Exploitation of Very Small Particles to Enhance the Probative Value of Carpet Fibers

There is a fundamental limitation to the probative value of carpet fibers because their characteristics are determined by manufacture. As mass-produced commodities, probative value is limited to class associations. Our prior NIJ-funded research has demonstrated that very small particles (VSP), adhering to surfaces of individual carpet fibers, can link these fibers to their source carpet. These smaller particles accumulate post-manufacture, during the carpets use. This removes the fundamental limitation to probative value and provides an independent means to quantitatively test carpet fiber associations. This is a new approach, of revolutionary significance for the analysis of carpet fiber evidence, which will increase the probative value, provide a quantifiable measure of this value, and lay the groundwork for a similar approach to other traces. Proof of principle was established using carpet fibers cut from multiple areas of multiple carpets and a set of well-defined, frequently occurring particle types. Research is now required to refine this process, apply it under realistic casework conditions, and deliver a working prototype. Program goals are to (1) refine the process for exploiting (VSP) to associate residential carpet fibers with their source carpet, (2) apply this process under realistic casework conditions, and (3) deliver working prototype methods for collection, analysis and interpretation of this evidence. There are four specific technical objectives. 1. Apply and test the methods under realistic conditions of carpet fiber transfer and crime scene sampling. Using a simple, efficient collection kit, practitioners will collect samples from 100 crime scenes in 8 to 12 jurisdictions across the United States. Using protocols developed previously for VSP recovery, along with existing methods for computer-assisted SEM analysis, each of the 200 samples (100 mated pairs) will be analyzed. 2. Expand and improve VSP target particle type classification criteria based on analysis of forensic performance characteristics. Using between-carpet data from Objective 1, and existing within-carpet data, Target Particle Types (TPTs) will be defined using unsupervised clustering, resulting in a set of compositional variables to describe VSP profiles. 3. Design, develop, and deliver a practical method for the measurement of the degree of correspondence between two VSP profiles. Correspondence measures for the VSP profiles, expressed as occurrences of TPTs, will be developed using mathematical algorithms: classic distance algorithms as well as non-parametric statistics, artificial neural networks, and distance-based machine learning. 4. Design, develop and deliver practical, quantitative measures of the probative value of this degree of correspondence. The correspondence measure will be determined for each of the 100 mated pairs (same source) and for each of the 4850 unrelated pairs (different sources). Population parameters will be estimated within specified tolerances, and the probabilities of scores exceeding any given threshold value will be determined for the hypothetical cases (1) where pairs are from the same source, and (2) where pairs are from different sources. Contrasting support for these alternative hypotheses will provide a quantitative measure of probative value. A fifth program objective is a series of workshops to facilitate full discussion and considered response from the forensic science practitioner and researcher communities. Capabilities, limitations, and potential of this fundamentally new approach will be covered and the resulting feedback will be analyzed in the final report. The program will result in prototype methods for the collection, analysis, comparison and quantitative interpretation of carpet fiber evidence using profiles of very small particles found adhering to their surfaces. The methods will be ready for validation processes within crime laboratory settings and for next stage improvements. ca/ncf