ABSTRACT
A common aspect of many modern court proceedings, particularly in litigation involving toxic torts, product liability, contaminating environmental agents, and the like, is the presence of expert witnesses who provide evidence relevant to the matter at hand. Very often such evidence is given through statistical argumentations, possibly through estimated statistical models (that might, for example, assess a probability of causation), meta-analyses, or other methods of data presentation and interpretation developed through graphical or tabular means. To be defensible ethically, such evidence when presented statistically must avoid the various pitfalls documented throughout this book (for example, regression toward the mean, the ecological fallacy, confusing test “impact” with test “bias,” and “lurking” third variable confoundings). The issues involved in expert witness admissibility, however, are much broader than just in how data are presented. The scientific reliability and validity of the available evidence are of major interest and are subject to the Federal Rules of Evidence. Here, we discuss some of the issues involved in the admissibility of evidence and the proffering of expert witnesses, both as they are currently understood and practiced and how they have evolved historically over the years.1
