NCJ Number
52640
Journal
New Scientist Volume: 67 Issue: 964 Dated: (AUGUST 1975) Pages: 476-478
Date Published
1975
Length
3 pages
Annotation
A LIE DETECTOR BASED ON VOICE ANALYSIS IS DESCRIBED AND CAUTION IS URGED IN ITS USE. ALTHOUGH THE MANUFACTURER CLAIMS 90 PERCENT ACCURACY, THIS RESEARCH FINDS NUMEROUS POSSIBILITIES FOR ERROR.
Abstract
THE OPERATION OF VOICE ANALYSIS DEVICES AND THEIR USE AS SECURITY TOOLS ARE EXPLAINED. IT IS SUGGESTED THAT SUCH DETECTORS MAY ELIMINATE SOME OF THE ERRORS COMMON TO THE POLYGRAPH, BUT THAT THESE DEVICES HAVE THEIR OWN PROBLEMS. A TYPICAL TEST IS DESCRIBED, AND CHANCES FOR MISINTERPRETATIONS OF VOICE STRESS AT EACH POINT ARE EXPLAINED. IT IS ALSO CONCLUDED THAT THE DEVICES ARE NOT COST EFFECTIVE. THE EQUIPMENT IS EXPENSIVE, THE OPERATORS COMMAND A HIGH SALARY, AND COMPANIES MIGHT HAVE TO FOREGO MORE PROVEN SECURITY MEASURES. A TEST BY THE U.S. ARMY FOUND THE VOICE DETECTORS NO MORE EFFECTIVE THAN TRADITIONAL POLYGRAPHS OR OBSERVATIONS BY TRAINED PERSONS. AN APPENDED CASE STUDY REVIEWS A LAWSUIT BROUGHT BY 10 EMPLOYEES WHO WERE FIRED AFTER A VOICE ANALYSIS TEST. THE DEVICES ARE SAID TO HAVE GREAT POSSIBILITIES AS RESEARCH TOOLS. VARIOUS RESEARCH USES ARE BRIEFLY SUMMARIZED. (GLR)