THE REPORTED RESEARCH IS PART OF A COMPREHENSIVE RESEARCH AND DEVELOPMENT PROJECT SEEKING TO ESTABLISH A TAXONOMY BASED ON MINNESOTA MULTIPHASIC PERSONALITY INVENTORY (MMPI) PROFILES. THE SAMPLE POOL CONSISTED OF MMPI PROFILES FOR 2,500 OFFENDERS, 18 TO 27 YEARS OLD, AT THE TALLAHASSEE (FL) FEDERAL CORRECTIONAL INSTITUTION. REEXAMINATION OF THE ORIGINAL DATA RESULTED IN REVISION OF THE CLASSIFICATORY RULES. A COMPUTER PROGRAM WAS DEVISED FOR THESE RULES AND APPLIED TO A SET OF 50 PREVIOUSLY CLASSIFIED PROFILES. RESULTS WERE 70 PERCENT ACCURATE. FURTHER REVISION OF THE RULES RESULTED IN CORRECT CLASSIFICATION OF 91 PERCENT OF 35 ADDITIONAL PROFILES ANALYZED. MORE REVISIONS, INCORPORATING ESSENTIALLY CLINICAL IMPRESSIONS, WERE INCORPORATED OVER A 3- TO 4-YEAR PERIOD OF CONTINUED TESTING. IN THE COURSE OF THIS PROCEDURE A 10TH GROUP, WHICH WAS PREVIOUSLY UNCLASSIFIABLE, WAS IDENTIFIED AND RULES FOR IT WERE DEVELOPED. THIS RESULTED IN 96 PERCENT OF THE PROFILES TESTED BEING ASSIGNED TO ONE OF THE TOTAL POPULATION SO CLASSIFIED IS PRESENTED FOR EACH OF THE 10 SUBGROUPS. THESE RULES ASSUME THAT THE MMPI PROFILE HAS BEEN SCREENED FOR VALIDITY. THEY ARE DESIGNED FOR USE BY THE CLINICAL PSYCHOLOGIST FAMILIAR WITH MMPI INTERPRETATION. CONSIDERATIONS IN SORTING ARE BRIEFLY DISCUSSED, AND CHARTS AND REFERENCES ARE INCLUDED. FOR OTHER ARTICLES IN THIS STUDY, SEE NCJ 47696-47697 AND 47699-47700. (JAP)