The Structure of Legal Rules and the Analysis of Judicial Decisions

46 Pages Posted: 30 Jan 2007

See all articles by Jonathan P. Kastellec

Jonathan P. Kastellec

Princeton University - Department of Political Science

Date Written: January 30, 2007

Abstract

One of the most successful lines of inquiry in judicial politics has been the study of fact patterns, in which scholars have shown a general consistency between the facts of a case and judicial decisions across several areas of the law. These findings quantitatively illustrate how courts often follow legal rules, which define the relationship between case facts and case outcomes. In this paper, I argue that the methodological approach that political scientists have taken has left largely unstudied the true structure of legal rules and judicial decision making. After discussing why traditional statistical methods are too restrictive to capture the structure of legal rules and legal reasoning, I introduce the use of a largely unexplored method in political science - classification trees - to study the mapping between case facts and judicial outcomes. I argue that this method is more likely to capture the hierarchical and dichotomous nature of judicial decisions. Analyses of search and seizure and confession cases demonstrate that classification trees can increase our understanding of the structure of legal rules, as well as unify political science and legal conceptions of legal doctrine and judicial decision making.

Keywords: Judicial decision-making, legal rules, fact patterns, classification trees

Suggested Citation

Kastellec, Jonathan P., The Structure of Legal Rules and the Analysis of Judicial Decisions (January 30, 2007). Available at SSRN: https://ssrn.com/abstract=960190 or http://dx.doi.org/10.2139/ssrn.960190

Jonathan P. Kastellec (Contact Author)

Princeton University - Department of Political Science ( email )

Fisher Hall
Department of Politics
Princeton, NJ 08544-1012
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
583
Abstract Views
2,969
Rank
85,903
PlumX Metrics