Inference in forensic identification
Balding DJ., Donnelly P.
© 1995 Royal Statistical Society. The problem of quantifying the weight of evidence in forensic identification is addressed. The essence of the problem is abstracted in a simple paradigm, the analysis of which yields valuable insights and highlights important distinctions. A special case of this analysis gives a resolution of the so-called island problem. The paradigm is extended to assess the effects of several features which may be important in practical situations, such as possible alibis, heterogeneous populations of potential suspects and informative protocols for finding suspects. The analyses of the paper clarify several issues pertaining to the weight of evidence associated with matching deoxyribonucleicacid (DNA) profiles and raise some new concerns. In addition, established concerns regarding the incorrect interpretation of probabilistic evidence by juries are discussed in the DNA profile context.