Determining tool class macroscopically on bone from varying levels of force of hacking trauma
MetadataShow full item record
A component to some forensic cases is being able to identify tool class, whether it is a murder weapon or if a tool was used postmortem in dismemberment. The goal of the present study is to determine if it is possible to identify tool class macroscopically and/or what level of force was applied in cases involving hacking. Three hypotheses are proposed. The first is that the cutmarks would appear differently at the varying levels of impact force for the same implements, including patterns of fractures, number of fragments, size of fragments, and appearance of the cutmark, i.e., the kerf. The second hypothesis is that there would be observable macroscopic differences on the cutmarks between tool classes, and the ability to distinguish between tool classes will not be affected by the differences from various levels of force of impact. The third hypothesis is that these observable macroscopic differences can be used to create prediction tables that can be used for predicting tool class and the level of force applied. Using a device created to simulate hacking, the long bones of white-tailed deer (Odocoileus virginianus), a chef’s knife, cleaver, machete, and axe were tested at three different impact forces each. The author examined the hack marks on the bones quantitatively by measuring the kerf width and depth, number of fragments present, as well as qualitatively by describing any fractures present and the appearance of the entrance and exits. It was found that there is a statistically significant relationship between the implement and the entrance width (p-value = 7.27e-13). There is a statistically significant relationship between the force of impact and the entrance width (p-value = 5.57-06), overall entrance appearance (clean cut: p-value = 2.40e-06; chattered: p-value = 0.004), and conchoidal flaking (p-value = 0.025). There is also a statistically significant relationship between the implement and the level of force as a combined influence, as opposed to separate influences, and if the overall appearance of the entrance is chattered (p-value = 0.017). These relationships support the first two proposed hypotheses. Recursive partition and regression trees were created for each implement to determine what characteristics may be used to create prediction guides based on the collected data. The results of the experiment were used in the creation of an implement prediction guide and force of impact prediction tables. A blind test showed that the implement prediction guide was accurate 50% of the time and that the force of impact prediction tables were accurate 10% of the time. While this is low accuracy, it indicates that this research has potential to help with hacking trauma analysis as a baseline for future research, but is not applicable at this time, accepting the null hypothesis for the third hypothesis.