The study analyzed data on approximately 6,000 defendants. Eight jurisdictions were selected for detailed analysis of release practices and outcomes: Baltimore City, Md., Baltimore County, Md., Washington, D.C., Dade County (Miami), Fla., Jefferson County (Louisville), Ky., Pima County (Tucson), Ariz., Santa Cruz County, Calif., and Santa Clara County (San Jose), Calif. The delivery system for pretrial release decisions was also assessed in each site. The study found that 85 percent of arrested defendants secured pretrial release, 87 percent of released defendants appeared for all required court dates, and 84 percent of released defendants were arrest-free during the pretrial period. In addition, experiments were conducted in four jurisdictions (Pima County, Ariz.; Baltimore City, Md.; Lincoln, Nebr.; and Jefferson County, Tex.) to assess the impacts of pretrial release programs. Their impact on release outcomes was determined through comparisons of defendants processed by the programs with control groups not processed by the programs. The programs had a major impact on release outcomes but little effect on court appearance or pretrial arrest rates. To improve pretrial release practices, the study recommends that fugitives be identified and apprehended more effectively, that trial delay be reduced, and that alternative detention facilities be developed to reduce jail overcrowding. Less restrictive criteria for program release recommendation are also suggested, and it is recommended that postrelease followup activities be evaluated at the individual program level. Tables and endnotes are supplied, along with appended material presenting release, court appearance, and pretrial arrest outcomes by specific charges, a glossary, and about 90 references. (Author abstract modified)