The Challenge of Creating Reliable Datasets on the History of Homicides of Children

Randolph Roth, The Ohio State University

I would like to deliver a paper on the challenges that my colleagues and I have faced in our effort to create a reliable database of homicides in Ohio from 1959 to the present. My paper will focus in particular on the difficulty of gathering reliable data on homicides of children. Because Ohio is an open-record state, and because we have received wonderful support from the Department of Health, the Bureau of Vital Records, medical examiners, and law enforcement officers, especially in the Police Departments of Cleveland, Cincinnati, and Columbus, we have been able to gather data from numerous sources. We’ve discovered that the data available through the National Center for Health Statistics and the Supplementary Homicide Reports of the Federal Bureau of Investigation are badly flawed, and that the data on homicides of children are far worse that the data on homicides of adults. As we expected, the errors are far from random. To give three examples from our proposed paper. First, coroners for less affluent, rural counties routinely overlooked the findings of pathologists who discovered cases of fatal child abuse. They declared these cases “deaths from violence of unknown intent,” which meant in nearly every instance that the cases were not prosecuted. Second, the death certificates in some counties reported accidental deaths from firearms as homicides, while the certificates in other counties reported them as accidents. And third, missing ages in the Supplementary Homicide Reports were frequently recorded as zeroes, leaving the false impression of a rash of homicides of infants and newborns. In short, the data used by nearly every scholar of homicide are suspect. Only intensive, time-consuming examination of multiple sources can repair them.

No extended abstract or paper available

 Presented in Session 191. How to Count Criminals: Methods of Evaluation and Categorization