Correct. It’s also easy to see that I’m not trying to prove him wrong, as I have now stated thrice within this thread (including one direct response to one of your posts). Rather, I’m asking him to prove himself to be right.
I am not claiming that he’s wrong, nor am I saying that he’s right. He made the claim. It’s up to him to substantiate it.
The utility is use of resources. If you have ten full-time investigators, and a city of eight million people, you cannot open the phone book and start with the ‘A’ names. but you can begin in the middle of Hubbard Street and expand outwards. You haven’t ruled anything out – you’re simply rolling the dice, in the same way that a child abduction case causes the police to investigate known sex offenders in the area. They don’t rule out others, but experience has taught them that recidivism for child sex offenders is high and it is often fruitful to start with them.
So how do you know how large the sample size was?
I’m not asking you to “cite a negative.” I’m asking what evidence you studied to arrive at your conclusion. In the language of a high school math class, “Show your work.”
If I went out and spent two hours studying profiling and examining cases that might be relevant, I could come back here to explain them and only then find out that you’ve already studied and found them wanting for some reason. Be kind to your fellow Dopers and prevent them from doing research you’ve already done.
But I suspect you haven’t studied anything. Your evidence amounts to: “Profiling hasn’t helped solve any cases because it’s bunk. How do I know it’s bunk? Because it has never helped solve any cases. How do I know it has never helped solve any cases? Because it’s bunk. How do I know it’s bunk? Because it has never …”
Until we see an actual profile, or have authoritative data as to what one contains, your example is absolutely and totally useless. It may resemble an actual profile-it may not. There is an easy way to find out, but you don’t seem to want to go down that path for some reason.
It’s in Gladwell’s article. The data that John Douglas and Robert Ressler used to generate their conclusions was based on interviews with 36 killers. Not only was that sample small, it was also non-random (all subjects were in California prisons and were selected simply if they were willing to talk), and there was no scientific protocol for the interviews, just informal chats.
Gladwell’s article points out that the FBI is reevaluating the data and has found it to be inaccurate in some areas (Douglas’ whole division of killers into “organized” and “disorganized” types has been falsified, for instance.
Informal interviews with 36 self-selected subjects is not enough to generate statistically reliable predictions.
One which is claimed as a success for profiling is David Meirhofer:
… but I know nothing about that investigation other than the (mostly unsourced) wiki article.
I wasn’t asking you about Douglas and Ressler.
I said:
I said nothing about Douglas or Ressler in there. Nor was I even thinking about Douglas or Ressler.
But you’re starting with statistically invalid assumptions, so it’s irresponsible to roll the dice with them. The statistics said the DC snipers would be one white guy. It was two black guys. The statistics said that the Unabomber would be uneducated and unintelligent. He was a Mathematics professor with a PhD from Harvard.
See above.
Why?
I answered this concern:
What is the flaw in that reasoning?
Whether you were thinking about them or not, that’s where the data for psycholgical profiling comes from.
Your scenario does not even really involve profiling, just the obvious fact that killers are going to operate within some kind of reasonable distance from where they live.
One of the sources for the article was Tru Crime Library writer Katherine Ramsland, who wrote “The kidnapping of a 7-year-old girl yielded no physical evidence but a viable suspect, David Meirhofer. Yet he was well-groomed, courteous, and educated, and he passed a polygraph. Although they eventually got him for killing a young woman in the area, he would not confess to knowing anything about the little girl. The local police who had consulted the profilers were ready to pass on him, but the BSU team who had looked at the crime remained convinced that Meirhofer was the guy. After killing the child, whom he’d kept imprisoned for a period of time, he did confess and also added the murder of two boys. Then he committed suicide. The profilers were sadly vindicated.”
Apparently, the police already had a suspect by the time the profiler came in.
The wiki article claims profiling caught him, but then goes on to say he was caught after calling the mother of one of his victims and “she obtained enough information to help the FBI track him down.”
What does that have to do with the profile?
So if you were to learn of another study, one that involved more than 36 individuals, conducted with more rigor than casual conversations, and conducted by persons other than Douglas or Ressler, you would change your mind?
Well, I contend that this is what is meant by “profiling,” and again point out the decreasing distance between the later dump sites and the residences, a point not explained merely by “killers are going to operate within some kind of reasonable distance from where they live.”
So: if you were to learn of another study, one that involved more than 36 individuals, conducted with more rigor than casual conversations, and conducted by persons other than Douglas or Ressler, would you change your mind?
The Wiki article is too garbled to make any conclusions one way or the other about the role of profiling in that crime. I merely mention it because it is passingly referenced elsewhere as an example of FBI profiling success.
What is needed is a more fulsome account of the case.
It would depend on what was being studied, how representaive the sample was, and how rigorously it was controlled. I am confident that no study of serial killers exists which has yielded sufficient data to be able to make statistically reliable predictions about personalities and demographics..i.e. personally identifiable traits outside of the crimes themselves.
Then you are incorrect in your understanding of the word. “Profiling” is supposed to tell you something about the subject’s personalty and demographics, not just the obvious fact that he will be physically proximate to his crimes.
Already answered.
On what do you base this claim?
Why is it an “obvious fact” that serial killers will operate near their homes?
OK. The study I’m referring to was David Canter and Maurice Canter, published in 1997. “Encounter and Death: The Spatial Behavior of U.S. Serial Killers,” Policing: International Journal of Police Strategy and Management, Vol. 20, pp 24-38. I don’t have an online link.
I ddin’t say they will operate near their homes, I said they will be physically proximate to their crimes. They have to be able to get to their crimes to be able to commit them, so assuming that they will operate within some kind of reasonably reachable distance of where they live would be fairly safe. That’s not psychology, just physics.
It seems pretty clear we’re going to watch this degenerate into a definitions dodge. A prediction is either “common sense” if true or “bunk” if it’s not.
This reminds me of Feynman’s jab at mathematicians. He was listening to a couple of topologists discussing some obscure claim, and one guy says, “It’s trivial!”
The other one says, “No, it’s not.”
“Sure it is. First, you have Heisen’s Continuity Principle, then you take a slice of the object and expand it through contour integration, then you …” blah, blah, blah, for ten minutes. At the end, the dissenter nods and agrees:
“Yeah, you’re right, it’s trivial.”
Feynman and the physicists joked that to mathematicians, “trivial” meant “proved,” and that therefore mathematicians could only ever prove trivial facts!
Same deal here. Any insight a profiler offers is either “common sense” or bunk. I don’t agree that it’s self-evident that killers’ dump sites decrease in distance from their residences over time, but that’s what Canter and Godwin’s study says.
David Canter and Maurice Godwin, rather.