Scientific testimonial standards for microbial forensic evidence

Abstract The last decade has seen a number of changes to expert testimonial standards in traditional forensic science fields, culminating in the Uniform Language for Testimony and Reporting (ULTR) guidance published by the Department of Justice. This chapter reviews issues that arose during the deliberations of the National Commission on Forensic Science before the adoption of ULTR standards, and their potential implications for Microbial Forensics testimony. A definition of error rate based on modern machine learning concepts can be used as a basis for understanding rigorous requirements for validation and the limits of testimony. Some habits of scientific discourse, especially hedging language, are possible pitfalls that weaken testimony. Statistically justified language for different kinds of microbial forensic inference is discussed for four examples: microbial identification, inferences about pathogen transmission, inferences about sources based on “morph” assays, and inferences about contamination and background when sensitive trace detection assays are used. There is a need for standards governing testimony when rigorous determination of error rates or other orthodox statistical measures of uncertainty is not possible.