Attorney General John Ashcroft’s recent call for federal prosecutors to pursue the strongest possible course of punishment and to avoid plea bargaining when dealing with criminal cases caused quite a stir. Some U.S attorneys claimed that the request would not change the way federal criminal cases are prosecuted while others resented what they perceived as a reduction in their discretion. Such controversy naturally raises the question of how Ashcroft’s order will be implemented, whether we will actually witness a decrease in plea bargains or whether the situation remains the same, due to other influences bearing upon the decision to prosecute to the fullest extent of the law. Principal-agent models are abundant in political science as they are useful for analyzing the political relationships between elected politicians and their appointed officials. More generally, we can utilize principal-agent models to observe the interactions between the highest units and lower ones in a particular hierarchical system, even when the highest units are not responsible for the appointment of those below them. For example, some studies have shown that judicial decision making on appellate courts often depends on the ideological makeup of the U.S. Supreme Court. We can also employ principal-agent models to analyze the relationship between the 93 U.S. attorneys around the country and the Department of Justice. There are numerous factors driving, not just the decision to prosecute, but also the degree to which to prosecute, but the systematic study of government lawyer decision making in political science has been remarkably scarce Naturally, such decisions will be predominantly legal-driven and by the likelihood of success in a given case, but many political questions are also at stake and remain unanswered. To what extent does the Attorney General influence federal prosecutors? To what extent does local politics influence federal prosecutors? To what extent do federal district and appellate court judges influence them? And to what extent is there a principal-agent problem within each U.S. attorney office between U.S. attorneys and U.S. assistant attorneys? The study of government lawyers remains neglected within political science partly due to a dearth of good data. However, information on criminal cases that go to trial in federal District Court are available through legal databases such as Lexis-Nexis while information on plea-bargains are available through the Justice Department and their various publications of statistics. Using the case as the unit of analysis, I plan to evaluate the effects of several variables on whether a criminal case is fully prosecuted or plea-bargained. In particular, I plan to pay special attention to the ideological similarities between the U.S. attorneys, the Attorney General, the state’s citizens and the District Court justices. Government lawyers often speak of neutrality in their decision-making, citing legal principles and case load as the primary factors determining their behavior. However, federal criminal cases can be politically charged and many other government actors and voters are interested in how they are concluded. More research into this area can greatly inform our understanding of legal behavior in American politics and of the relationship between law and politics. [ABSTRACT FROM AUTHOR]