This report begins with a discussion of prior research on wrongful convictions and then reports on hypotheses that relate to reversals in capital cases. These hypotheses are tested, and suggestions for policy reform are offered to help reduce future errors in capital cases. The information analyzed for both the sample and the comparison group was compiled through the use of a questionnaire, briefs, and petitions filed on appeal and postconviction, court opinions, and court files. For both groups, the study relied primarily on information gained through the questionnaire, which was sent to the person most familiar with the case. A total of 52 sample questionnaires and 46 comparison questionnaires were returned. Logistic regression was used for the analysis to allow comparison of the influences of the various independent variables on the dependent variable. The findings of this study strongly support prior research on wrongful convictions in capital cases. Allegations of perjury were statistically significant predictors of judicial reversals. The number of types of evidence presented at trial by the prosecutor also significantly affected the risk of a reversal. The author recommends that cautionary instructions be given to juries when the state relies exclusively on the testimony of a codefendant; that increased training and qualification for defense attorneys in capital cases be provided; and that time limits for producing newly discovered evidence be eliminated. 4 tables and 27 references