This study utilized Supplementary Homicide Report data compiled by the FBI. Several inconsistencies were found in the data. First, homicides could be placed in several of the categories in the Supplementary Homicide Report data. Second, the rules used to classify ambiguous homicides were not systematic. Third, the category 'unknown' was used too frequently in the data. To illustrate, there were 42 cases that were classified as robbery murders by either the Supplementary Homicide Report data or the replication study. However, only 20 were categorized in that manner in both studies. Reliability of the process of categorizing homicides could be increased by implementing relatively minor changes. Tables, notes, and 20 references. (Author abstract modified)