Offender Profiling

[Matt]

From Wikipedia (http://en.wikipedia.org/wiki/Offender_profiling#Definitions)



Offender profiling is a method of identifying the perpetrator of a crime (http://en.wikipedia.org/wiki/Crime) based on an analysis of the nature of the offence and the manner in which it was committed. Various aspects of the criminal's personality makeup are determined from his or her choices before, during, and after the crime. This information is combined with other relevant details and physical evidence (http://en.wikipedia.org/wiki/Physical_evidence), and then compared with the characteristics of known personality types and mental abnormalities to develop a practical working description of the offender.


We all know of the practice from TV crime dramas such a Cracker, Profiler, the X files. Whilst we as skeptics take the dramatised details from these shows with a pinch of salt, recognsiing a deus ex machina or two, we also know that psychological profiling really goes on, and it all sounds so plausible. Our faith in the system leads us to believe that it must therefore have some practical benefit.

At least that's what I thought Until googling for Ian Rowland lead me to Malcom Gladwell's New Yorker Article skewering this practice.

It's a must read so I'll wait whilst you at least skim it.

http://www.newyorker.com/reporting/2007/11/12/071112fa_fact_gladwell?currentPage=all

Back already? Then I'll continue. I realise that one article isn't the be all and end all of a subject. "Psychological profiling is bunk" plays much better in the press than "Interestesting questions raised about psychological profiling" and so I'd expect some journalistic bias. However I'd never considered before the degree to which our cherised crime fighting institution might have deluded themselves about the efficacy of certain techniques.

I wouldn't say from reading the article that it's all bunk, however it does raise the possibility.

Is psychological profiling simply the codification of predjudice? Street criminals are black, serial killers are white. If so how aware are investigators of the potential failings of these techniques. Do they offer practical benefits that outweigh the false potives and negatives.

What do others here think?

[ZERO]

I have one of the books mentioned in the article and accepted profiling as effective.

I was quite surprised to get the impression that it is little more than educated guess work.
It's interesting that the hits are remembered but the misses are not mentioned.
Sounds familiar. ???

This quote from Laurence Alison is quite damning.


the way the profiles were written --snip-- so full of unverifiable and contradictory and ambiguous language that it could support virtually any interpretation.

[Mulder]

I dislike cop shows that use profilers and/or mediums because they're always too ridiculous for words. It's nice to have your prejudices verified. :smiley:

[JJM]

I read John Douglas' book when it was new. In addition to the vagueness (as in the quote in post #2) typical of their profiles, I realized that Douglas had thousands of cases to draw from. Given that, picking a half-dozen that make him look really good should be easy.

[Graham Lappin]

I am not so skeptical on offender profiling as some others (putting on a tin hat and diving for cover).

I am not expert on offender profiling per se, but I would defend psychology generally as a genuine science. It applies the scientific method but at the same time recognises that it cannot be as rigourous as, say physics or mathematics. Providing it is understood that the error bars might be rather large, the application of criminal psychology does seem to make a genuine contribution. I suspect there are those that are good at it and those that are not so good at it, but that can apply to any science.

[JJM]

I am not so skeptical on offender profiling as some others (putting on a tin hat and diving for cover).

I am not expert on offender profiling per se, but I would defend psychology generally as a genuine science. It applies the scientific method but at the same time recognises that it cannot be as rigourous as, say physics or mathematics. Providing it is understood that the error bars might be rather large, the application of criminal psychology does seem to make a genuine contribution. I suspect there are those that are good at it and those that are not so good at it, but that can apply to any science.I have no quarrel with what you wrote. My point is that the "profiles" they come up with are pretty near useless, and their few "successes" are randomly mixed with a majority of failures.

Less than ten years ago, the murder of Jon Benet Ramsey (ca. 5 years old) in her home in Denver Colorado was BIG news. John Douglas (retired from the FBI) was hired to create a profile of the killer. I don't have it at hand; but others noted (and I agree) that it described about 10% of the population of Boulder. A description that narrows the field to several hundred thousand people is not useful. The killer has not been caught.

[JJM]

I confused "Boulder" and "Denver" Colorado in my previous post. I think both should be "Boulder."

I would also like to expand: even if a "profile" accurately describes a criminal, if it describes thousands that are not criminals- it is not useful.

[Graham Lappin]

I think I must concede the point. I cannot claim to have a knowledge of this area but assuming that we accept that psychology is a science, then if the track record of offender profiling does not come up with with evidence that it works, then we must accept that it does not. It's just like any other clinical trial. It may mean that the technique needs refining and that with a better understanding over time it improves.

[JJM]

I think I must concede the point. I cannot claim to have a knowledge of this area but assuming that we accept that psychology is a science, ...Psychology is a science. I am a chemist, (sadly) many of my colleagues think the "softer" sciences are "easy." I know better. The more difficult it is to properly control an experiment, the more difficult the science is.

I don't know much about psychology; but, in between psych and chem is a subject I know a bit about- biology. I am regularly awed by the results biologists produce.

[Allo Allo]

I found the article (link on first post) that Matt started this thread with, fascinating.What surprised me most was the similarity that profiling has with how psychics work. From the Movies I got the idea that offender profiling was scientific. I would think a lot of ordinary people like me think that way. :sad: This is an interesting discussion - please carry on - I'm listening! :smiley:

[Graham Lappin]

Psychology is a science. I am a chemist, (sadly) many of my colleagues think the "softer" sciences are "easy." I know better. The more difficult it is to properly control an experiment, the more difficult the science is.

I don't know much about psychology; but, in between psych and chem is a subject I know a bit about- biology. I am regularly awed by the results biologists produce.

We agree then - apart from the bit about "sadly" being a chemist

[tkingdoll]

I've read a couple of books on this subject and consider it pretty much educated guesswork.

[Pebble]

I confused "Boulder" and "Denver" Colorado in my previous post. I think both should be "Boulder."

I would also like to expand: even if a "profile" accurately describes a criminal, if it describes thousands that are not criminals- it is not useful.

I think this depends on what one is trying to achieve and on the certainty achieved by the technique. Given that in any given crime the number of culprits is likely to be = 1, then a technique that narrows ones search to 10% of the population accurately has a specificity of 90%, and is thus a valuable screening tool.
The questions that profilers would have to answer to show that they are worth employing is that their profiles are only very rarely wrong (if there are frequent false negatives, the screening value is lost), and that the profile adds value to the other evidence available. For example if we knew that the culprit drove a red '67 mustang, and there were only 100 such individuals, only 10 of whom fitted the profile, it could add value, while if 10% of the population fitted the profile, but 95% of mustang drivers also did - then the profile is useless.

[tolman]

I was quite surprised to get the impression that it is little more than educated guess work.
Does it matter that it's educated guesswork?

Even if it is educated guesswork, surely the important things are
a) Whether the guesses are generally better than those of decent police detectives (or, at least, whether the guesses of some identifiable subset of profilers are consistently better than those of decent detectives).
b) If a), then how much of that education could be easily acquired by a decent detective learning a few basic techniques and information about typical patterns of criminal behaviour.

I suppose if a department was lumbered with a detective who wasn't much good, even if profiling isn't *that* useful, it might give a way for assistance to be provided/enforced without the hassle that might ensue if the assistance came from another detective.

[ZERO]

Does it matter that it's educated guesswork?
That is not how it is portrayed. It is passed off as highly accurate and the misses are not mentioned.


Even if it is educated guesswork, surely the important things are
a) Whether the guesses are generally better than those of decent police detectives (or, at least, whether the guesses of some identifiable subset of profilers are consistently better than those of decent detectives).
I guess some numbers would have to be crunched to answer that, but the picture the article paints (to my mind at any rate) is that profiling in reality adds little.
At the end of the day, the particular criminal, accurate profile or not, has to be tracked down by good police work. No profile can provide a name and address.



b) If a), then how much of that education could be easily acquired by a decent detective learning a few basic techniques and information about typical patterns of criminal behaviour.Do criminals really slot into neat categories is another point raised?

[tolman]

That is not how it is portrayed. It is passed off as highly accurate and the misses are not mentioned.
How it's portrayed generally (in fiction and "How I solved case X" stories) may not be too important. After all, lots of psychics claim to have been useful to the police, but how many police forces actually use them?

Presumably, there are only a small number of cases where it might be useful (where there are a lot of potential suspects, some focussing of follow up can save time, and follow-up could lead to evidence to rule people in or out).

Even if (especially if?) it worked well on such cases, it might well end up being overapplied, especially if it attracted people with excessive faith in it.

[ZERO]

How it's portrayed generally (in fiction and "How I solved case X" stories) may not be too important. After all, lots of psychics claim to have been useful to the police, but how many police forces actually use them?
I think it is important insofar as it's misleading.
Just like psychics.



Presumably, there are only a small number of cases where it might be useful (where there are a lot of potential suspects, some focussing of follow up can save time, and follow-up could lead to evidence to rule people in or out).Small number indeed.

-the British Home Office analyzed a hundred and eighty-four crimes, to see how many times profiles led to the arrest of a criminal. The profile worked in five of those cases. That’s just 2.7 per cent,-I wonder how much time was wasted on the 97.3% chasing wrong information?


Even if (especially if?) it worked well on such cases, it might well end up being overapplied, especially if it attracted people with excessive faith in it.With a success rate as low as the above quote, over application may not be a worry in the UK.

[tolman]

I think it is important insofar as it's misleading.
Just like psychics.But misleading to *who*?
At least with psychics, they try and drum up personal publicity for their regular business by claiming to help the police solve crimes. Hardly any civilians have any potential need for a profiler, so it's not quite the same there.

A good analysis of a suggested technique wouldn't just look at number of applications vs. number of successes. It'd start out by trying to establish if any of those cases were ones where it was (or should have been) thought there was little potential of it succeeding.
How many cases were basically unsuitable, how many were cases where nothing else had worked (and nothing seemed likely to work), how many cases ended up unsolved, etc.

I doubt many people would turn to profiling if they actually thought they had some decent suspects, which could certainly lead to it being a last-ditch technique on a fair number of hopeless cases.


I wonder how much time was wasted on the 97.3% chasing wrong information? It would certainly be interesting to see an analysis of the failures - how many times a profilie was just stating the obvious, and how many times it not only didn't help (due to vagueness, or the investigation succeeding on some other track, or people just ignoring it, or the profiler saying "I can't really do much with the information you have"), but how many times it actually sent people off on wild goose chases, or directed police away from the guilty.

It'd also be interesting to know how it's supposed to have helped in the cases where it was claimed to have helped. What things did it say that people wouldn't have thought of themselves, and were they down to deep psychological insights, or just analysis of similar past cases?

[Matt]

PZ Weighs in on the same article
http://scienceblogs.com/pharyngula/2010/06/i_am_so_glad_im_not_the_only_o.php?utm_source=feed burner&utm_medium=feed&utm_campaign=Feed%3A+scienceblogs%2Fpharyngula+%28 Pharyngula%29&utm_content=Google+Reader

[Admin]

I've attached a copy (registered members can download) of a real life offender profile issued by the Baton Rouge police in Louisiana.

It is indeed not too dissimilar to psychic style information given out by 'psychic detectives'.

It also contains a lot of vague and varied information and so is ideal for retrofitting.

[Drop Bear]

Fora bit of an insight into the mentality of profiling,I recommend 'Stigma' by Erving Goffman and a look at labelling theory generally.

http://en.wikipedia.org/wiki/Erving_Goffman


Offender profiling remains controversial:

http://en.wikipedia.org/wiki/Offender_profiling#Controversies