Quantifying the Weight of Evidence from Forensic Fingerprint Comparisons

• Activity Number: 313
• Type: Invited
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Isolated Statisticians
• Author(s): Cedric Neumann*+ and Mina Yoo
• Companies: Penn State University and Penn State University
• Keywords: fingerprint ; weight of evidence ; likelihood ratio

Abstract

Forensic Science has been described as the "science of individualization". For most of the past century, forensic scientists have exclusively relied on the idea that nature doesn't duplicate itself to form categorical conclusions on the source of evidence recovered from crime scenes, and to report them with absolute certainty. The introduction of DNA profiling 20 years ago has led scientific and legal scholars to increasingly challenge the lack of scientific foundations of pattern-based evidence such as fingerprint, shoe impressions and tool marks. The determination of the source of prints recovered on crime scene is essentially a probabilistic inferential process, which can be supported by statistical models quantifying the weight of evidence of fingerprint comparisons. An ongoing research project, aiming at developing and validating a model to support fingerprint examiners in casework, has resulted in several iterations of such model. This presentation will report the latest development of our model and will address the implications of its use for forensic practice. More generally, some myth in the purpose of statistical research in forensic science will also be discussed.