To address the recent concerns and published research regarding erroneous exclusions, the Minnesota Bureau of Criminal Apprehension (MN BCA), in collaboration with 3M Cogent, Inc., conducted a study that led to the adoption of Automated Fingerprint Identification System (AFIS) technology in a novel “Case AFIS” application to reduce erroneous exclusions and catch more missed identifications. Using a Case AFIS approach, a process by which AFIS technologies were applied to specific cases, previously reported decisions were reviewed in this study for erroneous exclusions and missed identifications. This study shows that using the Case AFIS application would lead to more identifications and would be more efficient than the manual searching process.
Beginning in 2011, several articles were published demonstrating under various study conditions that fingerprint examiners make erroneous exclusions.1,2 In these studies they were seven times more likely to erroneously exclude than erroneously identify an individual. While the identification results were reassuring for many of our partners in the criminal justice community, in the forensic community the high number of erroneous exclusions was alarming. Within agencies, various policies were instituted to reduce the number of erroneous exclusions that may have been reported in case work. One of those policies was to review all identification and exclusion decisions. As a result more erroneous exclusions were discovered, and agencies became aware of the number of erroneous exclusions committed by their examiners. These errors usually resulted in a quality review, root cause analysis, and sometimes included removing examiners from case work while reviewing all of their previous work for a period of time. This creates costly repercussions for a laboratory as labor hours are expended while no new work is produced; it negatively effects morale and increases backlog accumulation.
As reviews of the root cause of erroneous exclusions began, a variety of contributing factors (human and latent-dependent) were proposed. Some agencies became proponents of implementing specific criteria that were required before examiners could render an exclusion decision.3 While none of these policies have been tested and validated in structured research, they may be effective at reducing erroneous exclusions by limiting the number of exclusion decisions that can be attempted. The MN BCA briefly explored these policies but ultimately abandoned them to return to the traditional practice of rendering exclusion decisions based on the examiner’s expertise. During this period, the idea that the best solution to the erroneous exclusion problem may not be to limit the examiners’ ability for rendering exclusion decisions but to provide a tool for catching erroneous exclusions by helping examiners quickly recognize similarities and more efficiently search difficult latent prints.
AFIS technology is the most obvious tool for this problem for several reasons. First, matching algorithms have significantly improved in the last decade to return matching candidates from databases with millions of individual records. If used in a case of narrowed donors to a database of known records of individuals directly related to the case—whether suspect, victim, or elimination—the results could only be improved. Second, AFIS technology can also be prompted to ignore level 1 information and focus solely on minutiae configuration. Third, it can quickly search any latent in 360 degrees in a single operation. This eliminates errors that can occur when examiners fail to properly orient the latent print, place too much reliance on a pattern type that has been distorted, or simply “miss” the corresponding similarities due to fatigue or inattention. AFIS technology offers a powerful remedy against the major contributing factors of examiner errors.