The Wikipedia article on the subject notes that many courts permit polygraph evidence under at least some conditions, but also observes that, “In 1991, two thirds of the scientific community who have the requisite background to evaluate polygraph procedures considered polygraphy to be pseudoscience.”
That was in 1991, over 25 years ago.
Is there a more solid consensus now, one way or the other?
What I have always understood is the used under the correct standards, polygraph results are better than pure chance but not anywhere close to perfectly reliable.
We did some fun/goofy type experiments with the polygraph back in college. The results were unreliable given the conditions we set up. But, I remember the prof saying that under strictly controlled conditions with trained examiner, the machine could detect physiological changes in the subject associated with emotional conflicts when they attempted to lie. This was for most normal people.
The problem is that some criminal types to not register any emotional changes when they deceive, because they lack any remorse or conscience over wrongdoing. Ones we often refer to as psychopathic personalities. Since a relative high proportion of these folks commit crimes, the reliability of the polygraph results is limited.
I think the historical reputation of the polygraph is probably more influential than use of the device itself…if someone is willing to take the test to “prove” their innocence that may convince others of their honesty without ever having to actually administer the thing.
One thing I never understood is the calibration. In all the movies and TV shows, they show the questioner asking preliminary questions, like what’s your name, and telling the subject to lie. Why would the subject be at all nervous about that and register a lie.
Compare that to “Do you know this person?” a legitimate they actually want an answer to and where the subject doesn’t think he’s ever met the person in question, but doesn’t think so. I’d think that the honest answer “no” wold register as much more of a lie.
Furthermore, they repeat questions which by itself would tend to make the subject nervous. ‘Why are they asking me this again?’ ‘Do they know something/’ ‘Have I forgotten something?’
A further wrinkle is this - there are only two answers: yes and no. If you’re a remotely contemplative person, you almost invariably want to answer with qualifiers. I’ve been through a bunch of polygraphs, and always show “significant responses.” I’ve never been failed, since I can legitimize everything in the post-poly review, but wow - do they sweat me out.
Child pornography? I’m an IT security guy, and it has been my misfortune to be assigned to these investigations.
Drugs? Yeah, I worked for a freakin’ narcotics team - never used them, but I’ve been around literally tons of the stuff.
Mishandle classified stuff? Yeah, I work in a SCIF, and have on occasion forgotten to lock one of the safes up - but we’re in a freakin’ SCIF; not like I was collaborating with the enemy.
Also, the majority of the questions are non-pertinent (“Have you ever lied to a loved one?”). That’s the real baseline set, but sometimes they go after you on them to shake you up.
There’s also a tradition among many polygraphers to order subsequent re-tests because the results are inconclusive. This can be good or bad - on one hand, you already know the line of questioning and may be more at ease; on the other hand, you feel like you failed the last time. That accelerated heart rate and increased sweat on your finger isn’t going to work in your favor.
I know this is GQ, and this is an opinion, but I’d read that polygraphs might never have been intended to be accurate. That’s it is a lot of scientific mumbo-jumbo and it gives the impression something is being measured, when really it’s just a trick to get people to think they’ll be caught. Just a more involved version of a lightbulb that lights when you “lie”, but it really is being switched by a hidden compatriot of the test runner. It’s all a trick.
I can never figure out why the CIA uses it. You’d think, of all people, they’d know how useless it is.
Polygraph machine manufacturers don’t like too cooperate with scientific studies. Traditionally they have insisted that the results are dependent on the skill of the operator, which is a clue to how little they are based on scientific principles. Noting the post above, many people certainly will crack under the pressure of a polygraph exam, even if a fake polygraph machine is used. But a good interrogator can get the same results.
My father worked at the Point Lepreau nuclear generating station in New Brunswick, Canada from 1980 until he retired a number of years ago.
There was an incident in which, as a poorly thought out prank, someone put a sample of heavy water into the crew’s Gatorade cooler. Naturally, they were all setting off radiation alarms at detection points when they went to leave at shift change.
Several protocols that should have prevented the incident were breached, and the subsequent investigation treated it very seriously.
Every single person on the crew was subject to extreme interrogation techniques, including polygraphs. Innocent men said they would rather lose their jobs than go through that again.
They are not "useless’ . Look I worked with a expert Polygraph guy, a FBI expert. He was a expert cold reader. He’d either know in advance if the guy was lying or determine that during the questioning- but not by the polygraph. He’d then point to a blip on the record and say “It says here you were lying.” Since the suspect knows he was lying, he often confesses right there.
But as for determining guilt or innocence, truth vs lies by reading the polygraph- yes they are useless.
*Polygraph tests are often portrayed as a reliable method for determine if someone is lying, but their usefulness is highly overrated. While polygraph defenders claim a 90% reliability rate, independent psychiatrists found it to only work 61% of the time. Obviously, if the polygraph test only has a slighter better than a 50/50 chance of detecting a lie, it’s totally worthless. However, what’s not so obvious is that even if it worked 90% of the time, it’s still useless.
Actually at that point, a suspect who knows the reliability of lie detectors, or has just read this thread, the proper response to that it to say, “No, it does not say here that I was lying, but you just admitted to me that you are a liar.”
People are complicated. Some are horrible at lying, and you will catch them easily. Some are good at lying, and you are not going to be able to catch them. Some aren’t lying, but are nervous enough being caught up in the situation that they appear to by lying.
With the sure fire “I know when someone is lying to me” method, you will catch amatuer criminals who are bad at lying, you will accidentally catch up innocent people who are nervous, and are not good at telling a convincing story, even if it’s the truth, and you will let go all the psychopaths and sociopaths that know how to game the system, and make you believe that they are telling the truth.
The only way to find out if someone is lying is to have independent evidence or corroboration of the events in question.
Or a lot of innocent people were convicted when they finally gave in to the pressure put on them by the police who were backed up by a so-called “expert” who told the police what they wanted to hear.
I suppose it depends on your definition of “useful” and “useless”. If closing a case and getting a conviction are what counts, then they are useful to cops and prosecutors. If figuring out who is innocent and who is guilty, and the procedure is neither accurate nor reliable, that would it in a worse than useless category.
NPR had a neuroscientist on a recent program in regards to lying. When a truthful person first tells a lie the amygdala on the fMRI lit up greatly, but afterwards, as the person got into the lying habit, no extra activity was noted – which explains how inveterate liars can lie so smoothly and sometimes for no real purpose. So lies by those persons would not show up on a polygraph.
Years ago, in a psychology class I took, volunteers were given polygraphs. My graph shot up dramatically after one question that I did not lie, but it was a question after one in which I did lie.
You wrote it better than I did, but that’s what I was getting at. The skilled operator is convincing people he knows when they are lying, so they confess.
As you noted, “usefulness” has different meanings.
Cold reading isn’t that accurate(fantasy shows like The Mentalist notwithstanding), and it is almost useless when the participant has been pre-chosen by the police and the questioning is done with only yes-or-no answers allowed. On the other hand, if the cold-reader is reading the police instead of the suspect, it is much easier to determine what the police think of the suspect and give an answer to them they he thinks they would approve of.
