This working paper, describes two distinct methodological approaches for examining racial disparity in the imposition of capital punishment.
In a 1999 report to the Special Master Judge David Baime, the authors asserted that there is no single statistical method that is sufficiently reliable to provide consistent evidence of racial disparity in death penalty sentencing. As a result, two methods are described that should be used in combination to assess race effects in capital punishment imposition. The first method involves multiple regression methods that isolate the effects of race variables at particular decisions points. The second involves a sorting method that describes cross tabulations illustrating the relationship between race and death outcomes within different combinations of data. In both methods, a limited number of variables were identified through the literature and through a judge survey. In the multiple regression method, the factors significantly related to race were identified and used to isolate the impacts of race on various criminal justice outcomes. In the sorting method, factors having strong relationships with the criminal justice outcomes were identified in order to explore the interaction between race characteristics, the identified factors, and the criminal justice outcomes. Results of these two diverse methods are similar in terms of the impact of race on death penalty sentencing; both methods indicate no effect for race of defendant or race of victim on death outcomes. Regarding race of defendant and advancement to penalty trial, the two methods revealed no evidence of a significant impact of race of defendant on advancement to penalty trial. Similar results were found concerning the impact of race of victim and advancement to penalty trials. The analysis revealed the importance of county variability in understanding advancement to penalty trial; this issue should receive further study before conclusions are drawn. Finally, the authors caution that the methods developed here should be used specifically to assess racial disparities in death sentencing; they were not developed to probe other models. Footnotes