A QUESTIONNAIRE WAS GIVEN TO A SET OF 1,119 HIGH SCHOOL STUDENTS. AMONG THE ITEMS IN THE QUESTIONNAIRE WERE DESCRIPTIONS OF 19 ILLEGAL ACTIVITIES, FOR WHICH EACH RESPONDENT WAS ASKED HOW MANY TIMES HE OR SHE HAD COMMITTED THE ACT IN THE PAST YEAR, THE NUMBER OF POLICE CONTACTS FOR EACH ACT, AND THE NUMBER OF ARRESTS. THE 19 ACTIVITIES WERE THEN REDUCED INTO 5 CLUSTERS, OR DIMENSIONS. IN ADDITION, RESPONDENTS WERE ASKED IF THEY HAD BEEN A VICTIM OF ANY OF FIVE OFFENSES IN THE PAST YEAR AND WHETHER THE POLICE HAD BEEN NOTIFIED THE VICTIMIZATION. AN EXAMINATION OF THE RELATIONSHIPS AMONG THE MEASURES OF SELF-REPORTED DELINQUENT BEHAVIOR, REPORTED POLICE CONTACTS, AND REPORTED ARRESTS SHOWED THAT THE RELATIONSHIP WAS NOT CONSISTENT ACROSS THE FIVE DIMENSIONS OF ACTIVITY. THE RELATIONSHIP BETWEEN THE THREE INDICATORS OF DELINQUENT ACTIVITY WAS STRONGER WHEN EXAMINED WITH TOTAL SCORES ACROSS ALL ACTIVITIES WITH REGARD TO JUVENILE VICTIMS OF OFFENSES, THE STUDY EXAMINED THE DECISION OF VICTIMS TO NOTIFY THE POLICE. ANALYSIS OF SURVEY RESPONSES FOUND A SMALL POSITIVE ASSOCIATION BETWEEN DELINQUENT CONDUCT AND VICTIMIZATION, AND A MIXED RELATIONSHIP BETWEEN NOTIFICATION OF POLICE AND VARIOUS TYPES OF DELINQUENT CONDUCT. IN ADDITION TO AN ANALYSIS OF THE CONSISTENCY OF MEASURES OF ILLEGAL CONDUCT AT THE INDIVIDUAL LEVEL, THE STUDY ALSO CONSIDERED THEIR CONSISTENCY ACROSS GEOGRAPHIC AREAS. IT WAS FOUND THAT THE CORRELATIONS VARIED MARKEDLY BY TYPE OF OFFENSE AND THAT OFTEN THE RELATIONSHIP WITH GEOGRAPHIC CHARACTERISTICS CHANGED AS THE TYPE OF MEASUREMENT CHANGED. THE QUESTIONNAIRE IS REPRODUCED IN AN APPENDIX, AND A BIBLIOGRAPHY IS PROVIDED. (AUTHOR ABSTRACT MODIFIED--VDA)