The Likelihood Ratio (LR) provides an index of how much a test result will change the probability or odds of having a condition after a known or assumed prior incidence rate (base rate). In the case of polygraph testing, the condition of interest is involvement in the issue under investigation. The LR+ tells us how much more likely it is that a person is lying than not, after failing a polygraph test, compared with the likelihood before he or she sat in the chair and completed the test. If a person produces a truthful test result, the LR-tells us how much more likely the person is to be telling the truth than before the test. LRs may also be used to compare the efficacy of two or more scoring and decision models, for a given or assumed base rate. The advantage of the LR, compared with traditional Bayesian metrics such as positive predictive value (PPV) and negative predictive value (NPV) is that the LR is inclusive of inconclusive results, and will provide information that more accurately generalizes to field settings.
