Objective Analysis of Impressed Chisel Toolmarks

Historical and recent challenges to the practice of comparative forensic examination have created a driving force for the formation of objective methods for toolmark identification. In this study , fifty sequentially manufactured chisels were used to create impression toolmarks in lead (500 toolmarks total). An algorithm previously used to statistically separate known matching and nonmatching striated screwdriver marks and quasi-striated plier marks was used to evaluate the chisel marks. Impression toolmarks, a more complex form of toolmark, pose a more difficult test for the algorithm that was originally designed for striated toolmarks. Results show in this instance that the algorithm can separate matching and nonmatching impression marks, providing further validation of the assumption that toolmarks are identifiably unique.     

A Historical,Detailed Toolmark Examination from 1916

Marc Bischoff received his PhD from the University of Lausanne, Institute of Police Science and Criminology, and later became the head of the Institute. His final examination paper for his doctorate degree constitutes a toolmark analysis and comparison between a suspected axe and questioned cuttings of wood trunks. Bischoff uses the scientific method to test the hypothesis he formulated. Dated 1916, this report includes his notes, measurements, photomicrographs of the questioned toolmarks and the test impressions, detailed sketches, technical discussion, results, and conclusions. Recognizing the historical value of Bischoff's work, it seemed important to get it translated. Photocopies of Bischoff's original work were obtained. The case report, translated from French, is presented here along with a brief history of the Firearm and Toolmark Identification field and Bischoff's role in the beginning of this field. Bischoff's work stands as testimony to the long tradition of the field of forensic toolmark analysis.     

Forensic identification science evidence since Daubert: Part I--A quantitative analysis of the exclusion of forensic identification science evidence

The U.S. Supreme Court decisions in Daubert v. Merrell Dow Pharmaceuticals Inc. and Kumho Tire Co. Ltd. v. Carmichael transformed the way scientific expert evidence was reviewed in courts across the United States. To gauge the impact of these rulings on the admission of forensic identification evidence, the authors analyzed 548 judicial opinions from cases where admission of such evidence was challenged. Eighty-one cases (15%) involved exclusion or limitation of identification evidence, with 50 (65.7%) of these failing to meet the "reliability" threshold. This was largely because of a failure to demonstrate a sufficient scientific foundation for either the technique (27 cases) or the expert's conclusions (17 cases). The incidence of exclusion/limitation because of a lack of demonstrable reliability suggests that there is a continuing need for the forensic sciences to pursue research validating their underlying theories and techniques of identification to ensure their continued acceptance by the courts.     

Angular Determination of Toolmarks Using a Computer‐Generated Virtual Tool

A blind study to determine whether virtual toolmarks created using a computer could be used to identify and characterize angle of incidence of physical toolmarks was conducted. Six sequentially manufactured screwdriver tips and one random screwdriver were used to create toolmarks at various angles. An apparatus controlled tool angle. Resultant toolmarks were randomly coded and sent to the researchers, who scanned both tips and toolmarks using an optical profilometer to obtain 3D topography data. Developed software was used to create virtual marks based on the tool topography data. Virtual marks generated at angles from 30 to 85° (5° increments) were compared to physical toolmarks using a statistical algorithm. Twenty of twenty toolmarks were correctly identified by the algorithm. On average, the algorithm misidentified the correct angle of incidence by ?6.12°. This study presents the results, their significance, and offers reasons for the average angular misidentification.     

The Development and Application of Random Match Probabilities to Firearm and Toolmark Identification

The field of firearms and toolmark analysis has encountered deep scrutiny of late, stemming from a handful of voices, primarily in the law and statistical communities. While strong scrutiny is a healthy and necessary part of any scientific endeavor, much of the current criticism leveled at firearm and toolmark analysis is, at best, misinformed and, at worst, punditry. One of the most persistent criticisms stems from the view that as the field lacks quantified random match probability data (or at least a firm statistical model) with which to calculate the probability of a false match, all expert testimony concerning firearm and toolmark identification or source attribution is unreliable and should be ruled inadmissible. However, this critique does not stem from the hard work of actually obtaining data and performing the scientific research required to support or reject current findings in the literature. Although there are sound reasons (described herein) why there is currently no unifying probabilistic model for the comparison of striated and impressed toolmarks as there is in the field of forensic DNA profiling, much statistical research has been, and continues to be, done to aid the criminal justice system. This research has thus far shown that error rate estimates for the field are very low, especially when compared to other forms of judicial error. The first purpose of this paper is to point out the logical fallacies in the arguments of a small group of pundits, who advocate a particular viewpoint but cloak it as fact and research. The second purpose is to give a balanced review of the literature regarding random match probability models and statistical applications that have been carried out in forensic firearm and toolmark analysis.     

Optimization of a Statistical Algorithm for Objective Comparison of Toolmarks

Due to historical legal challenges, there is a driving force for the development of objective methods of forensic toolmark identification. This study utilizes an algorithm to separate matching and nonmatching shear cut toolmarks created using fifty sequentially manufactured pliers. Unlike previously analyzed striated screwdriver marks, shear cut marks contain discontinuous groups of striations, posing a more difficult test of algorithm applicability. The algorithm compares correlation between optical 3D toolmark topography data, producing a Wilcoxon rank sum test statistic. Relative magnitude of this metric separates the matching and nonmatching toolmarks. Results show a high degree of statistical separation between matching and nonmatching distributions. Further separation is achieved with optimized input parameters and implementation of a “leash” preventing a previous source of outliers—however complete statistical separation was not achieved. This paper represents further development of objective methods of toolmark identification and further validation of the assumption that toolmarks are identifiably unique.     

Forensic identification science evidence since Daubert: Part II--judicial reasoning in decisions to exclude forensic identification evidence on grounds of reliability

Many studies regarding the legal status of forensic science have relied on the U.S. Supreme Court's mandate in Daubert v. Merrell Dow Pharmaceuticals Inc., and its progeny in order to make subsequent recommendations or rebuttals. This paper focuses on a more pragmatic approach to analyzing forensic science's immediate deficiencies by considering a qualitative analysis of actual judicial reasoning where forensic identification evidence has been excluded on reliability grounds since the Daubert precedent. Reliance on general acceptance is becoming insufficient as proof of the admissibility of forensic evidence. The citation of unfounded statistics, error rates and certainties, a failure to document the analytical process or follow standardized procedures, and the existence of observe bias represent some of the concerns that have lead to the exclusion or limitation of forensic identification evidence. Analysis of these reasons may serve to refocus forensic practitioners' testimony, resources, and research toward rectifying shortfalls in these areas.     

Assessment of Wear‐Related Features of the Kerf Wall from Saw Marks in Bone*†

Abstract: Analyses of saw marks in bone may yield important information about the class characteristics of saws used in postmortem dismemberments, yet little research has been directed at identifying saws’ individualizing characteristics. This study adds to existing saw mark analysis methodologies by examining wear‐related features of kerf walls using light‐ and environmental scanning electron microscopy. A crosscut saw and hacksaw were used to create sequences of 30 cuts in bone; these sequences reveal patterns of progressive loss of fine details of kerf wall morphology with increasing saw blade wear, because of the rounding of sharp points and edges. Nevertheless, diagnostic kerf wall features used to establish class characteristics persist despite these wear‐related changes. Unsuccessful attempts at statistical analysis of wear‐related changes, based on striae width and density, suggest these patterns are not readily quantifiable. Additionally, despite the scanning electron microscope’s superior imaging capabilities, it provided few practical, methodological gains over traditional light microscopy.     

Adapting the Chumbley Score to Match Striae on Land Engraved Areas (LEAs) of Bullets,

The same‐source problem remains a major challenge in forensic toolmark and firearm examination. Here, we investigate the applicability of the Chumbley method (J Forensic Sci, 2018, 63, 849; J Forensic Sci, 2010, 55, 953) (10,12), developed for screwdriver markings, for same‐source identification of striations on bullet LEAs. The Hamby datasets 44 and 252 measured by NIST and CSAFE (high‐resolution scans) are used here. We provide methods to identify parameters that minimize error rates for matching of LEAs, and a remedial algorithm to alleviate the problem of failed tests, while increasing the power of the test and reducing error rates. For 85,491 land‐to‐land comparisons (84,235 known nonmatches and 1256 known matches), the adapted test does not provide a result in 176 situations (originally more than 500). The Type I and Type II error rates are 7.2% (6105 out of 84,235) and 21.4% (271 out of 1256), respectively. This puts the proposed method on similar footing as other single‐feature matching approaches in the literature.