Quickfix,
You write:
Quote:
What was the question? I forgot already.
OK, now I remember. To Dr. Richardson, my optimism is based on continuing research, both at DODPI and universities trying such methods as brain-scanning, eye movement, and other non-traditional lie-detection methods. When the day comes that current polygraph technology can be replaced by better technology, I'll hop on the train. Lokking back at history, if I remember my DODPI history classes, polygraph instruments began with a glass jar filled with water, watching it rise and fall while a subject was questioned. I recall a three-channel instrument, with one pneumo tube; then came four and five-channel instruments with electronic pneumo and cardio channel; Lafayette had an adjustable discrotic notch, Stoelting had a "quality watch" feature which prevented manipulation of sensitivity settings by examiners. Circa 1993, Axciton and other manufacturers produced the computerized polygraph program based upon, if I recall correctly, algorithms developed by Johns Hopkins APL. Now Limestone is out with the next generation of computerized polygraph. Earlier movement bars are now replaced with Piezo movement pads. So, technology is advancing compared with what was the standard 75 years ago. But I think your assertions are more along the lines of methodology vice technology. Yes, I agree that the R/I technique is the least desirable and least accurate of all. PLCs have been around for decades, some techniques evolving from the original ZCT. TES, derived from the DLCT, which has been used for 30+ years, came along in 1993, after extensive research, and is widely used among the military services. The next technique developed may be extremely accurate compared to what we have now.
That's my source of optimism.
Regards
Absolutely none of the dependent variable measures you mention nor the computerized data acquisition and various scoring algorithms are worth a tinker's damn as long as the basic application is flawed. The relationship between relevant and control/comparison question responses has no similarity to analyte and control in an assay with true scientific control. No tinkering with dependent variables, data transformations, scoring algorithms, etc. will improve the state of things until major (to include basic theoretical understanding) advances occur with the independent variable (basic paradigm) side of the equation.
With regard to the TES (which you seem to indicate is still the mainstay of military personnel screening), it has been completely discredited and disavowed by Dr. Sheila Reed (the scientist in charge of the various DoDPI validation studies in the early 90's). The notion that our nation is being protected by this application/format is just plain silly/scary. I hope you understand why I don't begin to share your optimism stemming from the research and technological advances you cite.