Findings show that most of the working assumptions about the intermediate and longer range effects of victim/witness assistance projects are plausible assumptions, buttressed by anecdotes, but with weak or nonexistent support from systematic research and evaluation. For all project types (victim, witness, and victim-witness), good cross-site comparisons of process and impact, using uniform measures, are needed, and good single-site evaluations are still needed in the victim assistance area. At least two variations of victim project intervention deserve a comparative test: victim services with an onscene crisis intervention component and victim services without onscene intervention. Similarly, there are some major variations in witness services that should be examined: those projects or components providing notification services only; those projects supplementing notification with appearance support, restitution, counseling, and the like, and those jurisdictions in which no special witness services exist. A longitudinal study of the institutionalization experience of local victim/witness assistance efforts could provide many insights as well. Tabular data, 5 footnotes, and a selected bibliography of approximately 50 references are provided.