Anticipating the Future Based n Analysis of the Past: Intercity Variation in Youth Homicide, 1984-2007

The United States experienced two dominant trends in youth homicide over the past 25 years: an epidemic increase in the mid-1980s and early 1990s, followed by a sustained decline. By 2002, juvenile murder arrests had fallen to their lowest level since 1984. However since 2004, some U.S. cities have reported sharp increases in violence by youth. Arrests of youth for murder, robbery, and weapons offenses are up for the first time in a decade. Debate persists about whether youth violence is on the rise again. RTI International will partner with the Presley Center for Crime and Justice Studies to empirically determine characteristics of cities from which lethal youth violence can rapidly accelerate. This project, which leverages the research team's collective expertise in homicide and youth violence research, will conduct a city-level analysis of annual homicide trends for youth up to age 24 in the 100 largest cities in the U.S. from 1984-2007. In addition, a comprehensive data file representing youth lethal and non-lethal offending at the city-level over a 24-year period also will be constructed.

Construction of the data file will employ a Supplementary Homicide Report (SHR) adjustment procedure to address missing data in the SHR. This file will include key social and economic indicators for the period 1980, 1990, and 2000 and city-specific measures of drug market and gang activity. Data analysis will use a multilevel statistical modeling technique (Hierarchical Linear Modeling) to determine whether city characteristics identified in prior homicide research account for variation of these trends between cities. Specifically, the research team will estimate the annual temporal trends in youth homicide within cities during the period in which the U.S. experienced the dramatic rise and fall of youth homicide (1984-1999). The research team will then determine if the developed empirical model for this earlier time period adequately accounts for between-city variation of annual trends post 2000 (2000-2007). This will include determining whether the specified model applies equally well for juveniles (under age 18) and young adults (18-24) with further subdivisions by race and gender. The research team then will analyze whether the scope of the model can be extended to non-lethal youth violence, including robbery and assault, an issue unaddressed by previous homicide research. This study will enhance preparedness and prevention activities by identifying cities that are likely to experience significant increases in upcoming years based on past and current structural conditions and city-specific trends.

ca/ncf