The study used large single-equation probit models of pretrial arrest, failure-to-appear, and pretrial misconduct. These equations are more representative of the classification equations that have appeared in the literature than are other equation options. The data set was based on the extensive interviews and data collection efforts of the District of Columbia Pretrial Services Agency. Analyses of data on some 7,000 defendants found that urine test results do make a consistent, significant, incremental contribution to pretrial risk classification for arrestees in the District of Columbia. Analysis by type of drug shows that specific drugs and combinations of drugs are related in different ways to the risk of pretrial rearrest, failure-to-appear, or overall pretrial misconduct. 3 tables and 35-item bibliography.