The polygraph does not actually detect lies, because nothing can detect lies. Instead, it monitors and records reactions in the body of the person being tested. It checks how you’re breathing, how much you’re sweating, how fast your heart is beating, and so on. The theory is that when you tell a lie, your body reacts in a certain way, and that by detecting those reactions, the polygraph can determine when you’re trying to deceive the questioner. A seductive idea; to be able to sniff out a lie, right at its source! What power, and what convenience!
Fourteen individual analog studies:
– correct guilty detections ranged from 35.4 to 100 percent and averaged 63.7 percent;
– correct innocent detections ranged from 32 to 91 percent and averaged 57.9 percent;
– false positives ranged from 2 to 50.7 percent and averaged 14.1 percent; and
– false negatives ranged from O to 28.7 percent and averaged 10.4 percent.
As the FAS put it, “the polygraph detects deception at a rate better than chance, but with error rates that could be considered significant.” Not exactly a ringing endorsement. And then they give us this to chew on:
One area of special concern in personnel security screening is the incorrect identification of innocent persons as deceptive. All other factors being equal, the low base rates of guilt in screening situations would lead to high false positive rates, even assuming very high polygraph validity. For example, a typical polygraph screening situation might involve a base rate of one guilty person (e.g., one person engaging in unauthorized disclosure) out of 1,000 employees. Assuming that the polygraph is 95 percent valid, then, the one guilty person would be identified as deceptive but so would 50 innocent persons. The predictive validity would be about 2 percent. Even if 99 percent polygraph validity is assumed, there would still be 10 false positives for every correct detection of a guilty person.
10 people, falsely declared liars, for every liar found. And that’s for a 99 percent validity. Look again at the estimates for actual validity, and do the math for yourself.
Happily, some folks have paid attention, and have limited or banned the use of polygraphs. Here in the U.S., defendants and witnesses cannot be forced to take a polygraph test, and several states forbid employers from using the test on their employees. Unfortunately, government agencies are still allowed to polygraph their employees, so the test is still used as a pre-employment screen by the FBI and the CIA.
Let me emphasize the oddness of this. The polygraph is inaccurate enough that courts can’t force it on witnesses and private employers can’t force it on their would-be employees, yet the CIA and the FBI – two agencies who really ought to have top standards for using reliable tests – still use it to screen their employees. Two powerful agencies, whom we are asked to trust with our safety and security, are totally committed to using a bad test.
Sadly, this kind of behavior is typical. Whether it’s the bomb detector used by the UK and Iraqi government (among others) that’s no better than a dowsing rod, interrogation techniques that can compel anyone to confess to anything (never mind whether or not they actually did it), or the NSA’s habit of collecting huge amounts of data that don’t actually help prevent any attacks, the high and mighty have a proven track record of stupidity. Keep that in mind when they ask you to trust them and give them more power.
Edited to add: If you really want to have a good look at the shenanigans involving the polygraph, head over to AntiPolygraph.org. But before you go check that site out, keep in mind that the NSA is targeting that site (and possibly trying to entrap the folks behind it), so protect your privacy by using Tor, or TAILS if you want to be really safe.