This is the Final Technical Report of a study that solicited responses and conclusions from footwear examiners in the United States in order to infer accuracy, reproducibility, predictive value, and decision rules in forensic footwear examination and interpretation.
This research was conducted because there are few footwear reliability studies that report on the accuracy and reproducibility of conclusions drawn by examiners when evaluating the same case materials. This project prepared and distributed simulated case materials to 115 footwear experts. Of these participants, 77 completed all analyses, resulting in a total of 840 usable conclusions. These conclusions were used to compute several numerical metrics, including consensus; inter-rater reliability; accuracy; predictive value; and inferred decision rule support, coverage, strength, and confidence. The induced rules were evaluated as a function of strength, support, coverage, and confidence, and the highest conditional probability (the probability that a set of conditions will be reported given a specific conclusion/decision) was no more than 0.46. This suggests that although consistency in outcomes/conclusions is apparent across examiners comparing the same simulated case materials, the reasoning reported to justify these conclusions is more variable and requires closer examination in future white-box studies. Extensive tables and figures
Downloads
Similar Publications
- Estimation of Sex Assigned at Birth Using Dental Crown and Cervical Measurements in a Modern Global Sample
- Raman Spectroscopy for the Time since Deposition Estimation of a Menstrual Bloodstain
- How the Work Being Done on Statistical Fingerprint Models Provides the Basis for a Much Broader and Greater Impact Affecting Many Areas within the Criminal Justice System