THE RESEARCH DESIGN WAS LONGITUDINAL. INITIAL OBSERVATIONS WERE OBTAINED WHEN THE STUDY POPULATION ENTERED THE NINTH GRADE AND ADDITIONAL OBSERVATIONS WERE MADE ANNUALLY UNTIL THE USUAL TIME OF GRADUATION FROM HIGH SCHOOL. THE SELF-REPORT INSTRUMENT, SIMILAR TO THAT USED BY SHORT AND NYE (1957) AND BACHMAN (1971), WAS EMPLOYED AS THE BASIC MEASURE OF DELINQUENT BEHAVIOR. A SECOND MEASURE INVOLVED THE NUMBER OF POLICE CONTACTS (INVESTIGATION REPORTS) FOR PERSONS IN THE STUDY. A COMPARISON OF THE SELF-REPORTS AND POLICE REPORTS REVEALED THAT THERE WERE FEWER THAN 5 POLICE CONTACTS FOR EVERY 100 SELF-REPORTED OFFENSES. THE RATIO OF POLICE CONTACTS TO SELF-REPORT OFFENSES VARIED BY SEX, RACE, AND SOCIAL CLASS WITH MALES, MINORITY GROUPS (PARTICULARLY BLACKS), AND LOWER-CLASS SUBJECTS ENCOUNTERING A RELATIVELY GREATER RISK OF A POLICE CONTACT FOR EACH SELF-REPORTED ACT. THREE SEPARATE MEASURES WERE DEVELOPED: DROPOUT STATUS, ATTENDANCE STATUS, AND GRADUATE STATUS. PERSONS CLASSIFIED AS DROPOUTS AT THE END OF THE STUDY HAD A SUBSTANTIALLY GREATER NUMBER OF RECORDED POLICE CONTACT AND REPORTED CONSIDERABLY MORE DELINQUENT BEHAVIOR THAN DID THE GRADUATES. THE MEAN NUMBER OF POLICE CONTACTS FOR DROPOUTS WAS APPROXIMATELY FOUR TIMES HIGHER THAN THE AVERAGE FOR GRADUATES. A SIMILAR PATTERN EXISTS WITH RESPECT TO POLICE CONTACTS FOR SERIOUS OFFENSES, ALTHOUGH THE DROPOUT RATIO FOR FEMALES IS SOMEWHAT LOWER. THE POLICE CONTACT RATES FOR DROPOUTS AND GRADUATES APPEAR TO BE CONSISTENT WITH THE FINDINGS OF PREVIOUS RESEARCH. REFERENCES AND TABULAR DATA ARE INCLUDED. (KBL)