U.S. flag

An official website of the United States government, Department of Justice.

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Decision Accuracy for the Relevant-Irrelevant Screening Test: A Partial Replication

NCJ Number
Polygraph Volume: 43 Issue: 1 Dated: 2014 Pages: 20-29
Donald Krapohl; Terry Rosales
Date Published
10 pages
Because of some problems noted in the 2005 exploratory analysis of the accuracy of the Relevant-Irrelevant (RI) polygraph screening test, the current study replicated a portion of that study that involved global analysis of RI cases.
The study added positive evidence on the validity of the RI screening test as a screening instrument. Overall decision accuracy was greater than chance, thus, the RI screening test meets the minimum validity requirements of the American Polygraph Association's (APA's) Standards of Practice. Consistent with earlier research, this technique is effective in identifying deceptiveness, but significantly less so for truthfulness. This suggests that RI is appropriate as a component of a successive series of hurdles in screening. As a stand-alone method, the RI is recommended only for settings in which there is a high tolerance for false positive errors because of its weakness in detecting truthfulness. RI analysis is further handicapped by the lack of response benchmarks that other techniques have in the form of probable-lie and directed-lie comparison questions. Absent these benchmarks, RI data rely on more ambiguous rules or assessments, which lead to lower reliability and ultimately lower accuracy. The study involved four federally certified polygraph examiners with experience in conducting and evaluating the RI screening test. The testing had been conducted with Axciton computer polygraphs. All cases came from a federally funded project in the mid-1990s in which a security firm conducted RI screening cases of candidates for employment as contract security guards. A total of 733 confirmed cases constituted the population from which the study sample was drawn. 6 tables and 7 references