FBI File of Lie Detector's Creator

George Maschke writes, "It appears that the FBI considered William Moulton Marston (1893-1947), who invented the lie detector and created the comic book character Wonder Woman under the pseudonym Charles Moulton, to be a 'phony' and a 'crackpot.' He is alleged to have misrepresented the result of a study he conducted for the Gillette razor company in 1938, for which he reportedly received some $30,000, a handsome sum in those days. Despite these misgivings, the FBI today uses Marston's creation (the polygraph, not the Lasso of Truth) to guide investigations as well as to screen applicants and employees. You can download Marston's FBI file here (736 KB PDF)."

Comments

Re:A way out?

[CoverStory]

As shown recently on MythBusters, Cleve Backster the man that originated the comparison test used by most law enforcement agencies to determine the results of a polygraph test spent most of his career using those tests on plants.

From any interview given in 1997

... the imagery entered my mind of burning the leaf I was testing. I didn't verbalize, I didn't touch the plant, I didn't touch the equipment. The only new thing that could have been a stimulus for the plant was the mental image. Yet the plant went wild. The pen jumped right off the top of the chart.

If that won't convince someone about the accuracy of the test, I don't think TFA will.

Re:A way out?

[AKAImBatman]

Is there any chance that this could be used in court cases to challenge polygraph test results?

Polygraphs are already inadmissable as court evidence, and can no longer be used to screen employees. Pretty much the only area you'll run into them is in federal jobs requiring security clearance. Investigators also use them on occasion to determine if the suspect is misleading them during an investigation, but the results can't be held against the subject of the test.

The truth is that the polygraph is a form of psychological testing. The results are meaningless unless the "operator" is a well trained psychologist. Even then, he may be unable to extract the "truth" from you; partly because "truth" is a subjective matter. In addition, some people don't do well (or do TOO well) under stress testing. So the results can be bogus in those cases. Basically, polygraphs are unreliable at best, and should never be counted on for accurate information.

What if a high false positive rate doesn't matter?

[dircha]

According to the studies linked from the Wikipedia article (http://en.wikipedia.org/wiki/Polygraph) it seems that while the test has a high false positive rate, the false negative rate is lower than one would expect of random chance. Does anyone read it otherwise?

While I think it would be abhorrent to allow such a device to be used against a defendant in our criminal justice system, it the above is true it doesn't seem to me so unreasonable at all that it be used in the hiring of FBI and CIA agents and the like.

A better chance of keeping Russian and Chinese spies out of our security forces may very well outweigh turning away candidates incorrectly classified as deceitful.

Whereas in matters of criminal justice most seem to agree it is better that 10 guilty men should go free than that 1 innocent man should be condemned.

Also, I've always wondered whether this isn't really more of a "nervousness test" than anything else.

Re:What if a high false positive rate doesn't matt

[Dun Malg]

According to the studies linked...it seems that while the test has a high false positive rate, the false negative rate is lower than one would expect of random chance. ...if the above is true it doesn't seem to me so unreasonable at all that it be used in the hiring of FBI and CIA agents and the like.

As clearly demonstrated by the above mentioned stats, the problem is that polygraphs achieve their low false negative by basically lowering the thresholds, casting a wider net of "guilt" and snaring more innocent people. I can guarantee a 0% false negative rate-- so long as you let me declare everyone who walks in the door "deceptive". Polygraph is just theater. It's pretty much bog-standard interrogation techniques dressed up with a few electronic props to trick people into essentially admitting guilt.