Centrality Scores and Precedent Value in Legal Network Analysis

Courts commonly rely on precedents to guide their judgments. Centrality measures have been used to calculate precedence value in citation networks of judgments, yet it remains largely unknown whether and which centrality measures correlate well to precedent value. An analysis of European Court of Human Rights (ECtHR) judgments offers a unique opportunity to uncover this relationship, as the ECtHR publishes an importance score for its judgments and the branch of court that dealt with them. These scores, although not perfect, may serve as proxies for a case’s precedent value. Various network centrality measures correlated reasonably with these proxies, with Degree being a stable measure across the (sub)networks. An ordinal regression model with network centrality among other predictor variables performed reasonably when Importance Score was used as an outcome variable (F1 ≈ .655). The results support that network centrality, to an extent, indicates case precedent value, and that Degree seems to be a stable proxy for precedent value across different networks. The data, code, and additional results are made available.