Statistical Evidence and the Problem of Specification
As a number of modern sexual misconduct cases demonstrate, often there are multiple charges against a single individual under circumstances in which the proof of an individual charge may fall short of the required standard of proof, but in which it is clear – overwhelmingly, or beyond a reasonable doubt – that at least one of the charges is true, even if we cannot be sure which one. Building on earlier work by myself and Richard Zeckhauser, by Alon Harel and Ariel Porat, and by Ariel Porat and Eric Posner, among others, this paper, prepared for the World Congress on Evidential Reasoning at the University of Girona, offers a sympathetic examination of sanction imposition – in and out of formal legal proceedings – on the basis of aggregate probabilities, and addresses a series of common objections. And the paper suggests that the greatest value of aggregating low (or lower) probability charges may be greatest outside of the official judicial process. It hints as well, although inconclusively, at the larger question of why the law focuses on acts when it is imposing sanctions, rather than focusing on the actors who may have committed those acts.