|Research on criminal sentencing has typically focused on judicial discretion, although a comparatively smaller literature exists on prosecutorial discretion. This paper focuses on extra-judicial discretion, particularly that of the prosecution and federal probation officer, in federal criminal case processing. In federal case processing (as with nearly all jurisdictions in the U.S.), prosecutors have discretion over the criminal charging decision, and they are also potentially free to negotiate charge reductions with the defense as part of a plea agreement. A peculiarity of federal sentencing, however, is the calculation of what the U.S. Sentencing Guidelines call "relevant conduct." Relevant conduct incorporates factors such as the defendant's role in the offense, the amounts of drugs seized or property lost, harm to the victim, and so on to determine a "final offense level" which, together with a criminal history score, determines the guideline range under which defendants will be sentenced. Formally, the calculation of relevant conduct, the final offense level, and the guideline scores are the province of federal probation officers as part of their presentence reports to the court, although in practice federal prosecutors can potentially have considerable influence in this process (defense attorneys have an opportunity to object to the contents of the presentence report as well). Charge manipulation and relevant conduct calculations provide an opportunity to examine important non-judicial decisions that can substantially influence federal sentencing. We frame our analysis under a theoretical perspective that emphasizes the embeddedness of focal concerns of criminal punishment in the context of court communities.
We investigate charge manipulation and relevant conduct using retrospective case processing data from federal defendants sentenced in 1999. The data, made available by the Urban Institute, merge information from the Executive Office of U.S. Attorneys, Administrative Office of U.S. Courts, and the U.S. Sentencing Commission. Our key dependent variables are: 1) the difference in severity between arrest charges and conviction charges, or charge manipulation; and 2) the difference between the federal guidelines base offense level severity and the final offense level severity, or relevant conduct calculations. We use hierarchical linear modeling to analyz case- and offender-level influences, as well as interdistrict variation in charge manipulation and relevant conduct calculations.
(Return to Program Resources)