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 |