Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences


Forensic and Investigative Science

Committee Chair

Tatiana Trejos

Committee Member

Luis Arroyo

Committee Member

Patrick Buzzoni

Committee Member

James Curran

Committee Member

Keith Morris


Glass and paint particles are frequently discovered at crime scenes as valuable trace evidence. Often, these fragments are transferred from a crime scene to a victim or person-of-interest, and can subsequently be evaluated, analyzed, and classified according to their physical and optical measurements, and chemical properties, each of which can provide important information to the analyst to assist in the determining if a recovered particle is distinguishable or indistinguishable from a known source. Moreover, these minute materials can provide information about how the events took place, leading to the reconstruction of crime scenes and the identification of valuable clues for the investigation.

Currently, there are standardized methods for examining and interpreting glass and paint evidence. In the United States, most trace examiners use a frequentist approach, where the population value of the true parameter (usually the mean) is regarded as fixed, but unknown, and the data are regarded as random. Confidence intervals, based on the data, are calculated for this value. However, alternative models, such as a Bayesian framework are more widely used by forensic agencies abroad and are gaining interest within the U.S. system. One advantage of Bayesian interpretation is that it can provide a more encompassing approach by using likelihood ratios to formally quantify the weight of the evidence under alternative propositions. These likelihood ratios can be evaluated at the source and activity levels using background information from previously performed studies to determine the probabilities of transferring traces. This baseline information on random occurrence, transfer, and persistence of trace materials is a critical input for these models. Unfortunately, most background studies on this topic have varying demographic, socioeconomic, and geographic information that differs from the United States.

The paucity of fundamental information on the prevalence of trace materials on the general public relative to those involved in an alleged event also decreases the efficiency of decisions made while collecting evidence at the crime scene and formulating significant investigative leads. Ground knowledge on how common it is to find glass and paint traces in the general population, their relative incidence by material type (e.g., architectural vs. vehicle), or by location (e.g., shirt vs. footwear) are a few examples of questions that should be substantiated with empirically verifiable data. Consequently, the overall goal of this study was to obtain baseline data on the frequency of occurrence of glass and paint relevant to the U.S. territory to fill out this existing gap.

This long-term goal is to facilitate the future adoption of probabilistic interpretation models in the US criminal justice system by creating a baseline for the frequency of occurrence of glass and paint within the United States, followed by applying the Bayesian approach to the data and for the evaluation of mock cases at the source and activity levels, which will allow the use of likelihood ratios to formally quantify the weight of the evidence under competing propositions.

This project addressed essential factors never evaluated before in a single and systematic study. The study provides data from four different cities in two geographical regions of the United States, including small and metropolitan areas with diverse socioeconomic and demographic conditions. It also evaluates the frequency rates of these traces in different seasons. It considers factors that may influence the retention of glass and paint on apparel, such as modes of transportation and clothing and footwear worn. Finally, a full characterization of features of interest in the recovered traces by appropriate analytical techniques permits the evaluation of the relevance of glass and paint occurrence by major end-uses.

Tape lifts and sole scrapings (2,391) were collected from 510 participants and up to six clothing and footwear areas per individual. Glass fragments were analyzed via polarized light microscopy (PLM), refractive index (RI), micro-X-Ray fluorescence (µXRF), and scanning electron microscopy-energy dispersive spectroscopy (SEM-EDS), while paint specimens were examined by light microscopy and infrared spectroscopy (µFTIR). Overall, the background studies determined that:

a) Glass is less frequently recovered than paint in the general public. For example, paint residues were found in 24% of the overall background participants, compared to only 2.9% for glass.

b) It is uncommon to recover glass from an individual's upper and lower garments, but relatively more common to recover it from footwear. In contrast, recovering paint from the upper and lower garments was more common than footwear.

c) The parallel occurrence of glass and paint in a single individual was very rare (2 out of 510 individuals), and even more uncommon to find both traces in a single item (none of the garments or footwear contained glass and paint on the same item).

d) Most of the glass recovered from the background populations was classified as a container followed by sheet soda-lime-silicate, with some specialty formulations also encountered in the Houston and Pittsburgh collections.

e) A higher occurrence of glass and paint was found in the winter than in summer in a city where the average temperatures dropped in the winter by approximately 40 °F. The study indicates that differences in the clothing worn and the primary modes of transportation during these seasons affect the background of these traces in that background population.

f) The factor that was most significantly different across the four cities was the mode of transportation, while the clothing type was relatively similar, except for the winter collection set.

This research also used the conducted baseline studies to apply the Bayesian framework to two mock cases representative of reported crimes in the United States, which included collection, recovery, analysis, and interpretation. Using this framework, the weight of the evidence was assessed at both the source and activity levels by calculating P and S (background probabilities), and T (the activity probabilities).

During the background study collection, nail polish fragments were recovered from volunteers. Thus, this study extended its scope to assessing the evidential value of nail polish in forensics, specifically to evaluate the variability between gel nail polish and nitrocellulose-based (regular) polishes and variability within single sources (i.e., same container). This study determined the basis for classifying gel and regular polishes and their main features to distinguish them by visual, microscopical, and chemical properties. From 8,128 pairwise comparisons, among the 128 polish samples, 9 pairs remained indistinguishable from each other with micro-chemical testing. Attenuated total Reflectance-µFTIR (ATR-FTIR) further discriminated four additional pairs, leading to an overall combined discrimination of this dataset of 99.9%.

This research builds upon the essential understanding of the frequency at which trace particles are found on members of the general population who have not been involved in a crime. The data created from the background portion of this study is anticipated to offer the criminal justice system valuable knowledge for integrating more comprehensive models to assess the evidential value of trace evidence in U.S. courts.