U.S. flag

An official website of the United States government, Department of Justice.

Use of an Automated System to Evaluate Feature Dissimilarities in Handwriting Under a Two-Stage Evaluative Process

NCJ Number
255123
Date Published
Author(s)
Cami Fuglsby, Christopher Saunders, Danica M. Ommen, Michael P. Caligiuri
Agencies
NIJ-Sponsored
Publication Type
Article
Annotation
This study repurposed a commercially available automated system to generate empirical distributions for ranking feature dissimilarity scores among pairs of handwritten phrases.
Abstract
The two stage evaluative process is an established framework used by forensic document examiners (FDEs) for reaching a conclusion about the source of handwritten evidence. In the second, or discrimination, stage, the examiner attempts to estimate the rarity of observations in a relevant background population. Unfortunately, control samples from a relevant background population are often unavailable, leaving the FDE to reach this determination based on subjective experience. Automated handwriting feature recognition systems can perform both feature comparison and discrimination, yet these systems have not been subjected to empirical validation studies. The blinded results of the automated process of the current study were used to survey an international cohort of 36 FDEs regarding their strength of support for same and different writer propositions. The survey served to cross validate FDE decision making under the two stage approach. Results from the survey demonstrated a clear pattern of response consistent with ground truth. Predictive regression analyses indicated that the automated feature dissimilarity scores and the log of their cumulative distribution functions accounted for 72 percent of the variability in FDE opinions. This study demonstrated that feature dissimilarity scores acquired using automated processes and their distributions are closely aligned with FDE decision making processes supporting the heuristic value of the two stage evaluative framework. (publisher abstract modified)
Date Created: October 18, 2020