Stuart S. Nagel


The purpose of this article is to analyze a microcomputer program that can process a set of (1) prior cases, (2) predictive criteria for distinguishing among the cases, and (3) the relations between each prior case and each criterion in order to arrive at an accurate decision rule. Such a rule will enable all the prior cases to be predicted without inconsistencies, and thereby maximize the likelihood of accurately predicting future cases. To illustrate the program, this article uses five substantive fields, including the predicting of cases dealing with religion in the public schools, legislative redistricting, housing discrimination, international law, and criminal law.