PSYCHOLOGISTS ARE OFTEN ASKED TO DETERMINE THE DEGREE OF DRUG DEPENDENCY OF A GIVEN DELINQUENT. IN ORDER TO SHARPEN THE ACCURACY OF THIS JUDGMENT, 8 TESTS WERE SELECTED ON THE BASIS OF 10 CRITERIA, INCLUDING BREVITY AND PALATABILITY. THE TESTS ARE LISTED AND BRIEFLY DESCRIBED IN TABLES. NEUROPSYCHOLOGICAL TESTS WERE SELECTED BECAUSE PREVIOUS FINDINGS HAD SUGGESTED THAT HEAVY DRUG USE MIGHT BE ASSOCIATED WITH ORGANIC BRAIN PATHOLOGY. THE MEASURES CHOSEN WERE ADMINISTERED TO 102 MALE OFFENDERS AGED 12 TO 17 YEARS, OF VARIED ETHNIC BACKGROUNDS. CONTROLS WERE PROVIDED FOR INTELLIGENCE QUOTIENT, PREVIOUS HEAD INJURY, AND RESPONSE BIAS. TO FORM THE CRITERION OF DRUG USAGE LEVEL, A SERIES OF 4 CRITERIA WAS EMPLOYED TO DIVIDE THE BOYS INTO 2 GROUPS -- 68 HEAVY USERS AND 34 LIGHT USERS. HIGHLY DISCRIMINATING CUTOFFS FOR EACH OF THE DEPENDENT-VARIABLE TESTS AND FOR EACH AGE AND ETHNICITY SUBGROUP, YIELDED SIGNIFICANT DIFFERENTIATION FOR SEVEN OF THE EIGHT TESTS INVOLVED. A COMPOSITE 'DRUG INDEX' CORRECTLY IDENTIFIED 74 PERCENT OF THE SUBJECTS, AND WHEN THE MOST DISCRIMINATING TESTS FOR EACH AGE AND ETHNICITY GROUP WERE SELECTED, INSTANCES OF FALSE POSITIVES WERE REDUCED TO 4 PERCENT. FUTURE RESEARCH AND CLINICAL APPLICATIONS ARE DISCUSSED. SUPPORTING DATA AND REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED-VDA)