Originally posted by JeffHamm
View Post
Well, it was the head of the local police who emphasized the Geordie accent. The FBI dismissed the tapes as a hoax. Also, the FBI were suggesting a number of occupations where the offender could be expected to be driving around at night, postal delivery just being one example they listed (which is quite different from them saying he was a postal delivery worker to the exclusion of other occupations).
That being said, I'm not suggesting their profiles are particularly useful as it's hard to know exactly how to evaluate them. Obviously, some points would be expected to be correct, some not, but at what point does something that doesn't match turn into an error? If they suggest he's a loner, but it turns out he went out a lot, but had no "close friends", is he a loner? Is that correct or not? When is a profile considered "wrong"? How many errors, if we can define them, are "too many"? I've not looked for awhile, but I've never found a study where that sort of thing has been analysed. We hear, of course, about the "successes" in books and media, and we hear about the dismal failures (the Beltway Snipers, for example), but those are the extremes, and are also a very small percentage of the profiles written. What would be useful would be an analysis of a large number of them, to see if they are providing useful information. And if they do, is that information broad and general (like the spatial profiles are - broad regions to consider, with quantifiable and known rates of success and failure?), in which case drop the specific details that just create noise. Or are they little more than a summary of the general information about offenders (probably male, probably 20-40s, etc), in which case they boil down to a general description of offenders while purporting to get beyond that and differentiate this offender from the general description.
There are some things, of course, that one can infer from a crime scene. Certain aspects may require a fair amount of strength to accomplish, so the offender must have that strength. Or in the case of the Snipers, the offender must have marksmanship skills from somewhere. We debate these sorts of things too when it comes to notions about anatomical knowledge, only in this case the medical opinion (then and now) is divided as to how much was required! But if someone is consistently making long range shots on victims, that at least becomes clear - they have marksmanship skills. It doesn't mean, however, they had to be military (though of course that would be one way those skills could be obtained), only that military, hunting, target shooting, gun enthusiast, etc background might be something to look for in a person of interest.
Those sorts of things are pretty obvious though. Where behavioural profiling claims it can go, though, is to find more obscure aspects of behaviour from the crime scene and get to further details about the offender's skill sets, personality, life traits, etc. And I don't think that has ever been studied to determine it can actually do anything like that at all.
Even the spatial analysis area needs a proper large scale study of efficacy, but these are very difficult to do because getting accurate information on a large number of cases is both expensive and complicated. While I've done some small scale testing, as have others, which all show things work, there is nothing like a large data set to really put things to the test. Also, most routines do about the same, so the large broad patterns are being detected, and it will take a large data set to see if the routines can be improved beyond current performance levels.
- Jeff
Could you summarise that in under a million words?
Were the FBI guys taking out of their arses or not? Or don’t you know?
Comment