Racial Bias in Polygraph Testing

Is there racial bias in polygraph testing? The Department of Defense Polygraph Institute (DoDPI) released a study showing that innocent blacks are more likely to fail the polygraph than innocent whites. It was first presented at a group meeting of the federal polygraph research community in 1990. Later, the DoDPI director asked members to return the documents or destroy the portion which referred to racial bias studies.

What the study found was shocking, even if there hadn't been racial bias shown. For example, with 1,141 subjects in the study, less than 60% of the tests were accurate in determining guilt or innocence.

36.9% of whites who were non-deceptive were classified as such. To put that in perspective, this means that if you took the polygraph exam as part of the hiring process for a federal job, you had a 63.1% chance of being of being called a liar or getting an "inconclusive" result. Either result would eliminate you from consideration for the job, and likely for any other government position.

Even worse, only 23.5% of blacks who told the truth were correctly classified. Using another format (the MGQT polygraph format), only 14.6% of innocent blacks were correctly classified, compared with 33.3% of whites. It has long been suspected that examiners' biases and expectations influence the results, and this study seems to confirm that. The alternate explanation - that the polygraph just doesn't work as well for some groups of people - would simply provide another good reason to stop relying on the test.

(Source: antipolygraph.org, where the original suppressed report can be found.)

Continues here... What the Polygraph Association Says

Beat a Lie Detector Test | Racial Bias in Polygraph Testing