Saidme wrote on Aug 13^th, 2003 at 2:46am:

Skeptic Regarding specific issue testing (which is what I do).  Your coveted NAS study (which I have no regard for) states something to the effect that specific issue testing is "far above chance" but less than perfect.  Is far above chance 90%, 95%, 98%.   I think I would take any one of those numbers to the bank.  Let's even concede 85%, still pretty damn good.

The actual finding was:


Quote:

CONCLUSION: Notwithstanding the limitations of the quality of the empirical research and the limited ability to generalize to realworld settings, we conclude that in populations of examinees such as those represented in the polygraph research literature, untrained in countermeasures, specific-incident polygraph tests can discriminate lying from truth telling at rates well above chance, though well below perfection.

-- NAS Report, Executive Summary, p. 4

You'll note a couple of things.  First of all, they don't give a reliability rate, period.  Second, they caveat the statement regarding the populations that are being discussed (e.g. subjects "untrained in countermeasures").  Third, the NAS is, in the above, specifically talking about polygraph research -- exactly the sort you seem to disdain as unrepresentative of the polygraph in the real world.

Finally, if you take an 85% accuracy rate yet hold 100% confidence in your results, you're talking about barking up the wrong tree 15% of the time.  That may sound OK, but we're talking about an awful lot of misdirected investigations and wrongfully accused people.


Quote:

Let me caveat this next statement so your co-horts don't take this out of context.  The most recent (and more recent) shuttle disasters were horrendous.  America lost a lot of great astronauts in both instances.  The shuttle and the technology driving those shuttles is awesome.  The people working on them are extraordinarily bright.  Were the shuttles 100%?  Obviously not.  Would I volunteer to ride the next shuttle to space?  In a heartbeat.  Anytime you factor in the human element you've got room for error.  Polygraph is no different.  It's pretty damn good but face it, nothing's perfect.

Nope.  But we're always trying to make things better.  If that means a system is fatally flawed, we change it out.  There's no way the Shuttle would fly if engineers gave it only a 75% chance of reaching orbit.  And though your courage is remarkable, I sure as heck wouldn't fly under those conditions, either.

Skeptic