The truth about lie detectors

Quick Overview

Polygraph tests, despite claims of high accuracy in historical and fictional contexts, are scientifically unreliable because human lie detection accuracy is only slightly above chance (around 54% for truths and 47% for lies in lab settings), and practical applications like screening or interrogation are further compromised by counter-measures, examiner bias, and the high risk of false positives, as demonstrated by the failure of polygraphs to convict the guilty Aldrich Ames and the potential for innocent people to falsely confess under high-pressure interrogation techniques.

Key Points: Average human accuracy in real-time lie detection is only slightly better than chance, achieving about 54% correct judgments for truths and 47% for lies, according to meta-analyses of deception research. Fictional portrayals, such as those in Vanity Fair's lie detector series, often show near-perfect accuracy (Wonder Woman achieves 100% accuracy in the model), which contrasts sharply with scientific findings. The historical inventor of the polygraph, William Moulton Marston, claimed 94% accuracy in 1918, but later tests like the one on Jack Ruby failed to yield admissible evidence in court, leading him to call the device a "Frankenstein's monster" (2:31:44). Laboratory studies demonstrate that countermeasures like biting one's tongue (60% success rate in passing) or pressing a toe (60% success rate) can easily fool polygraph machines, even when subjects are guilty. Real-world screening applications, such as the Washington State Patrol's use, show significant examiner bias, with one examiner failing applicants at a 22.3% rate versus another failing only 7.6% of applicants. Interrogation techniques, especially when combined with the perceived authority of a polygraph machine, can lead to false confessions, as illustrated by the case of Helen Graham (8:35:Helen Graham confessed to a crime she didn't commit after interrogation). The fundamental issue is confirming ground truth; without it, even highly accurate lab results (like 91% accuracy on guilty subjects) are undermined by high false positive rates (55% accuracy on innocent subjects), leading to the false accusation of innocent people.

Raw markdown version of this recap