SUBJECTS SORTED 78 CRIMES INTO SELF-DEFINED CATEGORIES. THESE DATA THEN WERE USED TO GENERATE MEASURES OF DISASSOCIATION AMONG THE CRIMES. DISASSOCIATION MEASURES WERE USED AS INPUT FOR THE MULTIDIMENSIONAL SCALING ANALYSIS. RESULTS INDICATED THAT THE PERCEIVED STRUCTURE OF CRIME IS THREE DIMENSIONAL. INDEPENDENT RATINGS INDICATE THAT THE UNDERLYING PROPERTIES OF THE CONFIGURATION WERE PROPERTY-NONPROPERTY; PERSONAL-INSTITUTIONAL; AND HIGH FEAR-LOW FEAR. IMPLICATIONS FOR CRIMINAL JUSTICE PROFESSIONALS ARE CLEAR. FIRST, THERE APPEARS TO BE A THREE-DIMENSIONAL COGNITIVE STRUCTURE THAT BEST DEFINES LAYPERSONS' PERCEPTIONS OF CRIMINAL BEHAVIOR. SECOND, A CONCEPTUAL SYSTEM IS INDICATED WHICH PROVIDES A BASIS FOR USING DIMENSIONS OF CRIME AS PERTINENT INDEPENDENT VARIABLES RATHER THAN RELYING ON SPECIFIC CRIMINAL BEHAVIORS. THIRD, IT WAS OBSERVED THAT LAY PEOPLE TEND TO BLUR SPECIFIC CRIMINAL BEHAVIORS WHICH PROFESSIONALS HAVE TRADITIONALLY ISOLATED FROM EACH OTHER. AND LASTLY, RESULTS HAVE IMPLICATIONS FOR THE RELATIONSHIPS BETWEEN STIGMATIZATION AND CRIME. IT MAY WELL BE THAT PEOPLE REACT DIFFERENTLY TO CRIMES ON THE BASIS OF THE UNDERLYING DIMENSIONS. TABULAR DATA AND REFERENCES ARE INCLUDED. (AUTHOR ABSTRACT MODIFIED--MJW)