Bob Park's polygraph paragraph posted

[walthrup48]

Bob's recommendation at the end of this paragraph made me chuckle, so I'm posting it for anyone not already on the 'What's New' mailing list:

WHAT'S NEW Robert L. Park Friday, 18 Apr 03 Washington, DC

1. POLYGRAPH: DOE DECIDES TO SIMPLY REISSUE ITS OLD POLICY. The National Academy of Sciences completed its review of scientific evidence on the polygraph (WN 15 Dec 00). The NAS report, "The Polygraph and Lie Detection" (NAS Press, 2003), found polygraph tests to be unacceptable for DOE employee security screening
because of the high rate of false positives and susceptibility to countermeasures. Congress instructed the Department of Energy to reevaluate its policies on the use of the polygraph in light of the NAS report. DOE carefully reevaluated its policies and reissued them without change, arguing that a high rate of false positives must mean the threshold for detecting lies is very low. Therefore, the test must also nab a lot of true positives. Since that's the goal, the DOE position seems to be that the polygraph tests are working fine and false positives are just unavoidable
collateral damage. But there is still a countermeasures problem: anyone can be trained to fool the polygraph in just five minutes. WN therefore recommends replacing the polygraph with a coin toss. If a little collateral damage is not a problem, coins will catch fully half of all spies, a vast improvement over the polygraph, which has never caught even one. Moreover, coins are notoriously difficult to train, making them impervious to countermeasures.

Archives of What's New can be found at http://www.aps.org/WN

The plug:
To subscribe, send a blank e-mail to: join-whatsnew@lists.apsmsgs.org

 

[BobK]

That's great!

Seems like WN's approach would be a vast improvement.

 

[neutrino_cannon]

In classy pnp style I recomend we roll 2d6 to determine guilt.

 

[shanek]

In classic Paranoia! style I recommend we call them "commie mutant traitors" and just shoot the lot of 'em.

 

[davefoc]

There is a bit of a conundrum here I think: Polygraphs work a little if people don't know they don't work. So by continuing to use polygraphs they continue to work a little because people assume nobody would use them if they didn't work.

My own cut at it is that they are a load of crap and I don't want to try to defend the above thought very hard. It looks like the main reason they "work" at all is that the operators have a pretty good idea wheter the person is lieing or not based on the circumstances. This is of course completely useless unless the polygraph adds more reliability to that assessment and highly dangerous when what the polygraph adds is extra credibility to the conclusion when no extra reliability is added..

 

[George Maschke]

To quote from the National Academy of Sciences polygraph report (p. 8-2): "[t]here is essentially no evidence on the incremental validity of polygraph testing, that is, its ability to add predictive value to that which can be achieved by other methods."