Applications manager, Camscan USA Inc., Mars, PA
(Received 2 May 1989; accepted 14 June 1989)
Although the technique of scanning electron microscopy and energy-dispersive X-ray analysis (SEM/EDX) has gained widespread acceptance for the identification of gunshot residues (GSR), the time required (especially for manual searches) is still considered excessive. As a result, statistical considerations are commonly used to justify a reduction in the total specimen area to be searched. A detailed statistical analysis was presented in the Aerospace Corporation Final Report on Particle Analysis for Gunshot Residue Detection published in September 1977, and its conclusions have had significant influence in the forensic science community as concerns the determination of acceptable particle analysis search areas. A close examination of the Aerospace Corporation report has revealed a significant programming error which resulted in statistical probability errors ranging from 30% to well over ten orders of magnitude. Corrected results, presented in this paper, suggest that a great deal of caution should be exercised in applying statistical analysis to justify a reduction in search area. In particular, the probability of a false negative report increases rapidly as the number of gunshot residue (GSR) particles assumed to be present decreases. Since the investigator cannot know in advance the number of GSR particles present on a sample, the corrected Aerospace analysis suggests that statistical considerations may not provide sufficient justification for any significant reduction in the sample area to be searched.
Paper ID: JFS12875J