Author ORCID Identifier

https://orcid.org/0009-0009-9427-3854

Semester

Spring

Date of Graduation

2026

Document Type

Dissertation

Degree Type

PhD

College

Eberly College of Arts and Sciences

Department

Forensic and Investigative Science

Committee Chair

Keith B. Morris

Committee Member

Casey Jelsema

Committee Member

James Nolan

Committee Member

Casper Venter

Abstract

Ideally, the evaluation of relevant forensic evidence must corroborate the hypotheses that the prosecution and/or the defense holds regarding the commitment of a crime. Therefore, optimal forensic evidence would be the one that can be found all the time under the same circumstances, and cannot be found otherwise, making the forensic scientist able to evaluate this evidence given the hypotheses provided. Nevertheless, this is hard to find, as ideal evidence being only a concept.

In this project, the probabilistic interpretation of relevant forensic evidence investigated in criminal cases with the aid of Wigmore charts and Bayesian networks was proposed. Using these methods, a possible decision model can be determined to establish the relevant evidential value at a crime scene of a specific type of felony to assist the trier of fact.

In any crime scene, the events that have taken place leave traces of their contact and movement. To elucidate and recreate these events, the forensic scientist or the crime scene investigator, includes his intuition and experience to evaluate the evidence. The identification and evaluation of the most relevant evidence of the crime will allow the investigators to reconstruct the events in a more precise way. The value or relevance of the evidence is the effect of how this element provides strong support to make plausible the formulated hypotheses.

However, what would be the ideal approach to evaluate the significance of the evidence? Is one piece of evidence more important than the other one? For this reason, it is critical to assess the general impact the recovered evidence has on the propositions (penultimate probanda) that support the final hypothesis to be established (ultimate probandum), and specifically how this evidence supports the investigator’s assumptions, and vice-versa.

Wigmore charts are tree diagram representations designed to support legal arguments with standardized graphical representations of linked multiple premises that ultimately support the main hypothesis. On the other hand, Bayesian networks are tools for statistical modeling that represent a group of correlated uncertainties. Due to its graphical structure and probabilistic approach, these networks are appropriate to design systems focused on classification, diagnosis, and decision-making processes. The elaboration and validation of these Bayesian networks is performed by the software Netica™ by Norsys Software Corporation, whereas Wigmore charts are elaborated using i2® Analyst’s Software.

This work aimed to assess relevant evidence, especially contradictory evidence, in criminal cases with the usage of Wigmore charts and Bayesian networks. Specifically, their construction, their parameter estimation and conditional probabilities, and their model to evaluate the impact of evidence on the main hypothesis or the ultimate probandum. To understand the impact that the evidence will have over the main hypothesis, these graphical tools will aid by the evaluation of the posterior probabilities, allowing a better understanding of which piece of evidence will have a greater or lesser influence on the final proposition.

This graphical assessment and statistical analysis were carried out on criminal cases, such OJ Simpson case (1992), and the Crewe Murders (1970). Both cases were chosen due to their easy access to their well-documented pieces of evidence, as their flexibility of analysis through the development of Bayesian networks and Wigmore charts. Additionally, taking advantage of their evidence records availability, this assessment was performed and studied on the murders of Mared Malarik and Karen Ferrell (1970) where the evidence was classified and assessed by their category; motives, the opportunity to commit the murders.

This project delivered a theoretical framework that assessed the significance of the relationship between the statistical assessment of relevant evidence and the primary proposition of the court (ultimate probandum) for a given crime.

Download

Share

COinS