Cover Story

There's no easy way to test whether professional profilers are better than nonprofilers at identifying perpetrators. However, a few psychologists have tried.

In a 1990 study published in Law and Human Behavior (Vol. 14, No. 3), Anthony Pinizzotto, PhD--formerly of Georgetown University, now with the FBI's Behavioral Science Unit--and Norman Finkel, PhD, of Georgetown University, tested a group of FBI profilers, police detectives trained by the FBI, other police detectives, clinical psychologists and students. The researchers gave each group detailed case materials from two solved crimes (a murder and a rape), asked the participants to write profiles of the type of people likely to commit such crimes, and then compared the profiles with the actual, convicted offenders. The results were mixed.

The trained profilers wrote longer and more detailed profiles, and their profiles of the rapist were more correct than any other group's. In the murder case, however, they fared no better on average than the nonprofilers did.

In 2003, Australian forensic psychologist Richard Kocsis, PhD, published in the International Journal of Offender Therapy and Comparative Criminology (Vol. 47, No. 2) the results of a series of studies that replicated and expanded on Pinizzotto and Finkel's work. Kocsis used solved arson and murder cases to test groups of profilers, undergraduate science students, psychologists, police recruits, experienced police personnel, arson investigators, psychics and random control participants. Kocsis chose those groups because the nonprofilers represented different skills typically considered essential for profiling--investigative experience (experienced police personnel and arson investigators), behavioral knowledge (psychologists), logical reasoning (science students) and intuition (psychics).

This time, the professional profilers made more correct predictions about the offenders than any other group. But they weren't uniformly good at their jobs--they also had the highest statistical variation among any of the groups. Interestingly, overall, the science students did the second-best job. Kocsis says this indicates that a capacity for logical reasoning is a particularly important profiling skill.

The studies provide interesting and provocative food for thought, say some psychologists, but like any lab experiment, they have their limits.

"Profilers would say, 'That's not really the conditions under which we do our thing,'" says forensic psychologist Robert Homant, PhD, of the University of Detroit Mercy. "If I really believed that profiling had a lot to offer, these studies wouldn't convince me otherwise. And if I believed that profiling was just dressed-up astrology, they probably wouldn't convince me that profiling is good. It's a little too lacking in external validity."

--L. WINERMAN