Disentangling Criminal Profiling: Accuracy, Homology, and the Myth of Trait-Based Profiling
International Journal of Offender Therapy and
Comparative Criminology 2015, Vol. 59(3) 313 –332
© The Author(s) 2013 Reprints and permissions:
sagepub.com/journalsPermissions.nav DOI: 10.1177/0306624X13513429
ijo.sagepub.com
Article
Disentangling Criminal Profiling: Accuracy, Homology, and the Myth of Trait-Based Profiling
Richard N. Kocsis1 and George B. Palermo2
Abstract The scholarly literature over the past decade has chronicled a growing problem in the forensic technique colloquially called criminal profiling. The basis of this conundrum appears to originate from a concept referred to as “offender homology,” which presumes an inherent uniformity among offenders that is believed to underpin the analytic process incumbent to criminal profiling. Studies thus far conducted have apparently struggled to find evidence of offender homology, and based upon these findings arguments have been promulgated that various approaches to criminal profiling imputably labeled as “trait-based” are therefore not viable. Indirectly contradicting these arguments, however, have been studies testing profiler accuracy that have found evidence of individuals who appear to use trait-based methods but can nonetheless proficiently predict the characteristics of unknown offenders. Against this backdrop, the present article examines a number of tenets and disjunctions that appear to have arisen from research into offender homology and imputed to the practices of so-called trait-based profiling. The notion of whether trait-based profiling is, in fact, representative of profiling methods is examined and an integrative hypothesis proposed that attempts to resolve the quandary between offender homology and profiler accuracy.
Keywords criminal profiling, criminal investigative analysis, investigative psychology, profiler accuracy, offender homology
From modest beginnings, the technique colloquially known as criminal profiling has grown over time in both its use by law enforcement agencies and in terms of the
1Consultant Psychologist in Private Practice, Dee Why, New South Wales, Australia 2University of Nevada, School of Medicine, Henderson, NV, USA
Corresponding Author: Richard N. Kocsis, P.O. Box 662, Dee Why, NSW 2099, Australia.
513429 IJO59310.1177/0306624X13513429International Journal of Offender Therapy and Comparative CriminologyKocsis and Palermo research-article2013
314 International Journal of Offender Therapy and Comparative Criminology 59(3)
research activity it attracts (e.g., Kocsis, 2007a, 2009; Palermo & Kocsis, 2005). In this latter context, the literature over the past decade has chronicled a growing problem concerning the premise underpinning the practice and whether criminal profiling, as it is traditionally conceived, is in fact possible.
This conundrum appears to originate from a concept that has come to be known as the “homology assumption.” It should be noted that the broad proposition concerning the homology assumption is, in and of itself, not entirely unreasonable in that there is some basis to the suggestion that the technique of criminal profiling involves to some extent an underlying reliance, and thus assumption, on there existing to some degree a commonality in the characteristics of some offenders and the behaviors they may com- monly exhibit when perpetrating certain modes of behavior constituting a crime. It is this presumed dimension of commonality among some offenders which, to some degree, functionally underpins the retro-classification process that is described as “criminal profiling” (e.g., Kocsis, 2006a).
Following a somewhat rigid interpretation of these concepts, a number of studies have attempted to analyze various samples of offenders but have apparently struggled to find evidence of the anticipated commonality among them and their behaviors that would be congruent with the proposition of the homology assumption thought to underpin the technique of criminal profiling. Such findings have thereafter led to argu- ments that profiling methodologies that have been labeled as “trait-based”1 are unlikely to be functionally valid. While such arguments seem initially compelling, separate research investigating accuracy among profilers has rendered findings that appear to be incompatible with these propositions. Specifically, various studies have demonstrated the abilities of profilers to proficiently identify the characteristics of unknown offenders (e.g., Kocsis, 2006a, 2009, 2010, 2013; Kocsis, Middledorp, & Karpin, 2008; Pinizzotto & Finkel, 1990). Accordingly, the contradiction that emerges is that following the arguments promulgated from homology research, this proficiency observed among the profilers who use trait-based paradigms should not be evident as these methods are invalid as they are believed to be dependent on the unsupported concept of offender homology.