Joseph L. Gastwirth (George Washington University – Columbian College of Arts and Sciences) has posted Statistical Testing of Peremptory Challenge Data for Possible Discrimination: Application to Foster v. Chatman on SSRN.
Here is the abstract:
Ensuring that minority groups are treated fairly in the legal process is an important concern. The Castaneda v. Partida and Duren v. Missouri decisions enable courts to monitor the demographic composition of the selection of potential jurors using a variety of statistical techniques. This paper shows that Fisher’s exact test is appropriate for examining statistical data on peremptory challenges when Batson issues are raised. In addition to being a well-established method, it evaluates the challenges made by each party assuming the other side is fair. Thus, it is consistent with the Supreme Court’s statement in Miller-El that the defendant’s pattern of challenges is not relevant in determining whether the prosecution’s challenges were fair. Although one has the entire population of potential jurors and the number of peremptory challenges, which are regarded a sample from the venire, both the population and the sample are of small size. This limits the power of the test to detect a system in which the odds a minority member is challenged are two or three times those of a majority member. When data is available for similar or related trials, an appropriate method for combining the Fisher tests for each trial is noted. In every case where the Supreme Court found discrimination in peremptory challenges and the data is reported, even though the power of Fisher’s exact test is low, it found a statistically significant difference in the proportions of minority members of the venire and majority members removed. It also finds a statistically significant excess of African-Americans were challenged by the prosecutor in Foster. In a case where the Court did not find bias in peremptory challenges the test did not have sufficient power to detect a substantial disparity, so the Court properly did not give the statistics much weight.