THE DISTRICT OF COLUMBIA'S PRETRIAL RELEASE PROGRAM CONTAINS ELEMENTS OF BOTH THE CRIME CONTROL MODEL AND THE DUE PROCESS MODEL. THE STUDY INVOLVED MULTIVARIATE BEHAVIORAL ANALYSIS BASED ON ECONOMIC CONCEPTS USED IN THE LANDES LOGIT MODEL. RESULTS INDICATED A LACK OF CORRESPONDENCE AMONG SETS OF VARIABLES PREDICTING IMPOSITION OF BOND, FAILURE TO APPEAR IN COURT, AND PRETRIAL ARRESTS. ONLY EMPLOYMENT STATUS HAD A CONSISTENT EFFECT: EMPLOYED DEFENDANTS WERE LESS LIKELY TO BE HELD ON BOND, TO FAIL TO APPEAR, AND TO BE REARRESTED BEFORE TRIAL IF RELEASED. LOCAL RESIDENTS DID NOT HAVE BETTER APPEARANCE OR REARREST RECORDS THAN NONLOCALS. THE DATA INDICATED THAT (1) GOALS OF THE DISTRICT OF COLUMBIA RELEASE SYSTEM ARE UNCLEAR, (2) THE SYSTEM NEITHER ATTEMPTS TO MINIMIZE FAILURE TO APPEAR FOR TRIAL NOR PRETRIAL CRIME, AND (3) THE SYSTEM DOES NOT NOT SELECTIVELY RELEASE THE LOWEST RISK DEFENDANTS AND HOLD THE HIGHEST RISK DEFENDANTS. THE STUDY CONCLUDED THAT STATISTICAL ANALYSIS OF PREVIOUS CASES CAN ASSIST JUDGES IN ACHIEVING A MORE EFFICENT TRADEOFF BETWEEN RISK OF MISCONDUCT AND UNNECESSARY PRETRIAL INCARCERATION. HOWEVER, THIS ANALYSIS SHOULD NOT BE USED TO PREDICT INDIVIDUAL CASES AND IS LIMITED BY THE DATA AND METHODOLOGY USED. FINDINGS NEITHER SUPPORT NOR REFUTE THE POSITION THAT SPEEDIER TRIALS WOULD SOLVE THE PROBLEMS OF PRETRIAL CRIME AND FAILURE TO APPEAR. FURTHER RESEARCH ON THIS ISSUE IS NEEDED. FIGURES, TABLES, NOTES, AND REFERENCES ARE INCLUDED. (CFW)