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

Comparison of Evidentiary and Investigation Decision Rules: A Replication

NCJ Number
213953
Journal
Polygraph Volume: 35 Issue: 1 Dated: 2006 Pages: 55-62
Author(s)
Donald J. Krapohl; Barry Cushman
Date Published
2006
Length
8 pages
Annotation
This research sought to replicate Krapohl's research (2005) in comparing traditional polygraph decision rules to those he proposed for use in court and paired testing (Marin).
Abstract
Krapohl noted from previous research findings that the cutoff scores used by many investigative organizations have been significantly less sensitive to truthfulness than deceptiveness. These cutoffs, referred to in this paper as Investigative Decision Rules, tend to minimize false negative outcomes. Investigative Decision rules are useful in settings in which the costs of false negatives are high and those of false positives are low. They present a problem, however, in settings where the costs of errors shift. Krapohl devised a new set of decision rules, called Evidentiary Decision Rules, which he tested on four independent datasets. The new rules produced a more balanced accuracy than when the Investigative Decision Rules were applied to the same scores; and they permitted most scorers to meet the stringent Marin Protocol standard of limiting inconclusive decisions to 20 percent or lower while achieving an accuracy of 86 percent or greater with the remaining cases. In replicating Krapohl research with his new rules, the current study confirmed that the standard Investigative Decision Rules (+/-6 cutting scores and -3 spot scores) did not produce the balanced accuracy and low inconclusive rate of the Evidentiary Decision rules (two-stage process: -6 and +4 cutting scores followed by -3 spot score for those cases that would have otherwise been found inconclusive). The current study randomly drew 100 cases from the same large database of confirmed cases used by Krapohl and McManus (1999). Half of these field cases were from deceptive examinees, and the other half were from truthful examinees. The processing of the data was identical to that used by Krapohl. 2 tables, 1 figure, and 15 references