STUDIES USING SEX, RACE, AND SOCIAL CLASS AS CORRELATES OF DELINQUENCY DISPROVE THE NOTION THAT OFFICIAL AND SELF-REPORT METHODS PRODUCE CONTRADICTORY RESULTS. TO REACH A CONCLUSION OF DISAGREEMENT, EVALUATORS USE SEVERAL TECHNIQUES; THE MOST GENERAL IS THE ASSUMPTION THAT SELF-REPORTS AND OFFICIAL DATA TAP THE SAME DOMAIN OF BEHAVIOR. HOWEVER, SELF-REPORT INSTRUMENTS TYPICALLY TAP A DOMAIN OF BEHAVIOR OUTSIDE THE DOMAIN OF OFFICIAL DATA. WHEN THE DOMAIN LIMITATIONS OF SELF-REPORTS ARE RECOGNIZED (AND OTHER ILLUSORY TECHNIQUES ABANDONED), THE CONCLUSION OF GENERAL CONSISTENCY BETWEEN SELF-REPORTS AND OFFICIAL CORRELATES FOR SEX, RACE, AND CLASS EMERGES. THIS CONSISTENCY AND OTHER EVIDENCE FROM VICTIMIZATION SURVEYS, STUDIES OF THE RELIABILITY AND VALIDITY OF SELF-REPORTS, AND STUDIES OF BIASES IN CRIMINAL JUSTICE PROCESSING SUGGEST THAT BOTH OFFICIAL DATA AND SELF-REPORTS PROVIDE VALID INDICATORS OF THE DEMOGRAPHIC CHARACTERISTICS OF OFFENDERS, ESPECIALLY WHEN ATTENTION IS GIVEN TO THE DIFFERENT DOMAINS OF BEHAVIOR EFFECTIVELY TAPPED BY EACH METHOD. EXTENSIVE TABULAR DATA AND REFERENCES SUPPORT THE TEXT. (AUTHOR ABSTRACT MODIFIED--MHP)