Two questions were the focus of the research: (1) whether frequent and violent offenders had the same characteristics as frequent and non-violent offenders; and (2) whether there was specificity or generality in violence. Data were obtained from children of mothers who participated in the Philadelphia portion of the Collaborative Perinatal Project between 1959 and 1962. Study measures included family background, child development, and criminal offending variables. In general, findings suggested that frequent and violent offenders were not distinguishable from frequent and non-violent offenders and that involvement in violence was primarily a function of increasing offense frequency; that is, there was no tendency to specialize in violence. Implications for future theoretical research are addressed. 100 references, 15 notes, and 4 tables