FITS—Functional Implementation of Thorough and Systematic Approaches for Fracture Examinations

Tuesday, June 4, 2024 | 1:00 PM - 2:00 PM Eastern
Duration: 1 hour

Overview

The separation of materials such as tape, plastics, and textiles from their original source frequently occurs during violent activities, leaving distinct patterns along the fractured edges. These features assist examiners in determining if two pieces were once joined together, also known as a "physical fit." This webinar presents the findings of an NIJ-funded project aimed at assessing the strength of trace evidence fracture fits through a practical, comprehensive, systematic, and quantifiable approach.  

The presentation covers three main outcomes: 

1. First, a systematic method is outlined for comparing and documenting fracture fits of common trace materials such as duct tapes, textiles, and automotive plastics, using human-based protocols and automated computational algorithms. This study describes material-specific relevant features for various materials and the use of reporting templates to facilitate thorough and reproducible documentation of an analyst's decision-making process while minimizing bias risks. It also establishes criteria for utilizing quantitative metrics, such as the edge similarity score (ESS), which estimates the quality of a fit, and the feature prominence score (FPS), which captures the relative features' importance in each comparison. 
2. Second, an extensive database of nearly 9,000 samples was built to evaluate performance rates in this field and assess the weight of a fracture fit using similarity metrics and probabilistic estimates. 
3. Third, the evaluation of the methods through inter-laboratory collaborations that can establish consistency base rates. 

We also share our experience on how our partnership of forensic researchers, computational material science physicists, statisticians, and practitioners helped answer fundamental questions such as: 

• Do all physical fits hold the same probative value?  
• Which individual or class characteristics can be evaluated in fractured edges to assist the forensic examiner during a physical fit examination? 
• Are these features dependent on the fractured or separated material? 
• Which factors influence the occurrence of these features and the quality of a physical fit? 
• What are the performance rates of physical fit examinations? Are the performance rates dependent on the type of object? 
• Can quantitative metrics demonstrate the quality of a fit and be used for the probabilistic interpretation of the evidence? 
• Can computational and mathematical models be used to complement human-based examinations? 
• What strategies can be developed to minimize potential bias and subjectivity during the forensic examination of physical fits? 
• Can standardized protocols be developed for the examination, documentation, and interpretation of physical fits through the assessment of the method via large datasets and interlaboratory studies? 

This study is anticipated to transform current trace evidence practice by providing harmonized examination protocols and decision thresholds; effective mechanisms to ensure a transparent and systematic peer-review process and interlaboratory testing; and a quantitative basis that substantiates the evidential value of fracture fit conclusions. We seek future collaborations to assist with implementing the methods in forensic agencies. 

Detailed Learning Objectives

1. Attendees will identify how the proposed approach can be used to standardize comparison criteria, streamline independent review process, demonstrate the evidence in court, or as a training tool. 
2. Attendees will identify the material-specific individualizing features for duct tapes, textiles, and automotive polymers and the factors that can influence error rates in fit examinations. 
3. Attendees will recognize the challenges in physical fit examinations and how the information derived from this research can be used to substantiate the scientific foundations in the field. 

Presenter

• Tatiana Trejos, Ph.D. | Associate Professor, Department of Forensic and Investigative Sciences, West Virginia University