TO TEST THE HYPOTHESIS THAT A SUCCESSFUL PREDICTIVE SCHEME CAN BE CONSTRUCTED IN THIS WAY, CATEGORIES OF CRIMINAL CHARGES WERE CONSTRUCTED BASED ON SIMILARITIES OF TEMPERAMENT OR ECONOMIC MOTIVE BELIEVED TO EXIST BETWEEN THE CRIMINALS WHO COMMIT THEM. IN GENERAL, PERSONS WITH SIMILAR CRIMINAL ECONOMIC PERSPECTIVES WERE GROUPED TOGETHER AND CRIMES GENERALLY COMMITTED BY PERSONS OF SIMILAR TEMPERAMENT WERE GROUPED TOGETHER. THREE CATEGORIES OF CRIMINAL CHARGES THOUGHT TO ENTAIL LOW BAIL RISK AND FOUR THOUGHT TO ENTAIL HIGH BAIL RISK WERE CONSTRUCTED. TO SEE IF PERSONS CHARGED WITH THESE CRIMES DID IN FACT HAVE THE PREDICTED HIGH OR LOW PROPENSITIES TO FLEE PROSECUTION, A BRIEF EXPERIMENT WAS UNDERTAKEN. A SAMPLE WAS CONSTRUCTED COMPOSED OF ALL OF THE CRIMINAL CASES DOCKETED IN PEORIA COUNTY, ILLINOIS, DURING THE FOUR RANDOMLY CHOSEN MONTHS OF JANUARY, FEBRUARY, MARCH, AND JULY, 1972. OF THE 1,088 PERSONS IN THE SAMPLE WHO WERE RELEASED FROM CONFINEMENT PRIOR TO THEIR TRIALS, 79 WERE IDENTIFIED AS HAVING WILLFULLY FAILED TO APPEAR. OF THREE CATEGORIES OF PROPOSED GOOD RISKS, ONE CATEGORY, THAT OF AMATEUR CRIME, CONTAINED NO FAILURES TO APPEAR, AND A SECOND (CRIMES COMMITTED BY PROFESSIONAL CRIMINALS) HAD A FAILURE-TO-APPEAR RATE LESS THAN HALF THAT FOR THE SAMPLE AS A WHOLE. OF THE FOUR PREDICTED HIGH BAIL RISK GROUPS, ONLY ONE (CRIMES OF THEFT AND DECEPTION) HAD A SIGNIFICANTLY HIGHER FALURE-TO-APPEAR RATE THAN MERE CHANCE VARIATION FROM THE AVERAGE. THEREFORE, THOUGH THERE WERE SOME SIGNIFICANT DIFFERENCES, THE RELATIONSHIPS BETWEEN CRIMINAL CHARGES AND THE PROPENSITY TO FLEE WERE SHOWN TO BE WEAK. IT IS CONCLUDED THAT EVEN THOUGH THE SET OF CLASSIFICATIONS PREDICTED GOOD BAIL RISKS, AND IDENTIFIED SOME OF THE VERY BAD RISKS, IT COULD NOT BE USED AS AN ALTERNATIVE TO A PREDICTION SYSTEM BASED ON PERSONAL INTERVIEWS. (AUTHOR ABSTRACT MODIFIED)...ELW