H. J. EYSENCK (1977) HAS HYPOTHESIZED THAT ANTISOCIAL BEHAVIOR IS MORE LIKELY TO BE EXHIBITED BY PEOPLE WHO SCORE HIGH ON THE PERSONALITY DIMENSIONS OF EXTRAVERSION, NEUROTICISM, AND PSYCHOTICISM. A SCALE OF CRIMINAL PROPENSITY WAS CONSTRUCTED BY SELECTING THOSE EXTRAVERSION, NEUROTICISM, AND PSYCHOTICISM ITEMS WHICH BEST DIFFERENTIATE ADULT CRIMINALS FROM CONTROLS IN THE STUDY SAMPLES, THEREBY SHOWING THAT DIFFERENCES EXIST IN THE DEGREE TO WHICH PERSONALITY ITEMS DISCRIMINATE BETWEEN SUCH GROUPS. THE CRIMINAL PROPENSITY SCALE FROM THE JUNIOR EYSENCK PERSONALITY QUESTIONNAIRE WAS ADMINISTERED TO FIVE GROUPS OF ADOLESCENT BOYS. SUBJECTS WERE CLASSIFIED INTO 4 GROUPS OF 20 EACH BY SCHOOL STAFF RATINGS AND A SELF-REPORT QUESTIONNAIRE OF ANTISOCIAL BEHAVIOR. THE FIFTH GROUP OF 20 WAS COMPRISED OF DELINQUENT BOYS IN DETENTION. WELL-BEHAVED BOYS, LOW IN ANTI-SOCIAL BEHAVIOR, OBTAINED SIGNIFICANTLY LOWER SCORES THAN ALL OTHER GROUPS. THE HIGHEST SCORES WERE EARNED BY DELINQUENT BOYS AND BADLY-BEHAVED SCHOOL BOYS HIGH IN ANTISOCIAL BEHAVIOR. RESULTS GENERALLY SUPPORT THE SUGGESTION THAT A SCALE COULD BE CONSTRUCTED FOR DISCRIMINATING BETWEEN WELL-AND BADLY-BEHAVED GROUPS BY COMBINING THE MOST DIAGNOSTIC ITEMS FROM THREE EYSENCK SCALES. THE CRIMINAL PROPENSITY SCALE MAY ALSO PROVE TO PREDICT THE FUTURE PROBABILITY OF DELINQUENT BEHAVIOR IN BOYS WHO HAVE NOT YET OFFENDED. TABULAR DATA AND REFERENCES ARE PROVIDED. (RCB)