Written by Samantha Kitt
There’s very little skepticism offered about criminal profiling within hit TV programs like Criminal Minds and Mindhunter – which, by the way, are both 100% worth a watch – however this has led to major public misconceptions about the (lack of) accuracy of criminal profilers. For those of you unfamiliar with this practice, criminal profilers are special investigators that infer an unsub’s (unknown subjects) personality, behavioural, physical and demographic characteristics based on crime scene data and evidence (Douglas, Ressler, Burgess, & Hartman, 1986).
Credit where credits due, shows like Criminal Minds and Mindhunter don’t get everything wrong, some aspects do accurately depict the way professional criminal profilers work. For example, their Behavioural Analysis/Science Units work to identify the unsub by determining the modus operandi (the MO: how) and whether or not they had a ‘signature’ (why), which are needed to fulfill the unsub’s emotional and/or psychological desires and fantasies (Turvey, 2011). Kind of unsettling, I know. These TV depictions were also right that profilers use analytics to predict the behaviour and traits of future criminals, as well as interviewing and researching convicted criminals to help predict future behaviour.
However, they do get some things wrong, and specifically, the accuracy of criminal profiles. Introducing Snook, Eastwood, Gendreau, Goggin, & Cullen (2007), who combined the findings from 4 independent studies to determine the scientific credibility of profiling (using a total of 981 participants!). They compared the accuracy of criminal profiles created by (a) professional profilers and individuals with investigative experience, and (b) individuals with no investigative or profiling experience. The accuracy measure of the profiles was broken down into four coded classes: (1) cognitive processes (e.g., whether the unsub showed remorse, or what the unsub’s motive(s) were), (2) offense behaviour (e.g., whether the unsub took items, or ‘trophies’, from the scene, or whether the unsub was careful to protect their identity), (3) social status and habits (e.g., level of education, military-experienced, marital status, or alcohol and drug use), and (4) physical characteristics (e.g., hair colour, race, age, gender, or height).
So, what did they find? The experienced profiler-investigator group only fared a little better predicting physical attributes, and actually did a worse job of predicting cognitive behaviour, social status and habits. As for predicting offense behaviours, neither group outperformed the other.
I think it’s safe to say these findings are pretty surprising, right? No matter the occupation, you’d expect an ‘expert’, or someone with experience in that field, to outperform your average Jane or Joe with no experience, wouldn’t you? But let’s back up a minute, we’re getting ahead of ourselves – this study wasn’t perfect, there are some limitations we need to consider. Most notably, the participants’ profile was determined by their responses on a multiple-choice questionnaire (MCQ) which they had to fill out in full in a single session without any access to any resources or references. The MCQ is not a problem in and of itself, it was undeniably thorough and covered all bases of the profile, and if anything, it allowed for a better representation of the individual’s profiling competence as they were required to infer substantial detail about the offender. The problem lies in the fact that this setting can’t reasonably be applied to any context outside of the study. Not only was the profile put together in one sitting, but the study also only presented case studies of serial arson, and while the heat was on them (sorry, I couldn’t hold that one back) there wasn’t as much pressure to produce an accurate profile as there would be in a true-crime situation.
Still, the results are pretty interesting. Profilers and experienced investigators were unable to outperform generally untrained individuals. In theory, profiling has the capacity to be a valid investigative strategy, but in practice, it clearly still needs improvement.
References:
Douglas, J. E., Ressler, R. K., Burgess, A. W., & Hartman, C. R. (1986). Criminal profiling from crime scene analysis. Behavioral Sciences & the Law, 4(4), 401-421. https://doi.org/10.1002/bsl.2370040405
Snook, B., Eastwood, J., Gendreau, P., Goggin, C., & Cullen, R. M. (2007). Taking stock of criminal profiling: A narrative review and meta-analysis. Criminal justice and behavior, 34(4), 437-453. https://doi.org/10.1177/0093854806296925
Turvey, B. E. (Ed.). (2011). Criminal profiling: An introduction to behavioral evidence analysis. Academic press.