THE RISK STUDY INVOLVED THE CAREFUL CODING OF A LARGE NUMBER OF VARIABLES ON JUST OVER 2,000 PERSONS PAROLED IN MICHIGAN IN 1971. VARIOUS COMBINATIONS OF VARIABLES WERE TABULATED AGAINST BEHAVIOR ON PAROLE TO DETERMINE WHICH COMBINATIONS SEEMED TO BEST PREDICT VIOLENT BEHAVIOR AMONG ONE HALF OF THE SAMPLE. THE OTHER HALF OF THE SAMPLE WAS USED TO VALIDATE - SUCCESSFULLY - THESE RESULTS. THERE WAS FOUND TO BE A HIGH RISK OF VIOLENT BEHAVIOR ON PAROLE AMONG SINGLE PAROLEES ORIGINALLY CONVICTED OF AN ASSAULTIVE CRIME; AMONG SINGLE PAROLEES CONVICTED OF ROBBERY WHO HAD HAD JUVENILE COMMITMENTS; AND AMONG SINGLE MEN CONVICTED OF ROBBERY WHO HAD BEEN ARRESTED AT LEAST ONCE BEFORE REACHING THE AGE OF 15. ALSO HIGH RISKS WERE SINGLE MEN CONVICTED OF ASSAULTIVE CRIMES WHO HAD BEEN RAISED BY THEIR MOTHERS; AND SINGLE MEN CONVICTED OF ASSAULTIVE CRIME WHO HAD SERVED AT LEAST HALF OF THEIR SENTENCE IN INVOLUNTARY SEGREGATION. THE RISK OF VIOLENT CRIMES BY THESE GROUPS, IN THE ORDER PRESENTED, RANGED FROM 24 TO 44 PERCENT. LOW RISK GROUPS INCLUDE THOSE NON-SINGLE MEN (MARRIED OR DIVORCED AT THE TIME OF THE CRIME) CONVICTED OF NON-ASSAULTIVE CRIMES; NON-SINGLE MEN CONVICTED OF NON-ASSAULTIVE CRIMES WHO HAD HAD NO JUVENILE COMMITMENTS; AND NON-SINGLE MEN CONVICTED OF NON-ASSAULTIVE CRIMES WHO HAD HAD NO JUVENILE COMMITMENTS, AND WHO HAD NOT BEEN ARRESTED PRIOR TO AGE 19. RISKS FOR THESE GROUPS RANGED FROM 10.5 TO .5 PERCENT.