Although WhiteHouse.com did not post it, Gordon Barland did in fact complete a review of Ed Gelb's polygraph examinations of Larry Sinclair. WhiteHouse.com sent Barland's review to Mr. Sinclair, who has posted it on
his "Windows Live Space." AntiPolygraph.org has made word-searchable versions of Gelb's reports and Barland's review available here:
Gordon Barland's review shows Ed Gelb's statement in his reports that "The polygrams [polygraph charts] were 'blind scored' by another expert examiner" to be false (assuming Barland is the examiner to whom Gelb referred). In scoring Sinclair's polygrams, Barland was not
blinded with regard to any of the following salient details:
- The name of the original examiner;
- The name of the examinee;
- The nature of the examinee's allegations;
- The actual polygraph questions asked (as opposed to just their type -- relevant, control/comparison, irrelevant, etc., which is all that is needed to score the charts);
- The decision(s) rendered by the original examiner
That Gelb, a past-president of the American Polygraph Association, could consider Barland's review to have been "blind" helps illustrate just how far removed polygraph practice is from the scientific method.
Barland's report makes it clear why Gelb failed to mention the results of any computerized scoring of his polygraph charts. Although Barland, in his non-blinded review agreed with Gelb's hand scoring of the charts, in the one case (regarding the drug allegation) where Barland ran a computerized scoring algorithm on the chart, contrary to Gelb and Barland's hand-scored finding that Sinclair had failed, the computer determined that he had passed (and with flying colors at that)! Barland writes (at para. 6):
Quote:I scored the printout of the second series of charts (regarding cocaine), but was not satisfied with the quality of the electrodermal channel on one of the charts. When I received the digital data and optimized the channel, I used the Federal 7 position scale and the 2007 DACA reaction criteria to evaluate the charts. I scored the charts as -7 (Deception Indicated). I also evaluated the second series using the computer algorithm PolyScore (v. 6.0). It evaluated the charts as No Deception Indicated, and calculated the probability of deception as being less than .01 on a scale from .00 to 1.00. This was inconsistent with my numerical analysis. This is a relatively uncommon occurrence. The DACA guidelines indicate that when there is conflict between the examiner's or reviewer's score and Polyscore, the human score takes precedence. The computer algorithms are considered to be useful supplements, but they are not definitive, I therefore concur with Mr. Gelb's conclusions that Mr. Sinclair showed indications of deception on both test issues.
So the PolyScore algorithm (that Ed Gelb so
hailed in his polygraph report for Wendy Ellis) found Sinclair
truthful with a
less than 1% probability of deception! But in this case, Gelb and Barland (who cannot have been unaware of the firestorm of controversy that would have resulted had they found Sinclair non-deceptive with regard to this question) somehow reached a completely opposite conclusion!
Gelb did not provide Barland with the computerized data for the examination on Sinclair's sex allegations, and thus he was not able to run PolyScore on them.
Pages: 
Larry Sinclair and the Polygraph (Read 91833 times)
