Experiments were conducted in Pima County, Ariz.; Baltimore, Md.; Lincoln, Neb.; and Jefferson, County, Tex. Defendants processed by a program were compared with a control group of nonparticipants. The two groups were selected concurrently using random assignment that provided individual defendants with an equal probability of selection into either group. A comparison of the release decisions made for the experimental and control groups permitted analysis of the program's impact on the overall rates of release, the speed of release, types of release, and the equity of the release process. Outcomes were measured by differences in failure-to-appear (FTA) rates and differences in pretrial criminality rates as reflected in rates of pretrial rearrests and convictions for the rearrests. Findings show that the pretrial release programs resulted in higher release rates, higher rates of nonfinancial release, speedier release, and greater equity of release in four of the five experiments. Improvements in release outcomes were not offset by higher rates of failure to appear or pretrial criminality for program participants. There is no evidence that more intensive postrelease followup has a positive effect on defendants' failure to appear or pretrial criminality rates. There may be great potential for programs to affect release outcomes through the use of less restrictive release recommendation criteria, without increasing FTA or pretrial criminality rates. Pretrial arrest costs were much higher for the control group than the program group in the three experiments that included defendants charged with felonies. This difference was largely due to much harsher sentences for the defendants convicted of pretrial arrests in the control group compared to the experimental group. The continued support of pretrial release programs is recommended, and ways to improve programs are suggested. Tabular data are provided, and supplementary tables and descriptions of research methodology are appended. For the overall study's introduction, volume I, volume III, and the summary, see respectively NCJ 81061-62 and NCJ 81064-65.