U.S. flag

An official website of the United States government, Department of Justice.

Interpretation of chemical data from glass analysis for forensic purposes

NCJ Number
300960
Journal
Journal of Chemometrics Dated: 2020
Author(s)
Date Published
2020
Annotation

This article reports on a study that demonstrates the utility of likelihood ratio (LR) calculations that use novel datasets of glass samples of known manufacturing history.

Abstract

The aims of evaluating forensic evidence are to provide a transparent, coherent, and unbiased opinion of the value of the evidence to fact finders. Measurements from glass evidence in a hit-and-run case, for example, can assist in deciding whether a particular vehicle was involved in the accident. The evaluation involves the comparison of the physical, optical, and chemical properties of the glass recovered from the broken window with glass fragments suspected of originating from the window. A standard method (ASTM E2927 16e1) describes a consensus-based approach to sampling, sample preparation, quantitative analysis and “match” criterion for comparison of chemical properties. The result is a binary decision of either finding a difference in the elemental composition (exclusion) or a failure to exclude, based on elemental composition. In the reported study, the LRs calculated from comparing glass manufactured at three plants over relatively short periods (over 2-6 weeks) ranged from low values when the glass samples were manufactured at different plants or to very high values when the glass samples were manufactured on the same day. Although the glass samples being compared may not have originated from the same broken window source, they did exhibit chemical similarity within the lower and upper bounds and the LRs presented in this article for the first time. They closely correlated chemical relatedness to manufacturing history, specifically the time interval between production. The study results support the use of the match criteria recommended within ASTM E2927-16e1 and provide a data-driven path forward for expanding the interpretation of glass using LRs. (publisher abstract modified)

Date Published: January 1, 2020