454 ChapTER 13 Emotion, Stress, and Health
not for their accuracy but because they hope to
scare people into telling the truth and induce sus
pects to confess—by telling them that they failed
the test (Leo, 2008).
Because of the unreliability of the polygraph,
researchers are trying to find other ways of mea
suring physiological signs of lying. One approach
uses a computer to analyze a person’s voice, on the
assumption that the human voice contains telltale
signals that betray a speaker’s emotional state and
intent to deceive. But research has mostly yielded
negative or inconclusive findings (Harnsberger et
al., 2009; Leo, 2008). Like the polygraph, voice
analyzers detect physiological changes that may
indicate fear, anger, or other signs of stress rather
than lying, and they often falsely label true state
ments as lies.
Other researchers are using bran scans, such
as fMRI, to test the hypothesis that when people
are lying, they leave “brain fingerprints”—brain
activity that reveals guilty knowledge of a crime.
Some companies claim that they can predict with
better than 90 percent certainty whether someone
is telling the truth. Don’t buy it. Areas of the brain
that light up on an fMRI when people are allegedly
lying are also those involved with many other cog
nitive functions, including memory, selfawareness,
and selfmonitoring (Greely & Illes, 2007). And
because of the normal variability among people in
their autonomic and brain reactivity, innocent but
highly reactive people are still likely to be misla
beled guilty by these tests (Stix, 2008).
To date, efforts to find physiological markers
of lying have produced unreliable results because
they rest on a faulty assumption: that there are
inevitable, universally identifiable biological signs
that reveal with high accuracy when a person is
lying. We’re telling the truth!
Watch the Video Special Topics: detecting Lies
at MyPsychLabout will therefore have increased activity in the
autonomic nervous system while responding
to incriminating questions: a faster heart rate,
increased respiration rate, and increased electri
cal conductance of the skin.
Psychological scientists, however, regard poly
graph tests as invalid because no physiological
patterns of autonomic arousal are specific to lying
(Iacono, 2001; Leo, 2008; Lykken, 1998). Machines
cannot tell whether you are feeling guilty, angry,
nervous, amused, or revved up from an excit
ing day. Innocent
people may be tense
and nervous about
the whole procedure.
They may react to
the word bank, not
because they robbed
a bank but because they recently bounced a check;
in either case, the machine will record a lie. The
reverse mistake is also common: People who are
motivated to escape detection can often beat the
machine by tensing muscles or thinking about an
exciting experience during neutral questions.
The polygraph will correctly catch some liars
and guilty people. The main problem is that it
also falsely identifies many innocent people as
having lied (Saxe, 1994). (See Figure 13.2.) For
this reason, polygraph results are inadmissible in
most courts. But some government agencies and
most police departments continue to use them,Percentage0102030405060Guilty
people judged
innocentInnocent
people judged
guiltyFIgURE 13.2 Misjudging the Innocent
This graph shows the average percentages across three
studies of classifications by lie detectors. Nearly half
of the innocent people were classified as guilty, and a
significant number of guilty people were classified as
innocent. The suspect’s guilt or innocence had been
independently confirmed by other means, such as by
admissions of guilt by the actual perpetrators (Iacono &
Lykken, 1997).About “Lie Detectors”Thinking
CriTiCallySidney Harris/ScienceCartoonsPlus.com