NCJ Number
90986
Journal
American Criminal Law Review Volume: 21 Issue: 1 Dated: (Summer 1983) Pages: 55-79
Date Published
1983
Length
25 pages
Annotation
Numerical evidence regarding the probability that a certain event will or will not occur or regarding the interpretation of numerical probabilities through Bayesian techniques should be admissible in criminal trials provided appropriate safeguards are used.
Abstract
A valid, objective study should underlie any numerical evidence. The proponent should also provide experts in both statistics and any scientific discipline used in the study. These experts should be sufficiently familar with the study to testify regarding its validity. The proponent should also provide advance notice of the numerical evidence to be introduced and the purposes for which it will be used at the trial. The government should provide funds for employing experts for the defense of indigent defendants. The court should also require the proponent to show that a high probability exists that any probability estimates introduced at trial are not too low and are therefore not unfair to the defendant. The average juror can understand numerical evidence which is properly interpreted. Both a 1 in 4,500 probability, as demonstrated in the Gaudette Keeping study on hair matching, and an incomplete identification are relevant and should be admissible, but neither can provide conclusive proof of the defendant's guilt. Proper use of numerical techniques can give the criminal justice system a valuable and reliable source of evidence and can promote public confidence in the accuracy and fairness of the trial process. Judicial decisions entailing the use of numerical evidence and footnotes which contain references are provided.