Director, Bureau of Legal Dentistry, Vancouver, British Columbia,
Doctoral student, The University of Liverpool, Liverpool,
(Received 15 February 2001; accepted 12 March 2001)
U.S. courts have stated that witnesses must be able to identify published works that define operational parameters of any tests or procedures that form the basis of scientific conclusions. Such works do not exist within the field of bite mark analysis. As the most commonly employed analytical technique in bite injury assessment, this study defines quantifiable variables for transparent digital overlays. A series of ten simulated, postmortem bites were created on pigskin and, with accompanying overlays, assembled into cases. Using two separate studies with four examiner groups, the study defined values of intra- and inter-examiner reliability, accuracy, sensitivity, specificity, and error rates for transparent overlays. Methods and statistical treatments from medical decisionmaking and diagnostic test evaluation were employed. Forced decision models and receiver operating characteristic analyses were utilized. Sensitivity and specificity values are described, and the results are consistent with other dental diagnostic systems. It was concluded that the weak inter-examiner reliability values explain the divergence of odontologists' opinions regarding bite mark identifications often stated in court. The effect of training and experience of the examiners was found to have little effect on the effective use of overlays within this study. The authors conclude that further research is required so that the results of the current study can be placed into context, but this represents a significant first step in establishing the scientific basis for this aspect of forensic dentistry.
Paper ID: JFS15160J