I came across a great and brief article written by Richard Kocsis.
Richard Kocsis has written over 70 articles and works as a psychologist and criminologist.
In this article published in The Lancet, Kocsis reflects in 2004 on the scientific evidence of profiling. Since then the evidence base has grown somewhat, but Kocsis argues for more critical assessments that would help place offender profiling well within the definition of established science.
Kocsis also mentions a major study he did where several people from varying fields contributed their knowledge to a problem and profiling criminals. By the conclusion of the study, profilers came out on top even above seasoned detectives. However, Kocsis argues that more still needs to be done to ensure the high accuracy of profiling statements.
One of the outcomes of that study was the realization that not all profilers are created equal, there was a large amount of variance between the profilers, and it is this which Kocsis sees where the danger lies.
If you're unfamiliar with Kocsis as I was, might be worth checking out his work, especially since his work in this area has been so prolific.
Article: 'Profiling the criminal mind: does it actually work?'
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(04)17623-X/fulltext
Thanks for bringing this to my attention, @Daniel Sumner. It’s quite interesting to see the notion that the physical features reflect people’s personality prevails since ancient times.
On another note, the view that "Proficient profiling was most likely to arise from an individual's fundamental capacity for logical and objective reasoning" got me thinking about whether empathy is a good predictor for proficient profiling as well.