THE CONVENTIONAL OFFENSE RATE IS THE NUMBER OF OFFENSES THAT OCCUR IN A GIVEN UNIT DURING A STIPULATED PERIOD DIVIDED BY THE ESTIMATED POPULATION OF THE UNIT DURING THAT PERIOD. THE CATEGORICAL OFFENSE RATE EQUALS THE NUMBER OF PEOPLE IN A GIVEN POPULATION WHO HAVE COMMITTED A GIVEN OFFENSE DIVIDED BY THE TOTAL POPULATION. THE REPETITIVE RATE, EXPRESSED IN TERMS OF THE FREQUENCY WITH WHICH OFFENDERS IN A POPULATION HAVE COMMITTED A GIVEN OFFENSE, IS OF USE IN EXAMINING THE RESTRICTIVE EFFECT OF A THREATENED PUNISHMENT. THE RECIDIVAL RATE, EQUAL TO THE NUMBER OF CRIMES COMMITTED BY AN OFFENDER AFTER HIS OR HER PUNISHMENT DIVIDED BY THE NUMBER OF OFFENDERS WHO HAVE BEEN PUNISHED OF THE CRIMES IN QUESTION, IS OF USE IN RESEARCH ON SPECIFIC DETERRENCE. THE FOUR TYPES OF RATES ARE CALCULATED ON THE BASIS OF SELF-REPORTED OFFENSE DATA FOR 3,268 HIGH SCHOOL STUDENTS IN SIX SCHOOLS. FINDINGS ARE PRESENTED ON RELATIONS AMONG TYPES OF OFFENSES AND AMONG SCHOOLS. THE ANALYSIS SHOWS THAT NOT ONE OF THE FOUR TYPES OF OFFENSE RATES CAN BE TAKEN AS A SUBSTITUTE OR PROXY FOR ANOTHER. NO CORRELATION BETWEEN ANY TWO TYPES OF RATES IS POSITIVE AND SUBSTANTIAL BOTH AMONG OFFENSES IN THE SAME POPULATION AND FOR THE SAME KIND OF OFFENSE AMONG POPULATIONS. THE EVIDENCE SUGGESTS THAT, IN FORMULATING THEORIES ABOUT CRIME, CRIMINOLOGISTS SHOULD NOT REFER TO 'THE' CRIME RATE. THE FINDINGS DO NOT SUPPORT THE ASSUMPTION THAT DATA ON SELF-REPORTED OFFENSES ARE MORE RELIABLE THAN OFFICIAL STATISTICS. THE FINDINGS DO MAKE CLEAR THAT ONLY THROUGH RESEARCH ON SELF-REPORTED OFFENSES CAN CRIMINOLOGISTS USE SPECIAL RATES.