ABSTRACT
Focusing on U.S. law and criminal cases, this chapter describes the legal framework for expert testimony involving statistical assessments in litigation. It introduces the principles governing the qualifications of statistical and other experts, the different ways in which an expert witness can assist a judge or jury, the requirements for admitting expert testimony generally, and the special rules (such as those announced in Frye v. United States and Daubert v. Merrell Dow Pharmaceuticals, Inc.) for determining the admissibility of scientific expert testimony. It distinguishes three uses of statistical analysis in litigation-as part of a scientific procedure for obtaining test results or conclusions; as a device for assessing the value of scientific test results (or other evidence relevant to the events in a case); and as a way to help show that scientific methods are valid and reliable. It examines how the principal rules and cases on admissibility of scientific evidence apply in these situations, with special attention to the meaning of the "error rates" that courts consider in connection with scientific findings and how some conditional error probabilities can be estimated. In the course of this exposition, it briefly describes the legal concept of probative value of evidence and its relationship to error probabilities, likelihoods, and Bayes factors.
