THIS RESEARCH ATTEMPTED TO DETERMINE THE CHRONOLOGICAL AGE AT WHICH CHILDREN BEGIN TO RATE THE SERIOUSNESS OF OFFENSES IN THE SAME MANNER AS DO ADULTS WITH MIDDLE-CLASS VALUES. THE FOLLOWING WERE STUDIED: (1) OFFENSE RANKING, (2) ABSOLUTE AND RELATIVE WEIGHT GIVEN TO EACH OFFENSE, AND (3) DEGREE OF CONSENSUS ABOUT THE SERIOUSNESS OF EACH OFFENSE. VARIOUS DEMOGRAPHIC VARIABLES WERE EXPECTED TO CORRELATE WITH ADULT VALUES. THE SAMPLE CONSISTED OF 807 SCHOOL CHILDREN AGED 5 TO 18. THE SELLIN-WOLFGANG LIST OF OFFENSES WAS USED AS THE PRIMARY INSTRUMENT IN THIS STUDY, EMPLOYING TWO FORMS OF THE SCALE, ONE FOR THE LOWER AGE CHILDREN AND ONE FOR THE OLDER CHILDREN. IT WAS FOUND THAT IN ABOUT THE SEVENTH OR EIGHTH GRADES CHILDREN'S RATINGS OF CRIME SEVERITY BEGIN TO BE MORE LIKE THOSE OF ADULTS AND OLDER CHILDREN. FEMALES AND MALES TENDED TO DEVELOP CRIME SEVERITY DISCRIMINATION PATTERNS APPROXIMATELY AT THE SAME AGES, ALTHOUGH THERE WAS SOME EVIDENCE THAT FEMALES DEVELOP ADULT PATTERNS SLIGHTLY EARLIER. IT WAS ALSO FOUND THAT CHILDREN DO NOT RANK PROPERTY CRIMES, CRIMES AGAINST THE PERSON, STATUS OFFENSES, WHITE COLLAR CRIMES, AND OTHER NONINDEX CRIMES THE SAME AS DO ADULTS WITH MIDDLE-CLASS VALUES. FINALLY, UNEXPLAINED DIFFERENCES WERE FOUND FROM SCHOOL TO SCHOOL. ADDITIONAL RESEARCH IN THIS AREA COULD INCLUDE THE COLLECTION OF BACKGROUND VARIABLES OF RESPONDENTS, ADDITIONAL STIMULI, VARIATIONS IN WORDING OF STIMULI, AND OTHER METHODOLOGIES TO TEST PERCEPTION OF CRIME AT THE VARIOUS AGE LEVELS. TABLES AND A BIBLIOGRAPHY ARE INCLUDED. (MLC)