Date of Graduation

2016

Document Type

Thesis

Degree Type

MS

College

Eberly College of Arts and Sciences

Department

Mathematics

Committee Chair

Jacqueline Speir

Committee Co-Chair

Casey Jelsema

Committee Member

Robert O'Brien

Abstract

Crime scene analysts are the core of criminal investigations; decisions made at the scene greatly affect the speed of analysis and quality of conclusions that directly impact the successful resolution of a case. If an examiner does not recognize the pertinence of an item on scene, it cannot be used to support his or her theory regarding the crime. Conversely, unselective evidence collection will include irrelevant material, thus increasing a forensic laboratory's backlog and sending the investigation into an unproductive and costly direction. Therefore, it is critical that analysts recognize and properly evaluate forensic evidence that can support or disprove hypotheses and accurately reconstruct events. With this in mind, the aim of the study was to determine if quantitative eye tracking data and qualitative observations could be used to distinguish investigator expertise. In order to assess this, 32 participants were recruited and categorized as experts or trained novices based on their educational and practical backgrounds. Each volunteer then processed a mock crime scene while wearing a mobile eye tracker, wherein visual fixations, duration, search patterns and qualitative reconstruction accuracy were evaluated. Quantitative eye tracking data were compared using pursuit percentage on areas of interest (AOIs), Earth Mover's Distance (EMD) and the Needleman-Wunsch (N-W) algorithm. Results indicate significant group differences (Mann-Whitney U test, alpha = 0:05) for two out of 14 AOIs selected for analysis in this study. In addition, significant group differences were also detected for both search duration on specific AOIs (EMD), as well as search sequence (N-W), wherein experts exhibited more dissimilar search durations, but more similar search sequences than their novice counterparts (with possible implications regarding hypothesis-based scene reconstruction). Finally, hierarchical and k-means clustering based on multivariate AOIs suggest that latent variables may be present, which is the topic of future research. In addition to the quantitative visual comparisons, each participant's reconstruction skill was assessed using a 22-point binary system. Significant group difference (Mann-Whitney U test, alpha = 0:05) was detected as a function of total reconstruction accuracy. However, the total number of cases processed does not correlate linearly (or well) with total reconstruction score. Equally of interest was the fact that canonical discriminant analysis (CDA) provided good group separability as a function of participant education (which was found to be more important for this cohort than professional development). Overall, results from this study found significant and interesting group differences, but likewise revealed the complexity associated with using gaze behavior as a means of assessing cognitive processes.

Share

COinS