Miami-Dade Research Study for the Reliability of the ACE-V Process: Accuracy & Precision in Latent Fingerprint Examinations

This research report presents an empirical study that evaluated the reliability of the Analysis, Comparison, and Evaluation (ACE) and Analysis, Comparison, Evaluation, and Verification (ACE-V) methodologies in latent fingerprint examinations. The foundation of latent fingerprint identification is that friction ridge skin is unique and persistent, and these features can positively identify the individual that produced it. The study reports the results of four categorical opinions: identification, exclusion, inconclusive, and no value decisions. Tests were assembled using 80 latent prints with varying quantity and quality of information from ten known sources, and were distributed to 109 latent print examiners across the United States. Participants had at least one year of latent print examination experience and employed the ACE methodology when comparing unknown latent prints to known sources. Responses from the participants yielded 5,963 sufficiency determinations, 4,536 ACE decisions, 532 ACE-V decisions, 1,311 repeatability decisions, 326 ACE decisions under biased conditions, and 333 repeatability decisions under biased conditions. The study took into account inconclusive responses in determining error rates and established a False Positive Rate (FPR) of 3.0% and False Negative Rate (FNR) of 7.5% for ACE examinations, as well as a FPR of 0.0% and FNR of 2.9% for ACE-V examinations. Participants were able to reproduce a correct identification 94.2% of the time and not reproduce an erroneous identification 100% of the time. Participants repeated their previous correct identifications 94.6% of the time and did not repeat their previous erroneous exclusions 93.1% of the time. Under biased conditions, participants were able to reproduce a correct identification 73.0% of the time and not reproduce an erroneous identification 96.5% of the time. Under biased conditions, participants repeated their previous correct identifications 93.2% of the time and did not repeat their previous erroneous exclusions 85.2% of the time.