SACSI was a multipronged anti-violence initiative in ten U.S. cities built upon theory and research regarding problem-oriented policing (Goldstein, 1990), community policing (Rosenbaum, 1994), repeat violent offenders (Kennedy, 1997), and partnerships (Roehl

et al., 1996). Adapting the process used in Boston, the SACSI model had several distinguishing characteristics: multi-agency collaboration, integration of research into program planning and implementation, and strategic problem solving, all under the leadership of the U.S. Attorney’s Office. In Boston, a multi-agency planning group developed coordinated problem-solving strategies using detailed information about severe juvenile homicide and gun-related crime problems supplied by a research partner and law enforcement officers. Boston’s signature strategy-convening offender notification or “lever-pulling” meetings2 with high-risk offenders that were designed to deter juvenile crime through a combination of warnings of swift and sure enforcement and prosecution for any violence and the provision of social and vocational services-seemed to be a solid success. But it was Boston’s collaborative, data-driven, problem-solving process that SACSI sought to emulate, not its central intervention strategy. The SACSI approach had much in common with previous collaborative problem-solving efforts, but the integration of a local research partner into the core planning group set it (and Boston) apart from its predecessors.