Polygraph supporters love to trot out that old line, “Some medical tests aren’t 100% accurate, but we still use them.”  The implication, of course, is that even though the polygraph isn’t 100% accurate (and, in all fairness, I’ve never encountered a polygrapher who claimed that it was 100% accurate) it is still useful.

Please allow me to explain, in the same vernacular, why this comparison is misleading and false, and why the polygraph has no utility whatsoever, except of course for getting nervous test subjects to sometimes admit to acts which otherwise might have gone undetected.

Let’s assume that a medical test exists for the XYZ disease, but this test is only successful at detecting the presence of the disease about 75% of the time.  The other 25% of the time, the test will not be positive, even if the subject does in fact have the XYZ disease.  Some people would conclude that the test is useful, even though it is not 100% accurate.

However, there is some crucial data missing.  In order to determine if the test for the XYZ disease is actually useful, we would have to know if it ever showed a positive test result for a person that did not have the XYZ disease.   

If the test never showed a positive result for a subject that did not have the XYZ disease, then it is indeed a useful test.  It is useful because after the test is completed the doctor (and the patient) may draw a definite conclusion at least part of the time.  If the test result is positive the doctor could conclude with 100% accuracy that the patient is suffering from the XYZ disease.  Negative test results would be less conclusive, since the doctor would not be able to rule out the presence of the disease.   

However, if the test sometimes showed a positive result for the presence of the XYZ disease when the disease was not actually present, the utility of that test would be zero.  It wouldn’t matter if this occurred in 1% of the tests or in 99% of the tests; the mere possibility of a false positive is sufficient.  It would be a useless test because no matter what the result, there would be no legitimate conclusions that could be drawn from the test data.  A positive result would mean that the person has the XYZ disease or that the test is a false positive.  A negative result means that the person doesn’t have the disease or that this is part of the percentage of cases where the result will be negative even though the disease is present.   

In other words, a positive result would not confirm the presences of the disease, and a negative result would not confirm the absence of the disease.  The doctor and the patient wouldn’t know any more after the test than they did before the test, which is as good a definition for the term “a useless test” as I can think of.

Compare this fictional test to the polygraph exam.  I have never encountered a polygraph examiner who attempted to claim that the polygraph was 100% accurate.  This lack of complete accuracy means it is logical to conclude that there is a chance (however large or small you believe it be) that a deceptive subject can still be scored “No Deception Indicated” on their polygraph.   

I have also never encountered a polygraph examiner who claimed there is no such thing as a false positive.  This means it is logical to conclude that there is a chance (however small or large you believe it to be) that a truthful person could be deemed deceptive on their polygraph exam.  This is the crucial point that renders the polygraph worthless.

Polygraph supporters would have you believe that their admission of less-than-complete accuracy subsumes false positives along with false negatives, and that there is no substantive difference between the two.  That is simply not true.  Just as with the fictional medical test for the XYZ disease, once the possibility of a false positive exists, the ability to draw any conclusions from a positive or a negative test result no longer exists.

Absent a damaging admission from the test subject, the examiner does not know any more at the end of a polygraph exam than he/she did at the beginning.  A “No Deception Indicated” score means that the subject was telling the truth, or that they fall into the (small or large, depending on your source) percentage of people for which the polygraph is not 100% accurate in detecting deception, resulting in a false negative.  A “Deception Indicated” score means the person is lying, or that they fall into the (small or large, depending on your source) percentage of people the polygraph will show as deceptive even though they are being truthful, resulting in a false positive. 

In other words, a “No Deception Indicated” score on a polygraph does not confirm that the subject was being truthful, and a “Deception Indicated” score does not confirm that the subject was being deceptive.  The polygraph examiner doesn’t know any more after the test than they did before the test, which is, again, as good a definition for the term “a useless test” as I can think of.

Quote:


You are absolutely correct, George.  I was trying to address an often-heard argument by using the data the pro-polygraph crowd already admits to.

I have read more than a few times on this and other message boards that:
The polygraph is not 100% accurate.
There is a chance for a false positive.
None of that matters because it's still good and useful.

I was simply trying to demonstrate that, given the facts readily admitted to by most, if not all, polygraph supporters, the polygraph's utility is limited to obtaining a confession from nervous test subjects.  If, at the end of the polygraph test no damaging admission has been made, the utility of the polygraph has been fully expended.

Good Day Sgt and Mr Moderator,

In fact, Polygraphy IS based on science.
New fact: It is not an exact nor a precise science - and that's where all the problems begin.

The polygraph records scientific data. No argument about that.

However, what the DOD industry does with that data is unscientific.

Using polygraph data as a means to decide whether a person is truthful or not is not science. Using a polygraph in that endeavour is a learned skill.

An educated, intelligent examiner possessed of good investigative skills can use the polygraph as a prop to elicit confessions. That doesn't always work and when it doesn't then the subject should be given all benefit of doubt and be called NDI.

Maybe INC's & NDI's should both be renamed as NAM's (No Admissions Made)
If a call of DI is made, it should be accompanied by an Admission, otherwise its NAM.
and DI can be renamed DIAM (Deception Indicated & Admission Made)

Glossary:
DOD = Detection Of Deception
NDI  = No Deception Detected
INC =  Inconclusive

Quote:



Thats what I inferred in the first instance

Of course physiological responses can be measured and recorded with a high degree of accuracy.  Human technology has made great strides in the past century.

Yet, our psyches remain bound to a more primal state: We seek determinism in our complex and sometimes confusing lives through a wide variety of mechanisms.  For example, humans cling stubbornly to a wide variety of religious beliefs, which in most cases are mutually exclusive.  The human need to somehow determine the truth--or a psychologically acceptable alternative--and obtain solace thereby provides ample opportunity for those posing with gadgetry and pseudo-rationale to profit financially.

A group of such posers has emerged in our society.  These persons seek to profit by linking measurable physiological stress responses with detection of deception.  While a polygraph machine can measure indications of stress, neither the machine nor it’s manual or software assisted operator can construe the real reason(s) why a human might be experiencing stress.  To claim otherwise perpetuates an elaborate con.

In the future it may occur that P-300 brain waves or other techniques are postulated of use in deception detection.  Yet, no deception detection technology can ever be validated scientifically: One cannot prepare a sample of statistically valid size containing known human truth and deception of varying degrees for which there are serious consequences to innocent humans who test positive for deception.  Any such experimental plan constitutes a violation of human rights.

Lloyd Ploense