Evaluating reporting practices in fingerprint comparisons using information theory: five response categories are better than three.

This article evaluates the effectiveness of different categorical decision scales used by forensic fingerprint examiners to communicate comparison results, focusing on two-, three-, and five-response-category scales. Using a Bayesian signal detection theory (SDT) model combined with the measure Expected Information Gain (EIG), which quantifies how much information examiner decisions provide about whether fingerprint pairs share the same source, the study finds that the five-response-category scale (e.g., exclusion, support for different source, inconclusive, support for same source, identification) yields higher information gain than scales with fewer categories. The five-category scale also encourages examiner decision thresholds closer to optimal, is more robust to risk-averse behavior and threshold misplacements, and maintains its advantages across a wide range of base rates of mated pairs. The findings support adopting a five-response-category scale in forensic fingerprint examination, as larger scales offer diminishing returns and may introduce interpretive challenges.