An evaluation of victim assistance programs is presented that examines the problems of doing research in a changing environment and the roles of program evaluation in victim/witness assistance programs as they evolved during the past decade. A study conducted under the National Evaluation Program (NEP) of the National Institute of Justice is outlined including the study approach, findings, implications of the current state of knowledge about victim/witness assistance for policymaking, and future directions for evaluation research. Community service sanctioning projects are emphasized in a paper that describes the procedures for arriving at a composite evaluable model, some initial impressions regarding community service sanctioning projects, and a preliminary draft of a composite project. In addition, a program is described that tests the effectiveness of alternate methods of handling requests for police service by the Wilmington, Del., police department. A 1978 evaluation of community anticrime programs is presented that assesses specific functions that residents and resident groups can best perform in the process of crime prevention, the structural characteristics of resident groups that limit their crime prevention effectiveness, the impact of LEAA grants on the development of such groups, and the other needed types of support to these groups. A final paper defends and examines the effects of targeted prosecution of 'career criminals' through an intensive analysis of program processes and program effects in four jurisdictions (New Orleans, La., Franklin County, Ohio, Kalamazoo, Mich., and San Diego, Calif.) Tables, figures and footnotes are included; references are provided for each chapter. A Criminal Justice Program Evaluation Agenda is appended. For separate papers, see NCJ 75713-18.