posted 12-12-2007 10:43 AM
Buster,If you can send that article to me, I'll post a link for download.
I'm not aware of much. Stan Abrams did a couple of things during about the mid 70s, I think those pertained to mentally retarded and/or psychiatric patients. There was an articles from 1981 in the journal, which I think we should be VERY cautious about using. It is mostly a syllogistic discussion involving a host of careless assertions and unsupported conclusions.
Abrams' findings were not good - which is not surprising, because those folks are what we call outliers (their functional characteristics are well outside of the normal range). It is not theoretically simple (or perhaps even feasable or desirable) to have a test that can at once manage the strata of functional characteristics across normal and outlier population groups.
For example: autistic and other mentally retarded persons do quite poorly on IQ tests, but IQ tests are not sufficient of themselves to diagnose autism (and cannot even be used by themselves to diagnose mental retardation - that's why we use them in combination with tests of adaptive functioning). Other tests, like the Gilliam Autism Rating Scales (GARS) are much better at describing the range of adaptive and functional skills for autistic persons. However, the test results of a non-autistic person on the GARS would be entirely uninformative, even if they get a perfect score. OK, this is a gross example, but it makes the point, and there are smaller versions of this situation in many testing contexts.
All I'm suggesting is that we not try to be cavalier or irresponsible about the capabilities of the polygraph with exceptional persons or persons whom we know are outliers to the normal range of functioning for the populations on which the polygraph is normed or validated.
At present, our polygraph methods were developed with populations including non-psychiatric criminal suspects, non-psychiatric volunteers for university lab studies, cohorts of presumably non-psychiatric persons for gummit research. In the case of OSS-3, our cohorts include non-psychiatric LE applicants, and non-psychiatric sex offenders in community supervision and treatment programs.
Rule of thumb: normative data and normative decision models/thresholds pertain to normal functioning individual from the intended populations represented by the study samples The application of normative decision models to known outliers should always be regarded with caution. I think you'd find a tough argument, against some opposing counsel's expert, to include such polygraph results in a rights-affecting legal decision. The big hazard is that any adverse legal decision, or even bad experience, affects polygraph as a whole. Just look at the Ohio decision regarding the 12 year-old with a neurologically based learning disorder.
r
------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)