How do you decide how dangerous a sex-offender is? Certainly all cases of sexual assault are appalling, but clearly some incidents are worse than others. In some places, teenagers who photograph themselves naked and send the pictures to their friends can be prosecuted as purveyors of child-pornography. While we may want to intervene in these cases, surely the action shouldn't be as drastic as when we're dealing with an adult who's a serial child rapist.
There are miles of gray area between these two extremes, and psychologists are often called on to make the tough judgment of how dangerous a individual might be. One common test is to attach a monitor to the offender's penis and then show them images of children and adults. In principle, true pedophiles will be more aroused by the children's pictures. But a convict applying for parole has a good reason to try to fake his response, and some people are inevitably misclassified, with potentially disastrous results.
Other methods, such as the Implicit Association Test, have also been tried, but these are also potentially subject to manipulation. So a team led by Anthony Beech decided to see if a different test could be used: The Rapid Serial Visual Presentation test, or RSVP. As we discussed on Monday, in an RSVP test, a distracting word or image is presented in a series of similar displays. If the viewer's attention is attracted by the distractor, he or she is more likely to miss a later image. As an example I've modified Monday's task. Can you spot the words naming a color (like blue, red, or green)? Ignore all the other words.
Click here to view the movie (QuickTime required)
Instead of words, Beech's team used photographs. They recruited convicted child molesters and other non-sex-related felons (from British prisons) to volunteer for their test. The volunteers were looking for four types of pictures: children or animals (the distractors), and chairs or trains (the targets). The rest of the pictures were neutral scenes or objects. The photos flashed by at a rate of ten per second, in sets of 11. At a random point one of the distractors would appear, just like in the example above. Two to three images later, the target appeared. Then respondents had to say what the distractor was, what the target was, and which direction the target was facing (left or right).
There's one other key to the study. Two types of child molesters were used in the study: intrafamilial, those who had attacked children in their own families; and extrafamilial, who had attacked children outside of their family. Extrafamilial molesters are considered more dangerous, generally preferring children to adults, while intrafamilial offenders prefer adults or have no preference. In this study, the extrafamilial group had attacked significantly more children compared to the intrafamilial group. How did they do on the test? Here are the results:
The child molesters did significantly worse when the distractors were children than when the distractors were animals. For the non-sex-offenders, there was no significant difference. So this test might actually be useful in identifying potentially dangerous criminals.
On the other hand, the researchers point out that there are some potential problems. Others who are interested in kids, like teachers, new parents, and grandparents, might also be distracted by pictures of kids. They've done some preliminary research suggesting this is not the case, but perhaps other non-dangerous populations would be mistakenly identified using this test.
Also, while the test might work for identifying groups of dangerous offenders, it's possible that it's still not reliable for determining whether a single individual is dangerous. It doesn't distinguish between intra- and extra-familial child molesters.
Finally, it seems to me that it would be possible to fake the results of this test -- all you'd have to do is deliberately mess up some of your responses to the animal distractors, and you'd appear to be just as accurate when the distractors were children.
Still, it's interesting to see that the attentional blink affects people differently depending on their interests (or deviancies, as the case may be). While it might be difficult or even impossible to use this test as a means of accurately figuring out which individuals are most dangerous, the test does offer some insight into how the mind works -- even when that mind is behaving in a way that is repulsive to (nearly) all of us.
Anthony R. Beech, Ellis Kalmus, Steven P. Tipper, Jean-Yves Baudouin, Vanja Flak, Glyn W. Humphreys (2008). Children induce an enhanced attentional blink in child molesters. Psychological Assessment, 20 (4), 397-402 DOI: 10.1037/a0013587
- Log in to post comments
"It doesn't distinguish between intra- and extra-familial child molesters." Maybe there isn't that much difference. Maybe people who abuse their own kids just have some other skills, such as being able to find a partner, or conceal their behaviour, or some additional lacks, such as having a protective instinct for "one's own."
http://sexoffenderissues.blogspot.com
Watch this video. It's called "Experiment on Homophobia!" And it states that most who are homophobic, usually have responses to the gay videos. And I am willing to bet, if 100% of the population took a similar test, all would be deemed a dangerous sexual predator.
http://www.youtube.com/watch?v=LmWwGZnX1Lw
We need to know both the average accuracy, AND the standard deviation. The information presented in the graph is useless without knowing the standard deviation. Otherwise, we have no idea how reliably discrimination can be done with a sample size of one.
http://sexofenderissues.blogspot.com
So why aren't my comments being posted? You don't agree with freedom of speech?
We need to stop the hysteria of using sex offenders as everyone's scapegoat, move away from being tough on crime, and be smart on crime.
When are we going to work on prevention?
Nothing about any of these laws will prevent another crime from being committed, and if you think it will, well, you are living in Wonderland.
Maybe I am badly misunderstanding the graph, but it looks to me as though the main difference between the groups is that the sex offenders are much less distracted by animals than non-offenders are. Indeed, the allegedly worse, extrafamilial offenders seem to be only very slightly more distracted by children than are the non-offenders (of course I can't judge significance from the information given). I can't imagine why sex offenders would be less distracted by animals, though (or less distractible in general), so I am inclined to think the results are nonsense. Is there something complex going on in the statstics that I am not getting? Why should it be the difference between animal vs. child distraction levels that matters rather than the differences between how much each group is distracted by child images?
According to neighborhood gossip (always reliable, no?) a man who has lived down the street from us for more than 20 years has been charged with pedophilia for looking at photos of children online.
The gossip is that he was not able to attend his daughter's wedding because there would be children at the event. Since he hasn't been convicted yet and isn't in jail, I wonder about this.
I don't think he's dangerous. He lives near the end of a dead end street where numerous children have played in the creek and woods, including my own. My youngest daughter was one of those who could not help but tell me everything. She was a joy to raise, I tell you.
The only complaint she had about anybody on the this dead end street was the guy next door who was "mean" and wouldn't buy girl scout cookies. Apparently he shut the door in her face.
How can we draw the line between fantasies (fueled by photos, perhaps) and the person who will actually act on such desires?
Acknowledging that this may be TMI, I have had fantasies about Sean Connery that would most certainly get me arrested if I tried to act them out. (Geez, that's dating me, isn't it? Harrison Ford is pretty hot too.)
A dangerous pedophile is one likely to act. One not likely to act... how dangerous?
This guy down the street from me doesn't concern me a great deal. He had ample opportunity to act and apparently didn't. Why should he be punished as severely as one who lacked his control?
Shaky ground, I know...
...it looks to me as though the main difference between the groups is that the sex offenders are much less distracted by animals than non-offenders are.
That's the first thing I noticed too. I haven't looked at the article, but according to what is reported here, it seems the researchers focused their analysis on the animal/child accuracy difference (e.g., target accuracy with animal as distractor - target accuracy with child as distractor) for each felon category, and not target accuracy differences for each felon category when the distractors were either animals or children. The latter analysis (going by the graph), would seem to suggest a main effect associated with felon responses to animals, and less so with their responses to children.
I can't imagine why sex offenders would be less distracted by animals, though (or less distractible in general), so I am inclined to think the results are nonsense.
The abnormal psych literature notes that serial killers and sex offenders have a history of harming and/or killing animals prior to their subsequent offenses, so perhaps that's the connection. I suspect the reduced distraction to animals evidences a conscious rejection of (normal) socioemotional attachments to animals (and likely people), with replacement with the more dysfunctional sexual attachment in sex offenders.
Others who are interested in kids, like teachers, new parents, and grandparents, might also be distracted by pictures of kids. They've done some preliminary research suggesting this is not the case, but perhaps other non-dangerous populations would be mistakenly identified using this test.
A way to distinguish between 'interest' that is socioemotional vs. sexual might be useful if it hasn't been done. As an example, presumably there is a counterpart for eye gaze (also a measure of attention) associated with socioemotional interest (e.g., eye/face staring), vs. eye gaze associated with sexual interest (e.g., cleavage staring) that could be used to determine whether sex offenders have inappropriate eye movements relative to non sex offenders.
On the other hand, the researchers point out that there are some potential problems. Others who are interested in kids, like teachers, new parents, and grandparents, might also be distracted by pictures of kids. They've done some preliminary research suggesting this is not the case, but perhaps other non-dangerous populations would be mistakenly identified using this test.
Yes, I was thinking of subjects who had children who had died. The RSVP task is sensitive to emotional responses of any valence, and my guess is that parents with children who had died would be much more distracted by pictures of children than other subjects. And we certainly wouldn't want to tag such people as pedophiles.
So you're right that it's somewhat interesting, but I don't think I'd ever be comfortable with such a test actually being used for any type of law enforcement or in a court of law.
The RSVP task is sensitive to emotional responses of any valence, and my guess is that parents with children who had died would be much more distracted by pictures of children than other subjects.
It would be interesting to see if the IAT, RSVP, and emotional Stroop tasks could all be used as a means to create psychological profiles that could differentiate. A google search reveals a few studies focused on the use of the emotional Stroop for sexual offender profiling.
This is very interesting. So in the video, what word was the distractor and what does it mean in terms on this? I missed the middle color so clearly there was some distractor. I'm also a teacher, so if it was child-related that may be why I missed it?
The issue with this, and the penis thing, is that it is not a crime to have paedophillic impulses: it is a crime to act on them. The "future crime" thingo in that book/movie was a parody of exactly this type of thinking.
I agree 100%. Thoughts, feelings, ideas, impulses. They are harmless to society. It's what we do with these feelings that matters.
The only potential harm done is to the person possessing these feelings or impulses. One person may find it extremely stressful, interfering with their everyday functioning or other activities. This person can seek psychological help if he or she feels it would be beneficial to them. Another person may have similar pedophilic tendencies, but doesn't allow it to cross over into their daily life, and doesn't experience any real psychological side-effects.
Neither of these people did anything wrong; the first actively seeks help, the second could be anyone you know. Neither deserves any kind of incarceration or punishment.
@8
Yes, I was thinking of subjects who had children who had died. The RSVP task is sensitive to emotional responses of any valence, and my guess is that parents with children who had died would be much more distracted by pictures of children than other subjects. And we certainly wouldn't want to tag such people as pedophiles.
And (non-pathological) people who wanted children but couldn't/didn't have them for some reason.
The issue with this, and the penis thing, is that it is not a crime to have paedophillic impulses: it is a crime to act on them. The "future crime" thingo in that book/movie was a parody of exactly this type of thinking.
I think that was Minority Report with Tom Cruise.
At best, the indicated tests could serve to identify at-risk-for-offending individuals, which could subsequently serve as a screening tool for persons applying to work in occupations involving daily interactions with underaged children. Certainly this might be useful in reducing situations such as high school teachers having sexual relationships with their students.
There already is a Minority Report-like technology called BrainFingerprinting, that is being used as a lie detector device. It's based on identifying persons presumed to have committed a crime, by identifying brain activity associated with memory traces of objects associated with the crime scene. If the person shows brain activity associated with crime scene objects (and no one else knows about these objects except the investigators), the conclusion is that the person (i.e., their mind) must have been at the scene.
It seems possible that Brain Fingerprinting could be adapted for sex crimes.
I don't think the application of this is "future crimes." It's in making decisions about convicted felons. A pedophile is convicted and sentenced to 10-25 years. After 10 years, he's been well-behaved and a model prison citizen. He received years of treatment for his psychological problems. Is he "cured" of pedophilia? Is it safe to release him on parole? Or is he a hazard who should stay in prison for the maximum specified in his sentence?
I don't think this test is the answer, but if there was an infallible test, wouldn't it make sense to use it?
I don't think the application of this is "future crimes." It's in making decisions about convicted felons. A pedophile is convicted and sentenced to 10-25 years. After 10 years, he's been well-behaved and a model prison citizen. He received years of treatment for his psychological problems. Is he "cured" of pedophilia? Is it safe to release him on parole? Or is he a hazard who should stay in prison for the maximum specified in his sentence?
Um... as far as I know, the consensus among professional psychologists is that pedophilia, like homosexuality, is incurable. The only thing you can do is to reinforce self-control and practice the recognition (and rejection) of self-deceptions like "that little girl wears tight shorts and smiles at me, so she wants to have sex with me".
Some time ago I read about a project where pedophlies (who have not molested any children) can volountarily get this kind of treatment. The best thing you could do with the test described above is probably to encourage people who test "positive" to get such treatment, though I can't think of a good way to implement this since you'd have to test the general populace.
This is only a test.