A signature characteristic of the federal judicial system is its hierarchical structure; that structure shapes the goals of the judges within it, and induces concerns about oversight (for lower courts) and compliance (for higher courts). Such factors, in turn, interact with both case- and judge-level influences on decision making. A complete characterization of judicial decision making in the federal system must therefore allow for the influence of case-specific, judge-specific, and hierarchical effects, with the latter considering both "upward-looking" (oversight) and "downward-looking" (compliance) behavior; ideally, such a model must also measure those factors consistently across the three levels of the hierarchy, and allow for the magnitude of those effects to vary across divergent courts. We develop such a model, and analyze its implications through an analysis of the actions of all participating federal judges in each of the lower court proceedings for all cases decided in the U.S. Supreme Court which originated in a federal district court from 1953-2005. By coding data at the various levels of the judicial hierarchy in each case, our approach allows us to examine judicial voting within a given case context, and thereby holds constant the range of case-specific factors that might otherwise influence judicial decisions. In addition, recent advances in generalized hierarchical and mixed models allow for a flexible specification of the interaction of judge- and court-specific influences, including those related to hierarchy effects due to judges' potential concerns with oversight, compliance, or both. Finally, the end result of our data collection will be a comprehensive accounting of every decision made by every federal judge in every case decided by the U.S. Supreme Court over the past five decades; those data will thus offer scholars an unparalleled resource for the study of judges' decisions in the U.S. federal courts.
