I mean, 52% of wrong convictions come from eyewitnesses, yet they are still used as accurate in court. I’m not saying that lie detectors should be used as stand-alone evidence, but, 86% accuracy is still pretty high all things considered.
It is a useful investigative tool. Most people can’t pull off a lie when interrogated by experts, but not that many experts are available. People believing the machine is as good as an expert can be uncovered. Also, people who are telling the truth will be considered as liars some times, and liars will be considered truthful other times, but that’s also what happens when experts and non-experts are making the call without the machine.
I want to emphasize that lie detector machines do not detect lies ever. The operators of those machines decide whether or not they have detected a lie and there is no credible evidence that those operators are 87% accurate in detecting liars based on data produced by the machine. If they could do that then the machine would have a red LIE LED and a green TRUTH LED right on top and the operator would be unnecessary.
Psychology 101 has the cops interrogating a suspect (without a lawyer). They tell him they have a machine which will detect whenever he lies (in reality, a photocopier with a sheet of paper with “HE’S LYING” written on it). Whenever the cops do not like his response, one of them surreptitiously pushes the “Copy” button. Sure enough, it did not take long to elicit a confession.
Sigh.
It is always useful to read the entire thread first.
While clearly they are vastly over hyped. I’m not sure any if the studies mentioned on the thread justify the term “pseudoscience” or “junk science”. e.g, from the NRC study on wikipedia:
polygraph tests can discriminate lying from truth telling at rates well above chance
Thats not pseudoscience. Homoepathy advocates would give their eye-tooth for those kind of numbers.
Whether they justify the reliance thats put on them is another question (and whether we should put even more, according to the OP). But its not complete bullshit.
Just to note: “52% of wrong convictions come from eyewitnesses” does not mean that the “wrong” rate of eyewitness testimony is 52%. It means that, of wrongful convictions, in about half of those, a mistaken or incorrect testimony by an eyewitness was involved in the conviction.
I am not claiming any facts beyond that it was told in an (old, and known to be outdated) Psych 101 textbook.
There actually are limited circumstances where the results of a polygraph can be admissible evidence in court, at least in some states. And there are circumstances in which certain types of eyewitness testimony, DNA evidence, fingerprints, and even confessions are not admissible. I can’t tell if OP is unaware of this, or if he means that, because nothing is ever perfect, we might as well ignore both science and the Constitution and compel criminal defendants to answer questions while hooked up to a polygraph.
I’d argue from a mathematical standpoint, that’s off. Assuming true negative = 75%, true guilty 85%, false guilty = 19%, and false innocent = 10%: You are convicted only if you get a false positive, or a true positive. It’s been too long since my last statistics class, so I can’t crunch the numbers reliably, but unless the people taking the test are largely in the group of guilty people, versus a random sampling, you won’t hit 75% bad outcomes.
Now, one presumes that anyone asked to submit to a polygraph is asked to do so because there is at least some reason to believe that they might have done the deed. Which would, indeed, skew the population toward those who actually might be guilty. Still, I agree: I wouid not be eager to take a polygraph as part of a criminal process.
Anecdotes regarding my own experience with polygraphy: A few decades ago, my employer would proactively clear people to have them available to work on projects with clients who required it. Something they no longer do, by the way, which leads to an interesting “chicken or egg” scenario: they get projects that require cleared staff, but you can’t get cleared unless you are joining such a project.
Anyway: talking with coworkers who went through the project: some were met with amazement and disbelief that they claimed to have never done drugs (as I myself was and am in that boat, I found it plausible). Another fellow was told to just give up, after he failed 2 polygraphs - he just reacted badly somehow. From what I knew of him, he was unlikely to have been up to any mischief - he just did not test well.
When I finally went through the process, I got the impression that they weren’t looking so much for outright lies. but more your reaction to things. I honestly don’t recall if the guy had me deliberately lie, e.g. “tell me the sky is orange”. At one point, he asked me “what is my name?” (his name, mot mine). Being terrible with names, I had immediately forgotten his upon introduction, and I’m sure the needles bounced all over the place when I hesitated, laughed and admitted I didn’t recall.
So, I could speculate that a polygraph might be useful in investigating things - by watching someone’s normal reaction to, say, the weather versus political topics, and using that to guide additional questioning / investigation.
As a standalone tool, though… not so much.
As a tool to convince innocent people that they don’t have a chance in hell of beating a conviction it can cut police work quite a bit.
If it’s not pseudoscience tell me how a polygraph determines if someone is lying. Is there some magical lying vibration it can detect? Or is there some combination of normal body functions that determine when a person lying? If it’s the latter then what is that combination and why are polygraph operators needed?
And lest we forget, in addition to routine polygraphs Aldrich Ames passed two CIA polygraphs specifically looking for the mole in the CIA, i.e., him, during which time he
had compromised more highly classified CIA assets than any other officer in history until Robert Hanssen’s arrest seven years later in 2001.
Well, a coin flip will eliminate 1/4 of potential miscreants. (For want of a better term). The polygraph allegedly will eliminate 3/4 or 1/2, depending who you believe. Think of it as a filter - some get through, many don’t. Since the default is “we have no better means” it is worth a try for the CIA or FBI. Plus, if the FBI reject 1/4 of perfectly good applicants then they simply have to recruit 33% more applicants than they would otherwise.
The critical point is - do you believe the statistics? Plus, what if the ones that get through are the worst of the bunch - not the ones that will flich a few dollars off the money during a drug bust, but the ones who will be active moles for a cartel? The ones who have some small misdeed to hide are perhaps more susceptible to believing the polygraph might work (the placebo effect? Or what would it be called?)
Certainly, I would not put my trust in anything that could send 10% of innocent defendants to jail. Yes, nowadays most people are charged because there’s an acceptable amount of evidence - but if a polygraph were mandatory, hen perhaps the prosecutors would be less gun-shy (like police) about simply “round up the usual suspects”, send them all to the polygraph and “let God sort them out”, to mix quotes.
Yes, exactly what I said, only said better. We do wonder whether operator bias plays any part in interpreting results which are by no means definitive. These are usually people whose living involves doing polygraphs, so they have an incentive to decide the results the way they think the case deserves to be decided, rather than being contrary to the rest of the investigative team. They can’t exactly be impartial or double-blind if they are asking detailed questions:
“Do you own a gun?”
“Did you shoot Fred on the night of July 4th?”
“Did you take a cab from 4th and Vine that night?”
“Did you discard a gun?”
“Do you know where the body is buried?”
Not hard to figure out what the prosecution thinks of the test subject’s situation, while interpreting results.
It is always useful to read the entire thread before responding to a post that suggests that it is always useful to read the entire thread.
Hint for @DPRK:
This is probably amusing, but who’s going to bother to read the entire thread to find out?
Something I have to live with is that a fairly close relative made his career in large part as a purported polygraph examiner. Was fairly big in advancing the acceptance of polygraphs in his particular state, and was also a one-time president of the American Polygraph Association. On the one hand, he seemed very sincere about it all, his life’s work, on the other, though I loved him deeply, I know it’s all a farce. His life’s work was a pseudoscience. Like making a career as a forensic hypnotist (or should that be forensic fortune teller?). Although on the plus side he made most of his career as a defense expert for hire (I’m more forgiving of pseudoscience that might help to keep people out of prison), on the other he trained police forces in the use of polygraphs for much of his later career/life.
In answer to the thread title question “Why is lie detector technology not considered viable in a USA court of law?”, I’m going to add a bit of common sense:
Juries are supposed to convict if guilty beyond a reasonable doubt. If the prosecution based their argument on the polygraph test, and the defense showed these stats (at best, 75% accuracy), isn’t that a reasonable doubt? At that point, why waste the money on the test?