RELATIONAL PROFILING: AN INVESTIGATIVE PSYCHOLOGICAL APPROACH TO IDENTIFYING VICTIM-SUSPECT RELATIONSHIP IN SOLVED HOMICIDE CASES FOR APPLICATION TO UNSOLVED CRIMES IN ENGLAND AND WALES

Although the relationships between homicide perpetrators and their victims have traditionally been recorded and reported by homicide research (e.g. family, friend, lover), few have attempted to connect these pre-existing relationships to victim and perpetrator behaviour during the homicide transaction. The current research is the first known to utilize more than 10 perpetrator action variables in a single relational analysis and more, the first to create comprehensive behavioural profiles of perpetrators by relationship typology (e.g. stranger, acquaintance, close) and relationship status (stranger, active, estranged) based on the presence or absence of 62 action variables within homicide crime scenes. Traditional methodologies for analysing behavioural information about homicide (e.g. multidimensional scaling, regression analyses, and qualitative comparative analysis) were not able to compute correlational data of this magnitude given the smaller sample size. The data sample for this project was derived from copies of 64 complete homicide police case files, previously collected from city and county police departments across the England and Wales spanning the years of 1985-1991, accessed from the Canter Archives at the University of Huddersfield, United Kingdom in 2011 (See Appendix D).The final sample included 87 suspects and 69 victims and it was the task of the current set of studies to compute how all 62 of the recorded homicide scene action variables correlated to relationship type and status between perpetrators and victims. Thus, for the first time in recorded homicide research, Fisher’s Exact Test was implemented in order to increase internal validity and pave the way for a more directive approach to psychological homicide research, coined for this paper as “Relational Profiling.” The purpose of the analyses across the five studies within the current dissertation wereto empirically establish to what degree the actual victim-offender relationships, the relational role the victims played for their offenders (Canter & Heritage, 2000), and relationship status between offenders and their victims impacted the outcome of crime scene actions in English and Welch homicides. Findings from Study 1 established external validity for the project with a comparative analysis of sample statistics to historical homicide statistics. Studies 2 and 3 supported the original hypotheses that homicide scene actions would connect back to relational circumstances between victim and convicted suspect. It was found that the relationship type between victim and convicted suspect prior to the homicide event (stranger, acquaintance, and close) did not have as strong as an impact on homicide transactions as was expected from prior research. Study 3 utilised the same stringent methodological parameters from the previous study, only the analyses controlled for relationship status (stranger, active, and estranged). The impact that this shift in categorization of known relationships had on the representation of crime scene action variables was remarkable and further validated this novel methodology for small homicide sample sizes. Studies 4 and 5 finalized the analysis by testing a long held theory in the field of Investigative Psychology, hypothesizing that the Narrative Action System framework (Canter& Youngs, 2009a) could further differentiate offenders in the way that they related to their victims (as objects, as vehicles, and as persons) (Canter & Heritage, 2000) as evidenced by behaviour. The methodological way this theory had been previously tested was, for the first time in recorded research, reversed, finding support for only the victim-as-object narrative. The implications of the results are discussed at length in the final Chapter, followed by a discussion of the limitations of this project and suggestions for future research into relational profiling.