SUBJECTS WERE 13 TO 17 YEAR OLD 'UNMANAGEABLE' DELINQUENT GIRLS WITH CHRONIC AND SEVERE BEHAVIOR PROBLEMS. THE FOUR-STAGE, SEQUENTIAL TREATMENT PROGRAM WAS RUN IN THE FOLLOWING ORDER: PHASE ONE BEHAVIOR MODIFICATION (A SOPHISTICATED TOKEN ECONOMY PROGRAM BASED ON POSITIVE SOCIAL BEHAVIORS AND LACK OF ANTISOCIAL BEHAVIOR); PHASE TWO BEHAVIOR MODIFICATION (A TOKEN ECONOMY BASED ONLY ON SPECIFIED POSITIVE SOCIAL ACTS); PHASE THREE BEHAVIOR MODIFICATION PLUS PEER TRAINING IN REINFORCEMENT THERAPY PRINCIPLES; AND PEER THERAPY ALONE. ONLY THE LAST OF THESE TREATMENT APPROACHES RESULTED IN BOTH THE ELIMINATION OF MAJOR INSTITUTIONAL ADJUSTMENT PROBLEMS AND SUCCESSFUL POST-INSTITUTIONAL ADJUSTMENT OVER A NINE MONTH FOLLOW-UP PERIOD. IT IS CONCLUDED THAT BEHAVIOR MODIFICATION MAY HAVE A DELETERIOUS EFFECT ON BOTH INSTITUTIONAL ADAPTATION AND POST-INSTITUTIONAL ADJUSTMENT, AND THAT POST-TREATMENT PERSISTENCE OF APPROPRIATE SOCIAL BEHAVIOR DOES NOT NATURALLY FOLLOW FROM EFFECTIVE CONTROL OVER INSTITUTIONAL BEHAVIOR. REFERENCES ARE INCLUDED. --SUMMARY IN FRENCH, GERMAN, AND SPANISH