Sun Mar 15th, 2009 at 03:45:32 PM EST
This diary follows up on an inconclusive discussion between JakeS and me regarding proof (or lack thereof) that the ICTY is a biased court. Here, I offer the results of a statistical analysis of the available data - mainly on war crimes in ex-Yugoslavia.
In order to complete a statistically significant test, I started out by collecting specific data on the 3 separate conflicts in ex Yugoslavia: 1. the Bosnian conflict, 2. the Croatian conflict and 3. the Kosovo conflict. I went to the ICTY list of indicted individuals and separated according to the specific conflict where they stand accused of war crimes. For 7 Serbs (like Slobodan Milosevic) and others (mostly members of the FRY government or army and navy) who were not accused of a specific war crime in Bosnia or Croatia, I associated their indictment with events in Kosovo.
This approach allowed me to isolate each war as an independent event, thereby eliminating the need to estimate which army killed which civilians.
Here are the results of this analysis:
These figure were then compared to the number of civilian casualties during each of these 3 wars. I limited the casualties to civilians because this is a priori what constitutes a war crime. Most of the civilian casualty figures were obtained from Wikipedia - which itself uses the ICTY, the Red Cross and for the Serbian civilian casualty figures in Kosovo, an EU funded project run out of Belgrade. You can easily find these estimates on Wikipedia. Here are the figures:
In addition to a comparison of Serb versus non Serb accused by the ICTY, it is also pertinent to assess how the ICTY's treatment of Serbs (and others for that matter) compares to other tribunals established to prosecute war criminals - namely the International Criminal Tribunal for Rwanda and the Nuremberg Trials. Here is the data that I found (on numerous sites for Nuremberg and the official ICTR site for Rwanda):
The total casualty figure for Rwanda is 800 000. I subtracted 100 000 for military casualties and used the result for the analysis. For Nuremberg, I used a ballpark figure of 10 million civilian casualties, although the true number is probably much higher. Whether the figures used for civilian dead during WWII should be higher or lower is up to debate, but given the ratio of indicted and sentenced to civilian dead, another figure would have only a nominal impact on statistical significance test.
Below is a table which offers a synthesis of the findings - per war and per ethnic group - all wars included (ex-Yugoslav wars, Rwanda and WWII).
Ind : Cas = Ratio of Indicted to Casualties
Snt : Cas = Ratio of Sentenced to Casualties
By comparing the ratio of the Serb mean of Indicted versus enemy civilian casualties to the other group (Bosnians, Albanians, Tutsis, Germans and Croats) we see that we have 5,13 more Serbs convicted per enemy civilian casualty than the other groups. A `t test' analysis of the figures gives us a result of 2,89 for 6 degrees of freedom. This is equivalent to a 97,5% rate of certainty that the bias real (ie. that it's not just a statistical fluke). NB. For all those who want the original Excel analysis, I'll be more than happy to mail it to you.
Below is a table which offers a synthesis of the findings - per war and per ethnic group - limited to the ex-Yugoslav wars.
This analysis shows that 3,18 more Serbs are convicted per enemy civilian casualty than Bosnian Muslims, Albanians and Croats. This `t test' gives a result of 2,28 with 4 degrees of freedom - or a certainty rate in the range of 95%.
A final comment on the casualty figures used for the war in Bosnia. Croats and Bosnian Muslims were lumped together (due to limited ICTY data) which puts my null hypothesis at a disadvantage in that the retained assumption (albeit forced and false) is that all Croat and Bosnian Muslim civilians were killed by Serbs - whereas in fact, there is ample evidence of ethnic cleansing and atrocities committed between Croats and Muslims (namely around Mostar).