More skilled, more cautious: Expertise and decision thresholds in fingerprint examination

Fingerprint examination relies on expert judgment to determine whether two impressions originate from the same source, yet variability exists across examiners evaluating identical evidence. The present study investigates the sources of this variability by distinguishing between examiner expertise and decision thresholds within a signal detection theory framework. Using data from two large-scale black box studies: the Latent Print Black Box 2 Study [1] and the Kibbey et al. dataset, we estimated examiner-level discriminability (d') and decision thresholds using an extension of signal detection theory with multiple decision criteria. Across both datasets, examiner's expertise as operationalized by (d') was positively associated with identification thresholds in the signal detection theory model, indicating that more skilled examiners adopt more conservative (risk averse) threshold when making same-source conclusions. In contrast, expertise was not associated with exclusion thresholds, suggesting that the threshold for different-source decisions operate independently of perceptual sensitivity. Additionally, we examined whether individual differences in personality predicted decision thresholds and found no statistically significant relationships, suggesting that these traits may exert influence primarily under conditions of ambiguity rather than in standard casework-like comparisons. This work extends prior signal detection theory findings by demonstrating that a change in discriminability (as measured by d') is accompanied by differences in decision-threshold placement, such that examiners with greater discriminability also tend to adopt a more conservative Identification threshold.