Voice Stress Analysis: Only 15 Percent of Lies About Drug Use Detected in Field Test

• NCJ Number: 221502
• Author(s): Kelly R. Damphousse, Ph.D
• Date Published: March 2008
• Length: 4 pages
• Publication Series: NIJ Journal
• Agencies: NIJ

Annotation

In order to test the ability of two voice-stress-analysis (VSA) software programs to detect deception, this study used them with 319 recent arrestees in the Oklahoma County jail in an effort to detect deception in interviews about their drug use.

Abstract

The VSA programs, which are purported by their developers and manufacturers to detect micro-tremors in the voice caused by the stress linked to deception, were found to be no better in detecting deception about recent drug use than flipping a coin. Both VSA programs had a low sensitivity rate, identifying an average of only 15 percent of the arrestees who lied (determined from a subsequent urine test). Although the two VSA programs had approximately a 50-percent accuracy rate in determining deception about recent drug use, the VSA study produced a higher rate of truthful answers about recent drug use (only 14 percent of the sample lied about recent drug use) compared with a control sample in the same jail in which recent arrestees were questioned about recent drug use but without the use of a VSA program. Apparently, being told that they would be monitored by a VSA program that would detect deception was effective in deterring arrestees from lying about their recent drug use. The average cost of the cheaper of the two programs tested is $11,500, not including the personnel expense in using it. Law enforcement administrators and policymakers should weigh all the factors when deciding whether or not to purchase or use VSA technology. In testing the two VSA programs, arrestees were asked questions about their use of marijuana in the previous 30 days and their use of cocaine, heroin, methamphetamine, and PCP in the past 72 hours. 14 notes