NCJ Number
139762
Journal
Journal of Forensic Sciences Volume: 37 Issue: 6 Dated: (November 1992) Pages: 1525-1533
Date Published
1992
Length
9 pages
Annotation
This article examines the solution of paternity disputes with the use of results from scientific analyses from a decision-theoretical perspective.
Abstract
Two alternative approaches to decisionmaking, the so- called "Bayes" and "Minimax" strategies, are described and discussed. If prior probabilities of paternity are exactly known, then "Bayes" decisions are independent of the source of evidence and optimal with respect to average losses caused by wrong decisions; however, the authors conclude the Minimax decisions, which depend on the employed test system but not upon prior probabilities, are more appropriate in paternity cases if equal prior good will towards disclaimed children and alleged fathers is demanded. The article further demonstrates that, when major evidence about paternity comes from multilocus DNA fingerprinting, prior probabilities must be known accurately for Bayes decisions to be superior with respect to average losses. The authors show that "quasi" Bayes decisionmaking, that is, adopting a neutral prior probability of 0.5 but leaving thresholds for decisionmaking unchanged, coincides with Minimax decisionmaking if multilocus DNA fingerprinting is used. 2 figures and 16 references