Using statistical models in estimating error rates for forensic firearm identification

Estimating error rates for firearm evidence identification is a fundamental challenge in forensic science. The recently developed Congruent Matching Cells (CMC) method provides applications to firearm evidence identification. To estimate error rates, appropriate statistical models are needed for the CMC values. In this paper, in addition to the binomial probability distribution the correlated binomial distribution is proposed. For an image comparison, the correlated binomial distribution can be applied to the cell pairs from CMC method. An application to an actual data set demonstrates that the correlated binomial distribution fits the relative frequency distribution of CMC values much better than the binomial distribution.