|
Abstract:
|
Forensic scientists collect and analyze trace evidence from crime scenes. Trace evidence may include fibers, cartridge casings, or bullet fragments, and the data are typically multivariate measurements such as the composition of the lead alloy in bullet fragments. Suppose traces also are found on a suspect. The question of interest is whether the crime scene and suspect evidence have a common origin. This often is answered in two steps. First, a match between the two sources of evidence is established. Then, the probative value of the match is assessed. A recent NRC report discussed the errors that may arise when establishing matches, but a more difficult problem is assessing the probability that the match occurred by chance alone. We focus on the second question and discuss the information and approaches that might be needed to decide on the probative value of evidence.
|
