Most of the data used for this study were developed as part of a National Evaluation of Pretrial Release. A major component of the study analyzed release practices and outcomes in eight jurisdictions throughout the country. Based on the very limited and poor information available, crime while on bail appears to account for no more than 10-15 percent of all crime (as measured by arrests) in most major urban areas. Further, most pretrial release programs were found to assess defendants on the basis of two broad types of factors: community ties and prior criminal justice system involvement. The most consistently important characteristics for distinguishing released from detained defendants were prior record, with three of the four indicators significant in at least four sites; and charge, which was important in all sites. Based on the available evidence, it is likely that more stringent release standards would cause substantial increases in detention, with its attendant costs for both the criminal justice system and defendants, while achieving only modest decreases in pretrial arrests. No relationship was found between rates of release and rates of pretrial arrest for the individual sites studied. Eight footnotes are provided along with appendixes containing tabular data from the study and examples of point systems at five of the sites.