Researchers have long been interested in the role of race in death penalty cases in the United States, often analyzing data using conventional causal modeling techniques. In response to concerns over omitted variables and statistical procedures, the current study re-analyzed a dataset on the role of race in capital cases in the State of Maryland. The re-analysis was conducted on all death-eligible homicides in Maryland between July 1978 and December 1999 and focused on three main research questions: (1) how appropriate were the statistical models employed in the original analysis; (2) how were the findings affected by different statistical methods; and (3) what can be concluded about the role of race in death penalty case processing in Maryland? The results revealed that the original logistic regression models employed were not robust and upon re-analysis, the racial effects were significantly diminished. It is possible to conclude, then, that the routine use of conventional causal modeling to analyze the role of race in capital cases can lead to fragile results. The authors recommend the use of a variety of regression diagnostics in addition to alternative statistical procedures to gain robust results. Tables, figures, appendix, references