THE CRIMINAL JUSTICE RECORDS OF 2,542 YOUTH COMMITTED TO DYS PROGRAMS DURING THE PERIOD FROM JULY 1, 1971, THROUGH JUNE 30, 1973, WERE MATCHED WITH THEIR RESPECTIVE (DVR) RECORDS. ADDITIONALLY, THE RECORDS OF 429 YOUTH ON PROBATION WERE OBTAINED FROM THE DYS FIELD OFFICES IN 4 MAJOR CITIES. THE TOTAL SAMPLE CONSTITUTED 69.03 PERCENT OF THE TOTAL DYS GROUP REFERRED TO DVR DURING THE STUDY PERIOD. A SAMPLE OF 1,313 COMMITTED YOUTH SUPERVISED BY DYS BUT NOT REFERRED TO DVR SERVED AS THE CONTROL GROUP. INDEPENDENT VARIABLES CONSIDERED IN THE HYPOTHESES WERE PERSONAL AND DEMOGRAPHIC CHARACTERISTICS, FAMILY RESOURCES, CRIMINAL HISTORY, AND DVR RECORD. THE DATA ANALYSIS INCLUDED USE OF THE CHI SQUARE, ONE WAY ANALYSIS OF VARIANCE, AUTOMATIC INTERACTION DETECTOR, MULTIPLE REGRESSION WITH DUMMY VARIABLES, AND SIMPLE PRODUCT INTERACTION TERMS. A STRONG CORRELATION WAS FOUND BETWEEN RECEIPT OF VARIOUS DVR SERVICES AND RECIDIVISM AS MEASURED BY RECOMMITMENT SUBSEQUENT TO FURLOUGH. DVR CASES WITH SIGNIFICANT SERVICES, AND CASES CLASSIFIED AS REHABILITATED HAD THE FOLLOWING RESPECTIVE DECREASING PROBABILITIES OF RECOMMITMENT: 0.041, 0.028, AND 0.009 AS COMPARED TO 0.096 FOR THE CONTROL GROUP. VARIABLE CHARACTERISTICS PREDOMINATING IN THE GROUPS STUDIED ARE DISCUSSED. WHILE SIGNIFICANT VARIANCE WAS OBSERVED FROM THE DATA, FACTORS OTHER THAN THOSE MEASURED, APPEARED TO BE MORE IMPORTANT THAN THOSE MEASURED, SO THE RESEARCH IS VIEWED AS 'DISCOVERY.' AREAS FOR FUTURE RESEARCH INVOLVE MORE DETAILED DATA ON OCCUPATIONS AND TRAINING, AS WELL AS RECIDIVISM. TABULAR DATA ARE INCLUDED, ALONG WITH REFERENCES. (RCB)