THE STUDY WAS CONDUCTED IN THE MORRIS COUNTY, NJ., PROBATION DEPARTMENT FROM MAY TO SEPTEMBER 1971 USING A SAMPLE OF JUVENILES RANDOMLY SELECTED FROM A POPULATION OF 75 PROBATIONERS, AGES 14 TO 16, WHO WERE NOT KNOWN DRUG ADDICTS OR INVOLVED IN OTHER TREATMENT PROGRAMS. THE SUBJECTS WERE RANDOMLY ASSIGNED TO SIX TREATMENT AND THREE CONTROL GROUPS. EACH TREATMENT GROUP MET FOR 75 MINUTES WEEKLY FOR A PERIOD OF 4 MONTHS. THREE COUNSELORS AND THREE PROBATION OFFICERS WERE RANDOMLY ASSIGNED TO THE TREATMENT GROUPS. THE COUNSELING MODEL EMPLOYED WAS THE COMMON PROBLEMS MODEL, WHICH ATTEMPTS TO DEAL WITH THE REAL PROBLEMS OF GROUP MEMBERS. THE TRAINED COUNSELORS WERE ALL RECENT RECIPIENTS OF MASTER'S DEGREES IN COUNSELING, WITH LIMITED EXPERIENCE IN GROUP EFFORTS. THE PROBATION OFFICERS POSSESSED BACHELOR'S DEGREES AND HAD NOT GROUP EXPERIENCE EXCEPT FOR A 4-WEEK TRAINING PROGRAM IN GROUP COUNSELING. PROBATION FAILURE WAS DEFINED AS AN OFFENSE RESULTING IN EXTENSION OF PROBATION OR INCARCERATION; THOSE SUBJECTS ADJUDICATED FOR ONE OR MORE OFFENSES COMMITTED DURING THE 4-MONTH TREATMENT PERIOD WERE CONSIDERED FAILURES. NO SIGNIFICANT DIFFERENCES WERE FOUND BETWEEN EXPERIMENTAL CONDITIONS IN REGARD TO PROBATION FAILURE. BOTH COUNSELING CONDITIONS TENDED TO HAVE FEWER FAILURES THAN THE CONTROL, BUT THE DIFFERENCES FAILED TO REACH THE EXPECTED LEVEL OF SIGNIFICANCE. THE FAILURE RATES WERE AS FOLLOWS: COUNSELORS, 5 PERCENT; PROBATION OFFICERS, 10 PERCENT; AND CONTROLS, 25 PERCENT. HOWEVER, THE DATA DO SUGGEST THAT PROBATION OFFICERS WITH PARAPROFESSIONAL TRAINING CAN BE NEARLY AS EFFECTIVE AS COUNSELORS WITH ADVANCED DEGREES. TABULAR DATA ARE APPENDED, ALONG WITH A BIBLIOGRAPHY. (AUTHOR ABSTRACT MODIFIED--KBL)