Criminal profiling is a relatively new scientific field making its way into law enforcement. The goal of criminal profilers is to gather evidence from crime scenes and victim reports to try to put together an accurate description of the offender.
Profiling has worked on multiple occasions. In one case, a profiler put together a probably description of the "mad bomber" of New York City, saying that he is most likely not married, foreign, and around the age of 50. This description turned out to be right on, and the cops found the offender a lot faster than they would have without the profiler's description.
But does criminal profiling actually work? Or is this just an anecdotal occasion? Robert Homant, PhD, of the University of Detroit Mercy thinks criminal profiling is lacking in external validity. When put to the test in an experimental situation, criminal profilers wrote more detailed descriptions of offenders, but they were not always more accurate than other group's descriptions were. They were given both a rapist and a murder case, and the criminal profilers gave more accurate descriptions of the offender in the rapist case, but no more accurate descriptions than other groups of the offender in the murder case.
In another experiment, criminal profilers did well at describing offenders, but did not prove to be consistently good at their jobs. The profilers showed the most variance in their ability to profile criminals out of any of the groups who were tested. The group who did the second best at profiling was a group of college students in science, which leads many to believe that overall logical reasoning is required to be skilled in profiling.