An Update on a World Wide Study of Bullets Fired From 10 Consecutively Rifled 9mm Ruger Pistol Barrels: Analysis of Examiner Error Rate Using Bayesian Statistics

Abstract

This technical note is the latest update on a continuing study, first designed and initiated by Brundage et al. over twenty years ago. This study was borne out of increased scrutiny of firearms identification in response to Daubert v. Merrell Dow Pharmaceuticals, Inc. in 1993. The purpose was to determine whether forensic firearms examiners were able to associate fired bullets with the barrels through which they had been fired. To date 792 participants from 36 countries have utilized over 240 test sets consisting of bullets fired through 10 consecutively rifled Ruger P-85 pistol barrels. Here we provide an update on the results of the ongoing “10 barrel test” up until the point in time of writing. To analyze the data thus far collected, a Bayesian approach was again selected. Posterior examiner error rates are estimated assuming only vague prior information. Given the data found over the course of this diverse decades long study, our most conservative estimate for examiner error rate has a posterior median of 0.03% with a 95% probability interval of [2×10-6 %, 0.1%].