Empirical Scoring System Multinomial (ESS-M)
ESS-M (Empirical Scoring System – Multinomial) is an important update to the Empirical Scoring System.
Science and scientific testing are a process of continuous research and study. The ESS-M is a product of ongoing development and research with the ESS, made possible by Lafayette Instrument Company.
ESS-M advances the polygraph in several ways. First, the ESS-M reference model is calculated under the analytic theory to the polygraph, and this is an answer to scientific criticism about the lack of a basic theory (which can be expressed mathematically). The ESS-M reference model serves as the likelihood function for the Bayesian analysis of polygraph results. Bayesian analysis, together with the use of odds as the statistical metric of interest, can provide for simpler and more intuitive discussion (compared to the difficult intuition for frequentist p-values) about the effect size of practical interest - the likelihood of deception or truth-telling. Finally, whereas the original ESS reference model did not include the vasomotor recording sensor, Bayesian analysis, together with the multinomial reference model, makes it possible to calculate the reference tables both with and without the additional vasomotor sensor.
Use of the ESS-M is virtually identical to the original ESS in terms of scoring features, score assignment, and decision rules. Improvements are mainly in the scientific, theoretic, and analytic foundations of the ESS-M reference model. For practical purposes only, the ESS-M cutscores have changed. Information in this guide explains the multinomial reference tables for those who are interested. For convenience, Lafayette’s LXSoftware includes a fully-featured ESS-M Report Generator that can automate the routine tasks associated with summarizing scores, executing decision rules, determining categorical and probabilistic results, and formatting a written narrative summary report.
Multiple experiments with archival datasets have shown the ESS-M to perform as well or better than the original ESS with event-specific/diagnostic exams and multiple-issue screening exams with 2, 3 or 4 relevant questions.
Nelson, R. (2017). Multinomial reference distributions for the Empirical Scoring System. Polygraph and Forensic Credibility Assessment 46(2), 81-115.
Nelson, R. (2016). Scientific (analytic) theory of polygraph testing. APA Magazine, 49(5), 69-82.
