Methodological considerations of determining dominance in multidimensional analyses of crime scene behaviours and offender characteristics

Abstract

Aimed at critically examining how offenders and offences are conceptualised, Investigative Psychology has seen a substantial increase in studies of crime scene behaviours and offender characteristics using multidimensional scaling (MDS) analyses that build thematic models of crime scene behaviours based on the co-occurrence variables and then testing the frameworks by determining if cases are dominant in a single thematic region. Although methods have been developed, issues that arise from using different methods and the effect on the results obtained have never been thoroughly discussed. Framed in the context of the evolution of how people have been classified in offender profiling research and elsewhere, this paper examines the established procedures used in MDS studies to determine dominance. Two methods—proportional versus the quantity of variables present—are compared across four levels of stringency within a framework of homicide offenders' criminal convictions. Results support the most common criterion used in prior studies of homicide and demonstrate that the subject of the framework or type of variables may dictate which method is more appropriate. The results strengthen the argument that a standard protocol for determining dominance should be adopted to increase consistency across studies to add to the development of Investigative Psychology. Copyright © 2009 John Wiley & Sons, Ltd.