A new analysis by the Transactional Records Access Clearinghouse (TRAC) — using case-by-case data updated through July 2012 — has found many cases in which there have been widely varying differences among the sentences issued by the judges within a federal courthouse. The findings about the courthouse variances parallels those of TRAC's earlier study on individual judges published in the October 2012 issue of the Federal Sentencing Reporter. Note that the Federal Sentencing Reporter study is based on data through September 2011.
For TRAC's new study, only courthouses with at least 2 judges who had sentenced 50 or more defendants from October 2006 to July 2012 were included. This restriction limited the study to 157 courthouses. Overall, records on a total of 420,339 cases were examined with federal judges who had each sentenced an average of 455 defendants during this period.
The courthouse with widest sentence gap in the nation was Baltimore, Maryland. Here the typical, or median, sentence for one of the eight judges who had sentenced at least 50 defendants was 17 months. At the other end of the scale in Baltimore was a judge whose typical sentence was 78 months. The remaining judges had median sentences that were arrayed between these two extremes. All but one judge had served throughout the study period (October 2006 - July 2012), and the omission of this judge did not change the sentencing range observed.
After Baltimore, the courthouses with the widest judge-to-judge differences were: Columbia, South Carolina (60 months); Philadelphia, Pennsylvania (54 months); Macon, Middle District of Georgia (51 months); and Norfolk, Eastern District of Virginia (51 months). The complete listing of the 25 courthouses in the nation with the widest judge-to-judge differences in their median sentences is given below (see Table 1).
All differences shown were highly statistically significant. This means that if cases were randomly assigned to judges at that courthouse, then such differences can't be accounted for by chance variation in the seriousness of the offenses or criminal histories of the defendants involved.
As in TRAC's earlier study, there were also a number of courthouses in which there were only modest differences among the median sentences judges there assigned. There were, in fact, twelve courthouses in which the judges had identical median sentences. Overall, there were 28 courthouses where judge-to-judges differences in median sentences were five months or less (see Table 2). Most of these 28 courthouses were relatively small, with just two or three judges each. The largest was Seattle in the Western District of Washington with seven judges.
 For more information about this TRAC database and how we developed it, see The Development of Judge-Specific Federal Sentencing Data at TRAC.