7 comments on “Someone\’s been selling the snake oil

  1. Isn’t it not that they don’t work, but that they don’t work reliably? Not good enough for Courts anyway.

    You can train yourself to fake them, apparently, but they are reliable enough that US Security organisations like the CIA use them routinely.

  2. I’ve used them, oddly enough. They do work, but don’t detect lies. ‘Comparative stress meters’ might be a better name for them. They let you measure stress levels quite well, even the voice analysis things. And this does map, imperfectly, onto truth/lie telling.

    So useful in-house as SMFS says, but shouldn’t be sued in legal proceedings.

  3. Yes, the CIA found them of great value in catching Aldrich Ames. Oh, wait, actually it was the complete opposite – he passed every test he underwent. The rate of false negatives and positives are both high enough to render them useless.

  4. Who cares if it works. It’s for sex offenders, so as much theater is required as possible. Maybe they should bring back the ducking stool.

  5. For what it’s worth, the only people I’ve known who have believed in lie detectors have been the most conspicuously stupid among my acquaintances.

  6. “Isn’t it not that they don’t work, but that they don’t work reliably?”

    No, they simply don’t work. At best they are a tool used by interrogators to cow suspects into admiting what the interrogator is sure of.

    They don’t seem to catch spies over here, since many of our moles took lots of these over periods of years and not only were never caught, never suspected.

    I’ve taken a couple (job interviews asking about criminal behavior not nat’l security) and passed. All it really takes is being calm and not care the slightest bit about what you’re being asked.

    Not to mention that they are full of false positives.

  7. Matthew L – “Yes, the CIA found them of great value in catching Aldrich Ames. Oh, wait, actually it was the complete opposite – he passed every test he underwent.”

    He passed two. That is not a lot. Yes, they can and so fail. The Green River Killer passed one for instance. But the CIA continues to use them.

    “The rate of false negatives and positives are both high enough to render them useless.”

    Useless for what purpose? If nothing else they are likely to have a deterrent value.

    6Agammamon – “No, they simply don’t work. At best they are a tool used by interrogators to cow suspects into admiting what the interrogator is sure of.”

    So they do work then? A German police chief was fired after he told a suspect in a kidnapping of a child that he had a specialist coming to see him, the suspect, and he was going to be slowly tortured at some length. The guy was so scared that he told the police where the body was. Of course he did not have someone coming to torture the guy but it still worked to get a confession.

    “They don’t seem to catch spies over here, since many of our moles took lots of these over periods of years and not only were never caught, never suspected.”

    Over here as in the US? Again the question should be not how many were caught but how many were deterred.

    “All it really takes is being calm and not care the slightest bit about what you’re being asked.”

    There are supposed to be big cultural differences in test-taking with Americans being a particularly well brought up lot that get very stressed when lying. I don’t know. If I was taking money from the KGB and I was being asked about it, I might be a little bit worried.

    “Not to mention that they are full of false positives”

    That is true. And hence not useful for Courts. Of course the question then becomes, do we want lie detectors that work? Suppose we had one – and the police came around to everyone’s house once a year and asked them if they had done anything wrong in the past year?

Leave a Reply

Name and email are required. Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>