The ethnographic data show the many different procedural stages leading to death classification. Also uncovered, was the complex web of informal rules, individual influences, institutional limitations, etc., making the classification system based more on “degrees of certainty” rather than a “true/false” statement. Homicide classification is classic criminological preoccupation, but criminology and criminal justice are only really ever concerned with the willful kind, which are murder and nonnegligent manslaughter. Other disciplines, such as epidemiology, do not operate such a differentiation and consider all homicides together, as one type of death, a violent one. This study adopted a broader, epidemiological approach to homicide, in order to understand classification effects at the death certification level. In order to achieve this goal, this article presents a procedural analysis of the classification of violent deaths based on systematic observations conducted in a medical examiner’s office in an urban area. Table, figures, and references