Empirical Probability Tables: International Dispute Cases
A previous chapter 1 described how some statistical techniques could be applied to synthesizing judicial precedents and to predicting the outcomes of future cases. For simplicity of presentation, the characteristics or variables in the precedent cases were made into either-or dichotomies. For example, the region of the country where the litigation-provoking incident occurred was dichotomized to read "originated in (1) the South, West, or District of Columbia, rather than (2) the East or Midwest." It is the purpose of this chapter to describe some related techniques that are particularly applicable to working with variables that consist of more than two nonnumerical categories. The present techniques represent an improvement not only in broadened applicability but also in predictive power, simplicity, and in utility for building a behavioral science of law. The article presupposes no prior reading, and technical matters have been relegated to the footnotes and to an appendix.
A second purpose of this chapter is to emphasize that the quantitative processing of cases can be useful to understanding why a set of cases was decided the way it was as well as to predicting future decisions. "Prediction" involves knowing how a case or entity will be positioned on a decision or dependent variable by knowing (1) how it is usually____________________