This article presents research into using glass fragment characteristics in criminal investigations.
The plausibility of finding glass fragments by chance in the general population becomes a relevant question when assessing glass recovered from someone suspected of being involved in an illegal activity implicating glass breaking. This research addresses this question in United States regions by searching the apparel of individuals assumed to be unconnected with crimes involving glass breaking. A large-scale study was conducted in the two metropolitan areas of Houston, TX, and Pittsburgh, PA, and the two suburban areas of Huntsville, TX, and Morgantown, WV. A total of 1,883 items of clothing and footwear were searched from 410 subjects, using tape liftings for garment areas and scrapings from the footwear soles. A total of eight glass fragments were recovered, five from the Houston set and one from each of the three other locations. No more than one glass fragment was recovered on a given item. The findings confirm that it is highly unlikely to observe glass fragments on the garments of individuals from the general population, and it is even less likely to find many fragments. Examination of glass by refractive index and micro-x-ray fluorescence spectrometry enabled their differentiation and classification by end-use. A zeta distribution and the R package fitPS developed for this study were used to estimate two parameters typically used for activity level likelihood ratio formulas: the probabilities of observing a given number of groups of fragments on target garments (P) and a particular number of fragments within an observed group (S). (Published Abstract Provided)