Quote from james_bond_3rd:
Actually, a National Academy of Science study found that polygraphs "would incorrectly classify 99.5% of âdeceptiveâ results (those telling the truth yet incorrectly deemed to be deceitful), and incorrectly classify 20% of deceitful subjects." That's pretty useless IMHO. Its utility is mostly in scaring the subject into telling the truth.
I recognize that the NAS study found that polygraphs weren't very good. But, that study did not do independent testing to my knowledge -- it only examined the results of third party studies to determine their scientific rigor.
Judges routinely require polygraph exams from persons on probation -- especially those persons convicted of sexual or domestic abuse.
Judges will also give heavy consideration to dismissing an ex-parte domestic abuse restraining order (which is the modern woman's silver bullet for forcing her husband out of the family home as preparation for a divorce action) if the defendant is willing to submit to a polygraph test.
In such cases, the TRO is based entirely on the unsupported testimony of the spouse claiming abuse, and the only way that the defendant can exonerate himself is to offer up the polygraph.
So, while it's only my anecdotal experience in this area, I will stick with my statement, that I can have nearly anyone who's charged with a crime or civil offense, attached to a polygraph and determine the truth of the charges, including the person's motive -- or lack thereof.
There are always those who are either so delusional or sociopathic, that they are immune to any emotion when lying. But, for the average person, the polygraph, in the hands of a qualified examiner, will produce extremely reliable results.