Polygraphs and eyewitness testimony - bad ideas?

Jeffry Ricker on the PESTS (Psychologists Educating Students to Think Skeptically) listserve alerted us to a post on polygraphing by Dan Greenberg on the Chronical Review. Greenberg points out that voluntarily undergoing a polygraph test (lie detector) is a bad idea due to a lack of scientific evidence for the reliability of such tests. He claims that "...it’s a voodoo device that can stain the innocent and exonerate the guilty."

Greenberg demonstrates how the device survived concerns about its scientific merits with an interesting quote from Richard Nixon: “I don’t know anything about polygraphs, and I don’t know how accurate they are, but I know they’ll scare hell out of people.”

One of the commenters on Greenberg's post points out that while screening polygraph tests may be suspect, directed issue tests are more accurate than eyewitness testimony. Which brings about the question, how accurate is eyewitness testimony? A Google search on eyewitness testimony brought me to an interesting example of memory misattribution.

An Australian eyewitness memory expert, Dr. Donald Thomson, was arrested for rape after he was identified by a rape victim as the rapist. At the time of the rape he was appearing on a live TV discussion about the unreliability of eyewitness memory. The raped woman had confused his face seen on TV with that of the rapist. The Australian police refused to accept his alibi, even though it was backed up by an assistant police commissioner who also appeared on the show, not to mention the studio audience and viewers! Which does not say much for the intelligence of the policemen involved (unless it was possible for Thomson to be in two places at once due to some weird quantum effect)! This incident from a book by Alan Baddeley, although I think it may have been mentioned before that by Daniel Schachter, both of them prominent memory experts.

Moral of the story? You agree to a screening polygraph test at your own peril. Consider perceptual distortions and unreliable memory when judging or challenging eyewitness testimony.