District Matters: An Analysis of Inter-District Variation in Federal Sentencing

Jeffery T. Ulmer, Pennsylvania State University
John H. Kramer, Pennsylvania State University
James Eisenstein, The Pennsylvania State University
Lisa L. Miller, The Pennsylvania State University

The Federal Criminal Justice System (FCJS) offers extraordinary opportunities to advance theory in the fields of law and society and sociology. It applies a single set of federal statutes using identical rules of procedure through the joint interaction of a common set of officials and organizations. It does so in the full diversity that characterizes the people, culture, and politics of the United States. The following question frames our paper: to what extent does the FCJS constitute a national system dispensing uniform justice, as opposed to 94 federal court communities whose operations and decisions reflect the diversity found in their population, politics, and culture? If there are key differences between districts, what organizational, political, or cultural factors explain these differences, and how are they related to case processing and sentencing outcomes that affect defendants, and ultimately, the nature of justice in the federal system? On one hand, one might expect little inter-district variation in the federal criminal justice system. For example, organizational and cultural factors promoting uniformity in federal criminal justice might prevent local variation. The Department of Justice might monitor the decisions of U.S. attorneys' offices to produce uniform prosecutorial activities, federal probation officers might fulfill their mission of ensuring faithful implementation of the Sentencing Guidelines, the U.S. Sentencing Guidelines might promote uniformity, and federal judges might accept the Guidelines and appellate court decisions regarding their application. Furthermore, the literature shows that federal sentencing differences by race, gender, and ethnicity persist, but there is as yet no evidence that they can be attributed to differences among districts. On the other hand, local variation is a persistent theme in state-level research on criminal courts. Indeed, sentencing guidelines were adopted by many states in part to reduce such discrepancies. Since political, economic, social, and cultural differences among the 94 federal district courts exceed those found in any single state's court, it is reasonable such differences emerge aong them as well. In fact, a handful of studies using federal data also suggest that substantial aggregate variation in sentencing outcomes may exist. We analyze aggregate- and individual-level data on case disposition and sentencing outcomes for all U.S. judicial districts, with a particular focus on inter-district variation in case disposition techniques and sentencing outcomes. These analyses use existing quantitative defendant and sentencing data from the U.S. Sentencing Commission, the Department of Justice, and the federal district-level combined data set recently created by The Urban Institute. The major methods of quantitative analysis are multivariate OLS, logistic, and HLM regression. Our major findings include significant inter-district differences in imprisonment, sentence length, and the likelihood of downward departures. Furthermore, these differences persist after controlling for district differences in caseload size and offense type composition, mode of conviction, offense severity, prior record, or defendant characteristics. We conclude by discussing our findings' implications for future research, and implications for theoretical understanding of federal criminal justice, law and society, and sociology.

(Return to Program Resources)

Updated 05/20/2006