Profiling Potentially Violent Youth: Statistical and Conceptual Problems

Abstract

Much attention is currently focused on the early identification of violence-prone adolescents. The U.S. Federal Government has even published a compendium of factors purported to be indicators of potential violence in youths. In this article we present a critical evaluation of efforts to identify potentially violent youths such as those proposed in the government's compendium. First, we review the purported early warning signs of violent youths and how these signs are being communicated to the educational community and to society at large. An evaluation of the use of such early warning signs follows with a review of how base rates of the indicators and target behaviors statistically impact judgment error rates. Differential consequences of specific types of judgment errors are discussed. Finally, we propose several broad recommendations intended to reorient our efforts toward maximizing safety in schools and help for troubled youngsters.