The current field of forensic practice heavily relies on human judgment when analyzing patterned sensory evidence found at crime scenes, such as tool marks and fingerprints. Trained observers compare the patterns in the evidence with exemplar patterns from the suspected source to determine if they are similar enough to have come from the same source. While these forensic pattern comparison disciplines have been valuable in criminal investigations, the potential for convicting innocent individuals has led to calls for caution and the development of better practices. The scientific community studying human information processing has responded to these calls, offering insights into sensory measurement, discrimination, and classification in a forensic context. By applying established theoretical and empirical approaches from sensory science, the vulnerabilities of current pattern comparison disciplines can be illustrated, and specific strategies for improvement can be proposed.
Full paper with PDF download. This paper was published on PNAS.org.
https://www.pnas.org/doi/10.1073/pnas.2206567119#sec-7