AAAS Scientific Responsibility, Human Rights and Law Program

Current Issue | Past Issues | About the Report

Report on Science and Human Rights

Fall 2002 Vol XXII, No. 2

Science on the Stand: AAAS Statistician Testifies at the Trial of Slobodan Milosevic

By Patrick Ball, Deputy Director
AAAS Science & Human
Rights Program

On March 13-14, 2002, I presented evidence for the prosecution in the trial of Slobodan Milosevic at the International Tribunal for the Former Yugoslavia (ICTY) in the Hague. The testimony was based on a statistical study I carried out with others concerning patterns of killings and migration in Kosovo during the conflict between Yugoslavia and NATO in March–June 1999. Most of the evidence presented in the trial came from eyewitnesses who described particular criminal acts. In contrast, the statistical evidence focused on a broader set of issues: the statistically permissible inferences concerning the likely cause of the migration and killings that could be drawn from patterns in the data.

The study began in April 1999 when I went to Albania with Fritz Scheuren, an eminent American statistician and current member of the AAAS Committee on Scientific Freedom and Responsibility, to investigate methods to quantify the abuses against ethnic Albanians in Kosovo that were being reported by numerous human rights organizations. While visiting the border between Albania and Kosovo, we discovered that the Albanian guards there were maintaining a registry of all the Kosovar refugees crossing the border at that checkpoint. We used this data, along with information from other sources, to create a statistical model of the number of people leaving each village in Kosovo over time.

Results of this study were published in 1999 as Policy or Panic? The Flight of Ethnic Albanians from Kosovo, March-May 1999.
We observed that people fled their villages in three waves. In each wave the number of people leaving their homes rose to a high point, then declined to a relatively much lower point. Subsequently, we undertook a second study, in cooperation with American Bar Association Central and East European Law Initiative. In the second report, we analyzed patterns of killings of Kosovar Albanians during the same period. In this study, we used a statistical technique known as “multiple-systems estimation” to combine data from four different sources to estimate the total number of people killed over time. The pattern of killings turned out to be very similar to the migration patterns. These similarities imply that the violations had the same cause.

There were three armed parties to the conflict in Kosovo—the guerrillas of the Kosovo Liberation Army (KLA), NATO, and the forces of the Yugoslav government—giving rise to three corresponding hypotheses for the origins of the migrations and killings. We obtained data on when and where NATO air strikes occurred and the ICTY gave us a collection of records (mostly from public sources) on the activities of the KLA. By comparing statistically the patterns of the NATO air strikes and KLA interactions with government forces, we showed that neither could have given rise to the killings and migrations.

This implies that Yugoslav government forces were the likely perpetrators, and other evidence exists to support this conclusion. We discovered that the dramatic decline in killings and mass migration at the end of the first wave coincided exactly with the announcement on April 6 of a unilateral cease-fire by the Yugoslav government, in honor of Orthodox Easter (which fell on Sunday, April 11, that year). Migration and killings immediately declined to their lowest levels in more than a month. During the same period NATO and KLA activities increased dramatically.
Milosevic is acting as his own attorney at the trial, and during cross-examination we came face-to-face to debate the evidence. Most of Milosevic’s questions were political, not statistical, and because they did not directly concern my testimony, the panel of three judges instructed me not to answer them. We will not know until the judges render their decision, at the end of a long trial that in subsequent phases will take up the conflicts in Croatia and Bosnia, the weight they assigned to the statistical evidence. However, the judges’ insightful questions regarding interpretation of the statistical evidence, suggest that they appreciate its value.

Our report is now “Exhibit 67” in the trial. We hope that the statistical analysis of patterns of migration and killings, combined with the testimony of eyewitnesses, will help the judges to establish responsibility for these gross human rights violations in Kosovo during the spring of 1999, thereby helping to bring justice to the region. ¨

• Video archive of Patrick Ball's testimony (March 13 and 14, 2002)