Technical note: Proposed method to objectively evaluate gunshot residue comparisons does not generalize to...

Area of Science:

• Forensic Science
• Analytical Chemistry
• Trace Evidence Analysis

Background:

• Previous likelihood ratio systems for gunshot residue (GSR) analysis focused on same-location sample pairs.
• Casework often requires evaluating GSR samples from different locations (e.g., hand-to-cartridge case), a scenario not previously addressed.
• Lack of ground truth for different-location pairs limited previous model validation.

Purpose of the Study:

• To assess the impact of sampling location limitations on a GSR elemental composition comparison system.
• To evaluate the system's performance using an experimental dataset with ground truth for both same- and different-location pairs.
• To determine the system's readiness for application in forensic casework.

Main Methods:

• Development of a likelihood ratio system for pairwise GSR sample comparison based on elemental composition.
• Utilizing an experimental dataset containing GSR samples from both same-location and different-location pairs with known ground truth.
• Evaluating model performance using the C_llr metric before and after introducing different-location pairs.

Main Results:

• A significant decline in system performance was observed when evaluating different-location sample pairs.
• The C_llr metric deteriorated sharply from 0.35 to 0.98, indicating reduced discriminatory power.
• Attempts to improve performance through system extensions did not yield significant enhancements.

Conclusions:

• The current GSR elemental composition comparison system is not robust enough for direct application in forensic casework due to sampling location limitations.
• The system's performance degradation highlights the need for further development to accommodate mixed-location sample comparisons.
• Future research should focus on addressing the challenges posed by different-location GSR sample analysis.