PAR Blog | Polygraph Tests: How Accurate Are They?

Polygraph Tests: How Accurate Are They?

February 15, 2011

Lying, it seems, is a very common part of human interaction. In their book 50 Great Myths of Popular Psychology (2010), Scott Lilienfeld and his colleagues cite studies in which college students and others in the community admit to lying once or twice a day, on average. But how hard is it to tell if someone is lying? Don’t liars give off tell-tale signs of their deceptions? In fact, research reveals surprisingly few valid cues of deception, and Lilienfeld asserts that “most of us are dead wrong about bodily cues that give away liars” (p. 116).

If people are poor judges of truthfulness in others, does technology offer a better solution? Is the polygraph, or lie-detector test, an accurate means of detecting dishonesty? Most Americans (67% in one study) believe that lie-detector tests are “reliable” or “useful,” and films and television programs tend to corroborate this belief with story lines that portray polygraph testing as infallible.

The science, however, tells a different story. Lilienfeld and his colleagues explain that rather than truthfulness, the polygraph machine simply measures physiological activity—and then it is up to the examiner to ask questions and interpret the results. Factors such as blood pressure, respiration, and sweating can offer clues to lying because they are associated with how anxious the examinee is during the test; however, anxiousness is not the exclusive domain of lying, and “an honest examinee who tends to sweat a lot might mistakenly appear deceptive, whereas a deceptive examinee who tends to sweat very little may mistakenly appear truthful” (p. 118).

Another problem is confirmation bias, that is, the tendency for polygraph examiners to see what they expect to see. Examiners may have a preconceived notion of the examinee’s guilt based on outside information. Further, information on countermeasures, or techniques to “beat the test,” is widely available on the internet.

Estimates of the accuracy of one popular form of the polygraph test, the Comparison Question Test, put it at 85% for guilty individuals and 60% for innocent individuals. “That 40% of honest examinees appear deceptive provides exceedingly poor protection for innocent suspects,” suggests Lilienfeld (p. 120).

Because of their limited validity, polygraph tests are rarely admissible in court, and federal law prohibits most employers from administering lie detectors. Yet the public perception is that polygraph tests are accurate measures of truthfulness. Are people simply vulnerable to the images they see in television and movies, or is there something else that makes us want a machine that can detect the truth?

We would like to hear your opinion on this topic, so please post a comment and let’s start the conversation!