U.S. flag

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

Dot gov

The .gov means it’s official.
Federal government websites always use a .gov or .mil domain. Before sharing sensitive information online, make sure you’re on a .gov or .mil site by inspecting your browser’s address (or “location”) bar.


The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Strengthening the Evaluation and Interpretation of Glass Evidence Using Statistical Analysis of Collection Sets and Databases of Refractive Index and Elemental Data

NCJ Number
Date Published
Jose Almirall
This is the Final Report on the findings and methodology of a research project with the goal of improving the value of glass-evidence analysis by developing objective and quantitative interpretation guidelines for the evaluation and reporting of glass-evidence analysis results.
The project first focused on the development of a reporting scheme based on previously reported likelihood ratio calculation by Aitken and Lucy, in conjunction with a “calibration” step previously reported by Ramos. The goal of this first part of the project was to develop a database derived from the analysis of glass using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). Typically, a match criterion is used to compare the known and questioned sample; however, match criteria have several disadvantages that can be overcome with an alternative approach. This approach is the likelihood ratio (LR). Two LA-ICP-MS glass databases were used to evaluate the performance of the LR. These were a vehicle windshield database (420 samples) and a casework database (385 samples). Compared to the match criterion, the LR achieved improved false exclusion rates and produced similar false inclusion rates. In addition, the LR limited the magnitude of the misleading evidence, providing only weak support for the incorrect proposition. The LR was also tested in a series of three inter-laboratory studies. Good correct association rates (94-100 percent) were obtained for same-source samples for all three inter-laboratory exercises. All different-source samples were correctly excluded with the LR. 3 tables, 1 figure, and 52 references
Date Created: November 24, 2019