A murder in Boston.

Abstract

In 2009 D was convicted for the 1990 murder of C in C's own apartment. In a post-conviction review DNA attributed only to D and C (and no others) was found underneath the fingernails of C. At trial, in a hearing for a retrial, and in a pending complaint to the Massachusetts Forensic Science oversight board, the statements of government witness' regarding the meaning of the DNA evidence at activity level were a topic of debate. In this paper, a Bayesian network (BN) evaluation of this evidence is presented. This BN uses the propositions that D was the attacker (H_p) versus an alternate proposition that he was not the attacker (H_a). The alternate, which was inferred from defense questioning, requires that transfer occurred from a social meeting 2 to 4 weeks earlier. The evaluation presented here suggests an LR of the order of 800. This analysis suggests that, while the original testimony may not have been prepared for in a formal manner, it was not misleading to a lay jury.