DNA Evidence: Probability, Population Genetics, and the Courts01 Jan 1993
To help meet the challenge of presenting properly performed DNA tests within the post-Daubert legal framework, this article outlines the statistical procedures that have been employed or proposed to provide judges and juries with quantitative measures of probative value, describes more fully how the courts have dealt with these procedures, and evaluates the opinions and the statistical analyses from the standpoint of the law of evidence.
Specifically, the article outlines the procedure used to declare whether two samples of DNA "match," and how shrinking the size of the "match window," as some defendants have urged, will decrease the risk of false matches, but will also exclude highly probative evidence of identity. It also demonstrates that a defendant's effort to show that a smaller match window would not permit the declaration of a match is irrelevant or misleading. Additionally, the article explains procedures for estimating the frequency of the incriminating genetic characteristics in various populations. These procedures have been the subject of an acrimonious debate, both in the courts and in the press, about the effect of "population structure." The population structure objection, which has proved so effective in court, applies most strongly to only a limited class of case, and therefore, courts have erred in excluding DNA evidence on the theory that the scientific community advocates that the most "conservative" procedures must be used in all cases. Finally, the article identifies more fundamental problems in the use of population frequency estimates, and advocates supplementary and alternative procedures that are essential if quantitative statements of the probative value of DNA evidence are to be admissible.