A Survey of Likelihood Ratio Method Development and Implementation AcrossMultiple Forensic Disciplines

• Activity Number: 107 - SPEED: Statistical Methods, Computing, and Applications Part 1
• Type: Contributed
• Date/Time: Monday, August 8, 2022 : 8:30 AM to 10:20 PM
• Sponsor: Section on Statistics in Defense and National Security
• Author(s): Lulu Chen* and Larry Tang and Jonathon Phillips
• Companies: University of Central Florida and University of Central Florida and National Institute of Standards and Technology
• Keywords: Forensics; likelihood ratio; score-based likelihood ratio; forensic evidence; handwriting; fingerprint

Abstract

The 2009 National Research Council report “Strengthening Forensic Science in the United States: A Path Forward” calls for the quantification of forensic evidence, and the 2016 President’s Council of Advisors on Science and Technology report in Forensic Science emphasizes on scientific validity and reliability of forensic methods. The likelihood ratio methods in forensics, including Bayesian likelihood ratio methods and score-based likelihood ratio methods, provide means to weigh the forensic evidence and update the prior odds. The reviews on the likelihood ratio methods in forensics have mainly focused on individual disciplines. The survey on the development and applications of likelihood ratio methods across disciplines has been scant. In this article we provide such a survey across four important forensic disciplines including glass, handwriting, fingerprint, and speaker recognition in the development, applications, and evaluation and comparison of likelihood ratio methods. A reference table is provided to relate the articles to their discipline and statistical methods developed in the articles.