Statistical analysis revealed clusters of deviance in comparison with findings on a sample of 81 nonoffenders. Five deviancy groups were clearly identified. These ranged from a very low deviancy group that had little difference from most of the non-offenders in scores on most of the questionnaires to a high-plus deviancy group in which questionnaire scores were extremely deviant on most of the measures. Further analysis of offense demographics revealed that the men with higher deviancy were at greater risk than were the lower deviancy men of reconviction for a sex offense, to have had a large number of victims, to have committed offenses both inside and outside the family, and to have committed offenses against both male and female children. Findings suggested that the use of psychometric tests may be a useful adjunct to any risk assessment. Tables, figure, and 24 references (Author abstract modified)