Evaluating Likelihood Ratio (LR) for firearm evidence identifications in forensic science based on the Congruent...

Area of Science:

• Forensic Science
• Ballistics
• Statistical Analysis

Background:

• Firearm evidence identification faces challenges due to subjective interpretations and lack of statistical foundation, as highlighted by NRC reports.
• Previous methods lacked objective measures for error rates and evidence weight, impacting legal proceedings.
• The National Institute of Standards and Technology (NIST) sought to develop a more scientific approach.

Purpose of the Study:

• To propose a likelihood ratio (LR) procedure for objective firearm evidence identification.
• To establish a scientific basis for evaluating the weight of firearm evidence.
• To enable quantitative error rate evaluation in forensic firearm analysis.

Main Methods:

• Development of the Congruent Matching Cells (CMC) method for automatic and objective firearm evidence identification.
• Application of a likelihood ratio (LR) procedure based on the CMC method.
• Evaluation using 9mm cartridge cases' breech face impression images with varying sample sizes and imaging techniques.

Main Results:

• Evaluated LRs for declared identifications were well above 10^6, indicating Extremely Strong Support for same-source propositions.
• Evaluated LRs for declared exclusions (3D images) were above 10^2, providing Moderately Strong Support for different-source propositions.
• The LR procedure demonstrated effectiveness across different sample sizes, imaging methods, and ammunition types.

Conclusions:

• The proposed LR procedure offers a scientific and objective framework for firearm evidence identification.
• This method provides a robust way to quantify the weight of evidence in firearm analysis.
• The findings support the use of LR in forensic science to address challenges in evidence interpretation.