Are Lie Detectors Accurate?

Are lie detectors accurate? Recently, in a report by the National Academy of Sciences for US Department of Energy, polygraph testing was found too inaccurate to be relied on in security screening. This is just one of many studies that shows how dangerous reliance on the polygraph machine can be. Let's take a look at just one example from page six of the report.

First, though, a quick explanation of one aspect of polygraph testing. It is important to understand that it can be done with greater or lesser sensitivity. If it is set for greater sensitivity, it is more likely to identify lies. However, it is also more likely to identify innocent people as lying. Less sensitivity means fewer false-positives (identifying innocent people as liars), but fewer liars detected also. Hence, how sensitive to make the exam is always an issue that has to be addressed by those who give the test.

Page six of the report looks at a hypothetical population of 10,000 government employees that includes 10 spies. It points out that "If the test were set sensitively enough to detect about 80 percent or more of deceivers, about 1,606 employees or more would be expected "fail" the test." By the way, the report also notes that this is an accuracy "greater than can be expected of polygraph testing on the basis of available research."

Okay, let's review. The test isn't that accurate normally, according to the available research. But if it was, you could get rid of 8 of the 10 spies - as long as you also fired about 1,600 loyal and honest employees. Does that sound useful?
Of course, it is noted in the report that you could take further measures to "separate the 8 spies from the 1,598 loyal employees caught in the screen." This presumably means more polygraph exams and interrogations. It also almost certainly means firing some innocent employees and getting rid of fewer than 8 of the ten spies.

The report continues; "If the test were set to reduce the numbers of false alarms (loyal employees who "fail" the test) to about 40 of 9,990, it would correctly classify over 99.5 percent of the examinees, but among the errors would be 8 of the 10 hypothetical spies, who could be expected to "pass" the test and so would be free to cause damage." In other words, you fire only 40 innocent people if the test is less sensitive, but then you catch only 2 of the 10 spies.

The report then says calmly:

"Available evidence indicates that polygraph testing as currently used has extremely serious limitations in such screening applications, if the intent is both to identify security risks and protect valued employees. Given its level of accuracy, achieving a high probability of identifying individuals who pose major security risks in a population with a very low proportion of such individuals would require setting the test to be so sensitive that hundreds, or even thousands, of innocent individuals would be implicated for every major security violator correctly identified. The only way to be certain to limit the frequency of 'false positives' is to administer the test in a manner that would almost certainly severely limit the proportion of serious transgressors identified.

"CONCLUSION: Polygraph testing yields an unacceptable choice for DOE employee security screening between too many loyal employees falsely judged deceptive and too many major security threats left undetected. Its accuracy in distinguishing actual or potential security violators from innocent test takers is insufficient to justify reliance on its use in employee security screening in federal agencies."

Are lie detectors accurate? Some may be in the future, but certainly the polygraph machine is not. No matter which way you use it, the polygraph test doesn't work well enough to bother using it. Unfortunately, neither this nor any other scientific research showing the flaws in polygraph testing is likely to stop the government from using it anytime soon (and I'll explain why this is so later).

Continues here... Racial Bias in Polygraph Testing

Beat a Lie Detector Test | Are Lie Detectors Accurate?