Schneier on Security
A blog covering security and security technology.
« Paul Kocher |
| EU Bans X-Ray Body Scanners »
November 17, 2011
Detecting Psychopaths by their Speech Patterns
The researchers interviewed 52 convicted murderers, 14 of them ranked as psychopaths according to the Psychopathy Checklist-Revised, a 20-item assessment, and asked them to describe their crimes in detail. Using computer programs to analyze what the men said, the researchers found that those with psychopathic scores showed a lack of emotion, spoke in terms of cause-and-effect when describing their crimes, and focused their attention on basic needs, such as food, drink and money.
To examine the emotional content of the murderers' speech, Hancock and his colleagues looked at a number of factors, including how frequently they described their crimes using the past tense. The use of the past tense can be an indicator of psychological detachment, and the researchers found that the psychopaths used it more than the present tense when compared with the nonpsychopaths. They also found more dysfluencies -- the "uhs" and "ums" that interrupt speech -- among psychopaths. Nearly universal in speech, dysfluencies indicate that the speaker needs some time to think about what they are saying.
I worry about people being judged by these criteria. Psychopaths make up about 1% of the population, so even a small false-positive rate can be a significant problem.
Posted on November 17, 2011 at 6:37 AM
• 65 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
This is a very dangerous trend that is spreading so fast: let's model everything with computers.
Another frightening example is the device, which has just started its pilot use in a city (can't remember now), to early-indentify criminal behaviour/intentions based on modeling the body language of people. Imagine you may be arrested for the "possible" crime you have *not* cmmited yet!
I don't understand why scientists persuade such extremely double sided techs? Techs whose -very probable- not-best-cases of use will yield disastrous results.
That's a well-founded worry. The PC-R test used to select the psychopaths in the first place is itself weakly validated for one of its principal uses -- assessing parole fitness in many State penal systems. In particular, the false-positive rate in the general population has not been meaningfully assessed --- although it is known to be dependent on who administers the test. In consequence of which, a potentially useful academic research tool has mutated into a "scientifically-tested" bureaucratic barricade that relieves parole panels from the necessity of actually assessing a fraction of convicted felons, while in actuality having very poorly understood characteristics.
The speech thing could very well turn out to reinforce this "success". Imagine assessing a penal population of tens of millions based on a test calibrated on a sample of 52 people. Do you think they'll try to validate their results on a statistically-significant independent sample? Me neither.
The problem with all such tests is if the test subject is aware of the test being carried out.
If they are they can modify their behaviour accordingly.
This sort of behaviour modification can be kept up for considerable periods of time and is apparently only a little more onerous than "voice training" whereby people learn to speak with a deeper voice as this generaly gives them more credability or authority.
I've seen a number of tests over the years for "psychopaths" and they all suffer from false positives and negatives partly due to the sample sizes and test methods used.
However atleast one gave rise to the (believable) finding that many senior executives exhibit many of the traits of psychopaths (you need to remember that "exhibiting a trait of" does not of necessity mean "you are").
So we now have a speech model for convicted psychopaths who have spent time in jail. Great, so we can spot convicted psychopaths who have spent time in jail in case we need to.
How this automatically applies to other groups (e.g. psychopaths who are not in jail) is unclear.
The doctor who developed the PCL-R is troubled about how the test is currently being used in the criminal justice system. http://www.npr.org/2011/05/26/136619689/... although I'm not sure if his problem applies here. Researchers can apply the test relatively unbiasedly when creating a population to study (of a population who scored affirmative on the PCL-R, x% show y behavior.) When the criminal justice system tests a subject with the PCL-R, they are putting the subject on a particular path within the system. The test wasn't designed to minimize the effect of researcher bias.
Order my new book: Speaking like a psychopath.
So now using the past tense for describing things that actually happened in the past is a sign of psychopathy?
This sounds like thoughtcrime (or maybe ... um ... slow thoughtcrime) to me.
Mind you there is a vaguely humorous side to this,
The main deciding points being given as,
1, Speaking in past tense.
2, Using "ums" and "uhs" a lot.
It sounds like many Classics Professors, when talking socialy...
I wonder how many other proffessions this would applly to ;)
Speech patterns involving a lack of emotion and cause-and-effect instead of emotional reasoning? Sounds like a scientist to me. Let's go, Captain Hammer; we need to stomp out those dangerous scientific leanings!
...point being, there can be more than one reason for attempting emotional detachment and ending up with that sort of speech pattern.
"So now using the past tense for describing things that actually happened in the past is a sign of psychopathy?"
Next we can look to the use of PAST tense to describe things that actually happened in the FUTURE as a sign of time-travelers hiding amoung us, and the use of FUTURE tense for PAST events as a sign of psychopathic time-travelers.
It was a brave, new world! Or will be. Is it?
@Carlo Graziani: That's exactly what I believe in too. The people in charge -from highest ranks to lowest- are looking for a scapegoat to blame in case something goes wrong; and what's better than a computer? --it can't defend itself.
Modeling is an utmostly complicated field in which even the slightest mistakes in training or defining the model parameters, can lead to awfully wrong results.
Depending on your assumptions, sample size, and filtering you can pretty much claim any hypothesis to be true with statistics and post reviews. The challenge is to let the statistics speak for themselves and not start with a preconceived notion that you found while perusing data. Perhaps, the late Nobel Laureate (Physics, 1965) Richard Feynman put it best with his car license plate analogy. When discussing that analogy, he would walk into a room and say that on his way to the lecture the most remarkable thing happened to him, he saw a car in the parking lot with the license plate “AXZ 579” (or whatever). Then he would say to his audience “What were the odds of me seeing that particular license plate?” and proceed to start a lengthy mathematical analysis on a chalkboard to calculate the odds. About halfway through he would stop, turn around, and ask his audience if they could see what was really going on. He would then state that since he already saw the license plate the odds were 100%!
The point being, if you really want to know whether something is valid or not you must first formulate a hypothesis and then conduct an experiment to prove or disprove the hypothesis. You cannot look at the data after the fact and draw a general conclusion from the way it is arranged. In other words, if they believe that you have developed a valid theory for detecting psychopaths from an examination of existing data, they should then acquire fresh sets of data to repeatedly test the hypothesis before they can claim that your observation is anything more than a fluke.
A sample of 52 convicted murderers.
My response: come back with a *real* sample size, with *real* control groups.
This is the kind of first-year student sampling error that convinces many scientists that psychology belongs over with literary theory. Using statistical formulae to produce pretty patterns is fun.
I wonder how autistic people would fare on this model? Bet a great many would flag as psychopathic.
It was Professor Plum, in the Gym, with the Keyboard ...
Could this be a modern version of the teachings of Cesare Lombroso?
"-- the "uhs" and "ums" that interrupt speech -- among psychopaths. Nearly universal in speech, dysfluencies indicate that the speaker needs some time to think about what they are saying."
I think they should expand their study pool. President Obama says "uh" more frequently than any public figure I've heard, even when reading from a teleprompter. Either "uh" is not necessarily linked to thinking about what to say or the POTUS exhibits severe psychopathic behavior.
So the statistics professor drowned in a river with an average depth of 30 centimeters.
Seriously: how can any research have any scientific value whatsoever with a sample size of 52 subjects out of about several thousands in the US alone ? That's just plain ridiculous.
Is it only me, thinking about Minority Report when reading this report?
Is it only me remembering Minority Report when reading this?
Yes, the rest of us are thinking about Jipi and the Paranoid Chip.
I've known more than one professor who fit that pattern.
Perhaps I should have been more careful around them.
I keep my emotions relatively in check, I talk about the things I do in terms of cause-and-effect (I'm a programmer), I use past tense when talking about things I've done in the past, and I often have to think about what I'm saying so I use a lot of "uhs" and "ums".
So apparently I'm a psychopath. Which is good to know - I've been wondering for a while now.
Interestingly, the "non-psychopathic" criminals' responses in the study sound very much like the kind of learned behavior (remorse, spiritual thoughts) that is widely reported to make a good impression on parole boards.
Well friggin "Duh". Isn't this kind of old news? I spend some time time studying violence and violent behavior and none of this strikes as new information.
I think I'd be classified as a psychopath according to quite a few of these tests if I answered them "naturally", even though I'm not (and I know I'm not, because not having real empathy and all that is a part of the definition). I'm a geek and somewhat a "loner". But I can quickly make friends with strangers if I find a common interest.
AFAIK, similiar traits are quite common for psychopaths as well (few "close" friends, can talk to and persuade strangers easily).
I'm also quite confident I could game the test to show any outcome I like, if I tried. I'd like to take a real such test to see.
Just to clarify, I'm not so much focusing on the speech patterns such as "ums" and "ohs", but rather some of the other traits the articles outlined. Not all pyschopaths will show this speech pattern.
For the budding psychopaths out there who want to "fit in" better with society, here's how to change your speech patterns.
1) Join Toastmasters along with someone you see frequently (a significant other, close friend, or coworker).
It's cheap, and they will teach you ways to improve your speech - how to recognize disfluencies, for instance.
2) Play a game with your partner where every time they hear you make a mistake, they say "ding!". That's all - just "ding!" every time they hear a problem.
(For what it's worth I've found that GF's are particularly good at noticing such flaws.)
It takes a week or two, but the constant feedback will eventually sink in and you'll be able to hold long conversations without saying "ah", "um", "you know", and so on.
3) Rhythm, meter, and pauses are more difficult. Find a newscaster whose vocal style you like and record one of their broadcasts.
It doesn't matter whether you agree with their point of view, only that you like their vocal variety. (You could choose Rush Limbaugh, for instance.) I chose Morley Safer.
Edit the broadcasts into individual sentences and rip these to a CD as individual tracks. While you are driving to work, play a sentence on infinite repeat. Recite the sentence along with the speaker over and over. Try to recite it exactly, mimicking the pauses and intonations.
You'll spend a few iterations just remembering the words. Once you know the words, your ear will start to pick out subtle emphasis and pauses used by the speaker. You'll start to learn when to pause (after prepositions, for instance), where to put emphasis to make a point, and so on.
When you get bored, switch to another sentence.
Don't do the mimicry thing more than a couple of weeks or you'll end up sounding *exactly* like the broadcaster. Switch to another one, mix it up a little.
As a side effect of all this, people will view your method of speech as more meaningful, you will be perceived as more reliable and confident, and people will give you greater respect.
It'd be interesting if PCs inserted "ums" & "ers" each time a poster/blogger paused while, um, typing, uh, keyboarding, er, keying in their wisdom.
Even with speech training you can look for clues in their vocabulary.
Words like 'vote for me', 'when elected' and 'my cabinet colleagues' suggest a degree of psychopathic behaviour.
People have "mirror neurons", which cause them to feel the state of others. It's apparently an adjunct to learning - when you see the results of someone else's action, you get a measure of learning from their results.
Chimpanzees don't have this mechanism.
One aspect of this which has been largely overlooked is the tendency of humans to act in large mobs. This was first studies by a student of Mesmer (of all people), who noticed that a mob of people appears to have a "mind of its own": people in the mob would throw themselves at armed police without regard to personal safety, in the manner of a larger organism sacrificing a minor injury in order to achieve a larger goal.
If you've ever attended a concert or sports event where you were "caught up in the mood", you have experienced this effect.
People tend to equate psychopathy with criminal behaviour, but that's not the complete description.
Psychopaths are merely the ones who lack empathy for others. This says nothing about their moral character, only whether they are emotional in their dealings with others.
Leadership requires several traits, among them a) having a vision, and b) not being swayed by the visions of others.
Case "b" above is important. You can't be an effective leader if you are always being swayed by the opinions and emotional badgering of others. Your vision will be constantly changing, you'll never get anywhere. It just won't work.
If there were no psychopaths there would be no leaders. This explains why most business owners score high on the psychopath scale - anyone who is easily swayed by the emotions of others is unlikely to stick to their vision.
Also, it's critically important to have some psychopaths in society simply to provide the dissenting view. Psychopaths stand outside of the mob, and tend to form their own opinions.
When Randall Flagg is egging the mob to kill the prisoners, it's the psychopath who steps up and yells "Stop! What we're doing is wrong!"
Try to draw a distinction between bad ethics, bad behaviour, and psychopathy.
I appreciated the point about UN-convicted psychopaths and detecting them. Has anybody analyzed politicians speeches?
Words like 'vote for me', 'when elected' and 'my cabinet colleagues' suggest a degree of psychopathic behaviour.
Me: "Siri -- listen to this and tell me if I'm talking to a psychopath."
Siri: "*ding* *ding* Run!"
"Psychopaths are merely the ones who lack empathy for others. This says nothing about their moral character, only whether they are emotional in their dealings with others."
Actually, it does. Without empathy there is no moral character to begin with.
I'm not so sure. I think "moral character" is as much about what a person says and does, as it is about whether they can feel true empathy for another person.
I don't think we can praise or condemn someone for anything they think or feel -- only for what they say and do. A psychopath might not get the feedback (a "feeling" or inner voice telling them to sympathize with other people), but if they still behave morally and don't hurt anybody, then there's no reason to criticize their moral character. I'd say "character" is entirely about how people behave, not about what goes on inside their head.
Holden: Describe in single words only the good things that come into your mind about... your mother.
Leon: My mother?
Leon: Let me tell you about my mother.
[Leon shoots Holden with a gun he had pulled out under the table]
"Actually, it does. Without empathy there is no moral character to begin with." - Posted by: Mario Vilas
I totally disagree.
Morals aren't about empathy. They are about understanding and choosing to adhere to society's agreed upon code of behavior. Or if you were religious you would say: following God's rules of behavior.
Plenty of people have empathy and are immoral. They feel the pain they cause others, but they do not have the self discipline to control their base impulses. On the other end of the spectrum you have people who are very cerebral and emotionally detached who are excellent law abiding citizens.
It's especially alarming to consider what might be done with this research when you consider that many speech dysfluencies (awesome word, btw) would probably also be observed in the falsely accused.
"Morals aren't about empathy. They are about understanding and choosing to adhere to society's agreed upon code of behavior. Or if you were religious you would say: following God's rules of behavior." -- Anon
I call B.S. There is a difference between morality and legality, and the two concepts are not necessarily coincident. If society thinks that keep dark-skinned people in chains is okay, and helping them find freedon is not, that does not make it so. Same for many other things, usually of lesser importance, society thinks is "okay".
Interesting. Because: define psychopath.
According to one study as much as one in 25 business managers may have psychopathic tendancies. It doesn't mean that they are murderers; just that they don't properly perceive discomfort or pain in others.
I was bully back in high school. Then I migrated to another country. My language skills wasn't very good, did not have many friends, spent a year or two with computer. This made me a really nice person (quite a few people told me about it). And here I am.. I am not sorry for what Ive done but I wouldn't do it anymore. I know i caused loads of pain, but it doesn't hurt me.
Anyway, I would like to take such a test ;) I bet I could get any result I want (don't you?). Especially if they are judging by the number of of "uhms" used.
I had always spoken only in the pluperfect tense, just in case.
Less "Minority report" and more "Minority report meets the Muppet Show". Or if you're in the UK, the Labour Party.
@ Bruce: "I worry about people being judged by these criteria.
Psychopaths make up about 1% of the population,
so even a small false-positive rate can be a significant problem.".
I agree. If 1% are identifiable as lacking in empathy,
and the confidence test for a sociological difference is at the 5% level,
and puttiing aside all else and assuming the test is completely sociologically accurate, even so:
that only five per cent are falsely assigned to a test result category,
then 1% x 5% or one in 200 would be diagnosed incorrectly as completely psychotic by the test,
if the test worked precisely as claimed.
More likely, one in twenty normals would be diagnosed as latent or para-psychotic,
i.e. having a clear yet unexpressed psychotic potential to decompensate later.
This [some factor]-as-destiny thinking may be applied to one in twenty persons.
The question is whether the diagnosis is made by a native analysis and assembly of signs and symptoms
forcing the conclusion on a skilled analyst of all psychopathic processes,
or a probabilistic screening and reporting and then a test instrument,
or merely a test instrument alone.
We can see this effect at work by comparing reported statistics for ADD diagnoses in children:
Psychiatrists required for diagnoses produces a single case in 1,000. [London]
General family practice M.D. plus psychological test battery criteria produce 1 in 100. [Boston]
Teacher plus psychologist diagnosis produces more than 1 in 10 children. [Los Angeles]
All these variances are found among the same school age children in the general population.
[This week a UCLA Neuroscientist claims the identified empathy gene for oxytocin receptors
in the neuronal synapses is present in only 1/3 of persons,
which would mean 2/3 lack this, and hence
have no barrier to psychotic behahior at the empathic level,
only at the cognitive level.
Hence, the one in 20 is likely about the correct order of magnitude
for those failing to internalizee a purely cognitive and complex lesson.]
Sociopaths know others think bad behavior it is wrong. Psychopaths do not know others consider it "wrong".
Both are empathy defecient.
Not all psychopaths are non-functional.
Some very functional sociopaths and psychopaths ingrain themselves into the structure of organizations,
which support them well in their limited perceptual functioning,
often by granting them power imbalances that allow them to externalize consequences.
ABC [from Oz, not Disneyland], reports a large fraction of sociopaths are attracted to large organizations,
and a large percentage of the top heads of large organization are sociopaths.
Killing someone is cognitively prohibited.
Exploiting others out of wealth is not. So, guess what else is happening in the world?
@Snarki: I would have to say that if someone were to accurately speak about the future, both near and long term, in the past tense and be correct at the same time, then there might be something their. And it can't be something like nostradamus, with vague names, time periods, and places that can be applied to anything (a good number of them can be easily applied to World War II, any further discussion of which will go into Godwin territory). I want to see specific names, places, events, and dates (preferably in ISO 8601 or one of the RFC formats).
Now, if this time traveler speaks with many "ums", then we may have a psychopathic time traveler who should be immediately taken into custody. After all, that's what the article says, right? or at least, that's what the masses will take home with them, and start using on their co-workers the next day to determine who is the psychopath that should be shunned and marginalized by the rest of the office/community.
Good point about the dangers of this approach Bruce. Checklists can exclude exculpatory. Checklists can focus on details without giving context. For example, let's say in a house search they say "look for anything about Arabs" and its a newspaper about the war, an object from another country that they aren't sure what it is, a bag of lentils with Arabic writing, etc. Ignorance and fraud thrive with the checklist approach. That's a big problem with dictionary style software that seeks key words/sounds. If the dictionary doesn't include or make exceptions for the words that are normal for the area targeted, say a group of reporters talking about the war on the phone/on the computer naming key people, you can have a lot of resources wasted and data corrupted.
If its in someone's financial interest to exploit the weakness of the checklist system in counter terror or just watching in data mining, you can really have a terrible mess on your hands.
"Psychopaths make up about 1% of the population, so even a small false-positive rate can be a significant problem."
Yes, but psychopaths make up 100% of the corporate population.
Probably more sociopaths, than psychopaths, in the corporate world ...
To say someone is a psychopath because of saying "um" a lot is a bit of a stretch...and aren't all poor people psychopaths b/c they're focused on where their next meal is coming from..?
I think # of psyco's are increasing though...
@Kingsnake " more sociopaths, than psychopaths"
Actually neither is a qualified diagnosis in the DSM. And they are pretty much the same thing described by two different word.
However psychopaths ARE more sucessful in a corporate enviornment (4% of CEOs are estimated psychopath) so I'm ordering someone's book.
I'd be very concerned if people drew too much from this paper--like, if they thought, absurdly, that they could diagnose psychopathy from a few paragraphs. That's not a critique of the paper itself, of course, but of a subset of readers.
Based on the comments, the term (psychopath) isn't well understood. It's a personality disorder characterized by social deviancy, lack of empathy, manipulativeness, impulsivity, affective dysregulation, etc. In other words, it's not just about the absence of remorse & incapacity to feel empathy, though those are prominent aspects.
The definition of sociopathy depends on who you ask. If you ask Dr Robert Hare (renowned expert in psychopathy), the answer is that "sociopathic" is a sociological term, not psychological, and that sociopaths are anti-social by society's standards but that they are normal within a sub-group; think of a gang member as an example.
Those w/ social deviancy will likely be diagnosed as having anti-social personality disorder. A subset, but not all, of those w/ that diagnosis will qualify as having psychopathy.
Narrative analysis is helpful when used responsibly by someone who knows what they're doing. Pertinent to this blog post, dropping personal pronouns generally indicates de-personalization from the subject at hand. Of course, there are plenty of reasons that might happen.. perfectly normal reasons having nothing to do w/ psychopathy or any other psychopathology.
in response to someone's post:
"If there were no psychopaths there would be no leaders. This explains why most business owners score high on the psychopath scale - anyone who is easily swayed by the emotions of others is unlikely to stick to their vision."
To summarize, I disagree. I've seen variants of this argument before, and it completely misunderstands psychopathy.
from what I've seen Obama tends to just stop talking to think
of what to say next, instead of actually saying, "um".
Amazing, they found that psychopaths acted exactly like... psychopaths.
It is important to keep in mind that psychopath != criminal. Psychopaths that are also criminals are the exception.
Great, more BS for the security snake oil peddlers!! Like we need more...
>"Nearly universal in speech, dysfluencies indicate that the speaker needs some time to think about what they are saying."
"Nearly"? Google Staccato. Also an indicator of Autism Spectrum disorders. This is an EXTREMELY blunt diagnosis instrument for picking out a Psychopath.
It is like saying "Psychopats are *most likely* wearing purple shirts" = no scientific meaning at all.
>"I wonder how autistic people would fare on this model? Bet a great many would flag as psychopathic."
Really badly, even if we are less common (1 in 1000). There are plenty of misdiagnosed people with either condition and plenty of psychiatrics that got their PHD with a box of serials!
Also, there are very CLEAR and defined diagnosis criterias (in DSM 4 and ICD-10) for Anti Social Personality Disorder vs the rest, one major factor is narcisism:
Psychopaths have a f**ked up perception of themselves and the world around them: Psychopaths usually describe other people as ants to play with and their self is the only important thing in the world. They also react badly when their self image is attacked ...or when you ask them for your money back.
I tried negotiating a software development contract with a former business partner - as we went along and i got more info about him, it became more and more obvious what i was dealing with and pulled out just in time.
Ichinin - Autistic and IT-Security professional.
couple more thoughts ...
let's give the researchers some credit here. I don't think they're saying that we can positively identify psychopaths via small samples of narrative patterns or that they've in any way supplanted more reliable means of diagnosis (such as the PCR).
they conducted an experiment, published it, & now various "science" article writers are sensationalizing it. how can anyone who reads this blog be surprised by that? my point here is that this sort of research is helpful and, taken w/in the proper context, can contribute to the broader body of knowledge on the subject. to be candid, the negative comments (toward the researchers) is misplaced.
I like Ichinin's descrip of pschopaths (it's candid, & dovetails w/ my own perception of them) ... quoting:
"Psychopaths usually describe other people as ants to play with and their self is the only important thing in the world."
Brutally honest, but yes, I agree.
In regards to the idea that "some psychopathic traits are good in certain contexts," I'd urge much more thought on this claim. First, let's not lump other behavioral characteristics (like machiavellian characteristics, a la the MACH test) w/ psychopathy. The former *can* be helpful (or harmful ... again, it depends on specifics). But the latter denotes social deviancy, narcissism, impulsivity, and importantly, lack of any empathy or remorse. I've read cases of (probable or confirmed) psychopaths defrauding individuals or companies; of "normal" middle-class people committing violent crimes for no apparent reason. Psychopaths are sometimes referred to as "alien" for their chilling lack of humanity. The most apparent commonality among psychopaths is the callous malevolence that they seem to all share. Damage can be described in individual terms (what one psychopath does) or in societal costs (the parasitic nature of psychopaths; the impact psychopaths have on people around them, sometimes for prolonged periods).
The idea that there are "good psychopaths" or that some have "good characteristics" fades quickly with study on the subject.
Gattica! Gattica! Gattica! Heh.
Luckily for us, a 1% false positive rate among politicians is impossible. Wait.....
@boog: I think there are probably a few more likely candidates over at TDWTF...
@Okian Warrior: I think having the wife say "ding!" to every verbal slipup I make would pretty quickly turn me into a psychopath if I wasn't already.
@garret: Is spelling "with" as "w/" and indicator of psychopathy, sociopathy, or just weirdopathy?
Hatshepsut - Thought I already answered that. Narrative analysis isn't for diagnosis of *any* psychopathology (mental disorder/illness). The subjects examined in the study had already been evaluated. The point was to identify patterns. The idea that a single abbreviation, word addition or omission, etc, is indicative of a particular disorder is complete nonsense. The researchers in the study make no such assertion.
The larger point is that the work of researchers shouldn't be vilified by people who have only read the output of "science writers" who are paid to sensationalize. It's unfortunate that not everyone agrees with that standard of behavior.
"I worry about people being judged by these criteria. Psychopaths make up about 1% of the population, so even a small false-positive rate can be a significant problem."
Nice try psychopath! We're on to you!
this smells of pseudoscience. there's always been a goldrush not for alchemy but developing some sort of 'lie detector' which unscrupulous law enforcement crackpots have used to unjustly convict or charge countless people.
i'm betting this guy's pseudoscience 'psychopath detection' would get thrown out of court by the most amateur of lawyers.
Schneier.com is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc.