Saturday, March 01, 2008


I'm doing some numbers crunching right now in writing an expanded analysis of JO'C's recent article, John F. O'Connor, Foolish Consistencies and the Appellate Review of Courts-Martial, 41 Akron L. Rev. 175 (2008). In the course of trying to work up some numbers, I discovered that I passed along a horribly, horribly, wrong statistic in an earlier post.

In this post, I noted that MAJ John A. Hamner's article, The Rise and Fall of Post-Trial—Is It Time for the Legislature to Give Us All Some Clemency?, Army Law., Dec. 2007, at 1, finds that clemency was granted in 1.7% of fully contested cases resulting in convictions. Id. at 16.

Here's the direct quote from the article: "From 2001-2006, convening authorities either disapproved or approved less-than-adjudged punishment in 155 cases where the accused pled not guilty. This reveals clemency was given at a rate of 1.7%." I picked up that statistic and simply relied on it. Bad move.

There are two problems with the stat, one minor and one major. The minor problem is that as the article previously suggests and as the article's Appendix A demonstrates, that stat was based on calendar years 2000-2006, not 2001-2006. Here's the major problem -- the 1.7% figure is horribly wrong.

MAJ Hamner tells us that "From 1 January 2000 through 1 December 2006 the Army tried 9081 courts-martial." Id. at 16. But that is every general and special court-martial case in the Army -- contested and guilty plea. (The annual reports on CAAF's web site are done by fiscal year, so they don't line up exactly with the article's stats based on calendar years, but it is apparent that the article's annual totals are EVERY case, not just contested cases.)

Here's the problem: 155 (the number of times clemency was granted in a FULLY CONTESTED CASE) divided by 9,081 (the TOTAL number of courts-martial) = .017. So the figure 1.7% was calculated using the wrong denominator. The proper equation is 155/THE NUMBER OF FULLY CONTESTED CASES RESULTING IN A FINDING OF GUILTY, and the article doesn't seem to tell us what that number is, but it is certainly far, far less than 9,081, which means the percentage of clemency in fully contested cases is far, far, greater than 1.7%

What percentage of cases is fully contested? Does something less than 20% sound right? Let's assume it's 20%. That would mean there were 1,816 fully contested cases over the seven-year period. What percentage of those results in a finding of guilty to something? 80%? That would give us 1,452 cases. So if that's the correct denominator, then Army convening authorities granted clemency in more than 10% of the fully contested cases in which they took action, not 1.7%

Does anyone have actual stats for the number of fully contested cases tried in the Army during calendar years 2000-2006 and the number of such cases resulting in some finding of guilty? If so, could you please provide them and we will compute the actual stat?

No comments: