Although federal circuit and district court judges are placed within a federal hierarchy, and receive legal and judicial training that emphasizes the importance of the judicial framework and its structure, such judges are also subjected to other pressures such as the types of litigants within the courtrooms as well as their local political environment. Furthermore, such judges are apt to form their own views about politics and legal policy and are often appointed by presidents who approve of their ideological leanings. Thus, federal courts are caught between competing goals such as their willingness to maximize their preferred legal policy, and their place within the judicial hierarchy. This dissertation applies hierarchy and impact theory to assess the importance of the judicial framework and its socialization, by analyzing both the judicial opinions and votes of federal circuit and district court judges in obscenity cases during a four-decade period (1957-1998). The research presented here finds the influence of higher court precedent to correspond in part with the conception of a judicial hierarchy. An analysis of citations of Supreme Court precedent (Roth v. United States (1957) and Miller v. California (1973)) in lower court majority opinions suggests low levels of compliance: lower courts at the circuit and district court level do not signal to the Supreme Court their acceptance of High Court doctrine; thus, except for 'factual' cases, most circuit and district court decisions do not comply formally with higher court precedent. An analysis of judicial votes, however, suggests that a Supreme Court doctrinal shift (to Miller v. California) influences lower court decisions only at the circuit court level. Further investigation suggests that Supreme Court precedent has a greater influence in circuit courts than in district courts: not only is the magnitude greater for circuit (versus district) court decisions, such results occur when controlling for such factors as the appointing president, regional variations, various constitutional claims and types of litigants. Thus, it appears that the influence of Supreme Court doctrine is much stronger in the circuit courts (only one step removed from the Supreme Court) than in district courts yet the hierarchy is influential nonetheless.

In this study I analyzed the 1089 desegregation outcomes in federal district courts that occurred between 1994 and 2004 in order to identify a) the legal and non-legal factors in the litigation process that predict pro-desegregation outcomes and b) the judicial patterns that impact the future of desegregation policy. Twenty-one legal and non-legal variables were analyzed via logistic regression analysis to identify factors that predict pro-desegregation outcomes. Only three predictor variables were statistically significant: Government Litigants; Region 3 (West) and Region 4 (Northeast.) Descriptive analyses of the data identified two trends in the pattern of litigation: The percentage of defendant wins increased after 1991 at a lesser rate than has been previously reported. I conclude that based on the results of both the quantitative and qualitative analyses the federal district courts are not a barrier to desegregation and can still be a part of a comprehensive desegregation strategy.

This thesis focuses on how well indicators of judicial ideology and institutional constraints predict whether a judge will vote to increase abortion access. I develop a model that evaluates a judge's decision in an abortion case in light of ideological factors measured at the time of a judge's nomination to the bench and legal and institutional constraints at the time a judge decides a case. I analyze abortion cases from all of the U.S. Federal District Courts from 1973-2004. Unlike previous studies, which demonstrate that the president and the home state senators are the best predictors of judicial ideology, I find that the Senate Judiciary Committee at the time of the judge's nomination is the only statistically significant ideological indicator. Also, contrary to conventional wisdom, Supreme Court precedent (a legal constraint) is also a significant predictor of judicial voting behavior in abortion cases.