infancy had clinically significant symptoms by the age of six but that not all of the children had aggressive symptoms (Dodge, Bates, and Pettit, 1990). Some symptoms, especially those of girls, were of an internalizing nature, directed inwards. Approximately 25 percent of the abused or neglected children became delinquent, a rate double that in nonabused controls. The antisocial behavior apparently begins very early. Children who have been physically abused tend to behave in more aggressive ways than children who have not (Cicchetti and Carlson, 1989; Widom, 1989a,b). Lewis (1992) described a lack of empathy in abused toddlers, which she suggested reflects their conditioned ability to insulate themselves from any stimuli that might evoke their own painful experiences. The long-term effects of physical abuse and neglect seem to be greatest among those who have aggressive parents and became aggressive themselves (McCord, 1983).
Indicators of socioeconomic disadvantage, such as poverty, overcrowding, and poor housing, have been shown to be associated with an increased risk of childhood conduct problems (Hawkins, Catalano, and Miller, 1992). Socioeconomic disadvantage appears to have its effect due to the aggregation of risk factors, that is, the number of life stresses and daily hassles that inevitably have an impact on family interactions and relationships in a way that changes what Rutter has termed the “under the roof culture” of the family (including family cohesiveness, parental responsiveness, and limit setting) (Rutter, 1985a). This effect occurs predominantly in urban settings. Offord and colleagues found that low income was one of the most significant risk factors for the onset of conduct disorder, but it had its effect on children aged 4 to 11, not on adolescents (Offord, Alder, and Boyle, 1986). In an exceptionally lengthy risk study—a 30-year follow-up to measures taken in the first five years of life—Kolvin and co-workers demonstrated that low socioeconomic status of a child's family at age 5 is a powerful risk factor (Kolvin, Miller, Fleeting, and Kolvin, 1988). Of the 31 percent of males who committed either a juvenile or an adult offense, only 5 percent came from the higher socioeconomic groups, I and II, whereas 26 percent came from group III and 42 percent from groups IV and V.
Sociological studies have shown that urban areas typically have higher rates of delinquency. In Rutter's classic studies comparing children living in the Isle of Wight with those in inner London, it was found that the rates of conduct disorder were twice as high in London. Within cities with high crime rates, there is marked variation by