INFORMATION ON 2,548 DELINQUENTS WAS PREPARED ON PUNCHED CARDS AND ARRANGED IN PREDICTION TABLES TO SHOW PAIRED ATTRIBUTES USING THE CONFIGURATIONAL ANALYSIS METHOD. CONFIGURATIONAL ANALYSIS PROCEEDS BY SUCCESSIVE BIFURCATION OF HETEROGENEOUS GROUPS INTO SMALLER HOMOGENEOUS SUBGROUPS FOLLOWING LINES OF MAXIMUM SUCCESS AND FAILURE AT EACH STEP OF THE SEPARATION INTO FACTORIAL PAIRS. OF 12 ATTRIBUTES TESTED FOR SIGNIFICANCE IN TERMS OF SUCCESS/FAILURE, ONLY 6 REACHED THE 0.01 LEVEL OF SIGNIFICANCE. AGE WAS DIVIDED INTO TWO GROUPS: YOUNGER, -- THOSE UNDER 16, AND OLDER, -- THOSE OVER 16. THREE CATEGORIES OF COMPLAINT TYPE WERE IDENTIFIED. VICTIM OFFENSES COVERED ANY CRIMES INVOLVING INJURY TO THE PERSON OR PROPERTY OF ANOTHER AND COVERED SUCH CRIMES AS ARSON, ASSAULT, ROBBERY, AND FORGERY. CONSPIRATORIAL OR CONSENSUAL OFFENSES INCLUDED THOSE INVOLVING NO VICTIM SUCH AS NARCOTICS OR SEXUAL OFFENSES. THE THIRD CLASS, PUBLIC ORDER OFFENSES INCLUDED ACTS SUCH AS VAGRANCY, PROSTITUTION, AND DRUNKENESS. RELIGION WAS BROKEN DOWN INTO PROTESTANT AND CATHOLIC. RACE VARIABLES WERE BLACK AND WHITE; AND SEX VARIABLES WERE FEMALE AND MALE AND WERE BROKEN DOWN IN TERMS OF SEX/AGE CATEGORIES. THE FINAL SIGNIFICANT VARIABLE, TYPE OF ACTION, WAS BROKEN DOWN INTO GROUP OFFENSES AND ACTS COMMITTED ALONE. THE STATISTICAL IMPORT OF EACH CONFIGURATIONAL VARIABLE WAS ANALYZED IN TERMS OF SUCCESS/FAILURE AS AN ISOLATE. SEX AND AGE WERE FOUND TO BE THE TWO VARIABLES MOST STRONGLY ASSOCIATED WITH RECIDIVISM: FAILURE RATE FOR MALES WAS 46 PERCENT AND 14 PERCENT FOR FEMALES. YOUNGER OFFENDERS FAILED AT A RATE TWICE THAT OF OLDER OFFENDERS. THE FAILURE RATE FOR THE ENTIRE SAMPLE WAS 32 PERCENT. BASED ON AN ANALYSIS OF COMBINATIONS OF VARIABLES, FAILURE RATE WAS HIGHEST (73 PERCENT) FOR OFFENDERS WHO WERE YOUNG, MALE, BLACK, PROTESTANT, AND ENGAGED IN ACTIONS WITH NO VICTIMS AND ALONE. FAILURE RATES WERE LOWEST (18 PERCENT) FOR THREE GROUPS OF FEMALE OFFENDERS AND TWO GROUPS OF OLDER MALE OFFENDERS. A TABLE SHOWING 18 GROUPS OF OFFENDERS CLASSED BY CONFIGURATIONAL ATTRIBUTES, TOGETHER WITH CASES AND FAILURE RATES PER GROUP IS PROVIDED. THE UTILITY OF CONFIGURATIONAL ANALYSIS AS A TOOL FOR PROGRAM EVALUATION AND FOR THE CONSTRUCTION OF A GENERAL TYPOLOGY OF DELINQUENCY IS BRIEFLY DISCUSSED. NOTES ARE INCLUDED. (JAP)