Statistical issues in DNA evidence.

Abstract

Some of the statistical issues involved in the interpretation of DNA evidence in courts, particularly focusing on the adversarial process as seen in the UK and the USA, are discussed. The appropriate way to consider the use of DNA evidence in courts is through the use of a likelihood ratio. In this, the relative likelihoods of the DNA evidence are calculated under two hypotheses, respectively that the subject is or is not the source of a DNA sample found at the scene-of-crime. This formulation allows, using Bayes' theory, the prior probability of guilt, based on the other evidence in a case, to be incorporated in a formal way into calculation of a final, or posterior, probability of guilt. The most controversial aspect of DNA evidence has concerned the calculation of the probability of a match between an innocent suspect and DNA from the scene-of-crime. The calculation is affected by subdivision within the population and the possibility of relatedness between the suspect and the true source of the scene-of-crime DNA. The qualitative importance of these effects will be discussed. In addition to the scientific questions concerning the likelihood ratio, there have been a number of other concerns about the use of DNA evidence, and other evidence presented in a statistical form, in courts. These include the question whether accurate or conservative values of the likelihood ratio should be presented, and the abilities of juries to deal with quantitative evidence.