Likelihood Ratio in Evidential Evaluation of Polygraph Examination

Abstract

The starting point of the paper is the observation that the likelihood ratio (LR) is not used in the evaluation practice of — so important in the field of internal security — polygraph examinations. Meanwhile, LR is the only scientifically justifiable parameter that shows the evidential weight of particular evidence. The authors present theoretical attempts to use LR for evidential assessment of the polygraph examinations value and subject them to criticism. The main objective of the paper is to present the LR calculation procedure in the context of interpretation of a polygraph examination result treated as evaluative expertise. The following assumptions are made: the analysis includes only comparison question techniques; examination results enable to include a relevant subject only in one of the three categories: deception indicated, no deception indicated, inconclusive; there are various ways to assign LR; in the course of LR assignment, the arbitrary adoption of the values of some variables is admissible. Several examples of LR calculations are presented in different tactical configurations of polygraph examinations. The significance of including the inconclusive results in the examination technique characteristics is analysed. The possibility of applying the cumulative LR is indicated, however, leaving this question open. Consequences of the LR application in the interpretation of polygraph examinations are also presented as an argument in the criminal analysis. Conclusions show that treating polygraph examinations as evaluative expertise opens a new perspective on this method of forensic identification and deserves to be continued; however, the issue of the evidential use of polygraph examination results, in the light of the evaluation made with the use of the Bayesian approach, requires a number of further discussions among lawyers and scientists.