Subjects were placed in three primary groups: (1) nonviolent offenders, composed of persons recently convicted of nonviolent offenses; (2) violent offenders, composed of persons recently convicted of violent crimes; and (3) a control/comparison group consisting of nonoffenders matched on demographic characteristics with the offender groups and a group of active duty police officers. for each of the primary groups and each of the 20 subgroups determinedby specific offense, mean MMPI profiles were determined, along with scores on Megargee's overcontrolled hostility scale and seven scales created by Foulds. One, two, and three scale high point code types were tabulated for the gorups, and a variety of configural relationships derived from the data were investigated to determine their ability to discriminate between different groups. While Foulds' and Megargee's scales were not shown to discriminate between the various groups, data supporting the conceptual framework of the overcontrolled hostility scale were found. The 4 and 278 code types are suggested as potential measures of Megargee's concepts. Four indicates characteristics of a psychopathic deviate and the 278 code shows in order of strength characteristics of depression, psychasthein, and schizophrenia. Tabulation of high point code types indicated that the most frequent among the offender population were 4 (psychopathic deviate), 489 (psychopathic deviate, schizophrenis, and hypomania), 278 and 49 code types. The data further indicate the utility of scales 8 (schizophrenia) and 6 (paranoia) in screening between offender and nonoffender groups. The importance of elevations on scale 8 (showing extreme alienation, bizarre fellings, and general dissatisfaction) in assessment of crimes of violence was demonstrated by a series of planned comparisons. findings implications are discussed. Tabular and graphic data and 200 references are provided. (Author abstract modified)