Evaluating the detectability of randomly acquired characteristics under crime scene–like conditions

Randomly acquired characteristics (RACs) on shoe soles may provide information relevant to the evaluative comparison between a crime scene footwear print and a known reference shoe, yet their detectability in realistic conditions remains uncertain. To address this question, 30 experiments were conducted in which participants wearing preselected shoes with numerous RACs spontaneously left shoeprints on different surfaces, while unaware of the study’s aims. In each trial, a trained field investigator located and collected the prints following standard protocols. A total of 302 shoeprints were analyzed, yielding 488 detected RAC instances out of 3464 RAC observation opportunities. Data were recorded at two levels. The first included crime-scene print-specific information: surface type (floor, wooden chair, or bristol board) and surrounding quality, linked to each RAC within the corresponding print. The second captured shoe-specific information, documenting for each RAC its outsole location, size, and type. RAC detection probabilities were estimated by logistic regression using Generalized Estimating Equations (GEE), to account for dependence among RACs on the same shoeprint. Results showed that RAC detectability is far from perfect, even under favorable conditions, and varies systematically with surface type, print quality, and RAC characteristics. Specifically, model based predicted detection probabilities for large RACs ((Formula presented) ) under favorable conditions range from 45% to 66% (by RAC type), whereas for small RACs ((Formula presented) ) under adverse conditions they drop to 0.4%–2.9%. These insights should guide the evaluation of the evidential value of both observed and unobserved RACs and inform the interpretation of footwear evidence in court.