law Behav. Sci. Law 28: 442–460 (2010) Published online 14 December 2009 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/bsl.916 * Correspondence to: Matthew Tonkin, B.Sc., M.Sc., Forensic Section, University of Leicester, 106 New Walk, Leicester, LE1 7EA. E-mail: matthewjtonkin@googlemail.com y School of Psychology, University of Birmingham, Birmingham, U.K. z Northamptonshire Police, Wootton Hall, Northampton, U.K. Copyright # 2009 John Wiley & Sons, Ltd.