Dark particulate developers remain the most widely applied means of visualizing latent fingermarks. Manganese cobalt oxide (MnCo2O4) nanoparticles were prepared by alkaline coprecipitation at 80 °C and characterized by XRD, FTIR, FESEM, and EDS, which indicated the formation of the cubic spinel phase of MnCo2O4 with a crystallite size of approximately 21 nm, a particle size of approximately 50 nm, and particle aggregation. Sebaceous and natural fingermarks on ten substrates were developed by powder dusting, visually graded, and scored using NIST Fingerprint Image Quality 2 (NFIQ 2). Sebaceous marks received grades of 4 or 5 on eight substrates, whereas natural marks achieved these grades on five substrates. NFIQ 2 scores correlated with visual grades (Spearman ρ = 0.69) and were higher for sebaceous marks (median 33.5 versus 15.0). Substrate texture inflated the automated minutiae counts, and microscopy showed particle deposition largely confined to the ridges.
How should a laboratory combine opinions when examiners disagree about whether fingerprint evidence supports a definitive conclusion? We examine this question through an archival reanalysis of 17,121 responses from 169 examiners across 744 fingerprint pairs in the FBI latent-print Black Box Study. Comparison-stage inconclusive rates increased from 17.2% for comparisons rated obvious to 53.8% for those rated very difficult. A one-level increase in difficulty was associated with 1.58 times the odds of an inconclusive response after accounting for mating status and clustering by examiner and pair. Examiner inconclusive rates ranged from 5.7% to 81.2%, correlated across mated and non-mated comparisons (r = 0.61), and varied more than expected under pair-conditioned permutations. These patterns establish persistent response tendencies but do not separate perceptual ability from decision criteria. In exact retrospective calculations across 738 eligible fingerprint pairs, requiring a strict majority of all panelists reduced erroneous definitive conclusions from 3.80% for one examiner to 0.58% for seven. A majority taken only among definitive responses increased coverage but raised error to 9.88%, mainly through false exclusions. No-value and inconclusive responses remain distinct decision stages; they have the same voting effect only under the two rules examined. Larger absolute error reductions occurred on pairs with more dispersed responses, although the entropy measure does not distinguish uncertainty from contradictory conclusions. The findings illustrate how aggregation rules change error and coverage in a historical response distribution. They do not establish an optimal verification procedure or validate these rules for contemporary expanded conclusion scales.
The log-likelihood-ratio cost ( C llr ) is the standard metric for validating automated likelihood-ratio (LR) systems in forensic science, justified by its being a strictly proper scoring rule. Several studies have developed calibration diagnostics, protocols and surveys of reported values; however, each presumes the reference set carries correct labels, whereas forensic reference sets rest on ground truth by proxy or examiner adjudication. Based on the literature, the sensitivity of C llr to such error is unquantified, because measuring it requires the true LR, unobtainable in real forensic domains. The aim of this study is to quantify that sensitivity against known true likelihood ratios, to locate where it is severe, to derive the optimal bound, and to provide a contamination diagnostic needing no clean-label baseline. The optimal system, computed in closed form from a specified generator, was scored against labels carrying error rates to 10 per cent and compared against two reconstructions of the Empirical Lower and Upper Bound (ELUB) procedure and four alternatives. The results show that at a mean absolute log 10 L R of 17.2 a 3.5 per cent error rate inflates the cost thirtyfold and scores it as worse than uninformative, whereas below about six inflation stays at or below 1.4-fold. Mislabelled items carry 62.8 to 99.6 per cent of the cost, and parameterising the bound by each arm's contamination predicts the empirical optimum almost eightfold more accurately than the nominal error rate does. Bounding is therefore necessary for valid C llr reporting on unverified reference sets.
The log-likelihood-ratio cost (Cllr) is the standard metric for validating automated likelihood-ratio (LR) systems in forensic science, justified by its being a strictly proper scoring rule. Several studies have developed calibration diagnostics, protocols and surveys of reported values; however, each presumes the reference set carries correct labels, whereas forensic reference sets rest on ground truth by proxy or examiner adjudication. Based on the literature, the sensitivity of Cllr to such error is unquantified, because measuring it requires the true LR, unobtainable in real forensic domains. The aim of this study is to quantify that sensitivity against known true likelihood ratios, to locate where it is severe, to derive the optimal bound, and to provide a contamination diagnostic needing no clean-label baseline. The optimal system, computed in closed form from a specified generator, was scored against labels carrying error rates to 10 per cent and compared against two reconstructions of the Empirical Lower and Upper Bound (ELUB) procedure and four alternatives. The results show that at a mean absolute log10LR of 17.2 a 3.5 per cent error rate inflates the cost thirtyfold and scores it as worse than uninformative, whereas below about six inflation stays at or below 1.4-fold. Mislabelled items carry 62.8 to 99.6 per cent of the cost, and parameterising the bound by each arm's contamination predicts the empirical optimum almost eightfold more accurately than the nominal error rate does. Bounding is therefore necessary for valid Cllr reporting on unverified reference sets.
The genealogy process is a key bottleneck within forensic investigative genetic genealogy, which is a novel approach to solving violent crimes and identifying human remains. Building on our earlier work, we develop FIGGsolver: an interactive algorithm that guides the genealogist through the tree-building process by specifying precise actions for the genealogist to take at each point in time. The algorithm has been used to solve actual cases and plans are underway to make FIGGsolver widely available to the forensic investigative genetic genealogy community. We describe the algorithm in non-mathematical terms and collect insights gathered from its construction and use.
