ALL OF THE SUBJECTS WERE ALLEGED TO HAVE COMMITTED A DELINQUENT ACT, BUT NONE HAD PARTICIPATED IN AN OFFICIAL COURT HEARING AND NONE HAD HAD MORE THAN TWO PRIOR POLICE CONTACTS. TO MAINTAIN THE NONCOERCIVE ASPECT OF THE PROGRAM, THE COURT AGREED TO FOREGO ADJUDICATION REGARDLESS OF A POTENTIAL SUBJECT'S DECISION TO PARTICIPATE OR NOT TO PARTICIPATE IN THE COUNSELING PROGRAM. THE SUBJECTS WERE ASSIGNED AT RANDOM TO ONE OF TWO GROUPS: A 3-MONTH OR A 9-MONTH COUNSELING RELATIONSHIP. AT THE BEGINNING AND COMPLETION OF THE COUNSELING, THE SUBJECTS TOOK A SELF-ESTEEM TEST. FOR BOTH GROUPS, SELF-ESTEEM IMPROVED OVER THE COUNSELING PERIOD. HOWEVER, FOR THE SHORT-TERM GROUP, THE IMPROVEMENT WAS SIGNIFICANT ONLY FOR TWO OF THE FOUR DIMENSIONS MEASURED (SOCIAL SELF AND SCHOOL-ACADEMIC SELF); FOR THE LONG-TERM GROUP, THE IMPROVEMENT WAS SIGNIFICANT ONLY FOR THE GENERAL SELF DIMENSION. THE SELF-ESTEEM OF THE SHORT-TERM GROUP WAS GREATER THAN THAT OF THE LONG-TERM GROUP ON THE DIMENSIONS OF SOCIAL SELF AND SCHOOL-ACADEMIC SELF. THE OPPOSITE WAS TRUE FOR THE DIMENSIONS OF GENERAL SELF AND HOME-PARENTS. THE SELF-ESTEEM SCORES OF BOTH GROUPS BEFORE AND AFTER COUNSELING WERE CONSIDERABLY LOWER THAN SCORES FOR THE GENERAL JUVENILE POPULATION. IN ANOTHER ASPECT OF ITS RESEARCH, THE PROGRAM ALSO FOUND SOME EVIDENCE OF A LINK BETWEEN INCREASED SELF-ESTEEM AND REDUCTION OF DELINQUENT BEHAVIOR. IMPLICATIONS OF THE FINDINGS ARE DISCUSSED. TABULAR DATA ARE INCLUDED. (LKM)