CONFORMIST OFFENDERS ARE THOSE WHOSE DELINQUENCY IS ANCHORED IN A CULTURAL MILIEU THAT IS IN CONFLICT WITH THE REST OF SOCIETY, WHEREAS NONCONFORMIST OFFENDERS ARE DEVIANT FROM SOCIETY AS INDIVIDUALS. THE STUDY TESTS FOUR HYPOTHESES: (1) THAT THE COMMITTED DEFECTIVE DELINQUENT POPULATION AT THE PATUXENT INSTITUTION IN MARYLAND ACTUALLY IS COMPOSED OF CONFORMIST AND NONCONFORMIST OFFENDERS AS WELL AS MENTALLY RETARDED OFFENDERS; (2) THAT CONFORMIST INMATES HAVE BEEN MORE EFFECTIVELY SOCIALIZED THAN NONCONFORMISTS; (3) THAT GROUP PSYCHOTHERAPY WILL BRING ABOUT CHANGES IN SOCIALIZATION AND OVERT BEHAVIOR AMONG NONCONFORMISTS; AND (4) THAT CONFORMIST AND NONCONFORMIST OFFENDERS WHO ARE RESOCIALIZED IN THE INSTITUTION WILL ADJUST BETTER IN THE FREE COMMUNITY WHILE ON PAROLE THAN THOSE WHO ARE NOT RESOCIALIZED. THE GOUGH SOCIALIZATION SCALE IS USED TO MEASURE DEGREE OF SOCIAL MATURITY, INTEGRITY, AND RECTITUDE. ADJUSTMENT IS MEASURED IN TERMS OF INSTITUTIONAL EMPLOYMENT, RULE INFRACTIONS, AND INCIDENT REPORTS. STUDY SUBJECTS INCLUDE 188 INCARCERATED MEN AND 75 PAROLED MEN. STUDY FINDINGS SUPPORT THE HYPOTHESIS THAT A PORTION OF THE DEFECTIVE DELINQUENT POPULATION AT PATUXENT ACTUALLY CONSISTS OF CONFORMIST OFFENDERS, I.E., NONEMOTIONALLY DISTURBED OFFENDERS WHOSE CRIMINAL CAREERS ARE DOMINATED BY PROPERTY OFFENSES. THIS FINDING CASTS DOUBT ON THE INSTITUTION'S CLASSIFICATION METHODS. ALMOST NO DIFFERENCE IN LEVEL OF SOCIALIZATION BETWEEN CONFORMISTS AND NONCOMFORMISTS IS FOUND. NONCONFORMISTS ACTUALLY SCORED SOMEWHAT HIGHER ON THE GOUGH SCALE THAN DID CONFORMISTS. HOWEVER, THERE WAS A DEFINITE TENDENCY AMONG NONWHITES FOR CONFORMISTS TO SCORE HIGHER THAN NONCONFORMISTS. AN ASSOCIATION BETWEEN EDUCATION AND LEVEL OF SOCIALIZATION IS FOUND. THE DATA INDICATE THAT EXPOSURE TO THE SPECIALIZED TREATMENT PROGRAM AT PATUXENT DOES NOT AFFECT EITHER SOCIALIZATION OR ADJUSTMENT LEVELS. DATA ON PAROLEES SUGGEST THAT CONFORMISTS ARE LESS FAILURE-PRONE THAN NONCONFORMISTS, BUT THAT ADJUSTMENT TO THE EXPECTATIONS OF INSTITUTION STAFF IS NOT A USEFUL PREDICTOR OF PAROLE OUTCOMES. IMPLICATIONS OF THESE AND OTHER FINDINGS ARE DISCUSSED. SUPPORTING DATA, STUDY INSTRUMENTS, AND A BIBLIOGRAPHY ARE INCLUDED.