|
Abstract: After describing the textbook, Statistics for Lawyers, by Michael O. Finkelstein and Bruce Levin, this review-essay reflects on the role of probability calculations in trials and appeals. It probes an unusual statistical argument involving a joint normal distribution for the time required to commit a murder. The argument was presented on appeal — and forcefully rejected in an erudite opinion — in Branion v. Gamly, 855 F.2d 1256 (7th Cir. 1988). This essay critiques the calculations of both counsel and the court of appeals. It suggests that the case is one in which the "probability evidence" was unconvincing for want of adequate data on the phenomenon that was modeled.
|