Is Criminal Profiling Dead? Should It Be?
A new study examines the art and science of a famous practice.
Posted Apr 04, 2019
We have all seen criminal profiling on hit TV shows and films like Criminal Minds, The Silence of the Lambs, Cracker, and Mindhunter, and heard the legendary tales of “profilers” catching elusive serial killers and other horrific offenders. In each case, profilers are depicted as having psychic-like talents used to predict features of an unknown suspect based upon behaviors at the crime scene alone. So, how accurate is this portrayal?
Consider the case of the Mad Bomber, who terrorized New York City in the 1940s-50s until police, at their wits’ end, asked psychiatrist James Brussel to develop a “profile” to help them catch the responsible offender. As the story goes, Brussel reviewed the files and then provided police an incredibly detailed description of the bomber: a middle-aged man of Eastern European descent living alone in Connecticut; a disgruntled former employee of Con Edison, the city’s utility company.
Oh, and one more thing. Before he left the police station, Brussel told the detectives, “When you catch him, and I have no doubt you will, he’ll be wearing a double-breasted suit. And it will be buttoned.”
A month later, George Metesky, an unmarried Polish immigrant living in Waterbury, Connecticut, previously employed by Con Edison until a dispute in 1931, was arrested by police as the mastermind behind the New York City bombings. When police came to the door, Metesky was in his pajamas and went upstairs to change. When he came back downstairs, he was wearing a double-breasted suit, fully buttoned, just as Brussel predicted. Pretty incredible, right?
Or is it? Stories like these leave us thinking that profilers are a godsend to police work, helping cops catch bad guys left and right. However, many experts in the fields of psychology, criminology, and law enforcement have raised serious concerns about the practice, particularly in light of research which suggests criminal profiles are often as accurate as a coin toss. This has led many experts to call offender profiling “less than useless”, a total “waste of police time”, and a few other choice words not fit to print.
So just how accurate are profilers at solving crimes? This important question, raised in this magazine in 1976 by former Psychology Today Editor Colin Campbell, is critical to determining if profiling is as beneficial as pop culture would lead us to believe. Or if, as Colin Campbell wondered, profilers were any better at predicting behavior than a bartender.
To answer this question, David Farrington and I decided to undertake a massive study on the state-of-the-art (and science) of offender profiling. To do this, we painstakingly searched for every book, study, report, chapter, and magazine article written on offender profiling in the four decades between 1976 and 2016. We found a total of 426 publications! We then read and coded each one in order to answer, for the first time, what profiles have been developed, how, by whom, and how accurate the profiles truly are. The results of this study, published in the December 2018 issue of Psychological Bulletin, were quite intriguing.
Contrary to popular belief, most of the work on criminal profiling was not conducted by members of the famed FBI Behavioral Science Unit. In fact, just 7% of all profiling publications were authored by FBI Special Agents and profilers. Most “profilers” are actually psychologists (43%) or criminologists (17%), and the majority of offender profiling publications came out in the last decade, between 2006 and 2016. None of those were authored by profilers or Special Agents in the FBI.
When looking at how these authors developed their profiles, it was clear there was a big divide. Some profilers used a scientific process involving statistical analysis of data on crimes and the responsible offenders. Others used a more clinical process where profiles were developed using what FBI profilers described as “experience, intuition, and educated guesswork.” In fact, our study found that over half of the profiling publications used no statistical analyses, while just over a quarter used advanced statistics and data to develop their offender profiles.
So how does this variation in the backgrounds and methods used to develop profiles impact profiling’s accuracy? While very few studies (two, to be exact) have measured the impact of offender profiling in the field, several studies examined profiling’s accuracy through other methods.
For instance, Anthony Pinizzotto asked police to evaluate all 192 profiles created by the FBI from 1971 to 1981. The police said that just 17% of the profiles directly aided in the identification of a suspect, but another 17% felt that the profiles “were not useful at all.” In a similar study from the United Kingdom, 83% of police officers felt that the profiles they received were operationally useful, but only 14% helped solve the case. In Canada, 40% of 51 surveyed officers who received offender profiles reported that a profile directly helped solve a case.
Results of the famous “Coals to Newcastle” study found that the predictions made by profilers were accurate about 66% of the time. However, the profiles led to an arrest in just 5 of the 184 cases. In other words, there was just a 2.7% success rate when the profiles were applied out in the field. But how many crimes would have been solved if police didn’t use the profiles?
To help answer this, we conducted the first scientific experiment on the effect of profiles used in active police investigations. To do this, we trained one police department to use new statistical profiles in their investigations, while three nearby departments continued their standard policing techniques. One year later, we compared results.
Our experiment found that the police department using the profiles solved over 260% more crimes compared to the departments not using the profiles. This experiment provides the best empirical evidence of the real-world effects of profiling, and they are quite positive. However, results were based upon statistically generated profiles and may differ for profiles developed using less sound statistical methods.
While there is still much more to be done to fully understand offender profiling’s true impact on unsolved cases, we are one step closer now to answering Dr. Campbell’s iconic question: Are profilers any better than a bartender at predicting human behavior?
While results are mixed, it appears that statistically generated profiles are more likely to lead to positive results than profiles created using subjective clinical methods alone. I hope the field continues on the current trajectory and transforms what some have referred to as a “pseudoscience” into a vibrant evidence-based discipline used to help police better understand criminal behavior—and make the world a safer place.
Campbell, C. (1976). Portrait of a mass killer. Psychology Today, 9, 110–119.
Copson, G. (1995). Coals to Newcastle? Part 1: A study of offender profiling. London, England: Home Office, Police Research Group.
Douglas, J. E., & Burgess, A. E. (1986). Criminal profiling: A viable investigative tool against violent crime. FBI Law Enforcement Bulletin, 55, 9–13.
Fox, B., & Farrington, D. P. (2012). Creating burglary profiles using latent class analysis: A new approach to offender profiling. Criminal Justice and Behavior, 39, 1582–1611.
Fox, B., & Farrington, D. P. (2015). An experimental evaluation of the utility of burglary profiles applied in active police investigations. Criminal Justice and Behavior, 42, 156–175.
Fox, B., & Farrington, D. P. (2018). What have we learned from offender profiling? A systematic review and meta-analysis of 40 years of research. Psychological Bulletin, 144, 1247- 1274.
Jackson, J. L., van Koppen, P. J., & Herbrink, J. C. M. (1993). Does the service meet the needs? (Report NSCR WD94-03). Leiden: Netherlands Institute for the Study of Criminality and Law Enforcement (NISCALE).
Pinizzotto, A. J. (1984). Forensic psychology: Criminal personality profiling. Journal of Police Science and Administration, 12, 32-40.
Snook, B., Taylor, P. J., & Bennell, C. (2007). Criminal profiling belief and use: A study of Canadian police officer opinion. Canadian Journal of Police and Security Services, 5, 1-11.