Roger Keith Coleman was arrested in Virginia for the 1982 rape
and murder of Wanda Fay McCoy. Many people felt that he had an
inadequate defense counsel and that there was convincing evidence
that proved his innocence, which his court- appointed lawyers---both
inexperienced---failed to introduce in court. The state's
expert claimed that two pubic hairs on the body were consistent with
Coleman's and it was unlikely that they had come from anyone else,
although that same expert was mistaken in his hair analysis on a
different murder case. There was also another man found guilty
of rape in the same small town who allegedly had confessed to the
murder while with one of his victims.
Coleman maintained his innocence and he asked for a polygraph test.
He was granted it, albeit under questionable circumstances, and the
state claimed that he failed it. On May 20, 1992, Coleman was
executed.
Jim McCloskey, of Centurion Ministries in Princeton, New Jersey, was
deeply involved in the Coleman case for over four years. His
non-profit organization is devoted to getting justice for convicted
prisoners who are innocent and he tried hard to get the evidence
reconsidered. He reports the shocking fact that that Coleman's
polygraph test was administered on the morning before he was
scheduled to die.
"When they offered that," he said, "we fully
recognized how insane it was giving a man a polygraph on the morning
of his execution, because his biological processes would be going
all over the place. We agreed to do it because we were
shooting for the moon. We were desperate and had no other
options. They did agree to allow us to have our polygrapher
present for the procedure, but they sandbagged us. They gave
us the wrong address, so we never saw the procedure."
Early that morning, Coleman was taken from his cell and driven to
Richmond, Virginia, where he was put into a room by himself.
He was tired and hungry, and disconsolate about what might happen to
him by nightfall. Without giving him anything to eat, the
polygraph was administered and the results analyzed. McCloskey
was told that he flunked. "We don't know whether he
really flunked or not. We don't know what questions were
asked. We've never seen anything on the results. If he
did flunk, it wouldn't surprise me given his state of mind, but we
have no idea how the test was even administered."
The problem with using a polygraph under such conditions, he adds,
is that the odds are stacked heavily against the person.
"You're strapped to a machine and you know that if you pass,
your life might be saved, but if you flunk or they just interpret it
as if you're flunking, it's the final nail on your coffin. If
you're nervous, how can they tell the difference between fear and
deception?"
Advocates of the polygraph claim accuracy rates as high as 99
percent, but critics argue that the techniques are based on
questionable psychological assumptions. In a 1981 study
of six polygraph interpreters, even the most experienced of them had
an error rate of 18 percent, and the least experienced made
classification mistakes 55 percent of the time. In another
study that used actual criminal investigation data, the accuracy
range was 63 to 76 percent. Examiners were consistently most
likely to label a truthful subject as untruthful rather than the
other way around. It seems that even if the machine can
achieve a much higher accuracy rate, the results depend more heavily
on the skill level of the examiner.
The problems cited most often are listed below:
The pressure that people often feel from the
idea that polygraphs will determine the truth can elicit
confessions, but among them are false confessions. There are
people who, through guilt or the desire for notoriety or some other
motive, feel compelled to admit to a crime they did not commit.
(Some famous murders, like the
Black Dahlia in Los Angeles in 1947,
have elicited dozens of confessions, and in that case, some were
from people who were not even born when she was killed.) Thus,
a confession that results in a conviction is not always about
accuracy.
The criterion of accuracy, known as the "ground truth,"
is difficult to pinpoint for evaluating a polygraph's assessment of
guilt. At times a panel of judges reviews all of the evidence
against which to judge the polygraph and in other cases, the results
are measured against a subsequent confession or the final judicial
outcome.
Analog studies in which some accuracy statistics are based
involve mock crimes that fail to adequately include the kinds of
emotional components that might occur in real situations.
Some deceptive subjects can pass a test by using cognitive or
physiological countermeasures to avoid detection. (There is even a
website that teaches people how to do this.) People with
antisocial personality disorder, for example, appear to have reduced
autonomic activity and show no remorse, so they may have an easier
time beating the machine. Others have devised specific tactics
that appear to have some effect. (Many polygraph examiners
claim to be aware of the techniques to control one's physiology, but
a 1994 study found that 50 % of subjects trained to fool the
polygraph examiner succeeded.)
-
Most polygraphers are not scientists and often receive no
follow-up information on the accuracy of their readings, so they
don't have sufficient feedback to self-correct.
-
Polygraphs read heightened physiological responses, such as fear,
which are not always about deception.
-
The polygraph analyst will come up with a certain percent of
erroneous interpretations, and some studies have shown that they
tend to err by calling an honest person dishonest than the other way
around. Some critics feel that this is partly due to the fact
that examiners are paid to get results and they feel pressured to
come up with a guilty person.
-
Not all polygraph situations follow the etiquette of allowing a
second test or a complaint filed to a panel of professionals for
review of a procedure. In the case of Roger Keith Coleman, not
only was his test performed under terrible conditions (no sleep,
extreme agitation, hungry, anxious), but also he was not allowed a
second test and no other professional was allowed to evaluate the
results. In cases where so much is at stake, both sides ought
to be allowed to administer the test so that results can be
compared.
|