« "What Is Relational Justice?" | Main | Eager for input on what to cover in webinar on "The First Step Act and Other Federal Sentencing Developments" »

March 24, 2019

Interesting new TRAC data on intra-courthouse judge-to-judge differences in sentences

The Transactional Records Access Clearinghouse (TRAC) at Syracuse University maintains lots of data on the work of federal courts and federal agencies. Seemingly inspired by the recent sentencing(s) of Paul Manafort, TRAC completed a "study of judge sentencing differences at 155 federal courthouses across the country" in which "the judge with the lowest average prison sentence was compared with the judge with the highest average sentence at each courthouse."  At this page, TRAC summarizes its findings this way:

Based upon case-by-case sentencing records, the Transactional Records Access Clearinghouse (TRAC) at Syracuse University found that half of federal judges served at courthouse where the average prison sentence differed by at least 23 months depending upon which judge handled the case.  Sixty-six of these judges served at six courthouses where the average prison sentence length differed by more than 48 months.

The Orlando courthouse in the Middle District of Florida with seven judges had a range of over 80 months between the judge with the shortest versus the longest average prison sentence.  This was followed by the Greenbelt courthouse in Maryland with over 64 months difference among the seven judges serving there....

To examine current sentencing differences at each of the 155 federal courthouses included in the study, read the full report [at this link].

Because TRAC is comparing average sentences for each federal judge directly without controlling for the specific caseloads of these judges, variations in average sentences could reflect caseload differences as much as judicial differences. But in the full report, TRAC reasonably notes that due to "the fairly large number of defendants sentenced by each judge, where there is random assignment of cases to judges then statistically speaking each judge should have closely comparable caseloads so that differences in the nature of the offenses and defendants' histories are roughly comparable."

Ultimately, this TRAC report provides a crude and incomplete account of intra-courthouse judge-to-judge differences because just one or two outlier judges could and would make a courthouse look bad in this TRAC accounting.  Still, it is interesting and useful to be reminded statistically of what all federal criminal justice practitioners know well, namely that most judges have their own distinctive and unique approaches to sentencing decision-making.

March 24, 2019 at 11:54 PM | Permalink


This statistic is meaningless. if one judge has three drug cases with safety-valve eligible defendants and another has three Armed Career Criminal gun cases, the average sentence will be wildly divergent. Bad data is worse than no data.

Posted by: defendergirl | Mar 25, 2019 9:30:41 AM

What about a complete County judicial corruption? Who can you complain to? Judge that referred to himself as God, Prosecutor that had no evidence and would not accept any, sheriff that did whatever the judge asked, County judge that was used and committed suicide, I think you get the picture!

Posted by: LC in Texas | Mar 25, 2019 8:38:57 PM

I have to agree with defendergirl. Without an analysis of statistical significance, or of how much variation would be expected by random chance, these data don't tell us anything.

For example, if a judge has a single case where they sentence the defendant to 200+ years, that would obviously drive their average up quite a bit. But did that sentence reflect the judge or the case?

Posted by: William Jockusch | Mar 26, 2019 9:37:37 PM

I largely agree, defendergirl and William, with your criticisms. As I mentioned in the mean post, this TRAC report provides a crude and incomplete account of intra-courthouse judge-to-judge differences. But If they were to refine their analysis and presentation of these data, it might still tell us something worthwhile (though a recent USSC report captured these ideas with much greater sophistication and effectiveness: https://www.ussc.gov/research/research-reports/intra-city-differences-federal-sentencing-practices).

Posted by: Doug B | Mar 27, 2019 9:39:06 AM

Post a comment

In the body of your email, please indicate if you are a professor, student, prosecutor, defense attorney, etc. so I can gain a sense of who is reading my blog. Thank you, DAB